Assignment 4

>businessresearchmethods

schi18939_FM_i-xxxii.indd 1 2/15/18 2:08 PM

The McGraw-Hill/Irwin Series in Operations and Decision Sciences

SUPPLY CHAIN MANAGEMENT

Benton Purchasing and Supply Chain Management Second Edition

Burt, Petcavage, and Pinkerton Supply Management Eighth Edition

Bowersox, Closs, Cooper, and Bowersox Supply Chain Logistics Management Fourth Edition

Johnson, Leenders, and Flynn Purchasing and Supply Management Fifteenth Edition

Simchi-Levi, Kaminsky, and Simchi-Levi Designing and Managing the Supply Chain: Concepts, Strategies, Case Studies Third Edition

PROJECT MANAGEMENT

Brown and Hyer Managing Projects: A Team-Based Approach First Edition

Larson and Gray Project Management: The Managerial Process Seventh Edition

SERVICE OPERATIONS MANAGEMENT

Fitzsimmons and Fitzsimmons Service Management: Operations, Strategy, Information Technology Ninth Edition

MANAGEMENT SCIENCE

Hillier and Hillier Introduction to Management Science: A Modeling and Case Studies Approach with Spreadsheets Sixth Edition

Stevenson and Ozgur Introduction to Management Science with Spreadsheets First Edition

MANUFACTURING CONTROL SYSTEMS

Jacobs, Berry, Whybark, and Vollmann Manufacturing Planning & Control for Supply Chain Management Sixth Edition

BUSINESS RESEARCH METHODS

Schindler Business Research Methods Thirteenth Edition

BUSINESS FORECASTING

Keating, Wilson, and John Galt Solutions, Inc. Business Forecasting and Predictive Analytics with ForecastXTM Seventh Edition

LINEAR STATISTICS AND REGRESSION

Kutner, Nachtsheim, and Neter Applied Linear Regression Models Fourth Edition

BUSINESS SYSTEMS DYNAMICS

Sterman Business Dynamics: Systems Thinking and Modeling for a Complex World First Edition

OPERATIONS MANAGEMENT

Cachon and Terwiesch Operations Management First Edition

Cachon and Terwiesch Matching Supply with Demand: An Introduction to Operations Management Fourth Edition

Finch Interactive Models for Operations and Supply Chain Management First Edition

Jacobs and Chase Operations and Supply Chain Management: The Core Fourth Edition

Jacobs and Chase Operations and Supply Chain Management Fifteenth Edition

Jacobs and Whybark Why ERP? A Primer on SAP Implementation First Edition

Schroeder, Goldstein, and Rungtusanatham Operations Management in the Supply Chain: Decisions and Cases Seventh Edition

Stevenson Operations Management Twelfth Edition

Swink, Melnyk, Cooper, and Hartley Managing Operations across the Supply Chain Third Edition

PRODUCT DESIGN

Ulrich and Eppinger Product Design and Development Sixth Edition

BUSINESS MATH

Slater and Wittry Practical Business Math Procedures Twelfth Edition

Slater and Wittry Math for Business and Finance: An Algebraic Approach Second Edition

BUSINESS STATISTICS

Bowerman, O’Connell, Murphree, and Orris Essentials of Business Statistics Fifth Edition

Bowerman, O’Connell, and Murphree Business Statistics in Practice Ninth Edition

Doane and Seward Applied Statistics in Business and Economics Sixth Edition

Doane and Seward Essential Statistics in Business and Economics Second Edition

Lind, Marchal, and Wathen Basic Statistics for Business and Economics Ninth Edition

Lind, Marchal, and Wathen Statistical Techniques in Business and Economics Seventeenth Edition

Jaggia and Kelly Business Statistics: Communicating with Numbers Third Edition

Jaggia and Kelly Essentials of Business Statistics: Communicating with Numbers First Edition

McGuckian Connect Master: Business Statistics

schi18939_FM_i-xxxii.indd 2 2/15/18 2:08 PM

>businessresearchmethods

Pamela S. Schindler Wittenberg University

thirteenthedition

schi18939_FM_i-xxxii.indd 3 2/15/18 2:09 PM

www.mhhe.com

BUSINESS RESEARCH METHODS, THIRTEEN EDITION

Published by McGraw-Hill/Irwin, a business unit of The McGraw-Hill Companies, Inc., 1221 Avenue of the Americas, New York, NY, 10020. Copyright © 2019 by The McGraw-Hill Companies, Inc. All rights reserved. Printed in the United States of America. Previous editions © 2014, 2011, 2008, and 2006. No part of this publica- tion may be reproduced or distributed in any form or by any means, or stored in a database or retrieval system, without the prior written consent of The McGraw-Hill Companies, Inc., including, but not limited to, in any net- work or other electronic storage or transmission, or broadcast for distance learning.

Some ancillaries, including electronic and print components, may not be available to customers outside the United States.

This book is printed on acid-free paper.

1 2 3 4 5 6 7 8 9 0 LWI/LWI 1 0 9 8 7 6 5 4 3

Bound: ISBN 978-1-259-91893-3 MHID 1-259-91893-9

Looseleaf: ISBN 978-1-260-21009-5 MHID 1-260-21009-X

Portfolio Manager: Noelle Bathurst Product Developer: Ryan McAndrews Executive Marketing Manager: Harper Christopher Content Project Managers: Erika Jordan and Angela Norris Buyer: Susan K. Culbertson Design: David Hash Content Licensing Specialist: Melissa Homer Cover Image: ©scanrail/Getty Images Compositor: MPS Limited

All credits appearing on page or at the end of the book are considered to be an extension of the copyright page.

Library of Congress Cataloging-in-Publication Data

Schindler, Pamela S., author.  Business research methods / Pamela S. Schindler, Wittenberg University.  Thirteen edition. | New York, NY : McGraw-Hill/Irwin, [2019] |  Earlier editions were co-authored with Donald R. Cooper.  LCCN 2017055982 | ISBN 9781259918933 (alk. paper)  LCSH: Industrial management--Research.  LCC HD30.4 .E47 2018 | DDC 658.0072/1—dc23  LC record available at https://lccn.loc.gov/2017055982

The Internet addresses listed in the text were accurate at the time of publication. The inclusion of a website does not indicate an endorsement by the authors or McGraw-Hill, and McGraw-Hill does not guarantee the accuracy of the information presented at these sites.

schi18939_FM_i-xxxii.indd 4 2/15/18 2:09 PM

To my soulmate and husband, Bill, for his sound counsel and unwavering support.

Pamela S. Schindler

schi18939_FM_i-xxxii.indd 5 2/15/18 2:09 PM

vi Preface

Walkthrough Addressing a Revolution in Student Learning. A transformation is taking place in many of our classrooms. During the last decade, more and more of our

students have transformed to visual—from verbal—learners. Visual learners need pictures, diagrams, and graphs to

clarify and reinforce what the text relates.

Integrated research process exhibits reveal a rich and complex process in a visual way.

31 fully integrated research process exhibits link concepts within stand alone chapters.

Each exhibit in this series shares symbols,

shapes, and colors with others in the series.

Exhibit 1-3 is the overview exhibit of the

research process.

Report Insights and Recommendations

ExplorationExploration

Analyze and Interpret the Data

Design the Research

Clarify the Research Question

Collect and Prepare Data

Data Collection Design

Sampling Design

Stage 1

Stage 2

Stage 3

Stage 4

Stage 5

Subsequent exhibits (like this one for

measurement instrument development)

show more detail in a part of this process.

>Exhibit 1-3 The Research Process

Investigative Questions

Revise

Revise

Revise

Pretest Instrument Elements

Instrument Ready for Data

Collection

Measurement Questions

P h

a se

1 P

h a

se 2

P h

a se

3

Develop Non-Question Elements

Develop Instrument

Pretest Individual Questions

Pretest Instrument

>Exhibit 12-1 Instrument Design within the Research Process

schi18939_FM_i-xxxii.indd 6 2/15/18 2:09 PM

1 How the various stages of and steps within the research process work together to complete a research project.

2 What decisions are made behind the scenes of a research project.

>learningobjectives After reading this chapter, you should understand . . .

An Integrated Example >chapter 17

“I approach each project with a new

insecurity, almost like the first project I ever did.

And I get the sweats. I go in and start working; I’m not sure where I’m going. If I knew where I was

going I wouldn’t do it.”

Frank Gehry

Award-winning architect

Stage 5: Research Reports: Supported Insights and Recommendations

>chapter 16 “If you’re navigating a

dense information jungle, coming across

a beautiful graphic or lovely data visualization is a relief.

It’s like coming across

a clearing in the jungle.” David McCandless,

British data journalist,

information designer, and author

1 How changes in business and technology are changing research reporting.

2 How to plan an audience-centric report using audience analysis.

3 Different report structures and why each is used.

4 The types and specifications of various types of reports.

5 The organizational frameworks for reports.

6 Visualization and how to effectively use different support materials.

7 The role of compilation, practice, and delivery in achieving audience effect.

8 The ethical considerations in reporting research results.

>learningobjectives After reading this chapter, you should understand . . .

David McCandless, “The Beauty of Data Visualization,” TED,

November 23, 2012, downloaded June 25, 2017 (https://www.ted.com

/talks/david_mccandless_the_beauty_of_ data_visualization).

LO13-1      The tasks of data collection.

LO13-2  The use of content analysis to postcode textual and verbal data.

LO13-3  The importance of editing raw data to assure it is complete, accurate, and correctly coded.

LO13-4  The exploratory data analysis techniques that provide visual representations of the data.

LO13-5 How cross-tabulation examines relationships between variables.

>learningobjectives After reading this chapter, you should understand . . .

Stage 3: Collect, Prepare, and Examine Data

>chapter 13 “A well-crafted, thoughtful

visualization makes the light bulb go off.

You just don’t get that with a spreadsheet.”

Dana Zuber, associate director of analytics Butler, Shine, Stern &

Partners Dana Zuber,

downloaded June 22, 2017 (http://get.tableau.com/asset/how-build

-dashboards-persuade-inform-and-engage.html?cid=70132000001

HCNA&ls=Advertisement&lsd=DBM%20-%20Retarget%20-%20

How%20to%20Build%20Dashboards&adgroup=Retarget%20-%20

Dashboards&distribution=DBM&creative=building&dclid=CMHOwqm

nnswCFUYagQodwwwKWQ#preview).

1 The nature of qualitative research and its distinctions from quantitative research.

2 The types of business decisions that use qualitative methods.

3 The variety of qualitative research methods.

4 The importance and responsibility of the interviewer.

5 Ethical issues related to qualitative research.

>learningobjectives After reading this chapter, you should understand . . .

“The only way to capture a deeply personal

insight, which will help you evoke that emotion in

consumers, is through qualitative.” Gia Calhoun,

global insights manager

Burt's Bees

Stage 2: Data Collection Design: Qualitative Research

>chapter 6

LO3-1 The question hierarchy.

LO3-2 The purposes and process of exploration.

LO3-3 How internal and external exploration differ.

LO3-4 The process and goal of research valuation.

LO3-5 The process and justification needed to budget for research.

LO3-6 Ethical issues at this stage of the process.

>learningobjectives After reading this chapter, you should understand . . .

“A beautiful question is an ambitious yet

actionable question that can begin to shift the way we perceive

or think about something—and that might serve as a catalyst to

bring about change.” Warren Berger,

consultant and author, A More Beautiful Question

Stage 1: Clarify the Research Question

>chapter 3

LO1-1   How business research and data analytics complement each other.

LO1-2 The language of professional researchers.

>learningobjectives After reading this chapter, you should understand . . .

“As big data increases, we see a

parallel growth in the need for ‘small data’

to answer the questions it raises.”

William C. Pink,

senior partner

Creative Analytics

Research Foundations and Fundamentals

>chapter 1

Responsive to industry changes.

Understand what is happening behind the scenes during a project.

Research reports are increasingly oral and all about storytelling.

Clean data is critical to effective analysis.

All researchers need qualitative skills.

The research question is the basis of effective research.

You can’t learn research without understanding the fundamentals.

schi18939_FM_i-xxxii.indd 7 2/15/18 2:09 PM

Rich with examples, this edition is a collaboration with dozens of researchers.

Snapshots are research examples from the researcher’s perspective.

>chapter 9 Stage 2: Data Collection Design: Survey Research

201

Managers often must make decisions about the future. These

decisions offer high uncertainty. Research is designed to reduce

the risk, but simply asking people to predict their own behavior,

attitude, or reaction hasn’t worked well; we are notoriously poor

at the task. For example, in 1985 when individuals were asked to

predict their acceptance of Coke’s planned reformulation, they

predicted incredibly wrong and it cost Coca-Cola millions.

Historically, researchers have used the consensus predic-

tion of experts (Delphi technique) to correct for the individual’s

poor predictive capabilities. However, not all situations offer a

logical panel of experts. James Surowiecki, in his The Wisdom

of Crowds, describes how a group of diverse individuals is able

to make decisions and predictions better than isolated individu-

als or experts. MIT researchers explain that people’s heightened

connectivity due to the Internet has brought about the “emer-

gence of surprising new forms of collective intelligence.” As

social animals, people are getting good at noticing what others

are doing, sensing why they might be doing it, and predicting

what they will do. In a PEW Research study, collectively Americans

predicted 37 percent of Americans were obese, a fairly good pre-

dictor of the actual 31 percent who were so diagnosed.

Marcus Thomas (MT) needed a research method that was

fast and that would overcome client skepticism about the inac-

curacy of self-reported anticipated versus actual behaviors for

its financial client. It chose to use a prediction market. “A predic-

tion market is like an online stock investing game,” explained

Jennifer Hirt-Marchand, associate partner and strategic insights

executive for MT. “Traders ‘invest’ virtual dollars in ideas, prod-

ucts, assets, etc. to be tested. Based on the investments they

make, traders can win greater incentives if they invest in the

winning idea than an incentive they might earn by completing

a survey alone. This ‘skin in the game’ is a critical component

of the methodology, as it fosters engagement and thoughtful-

ness on the part of the traders. Its strength is that is doesn’t

rely on asking individuals to make predictions about what they

would do in the future but rather what they think other people

would do.”

Using the services of a sample provider, an online survey

was sent to general population panelists. Likely participants

self-identified based on having an understanding of finance with

regard to estate planning, personal finances and investing, vaca-

tion planning, health care, etc. A thousand participants, known

as traders, were recruited from this group. While panel partici-

pants are compensated by the sample company for their regular

participation in research projects, panelists selected for this proj-

ect could earn additional compensation based on the accuracy

of their predictions. Those payouts would be determined by the

number of traders who invested in the “winning” group, as well

as the amount each trader invested in that group.

Through a continuation of the online survey, the selected

traders were first presented a written description of the new

financial service (each had previously agreed to a nondisclosure

agreement, as the product was in development). Then each

was provided six consumer profiles (called vignettes) one at a

time. Each vignette—developed based on consumer segmen-

tation from prior research and extensive secondary research—

represented a possible purchaser group. It included a narrative,

describing the group as people, along with photographs bring-

ing each group to life. Traders were each given $1,000 in vir-

tual money to invest in one or more vignette groups—the ones

they thought would be most likely to purchase the new finan-

cial service. In addition, through open-ended questions, each

trader was asked to explain the reasons why they believed each

vignette group would or would not purchase.

Using this methodology, Marcus Thomas identified three

segments based on the best-choice vignettes—including one

that seemed unlikely at the outset—for its financial client. The

vignette that represented the client’s core segment for its main

product line failed to rank high in the prediction research for

the new financial service. Media/engagement touchpoints and

messaging insights to reach the three best-choice groups were

developed based on this research. 

Some of the world’s leading consumer and industrial compa-

nies rely on Cleveland-based Marcus Thomas LLC to create or

refine their brands and drive customers to seek out and buy their

products and services. 

www.marcusthomasllc.com

>snapshot Internet Brings Prediction Research into 21st Century

©Ridofranz/Getty Images

>chapter 12 Stage 3: Measurement Instruments

309

If a topic deals with a sensitive subject, researchers may start the topic with buffer questions, designed to build rapport and put the participant at ease. These are broad, neutral questions on the topic that don’t require a participant to take a stand on the sensitive issue. For example, “In the last 30 days, have you personally used a streaming service to watch a movie?” before asking “Should anyone be able to access movies with graphic sexual content with streaming services?” In tests, sensitive questions that followed buffer questions have been shown to extract markedly different responses compared with when participants are directly asked a sensitive question without buffers.5

Facilitate Topic and Measurement Question Sequencing

The design of measurement instruments is influenced by the need to relate each question to the others in the instrument. Often, the content of one question (called a branched question) assumes other questions have been asked and answered in a certain way. In computer-based instruments or computer-assisted instruments, such branching is handled by internal coding of the initial question. The PicProfile indicates a typical branch question; it reveals the elimination of alternatives not chosen in one question when asking the second question, thus shortening the participant’s time.

Instructions also are a primary tool to facilitate sequencing Three types of sequencing result in skip directions. These instructions indicate where the participant or interviewer should go within the instrument—a question, topic or section— given one or a series of responses. These instructions can be embedded in the instrument (paper or computer-based) or provided to the interviewer. Computer-based instruments and computer-assisted interviewing make skipping fairly easy; once a pre-programmed response is entered, the computer automatically skips the participant ahead. The first type is a question- to-question skip: a question screens for experience or knowledge, and the participant is judged unable to answer the next question without it:

Example: In the last two weeks, have you used (product)? Yes No (If No, skip to Q3)

One of the attractions of using a web survey is the ease with which participants follow branching questions immediately customized to their response patterns. In this survey, participants were shown several pictures of a prototype vehicle. Those who responded to question 2 by selecting one or more of the attributes in the checklist question were sequenced to a version of question 3 that related only to their particular responses to question 2. Note also that in question 3 the researcher chose not to force an answer, allowing the participant to indicate he or she had no opinion (“Don’t know”) on the issue of level of importance.

>picprofile

2. Which of the following attributes do you like about the automobile you just saw? (Select all that apply.)

Overall appeal

Headroom

Design

Color

Height from the ground

Other

None of the above

3. For those items that you selected, how important is each? (Provide one answer for each attribute.)

Extremely

important

Neither

important

nor

unimportant

Not at all

important

Don't

know

a) Overall appeal

b) Height from the ground

c) Headroom

Next Question

PicProfiles use a visual cue to enhance the concept or key term.

schi18939_FM_i-xxxii.indd 8 2/15/18 2:09 PM

Images are worth more than 1,000 words; they serve as visual cues to anchor concepts in memory.

>part II The Design of Business Research

92

Early in panel development, panelists were offered $100 to

$150 to join and participate in qualitative studies. Historically,

these participants participated in longer engagements (on-

line qualitative, ethnography studies or face to face in-depth

interviews). As random digit dialing became less produc-

tive for recruiting respondents and the Internet became

more widely used, researchers started using panels—also

called communities—to recruit participants for quantitative

research. Quantitative respondents who are engaged for

much shorter periods are more likely to be paid $1 to $10 per

survey completed.

The professional respondent—one who takes repeated

surveys—was once considered a deterrent to quality research.

“As researchers, ” explained Jessica Broome, PhD, principal of

Jessica Broome Research, “we wanted to keep the ‘cheaters’ and

‘repeaters’ out of our studies, believing they biased results.”

As decision cycles shorten, the demand for better and more

timely information means attracting and retaining qualified par-

ticipants. During the last three decades, increasingly this means

researchers turn to panels, and by design, these participants

are asked to participate in numerous studies.

“We wanted to know ‘Who are these people willing to take

repeated surveys?’” explained Broome. “And given that some sur-

veys are overly long and others poorly designed, ‘Why do they

do this?’” Broome teamed with Kerry Hecht, Director of Research

Services, Recollective, a division of Ramius Corporation, to find

out what motivates panel respondents and if their motivations are

likely to reduce the quality of the information they provide.

The Qualitative Study

Broome and Hecht designed a multistage study that drew par-

ticipants from multiple panel providers, including Critical Mix,

Schlesinger Associates, and Swagbucks. “We started with a

5-day online qualitative community study with 20 people to

explore what got them started as a survey panelist and what

kept them going,” explained Hecht. “While money is a motivator

in keeping panelists engaged, they also shared the influences

of intrinsic motivators like fun, feeling useful, contributing to

important decisions, and participation being more interesting

than time spent on social media.”

For any particular study, panelists are often screened exclu-

sively on demographics. “Because participants derive intrinsic

benefits, panelists will sometimes fudge on screening informa-

tion in an attempt to be included in a study, ” shared Broome.

When a panelist doesn’t meet the desired demographic param-

eters, they are told “You don’t qualify” but are rarely told why.

“But once they are included, ” explained Broome, “participants

claim honesty drives their responses; they think lying on survey

questions would undermine the study.”

“Often panelists expressed feeling abused, misled, and dis-

respected. For example, when they are told about survey length,

they often felt deceived when a promised 15-minute survey took

45 minutes, or when the survey was not only long, but boring,”

explained Broome. Increasingly, panelists in quantitative stud-

ies are tech-savvy. “They understand what current technology

should permit a survey company to do—like eliminate the need

to ask demographic questions repeatedly in the same survey

process or use earlier answers to filter later questions, ” claimed

Hecht. “They basically think researchers can make the experi-

ence so much better.” “The research industry needs their in-

sights, but can treat panelists with disdain," claimed Broome.

The Quantitative Study

Broome and Hecht followed their qualitative exploration study

with Phase 2, a mobile-optimized quantitative study of 1,499

participants, also drawn from various panel providers and fielded

by Propeller Insights. Each panelist took a topical survey that

included a creativity assessment. Additionally, half the group

(750) took a VARK assessment. VARK assesses visual, aural/

audio, read/write, and kinesthetic learning preferences through

a series of learning scenario questions. Creativity was assessed

through a battery of 38 statements requiring agreement or dis-

agreement, as well as a checklist of 54 descriptors. “We dis-

covered that participants didn’t favor any one of the learning

approaches, nor were they outliers on the creativity assessment,”

shared Broome.

Who’s Really Taking Your Surveys?

>closeup

>chapter 5 Stage 2: Sampling Design

93

a particular software training strategy, we infer that others will also. The basic idea of taking a sample is that by selecting some cases in a population, we may draw conclusions about the entire target population.

There are several compelling reasons for using a sample (a subset of the target population) rather than a census (all cases within a population), including (1) lower cost, (2) greater speed of data collection, (3) availability of population cases, and (4) greater accuracy of results. The advantages of taking a sample over census are less compelling when two conditions exist: (1) a census is feasible due to a small target population and (2) a census is necessary when the cases are quite different from each other.5 When the population is small and variable, any sample we draw may not be representative of the population from which it is drawn. The resulting values we calculate from the sample are incorrect as estimates of the population values.

Lower Cost The economic advantages of taking a sample rather than a census are massive. Consider the cost of taking a census. The 2020 U.S. Census of the Population is expected to cost as much as $30 billion, barring any natural disasters and using 2010 costs as a barometer; the Census Bureau’s own estimate is $22billion.6 By any reasonable standard, continuing to take a census of the population every 10 years is unsustainable. Is it any wonder that researchers in all types of organizations ask, “Why should we spend thousands of dollars interviewing thousands of employees in our company if we can find out what we need to know by asking only a few hundred?”

Greater Speed of Data Collection  Due to the smaller number of cases in a sample, using a sample drawn from a target population will always take less time than conducting a census.

Broome and Hecht’s Phase 2 research revealed factors es-

sential for reducing respondent drop outs: Don’t ask screening

questions within a survey once a participant has answered a de-

tailed screener to qualify; merge the data. Participants don’t care

that different companies do different aspects of the research; they

want researchers to avoid duplication. Don’t ask essentially the

same question in different ways; once is enough. Use previous

question responses to determine later questions asked; panel-

ists are willing to share their thoughts, behaviors, and lives but

don’t want their time wasted. Respondents have different pref-

erences and styles; let them respond as they wish—with text or

video, exclusively on a mobile device or a computer. “They shared

lots of ideas for making surveys more engaging and interesting,

” claimed Hecht, “including, adding music or video, getting rid of

grid questions, and reducing survey time to 15 minutes or less.”

Using information extracted from an open question and ana-

lyzed with OdinText, Phase 2 research also revealed that “most

participants were a member of one or two panels but struggled

to remain engaged with their panels, ” shared Hecht. “Partici-

pants appreciate feeling like a valued part of a full process, not

just an interchangeable cog in a wheel. They like to participate

in studies that are interesting to them and about products or

services that are relevant to their daily lives—knowledge is a

powerful motivator.” Many also cited meeting interesting people

and hearing different viewpoints as motivation. “They also love

learning the results of studies they participate in and under-

standing why we as researchers do what we do or ask what we

ask, ” shared Broome.

Recently, a SurveyMonkey study also found panelists gave

thoughtful, consistent answers over time. It released results of a

1, 000-person international panel assessment study, using three

surveys with the same respondents, one in each of three se-

quential months, which checked for quality-reducing behaviors

like straight lining (repeatedly choosing the same answer choice

in matrix questions), poor open response validity (responding

with nonhelpful, gibberish answers), and whether they were un-

focused or not paying attention. Of the panelists, 97 percent,

97 percent, and 94 percent or more passed each of these tests,

respectively, with no difference between men and women.

And in terms of response reliability, over the three waves of

surveys, in 23 indicators SurveyMonkey tracked, only three

items showed significantly significant change among U.S par-

ticipants: time-to-complete, choice of the “other” response, and

attitude about “moral acceptability of alcohol use.”

Sources: jessicabroomeresearch.com; recollective.com; surveymonkey.com

>closeupcont’d

>chapter 2 The Research Process: An Overview

31

promotions, training experiences, project assignments, project leadership, new customers captured, and so on. Sophisticated businesses track these KPIs through digitized, perpetually maintained dashboards— a data visualization tool that shows current and a period of prior status on each metric, usually on one screen. Even small businesses can dashboard KPIs create with the software tools available.

Identifying opportunity-based management dilemmas is more time consuming and difficult. It requires monitoring obscure developments in a variety of industries, as well as emerging trends in your own industry. Those companies looking for opportunities are trying to be first movers; they want to capitalize on an environmental trend that others haven’t recognized. Jeff Bezos, chief executive officer (CEO) of Amazon, is always seeking these kinds of dilemmas: customers getting frustrated at not finding the solutions they sought at stores, companies needing a more reliable delivery option, readers wanting to take more books with them when they travel than luggage space permits, etc. The search for opportunity-based management dilemmas is full of risk, but when the approach pays off, it often pays big.

However, choosing one dilemma on which to focus may be difficult. Solving different dilemmas offers different rewards. One approach is to estimate the payoff of solving a dilemma, using this value to prioritize. This is more easily done for problems (e.g., estimating the cost of reducing customer returns) than for opportunities. Ultimately, to choose incorrectly puts a company or business on an unproductive path. As a manager, only practice makes you proficient at this task. For new managers, or for established managers facing new responsibilities, developing several management dilemma-to- research question hierarchies, each starting with a different dilemma, will assist in the choice process. To develop these, much exploratory research is used, tapping into published secondary sources and mining company data.

Exploration Seeking existing information is often used not only to identify dilemmas (e.g., identify industry standard to compare to company performance), but also to ask the right questions and better understand deci- sion options. Historic company data are increasingly being used in exploration as better tools to tap into digital data warehouses have become available.1 Much of this information may not be in search able databases; it may be in written reports, where accessibility is an issue. At this stage, it always pays the

We are attracted to experi- ments at an early age due to our unlimited curiosity. A great researcher fosters curiosity as an important skill.

© B

le n

d/ Im

ag e

S ou

rc e

A Closeup offers a more in-depth example.

schi18939_FM_i-xxxii.indd 9 2/15/18 2:09 PM

Using learning aids to cement concepts.

Discussion questions tie to learning objectives and come in four types. Relating questions to newsworthy businesses

makes them more relevant to students.

>part II The Design of Business Research

146

interview 129

semistructured 130

structured 130

unstructured 130

interview guide (discussion guide) 131

moderator 129

nonprobability sampling

pretasking 127

projective techniques 133

ambiguities and paradoxes 133

authority figure 133

cartoons or empty balloons 133

collage 133

completion/fill in the blank 133

component sorts 133

imagination exercises 133

imaginary universe 133

personification 133

projective drawing 133

visitor from another planet 133

laddering or benefit chain 133

metaphor elicitation technique

role playing 133

creative innovation role play 133

semantic mapping

brand mapping

sensory sorts 133

Storytelling 133

Thematic Apperception Test 133

Third-party projection 133

word or picture association 133

Write a letter 133

qualitative research 124

quantitative research

triangulation 143

Terms in Review

1 How does qualitative research differ from quantitative research?

2 What is data saturation, and how does it influence qualita- tive research?

3 What is the nature of data collected in qualitative research?

4 Why do senior executives feel more comfortable relying on quantitative data than qualitative data? How might a qualita- tive research company lessen the senior-level executive’s skepticism?

5 Distinguish among structured, semistructured, and unstruc- tured interviews.

Making Research Decisions

6 Assume you are a manufacturer of small kitchen electrics, like Hamilton Beach/Proctor Silex, and you want to deter- mine if some innovative designs with unusual shapes and colors developed for the European market could be suc- cessfully marketed in the U.S. market. What qualitative re- search would you recommend, and why?

7 NCR Corporation, known as a world leader in ATMs, point- of-sale (POS) retail checkout scanners, and check-in kiosks at airports, announced in June 2009 that it would move its world headquarters from Dayton (OH) to Duluth (GA), a suburb of Atlanta, after more than 125 years. An employer of 1,200 mostly high-salaried, professional workers in Dayton, NCR was enticed to move by Georgia’s offer of more than $56.9 million in tax credits; its fast-growing, educated 25- to 34-year-old population cohort; international offices for 10 European state governments; and the busiest international airport (Atlanta) in the world.

a. What qualitative research might NCR have done to reach this decision?

b. NCR will use its move to Georgia to downsize its world headquarters workforce. What qualitative research could help NCR determine which of its 1,200 employees will be offered positions in Duluth?

From Concept to Practice

8 Use Exhibit 6-6 to develop the recruitment screener for the research you described in your answer to question 5.

9 Conduct a focus group among students in your class on one of the following topics:

a. The department’s problems offering requirements and electives essential for meeting your graduation expectations.

b. Entertainment sponsored by your university to bring the community on campus.

From the Headlines

10 Lately, airlines have been having a rough time, in terms of legal actions and PR issues, with consumers openly expressing outrage at being bumped from—or forcibly re- moved from—flights. Design a qualitative study to reveal the suppressed (as opposed to surface) issues that are contrib- uting to this rage.

a. What are some of the surface issues?

b. Who will you want to participate and how will you recruit them?

c. What qualitative method(s) will you choose. Be specific about any exercises you will incorporate.

>discussionquestions

462

>part VI Stage 5: Report the Research

style (visual, auditory or kinesthetic) and the level of pathos, logos, and ethos needed to persuade the audience. Visualization includes selection and preparation of the best possible support materials to encourage the audience to embrace the research insights and recommendations. There are several types of support a researcher might use: facts, statis- tics, stories, demonstration, testimony/expert opinion, analogy, and metaphor. Facts and statistics are the core of most research reports, but these should not be the only support provided. Presenting statistics presents special challenges for researchers, given the possibility that the audience might misunderstand or misconstrue findings if incorrectly presented.

LO16-7 The last three activities within the content and style phase are compilation, practice, and delivery. Compila- tion includes not only the preparation of the written report to include all the support materials, but also the development of the oral presentation and its aids. It is the selection of the oral report within the report structure that determines the need for practice. During practice, the researcher determines the time it takes to explain various insights, marks a script or notecards to indicate pauses, experiments with various audience-engagement

exercises, cues audio and video clips, prepares responses to anticipated questions, stages the presentation environment (seating arrangement, lighting, sound, screen placement, etc.), and rehearses contin- gency plans for things that might go wrong. Practice is the difference between a stellar result in achieving audi- ence effect and disappointment.

LO16-8 Whether the researcher is an employee of the res- earch sponsor or an external supplier, to achieve and maintain an effective researcher–sponsor relation- ship, the report must fulfill several ethical respon- sibilities, including (1) the sponsor’s right to quality research, (2) the absence of researcher coercion, and (3) the sponsor’s right to findings nondisclosure. The first requires choosing visualization and data- reporting techniques and tools appropriate for the data collected and maximizing the sponsor’s value. The second requires explaining and maintaining the researcher’s role, living within the scope of the data collected, and not violating participant confidentiality. The third requires controlling the safety and distri- bution of any data reports, support materials, and presentation materials printed or maintained on the researcher’s or other cloud servers.

3-D graph 456

actionable insights 434

analogy 447

anchoring bias 438

area graph 451

audience analysis 437

audience-centric planning 436

auditory learners 445

bar graph 454

confirmation bias 438

conformity bias 438

data 434

data-centric planning 436

data clarity 451

demonstration 447

desired audience effect 436

ethos 446

executive summary 441

facts 446

findings nondisclosure 461

geograph 456

graph 450

infographic 459

information 434

insight 434

jargon 458

kinesthetic learners 445

language level 458

limitations 442

line graph 451

logos 446

loss-aversion bias 438

management report 440

metaphor 447

pathos 446

performance anxiety 459

pictograph 456

pie graph 451

predispositions 438

report framework 443

report structure 439

right to quality 460

scope 441

statistics 446

story 447

survivorship bias 438

table 449

technical report 440

testimony/expert opinion 447

tone 458

visualize 444

visual learners 445

whitespace 457

>keyterms

569

>glossary

3-D graphic a presentation technique that permits a graphical comparison of three or more variables; types include column, ribbon, wireframe, and surface line.

a priori contrasts a special class of tests used in conjunction with the F test that is specifically designed to test the hypotheses of the experiment or study (in comparison to post hoc or unplanned tests).

acquiescence bias a tendency for participants to agree with an item or statement within a measurement question that asks for levels of agreement/disagreement; occurs when they have less knowledge on a topic; more a problem for less educated or less informed participants.

action research a methodology with brainstorming followed by sequential trial-and-error to discover the most effective solution to a problem; succeeding solutions are tried until the desired results are achieved; used with complex problems about which little is known.

actionable insights insights aligned with key business goals and strategic initiatives that are novel, unusual, or unexpected and that lead to recommendations for specific decisions.

administrative question a measurement question that identifies the participant, interviewer, interview location, and conditions; generates nominal data.

after-only design preexperimental design that takes one measure- ment of DV after manipulation of the IV.

alternative hypothesis (H A ) an assumption that a difference exists

between the sample parameter and the population statistic to which it is compared; the logical opposite of the null hypothesis used in significance testing.

ambiguities and paradoxes a projective technique (imagination exercise) in which participants imagine a brand applied to a different product (e.g., a Tide dog food or Marlboro cereal), and then describe its attributes and position.

analogy a rhetorical device that compares two different things to highlight a point of similarity.

analysis of variance (ANOVA) tests the null hypothesis that the means of several independent populations are equal; test statistic is the F ratio; used when you need k-independent- samples tests.

anchoring bias an audience’s tendency to rely on its first impressions.

area graph a graphical presentation that displays total frequency, group frequency, and time series data; a.k.a. stratum chart or surface chart.

area sampling a cluster sampling technique applied to a popula- tion with well-defined political or natural boundaries; population is divided into homogeneous clusters from which a single-stage or multistage sample is drawn.

artifact correlations occur when distinct subgroups in the data combine to give the impression of one.

assimilation effect question order that encourages more similarity in responses between the first and the second question.

asymmetrical relationship a relationship in which we postulate that change in one variable (IV) is responsible for change in another variable (DV).

attitude a learned, stable predisposition to respond to oneself, other persons, objects, or issues in a consistently favorable or unfavorable way.

attitude scaling process of assessing a personʼs disposition (from extremely favorable disposition to an extremely unfavorable one) toward an object or its properties using a number that represents a person’s score on an attitudinal continuum range.

audience analysis an analysis of the expected audience for a research report.

audience-centric planning a research report orientation whose focus is on gaining the audience‘s embrace of data insights and recommendations; the resulting presentation is persuasive and tells a story employing statistics.

auditory learners audience members who learn through listening; represent about 20 to 30 percent of the audience; implies the need to include stories and examples in research presentations.

authority figure a projective technique (imagination exercise) in which participants are asked to imagine that the brand or product is an authority figure and to describe the attributes of the figure.

automatic interaction detection (AID) a data partitioning proce- dure that searches up to 300 variables for the single best predictor of a dependent variable.

balanced rating scale has an equal number of categories above and below the midpoint or an equal number of favorable/ unfavorable response choices.

bar graph a graphical presentation technique that represents frequency data as horizontal or vertical bars; vertical bars are most often used for time series and quantitative classifications (histograms, stacked bar, and multiple-variable charts are specialized bar charts).

Bayesian statistics uses subjective probability estimates based on general experience rather than on data collected. (See “Decision Theory Problem” at the Online Learning Center.)

behavior cycle how much time is required for a behavior and between behavior events; used with behavior frequency in developing behavior-based measurement questions.

behavior frequency how often a participant executes a particular behavior; used with behavior cycle in developing behavior- based measurement questions.

behavior time frame a factor needed for every behavior meas- urement question; an appropriate length is determined by a participantʼs behavior cycle and behavior frequency.

bibliography (bibliographic database) a secondary source that helps locate a book, article, photograph, etc.

bivariate correlation analysis a statistical technique to assess the relationship of two continuous variables measured on an interval or ratio scale.

bivariate normal distribution data are from a random sample in which two variables are normally distributed in a joint manner.

blind when participants do not know if they are being exposed to the experimental treatment.

Key terms are a valuable

refresher, in each chapter

and in the glossary.

Glossary reinforces the

language of research.

schi18939_FM_i-xxxii.indd 10 2/15/18 2:09 PM

264

>part III Measurement

Summated Rating Questions The Likert scale, developed by Rensis Likert (pronounced Lick-ert), is the most frequently used variation of the summated rating question. Questions based on summated rating scales consist of statements that express either a favorable or an unfavorable attitude toward the object of interest. The participant is asked to agree or disagree with each statement. Each response is given a numerical score to reflect its degree of atti- tudinal favorableness, and the scores may be summed to measure the participant’s overall attitude. Summa- tion is not necessary and in some instances may actually be misleading, as our caution below clearly shows.

In Exhibit 11-8, the participant chooses one of five levels of agreement. This is the traditional Likert scale because it meets Likert’s rules for construction and testing. The numbers indicate the value to be assigned to each possible answer, with 1 the least favorable impression of Internet superiority and 5 the most favorable. Likert scales may also use 7 and 9 scale points. Technically, such question is a Likert- type question as its construction is less rigorous than the process Likert created. However, the advan- tages of the 7- and 9- point scales are a better approximation of a normal response curve and extraction of more variability among respondents.

Conscientious researchers are careful that each item meets an empirical test for discriminating abil- ity between favorable and unfavorable attitudes. Originally, creating a Likert scale involved a procedure known as item analysis. Exhibit 11-9 provides the steps for selecting Likert statements (items) for the scale using item analysis. The values for each choice are normally not part of the measurement instru- ment, but they are shown in Exhibit 11-10 to illustrate the scoring system.

>Exhibit 11-9 How to Perform an Likert Item Analysis

Item analysis assesses each item (statement) in a Likert scale based on how well it discriminates between those people whose total score is high and those whose total score is low.

Step 1 Collect a large number of statements that meet the following criteria 

● Each statement is relevant to the attitude being studied.

● Each statement reflects a favorable or unfavorable position on that attitude.

Step 2 Select people similar to study participants (participant stand-ins) to read each statement.

Step 3 Participant stand-ins indicate their level of their agreement with each statement, using a 5-point scale. A scale value of 1 indicates a strongly unfavorable attitude (strongly disagree). A value of 5 indicates a strongly favorable attitude (strongly agree). The other intensities (2 (disagree), 3 (neither agree nor disagree), 4 (agree) are mid-range attitudes (see Exhibit 11-3).

● To ensure consistent results, the assigned numerical values are reversed if the statement is worded negatively. The number 1 is always strongly unfavorable and 5 is always strongly favorable.

Step 4 Add each participant stand-in’s responses to secure a total score.

Step 5 Array these total scores from highest to lowest; then and select some portion—generally defined as the top and bottom 10 to 25 percent of the distribution—to represent the highest and lowest total scores.

● The two extreme groups represent people with the most favorable and least favorable attitudes toward the attitude being studied. These extremes are the two criterion groups by which individual Likert statements (items) are evaluated.

● Discard the middle group’s scores (50 to 80 percent of participant stand-ins), as they are not highly discriminatory on the attitude.

Step 6 Calculate the mean scores for each scale item among the low scorers and high scorers.

Step 7 Test the mean scores for statistical significance by computing a t value for each statement.

Step 8 Rank order the statements by their t values from highest to lowest.

Step 9 Select 20-25 statements (items) with the highest t values (statistically significant difference between mean scores) to include in the final Likert scale

Researchers have found that a larger number of items for each attitude object improves the reliability of the Likert scale. As an approximate indicator of a statement’s discrimination power, one authority suggests using only those statements whose t value is 1.75 or greater, provided there are 25 or more participant stand-ins in each group. See Exhibit 11-5 for an example.

Source: Allen L. Edwards, Techniques of Attitude Scale Construction (New York: Appleton-Century-Crofts, 1957), pp. 152–54.

>part II The Design of Business Research

158

Observation Environment Observation studies can be designed for the field or laboratory. In business research, field studies may take place at a customer’s home, the shopping environment, an employee work area (plant, office, distri- bution center), a supplier’s location, and more. Field studies, offering a natural setting, are most likely to obtain unaltered behavior, especially when the observer isn’t directly involved.

Laboratory studies are most likely to provide data protection. When specialized equipment is needed for observation (e.g., eye-tracking cameras, heart rate monitors, galvanic skin response machines, etc) laboratory settings are often the choice. We’ve had some success with employing eye-tracking via a subject’s laptop and tablet cameras. Laboratory settings are obviously more expensive, usually involve smaller sample sizes, and pose more difficulties in recruiting subjects. Laboratory observations can be part of an experimental design.

If you read or watch the news, you’ll know that urban areas have

had a sharp increase in questioned—if not questionable—police

actions with regard to excessive use of force, and bystander vid-

eos have played an increasing role in judging police actions. In

2015, the U.S. Supreme Court, in Graham v. Connor, held that an

officer’s actions, however, “must be judged from the perspective of

a reasonable officer, rather than with the 20/20 vision of hindsight.”

In an article in The Atlantic, Seth W. Stoughton, a law professor and

former police officer, and Geoffrey Alpert, a professor of criminol-

ogy, both at the University of South Carolina; along with Jeff Noble,

a police consultant based in Orange County, California, write, “The

aversion to what officers derisively refer to as “ ‘second-guessing’

. . . [makes] officers less receptive to a critique of their actions . .

. [and] makes them reluctant to provide their own complete and

honest critiques.” Yet nationwide, we’ve seen a demand for police

to change and for police decisions to be more transparent, result-

ing in a clamoring for use of police body and cruiser cameras.

Do you believe you get a true picture of an incident when you

see a body-mounted or dash-mounted video? Stoughton, who

also consults with law enforcement agencies, has choreographed

a series of videos to demonstrate the answer to this question. His

use parallels an observation study based on respondents watch-

ing video footage of mock police incidents. Using a series of

chest-mounted or dash-mounted cameras and bystander videos,

he shows just how difficult it is to arrive at an accurate conclusion

when using only a police-cruiser or body-cam video.

Chest-mounted cameras during interactions or pursuits often

create jerky movements and wildly distorted images. Stoughton

calls this “deceptive intensity”; it creates the impression that the

officer is under attack when he might not be. In an interception

incident video, using a dash-mounted camera involving a fleeing

suspect and Taser use by an officer, accuracy is related to van-

tage point. The body camera doesn’t reveal the use of a Taser

or the absence of a gun, while video shot by a bystander does.

“When video allows us to look through someone’s eyes, we

tend to adopt an interpretation that favors that person,” explains

Stoughton, a psychological phenomenon known as “camera

perspective bias.” Stoughton’s research also reveals that the

degree to which the viewer trusts or distrusts the police influ-

ences his or her video interpretation. So while the bystander

might not know any facts leading up to the incident, his or her

camera has its own bias. He concludes video evidence inter-

pretation depends, therefore, on perspective as well as bias.

So in observation research, should we consider the camera

as an accurate, unbiased observer? See for yourself. Check out

the videos on Connect.

>snapshot Observation and Police Cameras

©Aaron Roeth Photography

Connect resources enrich and engage. Cases, video, sample projects,

templates, appendices, and more.

“How-to” exhibits and Appendices help students DO research.

schi18939_FM_i-xxxii.indd 11 2/15/18 2:09 PM

xii

It’s serious business to revise a book that has been a world- wide leading text for more than three decades, one that’s published in eight international editions and is published in eight languages. The process of text writing and the speed with which revisions are developed don’t often per- mit a complete overhaul. But for this edition, given major changes in how students and professors use text material and how professors are teaching, as well as major changes in the research industry, a fresh approach was necessary. To address major industry changes and both professor and student needs, my McGraw-Hill team and I came up with a process and plan to deliver what students are seeking while giving professors what they need. The approach involved the following:

• Reflect changes in student learning and teacher pedagogy:

• Separate

• Streamline and simplify

• Clarify

• Reflect the current state of the research industry:

• Rethink everything

• Collaborate

Reflect Educational Changes

A deep dive into educational articles, as well as comments from our reviewers and numerous teaching colleagues and students, present a picture of some major changes in stu- dent learning and professors’ use of textbooks.

• An increasing percentage of students don’t buy books; some professors don’t require this of them based on the way text material is or is not used in the course.

• Students often enter classrooms without a founda- tion for that day’s discussion or activities.

• Professors see students disengaging from important tasks of the learning process: self-preparation and self-learning.

• Some professors—even educational institutions—are choosing to craft their own books, drawing only those chapters from one or multiple textbooks that are critical for their instructional approach.

• Customized books omit topics and tools deemed unnecessary for a given course based on its number of credit hours or the chosen pedagogy. Students revealed that omitted material often leaves them with voids affecting their understanding.

• Professors want chapters that stand alone, so they may order chapters however they choose to teach the material.

• Many professors want chapters to focus on the essen- tial material of the topic of each chapter, not have that material spread intermittently throughout the book.

• Professors want any text-embedded examples to enhance clarity of a concept.

• Students want unnecessary material eliminated; to them, “unnecessary” means background and history of a concept, not practical examples.

• Students believe that in a course like business research, they should be able to DO research after they learn, not just be able to describe it.

• Students are expected to apply what is learned in their research methods course in more advanced classes; their research methods text needs to serve as a reference manual.

This revision accomplishes the above in the following ways:

• Separate. This edition:

• Eliminated elements that artificially tie chapters together.

• Reinvented the use of elements that share cross- chapter features.

• Streamline and simplify.

• At the book level, this edition:

◦ Changed the number and order of chapters.

◦ Addressed writing and reading level by choos- ing more widely accepted vocabulary, rather than jargon, to describe and explain; shorten- ing sentences; and employing a more approach- able journalist style.

• At the chapter level, this edition:

◦ Moved some material from one chapter to another where it had a more logical fit.

◦ Reorganized material in each chapter to make it flow more logically (always putting A before B),

◦ Removed from every chapter material that might be “nice to know” but wasn’t “critical to know.”

◦ Assessed the value of every exhibit and exam- ple and made appropriate changes.

◦ Removed multiple terms for the same concept.

◦ Held the list of Key Terms to research-specific terms.

>preface

schi18939_FM_i-xxxii.indd 12 2/15/18 2:09 PM

>preface xiii

• Clarify.

• Students are often frustrated with textbooks; books define concepts using other concepts stu- dents don’t fully understand. This edition rede- fines dozens of key terms to remedy this problem.

• Students want examples to be relevant to them. This edition chose business research projects to profile by choosing behaviors, issues, or brands students might know or embrace.

• Students want to understand; a definition alone doesn’t achieve this. In this edition, any concept important enough to include has more than a definition.

• Students prefer one term for a single concept; they find multiple terms aggravating and confus- ing. This edition uses one term for one concept, with alternatives relegated to the glossary.

• Students want the key terms list for each chapter to be comprehensive but to exclude terms that aren’t critical. After conferring with researchers and professors, several concepts were removed from the key terms lists and the glossary, while others—having achieved more common use in the industry—were added.

• Students want, and need, to be able to DO research. This edition delivers:

◦ One chapter for basic foundational concepts.

◦ One chapter with everything needed to craft measurement questions.

◦ One chapter with everything needed to develop a measurement instrument.

◦ One chapter each for each of the data collec- tion methods.

◦ One chapter for preparing data for subsequent detailed analysis.

◦ One chapter for reporting research.

◦ “How-to” guides for difficult tasks and helpful tips for other tasks in exhibits and in several new appendices (Better Tables, Better Reports, Sample Computer-Based Questions and Guide- lines for Mobile Q.).

◦ A new chapter, An Integrated Example, that provides an insider’s view of a research project from management dilemma to research report.

• Students need visual cues to process and retain information:

◦ This edition provides 66 new photo visual cues to help them remember material.

◦ This text uses a series of exhibits linked to the research process; these use common shapes

and colors. Every exhibit in the 31-exhibit process series of exhibits has been subtly or substantially redesigned; the series includes two new exhibits.

◦ Concepts are now strongly linked to the five stages in the primary research process model.

• Students and faculty alike want ways to assess student understanding of material. This edition offers a new resource: Connect. Connect provides students computer-based assessment exercises, which encourage practice and provide instant feedback on mastery of material. Use of Connect improves understanding and recall. Connect pro- vides instructors student- and class-level analytics to improve subject, class, and course decisions.

Reflect the Research Industry

As in prior editions, the use of various interim GreenBook Research Industry Trends (GRIT) reports guided the research for this revision. The 2016 GRIT Report is based on the largest study of research suppliers and users ever conducted.

• Rethink. This edition:

• Makes a clear distinction between research and data analytics.

• Removed topics that are no longer relevant.

• Reduced coverage of topics showing waning importance.

• Enhanced coverage of topics the industry has embraced.

• Redesigned exhibits to reflect industry changes; of the more than 200 exhibits in this edition, 31 are completely new, and an additional 55 have been updated or redesigned.

• Collaborate. In an industry that is changing very quickly, any revision depends on people on the front lines of research. Hundreds of emails and conversa- tions, numerous books and articles, and almost 100 webinars and presentations have influenced this edition.

• Of the 56 Snapshots, PicProfiles, and Closeups featured in this edition, 79 percent are new (32) or updated (12). Topics in these rich research stories cover cyber security, prediction markets, sentiment analysis, why data analytics isn’t delivering results, programmatic ad buying, millennials and hous- ing, the art of asking questions, using interviews to define the management question, learning from Pixar to tell research stories, automated second- ary data searchers, agile research, performance management research, who’s taking surveys, digital

schi18939_FM_i-xxxii.indd 13 2/15/18 2:09 PM

xiv >preface

transformation, use of smartphones, eye tracking, observation with body cameras, experiments in employee health, use of robots, experimental labs, gaming, packaging redesign, question banks, sur- vey engagement, infographic reports, coding word data, data insights, finding best practices, presenta- tion venues, and much more.

• Discussion questions, especially those labeled From the Headlines, cover Chipotle’s reputation, BMW and electric cars, Uber software that excludes neighborhoods and buildings, shifting jobs to robots, airline safety, Delta’s reorganiza- tion of LAX, Dolby’s experiments with theater light and sound, Mercedes-Benz and self-driving cars, Walmart and Nabisco’s Oreo O’s cereal, Kohl’s department store and Apple Pay, perfor- mance-enhancing drugs in the workplace, Toyota and public confidence, and more.

Keep the Features Adopters Love

• Critical Core Content. The materials adopters have loved for decades are still the core of this edition. In an attempt to make the book more flexible to current instructional methodologies, we haven’t abandoned what made the book an industry leader.

• Strong Learning Objectives and Summaries. Every chapter has new learning objectives. The summaries are comprehensive, knowing sometimes these are the only material a student has time to read before class.

• Multipurpose Discussion Questions. These can serve as review for students, as testing exercises, or as options for vibrant class discussions because many reflect real-business situations.

• Versatile Appendices. End-of-chapter and end-of-text appendices for information that, given the differing skills and knowledge of their students, professors may want to emphasize or exclude. New appendices relate to building Better Tables and offering tips on Better Reports; to address mobile and other types of computer-delivered measurement instruments, there is the appendix Sample Computer-Based Questions and Guidelines for Mobile Q. We retained end-of-chap- ter appendices related to More Effective Measurement Questions and Calculate Sample Size.

Professors sometimes use writing a proposal as an end-of-term project or testing exercise. As a result, Appendix A has been rewritten with three exercises in mind: writing a formal proposal, creating an RFP, and assessing a proposal submitted in response to an RFP. Other end-of book appendices offer a professional focus group discussion guide (B), cover nonparamet- ric statistics (C), and provide statistical tables (D).

Use the Cloud 

We offer a comprehensive set of teaching and learning resources for Business Research Methods for faculty in Instructor Resources within Connect and for students at www.mhhe.com/Schindler13e. You’ll find the following:

• Written Cases. Cases offer an opportunity to tell research stories in more depth and detail. You’ll also find cases about hospital services, lotteries, data min- ing, fundraising, new promotions, and website design, among other topics, featuring organizations like Akron Children’s Hospital, Kelley Blue Book, Star- bucks, Yahoo!, the American Red Cross, and more.

• Video Cases and Supplements. New to this edition is a video supplement about an experiment in observa- tion using body cameras; it should be ideal for dis- cussing error in observation research. Additionally, several short segments drawn from a two-hour meta- phor elicitation technique (MET) interview should be invaluable in teaching students to conduct almost any type of individual depth interview and to explain the concept of researcher–participant rapport. Four of our video cases were written and produced especially to match the research process model and feature noted companies: Lexus, Starbucks, Wirthlin Worldwide (now Harris Interactive), Robert Wood Johnson Foundation, GMMB, Visa, Bank One, Team One Advertising, U.S. Tennis Association, Vigilante New York, and the Taylor Group.

• Data Files. If your course doesn’t involve a project where students collect their own data, use one of the cases here that contain data.

• Sample Student Project. Visualization of the finished deliverable is crucial to creating a strong research report, or critique this one.

• Appendices. You’ll find helpful appendices within Connect: Bibliographic Database Searches, Advanced Bibliographic Searches, Complex Experimen- tal Designs, Test Markets, and Pretesting Options and Discoveries.

• Articles, Samples, and Templates. Students often need to see how professionals do research to really understand the research process. You’ll find a sam- ple EyeTrackShop report, a Nielsen report of using U.S. Census data, an Excel template for generating sample data displays, and more.

• Multivariate Analysis: An Overview is a chapter for the benefit of graduate students who use Business Research Methods.

• Instructor’s Manual (instructors only).

• Web Exercises. Due to the ever-changing nature of web URLs, you’ll find these exercises here.

schi18939_FM_i-xxxii.indd 14 2/15/18 2:09 PM

>preface xv

• Written and Video Case Discussion Guides.

• Additional Business Research Examples for discussion.

• Test Bank (instructors only).

Collaborators

Research industry collaborators are the lifeblood of this textbook writer. The following people collaborated directly on this edition or connected me with those who did: Andy Peytchev, Research Triangle Institute; Bella Tumini, Suja; Betty Adamou, Research Through Gaming Ltd.; Cassan- dra McNeill, GutCheck; Colin McHattie, iTracks; Dan Weber, iTracks; Daniel Enson, Toluna; David Harris, Insight and Measurement; Denise D’Andrea, Focus Vision; Edwige Winans, Marcus Thomas LLC; Elaine Arkin, research consultant; Eric Lipp, Open Doors Organization; Ilan Hertz, SiSense; Jane Boutelle, Digsite; Jennifer Hirt- Marchand, Marcus Thomas LLC; Jessica Broome, Jessica Broome Research; Justin Ohanessian, Sticky; Kerry Hecht, Ramius; Lance Jones, Keynote Systems; Lenard Mur- phy, GreenBook; Lisa Whestone, Gutcheck; Malgorzata Kolling, OdinText; Mark Bunger, Forrester Research; Matt Marta, GutCheck; Monika Wingate, Digsite; Nicola Petty, Statistics Learning Centre; Patricio Pagani, InfoTools; Pete Cape, SSI; Rob Ramirez, Schlesinger Associates; Robert W. Kahle, author; Tom H.C. Anderson, Anderson Ana- lytics and OdinText; Sean Case, Research for Good; Seth Stoughton, University of South Carolina;Stuart Schear, Robert Wood Johnson Foundation; and Zoe Downing, Focus Vision.

The following are just a few of the people who offered me ideas for new concepts, Snapshots, PicProfiles, and Clo- seUps for this edition: Andrew McAfee, MIT; Carlo Ratti, Senseable City Lab, MIT; David Kiron, MIT-Sloan; Didier Bonnet, Capgemini Consulting; George Westerman, MIT Sloan; Glenn Kelman, author; John Cendroski, TIAA; Julia Smith, AIG; Kevin Lonnie, KL Communications; Lenard Murphy, GreenBook; Martin Lindstrom, author; Michael Benisch, Rocket Fuel Inc.; Michelle Shail, TIAA; Nick Drew, Fresh Intelligence; Pamela Kirk Prentice, MIT- Sloan; Ray Poynter, NewMR; Richard Cassidy, AlertLogic; Sam Ransbotham, MIT-Sloan; Warren Berger, researcher and author; and William Pink, Millward Brown.

And to all those research collaborators who have sug- gested ideas, collaborated on cases or past Snapshots, Closeups, or PicProfiles, and continue to discuss the research industry with me, I’m grateful. These individuals include: Rachel Sockut, Innerscope; Erica Cenci, Brady PR for OpinionLab; Olescia Hanson, The Container Store; Cynthia Clark, 1to1 Magazine; Betty Adamou, Research Through Gaming Ltd.; Debra Semans, Polaris Marketing Research; Keith Chrzan, Maritz Research Inc.; Michael Kemery, Maritz Research Inc.; Christian Bauer,

Daimler AG; Kai Blask, TNS Infratest; Melinda Gardner, Novation; Keith Phillips, SSI; Nels Wroe; SHL; Ephraim (Jeff) Bander, Eye Track-Shop; Ron Sellers, Grey Matter Research & Consulting; Guadalupe Pagalday, Qualvu.com; Sandra Klaunzler, TNS Infratest; Steve August, Revelation; Kathy Miller, GMI (Global Market Insite Inc.); Takayuki Nozoe, NTT Communications Corporation; Janeen Hazel, Luth Research; Christine Stricker, RealtyTrac; Stepha- nie Blakely, The Prosper Foundation; Jennifer Frighetto, Nielsen; Andy Pitched, Research Triangle Institute (RTI International); Jeffrey C. Adler, Centric DC Marketing Research; Josh Mendelssohn, Chadwick Martin Bailey Inc.; Ruth Stan, SIS International Research; Sharon Starr, IPC Inc.; Keith Crosley, Proofpoint; Christopher Schul- theiss, SuperLetter.com; Hy Mariampolski, QualiData Research Inc; Julie Grabarkewitz and Paul Herrera, American Heart Association; Holly Ripans, American Red Cross; Mike Bordner and Ajay Gupta, Bank One; Laurie Laurant Smith, Arielle Burgess, Jill Grech, David Lockwood, and Arthur Miller, Campbell-Ewald; Francie Turk, Consumer Connections; Tom Krouse, Donatos Pizza; Annie Burns and Aimee Seagal, GMMB; Laura Light and Steve Struhl, Harris Interactive; Emil Vicale, Herobuilders.com; Adrian Chiu, NetConversions; Colette Courtion, Starbucks; Mark Miller, Team One Advertising; Rebecca Conway, The Tay- lor Research Group; Scott Staniar, United States Tennis Association; Danny Robinson, Vigilante; Maury Giles, Wirthlin Worldwide; and Ken Mallon, Yahoo!.

To our faculty reviewers, your insights, aggravations, challenges, frustrations, suggestions, and disagreements were very helpful. These encouraged me to examine every word, every sentence, and every concept and see better, clearer ways to engage students in the subject we all love. Reviewers for this edition’s revision are: Ahmed Al-Asfour, Ogala Lakota College; Zara Ambadar, Carlow Univer- sity; Don Ashley, Wayland Baptist University; Kristopher Blanchard, Upper Iowa University; Cristanna Cook, Hus- son University; Charlene Dunfee, Capella University; Ernesto Gonzalez, Florida National University; Wendy Gradwohl, Wittenburg University; Pam Houston, Ogala Lakota College; Yan Jin, Elizabeth City State University; Abdullah Khan, Clafin University; Tracy Kramer, North Greenville University; Rex Moody, Angelo State Univer- sity; Jason Patalinghug, University of New Haven; Glen Philbrick, United Tribes Technical College; Denel Pierre, Shorter University; Pushkala Raman, Texas Woman’s University; Charles Richardson, Clafin University; Marcel Robles, Eastern Kentucky University; Angela Sandberg, Shorter University; Brian Satterlee, Liberty University; Jonathan Schultz, Amberton University; Stefano Tijerina, Husson University; Greg Turner, Clafin University; Sam VanHoose, Wayland Baptist University; Greg Warren, Wilmington University; Beyonka Wider, Claflin University; and Ron Zargarian, University of Indianapolis

schi18939_FM_i-xxxii.indd 15 2/15/18 2:09 PM

Prior edition reviewers included: Scott Bailey, Troy University; Scott Baker, Champlain College; Robert Balik, Western Michigan University–Kalamazoo; John A. Bal- lard, College of Mount St. Joseph; Jayanta Bandyopadhyay, Central Michigan University; Larry Banks, University of Phoenix; Caroll M. Belew, New Mexico Highlands Univer- sity; Kay Braguglia, Hampton University; Jim Brodzinski, College of Mount St. Joseph; Taggert Brooks, University of Wisconsin–La Crosse; Cheryl O’Meara Brown, University of West Georgia; L. Jay Burks, Lincoln University; Mar- cia Carter, University of Southern New Hampshire; Raul Chavez, Eastern Mennonite University; Darrell Cousert, University of Indianapolis; David Dorsett, Florida Insti- tute of Technology; Michael P. Dumler, Illinois State Uni- versity; Kathy Dye, Thomas More College; Don English, Texas A&M University–Commerce; Antonnia Espiritu, Hawaii Pacific University; Hamid Falatoon, University of Redlands; Judson Faurer, Metropolitan State College of Denver; Eve Fogarty, New Hampshire College; Bob Folden, Texas A&M University–Commerce; Gary Grudintski, San Diego State University; John Hanke, Eastern Washington University; Alan G. Heffner, Silver Lake College; Ron E. Holm, Cardinal Stritch University (Director of Distance Learning); Lee H. Igel, New York University; Burt Kaliski, New Hampshire College; Jane Legacy, Southern New Hampshire University; Andrew Luna, State University of West Georgia; Andrew Lynch, Southern New Hampshire University; Iraj Mahdvi, National University; Warren Mat- thews, LeTourmeau University; Erika Matulich, University of Tampa; Judith McKnew, Clemson University; Rosema- rie Reynolds, Embry Riddle Aero University–Daytona; Randi L. Sims, Nova Southeastern University; Gary Stark, Northern Michigan University; Bruce Strom, University

of Indianapolis; Cecelia Tempomi, Southwest Texas State University; Gary Tucker, Northwestern Oklahoma State University; Marjolijn Vandervelde, Davenport University; Charles Warren, Salem State College; Dennis G. Weis, Alliant International University; Robert Wheatley, Troy University; Bill Wresch, University of Wisconsin–Oshkosh; and Robert Wright, University of Illinois at Springfield; and Ken Zula, Keystone College.

This revision incorporates the feedback of dozens of stu- dents who identified areas of confusion so that this edition could make concepts more understandable, who partici- pated in search tests, who worked on numerous research projects demonstrating where the book needed to include more information, and who provided reminders with their questions and actions that some aspects of the research process operate below their learning radar.

Through this 13th edition, I hope you and your students discover, or rediscover, how stimulating, challenging, fasci- nating, and sometimes frustrating this world of research- supported decision making can be.

Pamela Schindler

Many thanks to my McGraw-Hill team; without your assistance this revision wouldn’t have happened so smoothly: Chuck Synovec, Director; Noelle Bathurst, Portfolio Manager; Ryan McAndrews, Product Developer; Erika Jordan, Core Project Manager; Harper Christopher, Executive Market- ing Manager; David W. Hash, Designer; Daryl Horrocks, Program Manager; Sue Nodine, copyeditor; Elizabeth Kelly, proofreader; and Angela Norris, Senior Assessment Project Manager.

xvi >preface

schi18939_FM_i-xxxii.indd 16 2/15/18 2:09 PM

xvii

In its 13th edition, all chapters within Business Research Methods have been evaluated for currency and accuracy. Revisions were made to accommodate new information and trends in the industry, changing teaching pedagogy, and information about what teachers and students are looking for in their textbooks.

• The book’s chapter structure is slimmer and has been changed to reflect how teachers are teaching research and using the book; the book now has 17 chapters.

• A foundations chapter replaces the first three chapters.

• The chapters on data preparation and examina- tion have merged.

• A Research Reports chapter now merges informa- tion on oral and written reports, with empha- sis on oral reports to better reflect industry practice.

• A new Chapter 17, An Integrated Example, now provides an insider’s perspective of a research project. This example applies text practices and theory to one example from management dilemma to research report.

• Material has been reorganized to tie better to the modified research process model; there are now five parts, each part a match to a stage in the model. Part I, contains three chapters and establishes the foundations for what follows. Part II contains five chapters, all focused on research design and its vari- ous methodologies. Part III contains four chapters, all related to data collection and preparation. Part IV contains two chapters related to data analysis. Part V contains one chapter on research reporting. The part structure was designed to better reflect the research process as it is currently managed.

• Ethical issues are discussed, with their possible solu- tions, in every chapter, rather than in a stand-alone chapter, to reflect how teachers are using this material.

• Every section and every word has been examined for concept clarity and better student understanding; whole sections and whole chapters and appendices have been reconceived and rewritten.

• An emphasis has been placed on indicating solutions for problems or possible error sources, not just indi- cating or describing the errors/problems.

• Based on student feedback, an emphasis has been placed on providing sufficient information to “do” research, not just learn about research. Exhibits have been added to reflect how to execute a particular

practice, facilitating the experiential approach to teaching and learning business research methods.

• For clarity and to match a chapter’s new structure, numerous Exhibits are new (38), have been revised significantly (34), or have been slightly modified (8).

• Continuing examples no longer weave throughout the text; chapters can now be assigned in different order to fit any teaching pedagogy.

• Images (58) have been added or replaced, giving a visual cue for new Snapshots, PicProfiles, or new embedded examples.

• To reflect industry practices, the series of exhibits that reflect the research process and that are used as con- ceptual “thought flowcharts”—especially valuable for visual learners—have been reenvisioned and redesigned; new exhibits have been added to this process series.

• The Cases section contains an updated case-by- chapter-suggested-use chart.

• Continuing to provide rich examples from the research industry, 30 new Snapshots, five new PicProfiles, and two new Closeups have been added; two Closeups have been updated.

• Several new chapter-level appendices have been added to this edition: Better Reports (Chapter 16), Better Tables (Chapter 13), Sample Computer-Based Questions and Guidelines for Mobile Q (Chapter 11), and Sources of Measurement Questions (Chapter 11).

• The Glossary has been updated; 77 new terms ref- lect changes in industry practices and 27 additional terms were upgraded to key term status.

• The Instructor’s Manual contains additional research examples for discussion or testing.

• McGraw-Hill Connect® has been added to the book’s resources; Connect provides opportunities for both formative and summative assessment by providing students regular and consistent feedback, encourag- ing practice, and enabling them to move closer to mastery by improving understanding and recall. In- structors are provided student and class analytics, to improve teaching decision making. Assignable mate- rial within Connect for this edition includes multiple choice questions for homework for each chapter and test bank questions for online testing.

• SmartBook®, also assignable in Connect, is a digital ver- sion of our textbook that actively tailors that content to an individual student’s needs. It helps a student focus on the things they don’t know, helps them retain key con- cepts, is accessible on the go, and tracks student progress.

>detailedchangestothisedition

schi18939_FM_i-xxxii.indd 17 2/15/18 2:09 PM

• Student Resources/Faculty Resources within Con- nect contain new materials (sources, videos, exam- ples) and video showcasing an observation experiment using body cameras.

• Test Bank has been updated to reflect changes in content and organization.

• PowerPoint slide decks have been updated to reflect changes in content and organization.

For Each of the Chapters A detailed listing of chapter-by-chapter changes is provided here for your convenience.

• Chapter 1 This chapter was completely rewritten and has a new focus: the fundamentals or critical concepts students need to understand the remainder of the book. It combines material from 12e Chapters 1 and 3, with elements from Chapter 2. The follow- ing elements are new to this edition: chapter-opening quote (William Pink), the learning objectives and summary, a PicProfile on emerging trends based on the latest GRIT report, three new Snapshots (Big vs. Small Data, Research on Cyber Security, and Identify- ing and Defining Constructs), two revised exhibits, multiple images as visual cues, new embedded ex- amples related to Hobby Lobby and Siemens AG, one new key term (data blending), four new photos serving as visual cues, and new discussion questions. Several sections have been pulled and others moved to chapters with a better fit. Six snapshots and three exhibits have moved to the IM.

• Chapter 2 Previously Chapter 4, this chapter fea- tures a restricted and simplified research process exhibit and a new structure based on five stages of the research process, with material on propos- ing research moving to Chapter 3. The following elements are new to this edition: chapter-opening quote (Brad Smith, Microsoft), learning objectives and summary, three new sections (identifying and prioritizing dilemmas, research project timeframe, and ethical issues and responsibilities) and five restructured sections, a PicProfile on emerging trends in research design, a Snapshot (Research and Programmatic Algorithms), a revised snapshot on out- sourcing research, a new exhibit on Gantt chart of research project, four new images, five new key terms [key performance indicators (KPIs), dashboards, findings, insights, recommendations], and a new From-the-Headlines discussion question. The What is Good Research and Ethical Issues sections moved to this chapter from Chapters 1 and 3, respectively. Detailed management-research question hierarchy section was moved to Chapter 3. CPM chart was moved from Chapter 6.

• Chapter 3 Previously Chapter 5, this chapter is restructured and focuses on stage 1 of the research process: the management-research question hierar- chy and exploration to include valuing and budgeting research. The following elements are new to this edi- tion: chapter-opening quote (Warren Berger, author), learning objectives and summary, eight new sections, one modified and four new exhibits, three new Snap- shots (Housing and Millennials, The Art of Asking the Right Question, Using Interviews to Refine the Man- agement Question), eight key terms relocated from other chapters, and six new images as visual cues. The section on data mining was dropped to reinforce Chapter 1’s emphasis on research and data mining as different courses. Several key terms have moved to other chapters to reflect relocation of certain mate- rial. Several Snapshots have moved to the IM.

• Chapter 4 Previously Chapter 6, the emphasis of this chapter has changed to research design once the research question(s) and investigative questions have been determined and the decisions involved in research design, including those involved in sampling design, have been made. The following elements are new to this edition: chapter-opening quote (Nick Drew, Fresh Intelligence), learning objectives and summary, a new section (sampling design), two new Snapshots (TIAA Performance Management Overhaul, AIG and Research Design), a new CloseUp (How Agile Research Helped Prove the Value in a Packaging Redesign), two modified exhibits, embedded example on productivity and morale, five new images as vis ual cues, one new key term (single-methodology design), and modified discussion questions—including a new From-the-Headlines discussion question. Detailed sections on causation moved to Chapter 8, and focus groups moved to Chapter 6. Two Snapshots moved to the IM.

• Chapter 5 Previously Chapter 14, this chapter has a different structure based on the six steps of sam- pling design. The following elements are new to this edition: chapter-opening quote (Gerald Earl Gillum, American rapper and producer), learning objectives and summary, three new sections (sampling design, selection and recruiting protocols, and ethical issues and their solutions), one new snapshot (Who’s Tak- ing Your Surveys), one revised PicProfile on mixed access sampling, four new exhibits, four revised or modified exhibits, four new images as visual cues, and anew From-the-Headlines discussion ques- tion. Two key terms were moved here (case, target population).

• Chapter 6 Previously Chapter 7, the following ele- ments are new to this edition: chapter-opening quote (Gia Calhoun, Burt’s Bees), learning objectives and

xviii >detailed changes to this edition

schi18939_FM_i-xxxii.indd 18 2/15/18 2:09 PM

>detailed changes to this edition xix

summary, two new sections (qualitative sampling design, including incentivizing participants and interviewers as consultants, and ethical issues and their solutions) and one enhanced section (creative exercises), three new Snapshots (Digital Transfor- mation Revealed Using IDIs, IDIs Help Restructure Maritz Travel, Qualitative Research in the Era of Smartphones), two revised exhibits, three new im- ages as visual cues, and seven new key terms (data saturation, collage, completion/fill in the blank, role playing, creative innovation roleplay, story- telling, write a letter). Four Snapshots and PicProfiles moved to the IM.

• Chapter 7 Previously Chapter 8, this chapter has been reorganized to follow the observation research design steps. The following elements are new to this edition: chapter-opening quote (Katie Hafner, author), learning objectives and summary, two new sections (sampling design, ethical issues and their solutions), two new Snapshots [Visual Content Gets Sticky, Observation and Police Cameras (with video on the website)], three revised or modified exhibits, six new images as visual cues, two new key terms (memory decay, selective filtering), and a new From- the-Headlines discussion question. Several Snapshots and a PicProfile have moved to the IM.

• Chapter 8 Previously Chapter 9, this chapter has been reorganized, with evaluation of experiments moving toward the end of the chapter. The following elements are new to this edition: chapter-opening quote (Jeff Bezos, CEO, Amazon), learning objec- tives and summary, a new section (ethical issues and their solutions), a relocated section on causation (including two exhibits), four snapshots (Experi- ments in Improving Employee Health, Robotic Experi- ments, Zeotap Experiments with Mercedes-Benz, MIT SENSEable City Lab), two revised exhibits, four new images as visual cues, 14 new key terms (debriefing, after-only design, group time series design, history, instrumentation, maturation, nonequivalent control group design, one group pretest-posttest design, posttest-only control group design, selection, sepa- rate sample pretest-posttest design, static group com- parison design, regression toward the mean, testing), and a From-the-Headlines discussion question. Four Snapshots moved to IM.

• Chapter 9 Previously Chapter 10, this chapter has been reorganized. The following elements are new to this edition: chapter-opening quote (David Goldberg, CEO, SurveyMonkey), learning objectives and sum- mary, four new sections (classification of data col- lection designs, telephone survey trends, evaluation of survey design, and ethical issues and their solu- tions), two new Snapshots (Internet Brings Prediction

Research into 21st Century, Research Embraces the Smartphone), two new PicProfiles on emerging trends in survey research and declining response rates, two new and three revised exhibits, four new images as visual cues, updated statistics, three new key terms (acquiescence bias, probe, social desirabil- ity bias), and a From-the-Headlines discussion ques- tion. Two additional exhibits moved to this chapter (informed consent and IRB process).

• Chapter 10 Previously Chapter 11, The following elements are new to this edition: chapter-opening quote (David McCandless, author), one new exhibit, two revised exhibits, one new Snapshot (The Emo- tional Face of Research), four images as visual cues, and a From-the-Headlines discussion question.

• Chapter 11 Previously Chapter 12, this chapter is reorganized and rewritten to focus on measurement questions, rather than the scales on which they are based, in order to work better with the chapter on measurement instruments. The following elements are new to this edition: chapter-opening quote (David F. Harris, president, Insight and Measure- ment), learning objectives and summary, two new sections (instrument design, prepare the preliminary analysis plan) and one revised section (data entry), a new Snapshot (Toluna and Voss Measure Water) and a new PicProfile about Urban Dictionary, four new and three revised or modified exhibits, two ad- ditional exhibits moved from other chapters, some material on coding moved from another chapter, five new images as visual cues, five new key terms (attitude scaling, checklist, error of strictness, inter- view guide, scaling) and 17 key terms moved here from other chapters, modified discussion questions (including a From-the-Headlines question), and two new appendices (sample computer-based questions by scale type, sources of measurement questions).

• Chapter 12 Previously Chapter 13, this chapter has a new structure to work better with the chapter on Measurement Questions, with a stronger link to the preliminary analysis plan. The following ele- ments are new to this edition: chapter-opening quote (Kristin Luck, research consultant), learning objec- tives and summary, three new sections (instrument design, physical design, nonquestion elements), one new Snapshot (New Vehicle Survey), five new and three revised exhibits, four new images as visual cues, 15 new key terms (assimilation effect, behav- ior cycle, behavior frequency, behavior time frame, completion estimate, contrast effect, filter question, in- strument coverage, instrument scope, interview guide, measurement instrument, rapport, skip directions, skip logic diagram, social desirability bias), revised discussion questions, and a new From-the-Headline

schi18939_FM_i-xxxii.indd 19 2/15/18 2:09 PM

question. The Invoke PicProfile was moved to the IM.

• Chapter 13 As a merger of 12e Chapters 15 and 16, this chapter has a new structure and new con- tent. The following elements are new to this edition: chapter-opening quote (Dana Zuber, director of ana- lytics for Butler, Shine, Stern & Partners), learning objectives and summary, one new section (collect the data) and a revised section (coding), six revised and three new exhibits, two new Snapshots (How Might You Code Word Data, The Difference Between Data and Insight), updated statistics, 17 new key terms (coding scheme, context units, cross-tabulation, data collection, data validation, inter-rater reliability, intra- rater reliability, listwise deletion, data missing at random (MAR), data missing but not missing at ran- dom (NMAR), data missing completely at random (MCAR), pairwise deletion, predictive replacement, recoding, recording units, sampling units, survey activation), revised discussion questions, and new chapter appendix (Better Tables). Four Snapshots, a CloseUp, and a PicProfile were moved to the IM.

• Chapter 14 Previously Chapter 17, the following elements are new to this edition: one revised and one new exhibit, and three new images as visual cues.

• Chapter 15 Previously Chapter 18, the following elements are new to this edition: chapter-opening quote (Jeff Bezos, CEO, Amazon), one new Snap- shot, and three new images as visual cues. One exhibit (grammar and style proofreader results) was moved to the IM.

• Chapter 16 As a merger of Chapters 19 and 20, this chapter has a new structure emphasizing the

oral presentation. The following elements are new to this edition: chapter-opening quote (David McCand- less, British data journalist, information designer, and author), learning objectives and summary, four new sections (audience-centric planning, visualiza- tion specifically for the oral report, infographics, ethical considerations in reporting), a new Snapshot (Hitting the Wall is a Good Thing), a new CloseUp (Storytelling from Pixar Applied to Research), five new and 10 revised exhibits, an infographic image, 21 new key terms (predispositions, confirmation bias, anchoring bias, conformity bias, survivorship bias, loss-aversion bias, visualize, data clarity, actionable insights, audience-centric planning, data-centric planning, desired audience effect, graph, informa- tion, insights, limitations, report framework, report structure, table, tone, geography), revised discus- sion questions, and a new chapter appendix—Better Reports—with five new exhibits and five existing ex- hibits from prior chapters. Two items (constructing a story and overcoming the jitters) were moved to the IM.

• Chapter 17 This new chapter, An Integrated Example, provides an insider’s perspective of a research proj- ect. This example applies text practices and theory to one example from management dilemma to research report. The companies, Visionary Insights and BrainSavvy, might be fictional, but the research profiled in the example is very real. This chapter can be used throughout the course to review (or test) various concepts, or at the end of the course as the basis for a lively discussion or final exam.

xx >detailed changes to this edition

schi18939_FM_i-xxxii.indd 20 2/15/18 2:09 PM

xxi

>additionalresources

There is a wealth of information, samples, templates, and more within Connect for instructors, and at www.mhhe .com/Schindler13e for students.

Written Cases. Cases offer an opportunity to tell research stories in more depth and detail. You’ll find a new case, Marcus Thomas LLC Tests Hypothesis for Troy-Bilt Creative Development, complete with its online questionnaire, at the Online Learning Center. You’ll also find cases about hospital services, lotteries, data mining, fundraising, new promotions, and website design, among other topics, featuring organizations like Akron Children’s Hospital, Kelley Blue Book, Star- bucks, Yahoo!, the American Red Cross, and more.

Video Cases. We are pleased to continue to make available a first in video supplements: several short segments drawn from a two-hour metaphor elicitation technique (MET) interview. These segments should be invaluable in teaching students to conduct almost any type of individual depth interview and to explain the concept of researcher–participant rapport. Four of

our video cases were written and produced especially to match the research process model in this text and feature noted companies: Lexus, Starbucks, Wirthlin Worldwide (now Harris Interactive), Robert Wood Johnson Foundation, GMMB, Visa, Bank One, Team One Advertising, U.S. Tennis Association, Vigilante New York, and the Taylor Group.

Web Exercises. It is appropriate to do web searches as part of a research methods course, so each chapter offers one or more exercises to stimulate your students to hone their searching skills. Due to the ever-changing nature of web URLs, however, we offer these exercises in the Instructor’s Manual.

Articles, Samples, and Templates. Students often need to see how professionals do things to really understand, so you’ll find a sample EyeTrackShop report, a Nielsen report of using U.S. Census data, an Excel template for generating sample data displays, and more.

Sample Student Project. Visualization of the finished deliverable is crucial to creating a strong research report.

schi18939_FM_i-xxxii.indd 21 2/15/18 2:09 PM

▪ Connect content is authored by the world’s best subject matter experts, and is available to your class through a simple and intuitive interface.

▪ The Connect eBook makes it easy for students to access their reading material on smartphones and tablets. They can study on the go and don’t need internet access to use the eBook as a reference, with full functionality.

▪ Multimedia content such as videos, simulations, and games drive student engagement and critical thinking skills.

©McGraw-Hill Education

▪ Connect’s assignments help students contextualize what they’ve learned through application, so they can better understand the material and think critically.

▪ Connect will create a personalized study path customized to individual student needs through SmartBook®.

▪ SmartBook helps students study more efficiently by delivering an interactive reading experience through adaptive highlighting and review.

McGraw-Hill Connect® is a highly reliable, easy-to- use homework and learning management solution that utilizes learning science and award-winning adaptive tools to improve student results.

73% of instructors who use Connect

require it; instructor satisfaction increases by 28% when Connect

is required.

Homework and Adaptive Learning

Quality Content and Learning Resources

Over 7 billion questions have been answered, making McGraw-Hill

Education products more intelligent, reliable, and precise.

Using Connect improves retention rates by 19.8 percentage points, passing rates by 12.7 percentage points, and exam

scores by 9.1 percentage points.

schi18939_FM_i-xxxii.indd 22 2/15/18 2:09 PM

More students earn As and Bs when they

use Connect.

www.mheducation.com/connect

©Hero Images/Getty Images

▪ Connect Insight® generates easy-to-read reports on individual students, the class as a whole, and on specific assignments.

▪ The Connect Insight dashboard delivers data on performance, study behavior, and effort. Instructors can quickly identify students who struggle and focus on material that the class has yet to master.

▪ Connect automatically grades assignments and quizzes, providing easy-to-read reports on individual and class performance.

▪ Connect integrates with your LMS to provide single sign-on and automatic syncing of grades. Integration with Blackboard®, D2L®, and Canvas also provides automatic syncing of the course calendar and assignment-level linking.

▪ Connect offers comprehensive service, support, and training throughout every phase of your implementation.

▪ If you’re looking for some guidance on how to use Connect, or want to learn tips and tricks from super users, you can find tutorials as you work. Our Digital Faculty Consultants and Student Ambassadors offer insight into how to achieve the results you want with Connect.

Trusted Service and Support

Robust Analytics and Reporting

schi18939_FM_i-xxxii.indd 23 2/15/18 2:09 PM

xxiv

Preface xii

>part I Building the Foundation for Research 1

1 Research Foundations and Fundamentals 2

2 The Research Process: An Overview 26

3 Stage 1: Clarify the Research Question 45

>part II The Design of Business Research 69

4 Stage 2: Research Design, An Overview 70

5 Stage 2: Sampling Design 85

Appendix: Calculate the Sample Size 114

6 Stage 2: Data Collection Design: Qualitative Research 122

7 Stage 2: Data Collection Design: Observation Research 148

8 Stage 2: Data Collection Design: Experiments 170

9 Stage 2: Data Collection Design: Survey Research 194

>part III Measurement 225

10 Stage 3: Measurement Foundations 226

11 Stage 3: Measurement Questions 246

Appendix: Sample Computer-Based Questions and Guidelines for Mobile Q 283

Appendix: Sources of Measurement Questions 286

Appendix: More on Effective Measurement Questions 287

12 Stage 3: Measurement Instruments 293

>part IV Collect, Prepare and Examine the Data 319

13 Stage 3: Collect, Prepare, and Examine Data 320

Appendix: Better Tables 349

Appendix: Describing Data Statistically 352

>part V Analyze and Interpret Data 357

14 Stage 4: Hypothesis Testing 358

15 Stage 4: Measures of Association 395

>part VI Stage 5: Report the Research 431

16 Stage 5: Research Reports: Supported Insights and Recommendations 432

Appendix: Better Reports 465

>part VII Research Project Overview 475

17 An Integrated Example 476

>briefcontents

schi18939_FM_i-xxxii.indd 24 2/26/18 12:26 PM

>case index 507

>appendices 517 a Business Research Proposals and RFPs

(with Sample RFP) 518

b Focus Group Discussion Guide 538

c Nonparametric Significance Tests 540

d Selected Statistical Tables 547

References  558

Glossary 577

Indexes 592

>brief contents xxv

schi18939_FM_i-xxxii.indd 25 2/15/18 2:09 PM

xxvi

>contents

Preface xii

>part I Building the Foundation for Research 1

1 Research Foundations and Fundamentals 2

The Role and Process of Research 3

Research versus Data Analytics 3

The Research Process 7

The Language of Research 9

Concepts 9

Constructs 10

Operational Definitions 11

Variables 12

Hypotheses, Theories, and Models 16

>summary 22

>key terms 23

>discussion questions 24

>cases 25

2 The Research Process: An Overview 26

What Is Good Research? 27

The Research Process 27

Stage 1: Clarify the Research Question 28

Identify and Prioritize Dilemmas 29

Exploration 31

Value and Budget the Research 32

Stage 2: Design the Research Project 32

Data 32

Data Collection Design  35

Sampling Design 36

Stage 3: Collect and Prepare the Data 36

Collect the Data 36

Prepare the Data 37

Stage 4: Analyze and Interpret the Data 37

Stage 5: Report Insights and Recommendations 38

The Research Project Time Frame 38

Research Process Pitfalls and Solutions 39

Ill-Defined Management Problems 39

Unresearchable Questions 40

Overembracing Data Analytics 40

The Manager’s Hidden Agenda(s) 40

Favored-Technique Syndrome 40

Researcher Inexperience 40

Research Process Ethical Issues and Responsibilities 42

>summary 42

>key terms 43

>discussion questions 43

>cases 44

3 Stage 1: Clarify the Research Question 45

Stage 1 in the Research Process 46

The Question Hierarchy 46

The Management Question 48

The Research Question 50

Investigative Questions 51

Measurement Questions 51

The Exploration Strategy 52

Information Sources 53

Internal Sources  61

Evaluating Internal Sources 62

Value the Research 63

Justify and Budget for Research 64

Stage 1 Ethical Issues and Their Solutions 65

>summary 65

>key terms 66

>discussion questions 67

>cases 68

>additional content 68

schi18939_FM_i-xxxii.indd 26 2/15/18 2:09 PM

>contents xxvii

>part II The Design of Business Research 69

4 Stage 2: Research Design, An Overview 70

What Is Research Design? 71

Sampling Design 72

Data Collection Design 73

An Example 73

Design Dimensions 74

Design the Measurement Instrument 80

>summary 82

>key terms 83

>discussion questions 83

>cases 84

5 Stage 2: Sampling Design 85 Sampling Design 86

Define the Target Population and Case 87

Define the Population Parameters 89

Define the Sampling Frame 90

Define the Number of Cases 91

Sample versus Census  91

Lower Cost 93

Greater Speed of Data Collection  93

Availability of Population Cases 94

Better Quality Results 94

Accuracy 94

Precision 95

Sample Size 95

Define the Sampling Method 96

Probability Sampling 97

Simple Random Sampling 97

Complex Probability Sampling 98

Nonprobability Sampling 106

Convenience 106

Purposive Sampling 107

Snowball 108

Define the Selection and Recruiting Protocols 108

Ethical Issues and Their Solutions 110

>summary 111

>key terms 112

>discussion questions 112

>cases 113

Appendix: Calculate the Sample Size 114

6 Stage 2: Data Collection Design: Qualitative Research 122

What Is Qualitative Research? 124

Qualitative Research Design 126

Qualitative Sampling Design 127

Qualitative Data Collection Design 129

Interviews 129

Interviewers as Consultants 131

Creative Exercises 133

Individual Depth Interviews 134

Group Interviews 137

Focus Groups 139

Group Interview Drawbacks and Their Solutions  141

Recording, Analyzing, and Reporting Group Interviews 141

Combining Qualitative Methodologies 142

Case Study 142

Action Research 143

Merging Qualitative and Quantitative Methodologies 143

Ethical Issues and Their Solutions 144

>summary 145

>key terms 145

>discussion questions 146

>cases 147

7 Stage 2: Data Collection Design: Observation Research 148

Observation Research Design 151

Sampling Design 152

Data Collection Design 153

Content Type 153

Types of Observation 154

Observation Environment 158

schi18939_FM_i-xxxii.indd 27 2/15/18 2:09 PM

xxviii >contents

Data Collection Protocol 158

Observer Presence  161

Observer Concealment  161

Observer Involvement  162

Observer Training and Supervision  163

Evaluation of the Observation Method 163

Strengths 163

Limitations  166

Ethical Issues and Their Solutions 166

>summary 167

>key terms 168

>discussion questions 168

>cases 169

8 Stage 2: Data Collection Design: Experiments 170

Experiments and Causation 171

Causation Evidence 171

Causal Conclusion 171

Conducting an Experiment 174

Select Relevant Variables 175

Specify Treatment Levels 177

Control the Experimental Environment 177

Choose the Experimental Design 178

Select and Assign Cases 178

Pilot Test, Revise, and Pretest 180

Analyze the Data 181

Validity in Experimentation 181

Internal Validity 181

External Validity 183

Experimental Research Designs 184

Preexperimental Designs 184

True Experimental Designs 186

Field Experiments: Quasi- or Semi-Experiments 187

An Evaluation of Experiments 189

Advantages 189

Disadvantages 190

Ethical Issues and Their Solutions 190

>summary 190

>key terms 191

>discussion questions 192

>cases 193

>additional content 193

9 Stage 2: Data Collection Design: Survey Research 194

Data Collection Design: The Survey 196

Self-Administered Survey 202

Survey via Telephone Interview 207

Survey via Personal Interview 211

Error in Survey Research 213

Interviewer Error 213

Participant Error 215

Outsourcing Survey Tasks 217

Evaluation of Survey Research 218

Ethical Issues and Their Solutions 218

>summary 221

>key terms 222

>discussion questions 222

>cases 223

>part III Measurement 225

10 Stage 3: Measurement Foundations 226

The Nature of Measurement 227

Tasks of Measurement 227

What Is Measured 229

Measurement Scales 229

Nominal Scales 230

Ordinal Scales 232

Interval Scales 233

Ratio Scales 233

Recoding 234

Sources of Measurement Differences 235

Error Sources 235

The Characteristics of Good Measurement 237

Validity 237

Reliability 239

Practicality 242

>summary 243

>key terms 244

>discussion questions 244

>cases 245

schi18939_FM_i-xxxii.indd 28 2/15/18 2:09 PM

>contents xxix

11 Stage 3: Measurement Questions 246

Instrument Design 247

Prepare the Preliminary Analysis Plan 247

Confirm Communication Approach 249

Select Instrument Structure 249

Select or Craft a Measurement Scale 251

Research Objectives 252

Response Types 252

Number of Dimensions 252

Balanced or Unbalanced 253

Forced or Unforced Choices 253

Number of Scale Points 253

Participant Errors 254

The Nature of Attitudes 255

The Relationship between Attitudes and Behavior 256

Attitude Scaling 257

Rating Questions 257

Simple Category Questions 257

Summated Rating Questions 261

Semantic Differential Questions 263

Numerical/Multiple Rating List Questions 267

Stapel Scale Questions 267

Constant-Sum Questions 267

Graphic Rating Questions 268

Ranking Questions 268

Sorting Questions 271

Cumulative Questions 272

Find or Craft Measurement Questions 273

Question Coverage 273

Question Wording 273

Frame of Reference 274

Response Alternatives 275

Sources of Existing Questions  277

Pretesting 278

>summary 278

>key terms 280

>discussion questions 280

>cases 282

Appendix: Sample Computer- Based Questions and Guidelines for Mobile Q 283

Appendix: Sources of Measurement Questions 286

Appendix: More on Effective Measurement Questions 287

12 Stage 3: Measurement Instruments 293

Instrument Design 294

Phase 2: Gather and Sort the Pieces 294

Question Types 296

Pretesting: Scope and Coverage 299

Responsibility Fulfillment 299

Phase 2: Non-Question Elements 299

Non-Question Elements 302

Overcoming Instrument Problems 304

Responsibility Fulfillment 305

Phase 3: Organize the Pieces 305

Order Topics and Questions 306

Pretesting: The Instrument 311

Physical Design 311

Pretesting: Physical Design 314

Responsibility Fulfillment 314

>summary 315

>key terms 316

>discussion questions 316

>cases 318

>additional content 318

>part IV Collect, Prepare and Examine the Data 319

13 Stage 3: Collect, Prepare, and Examine Data 320

Collect the Data 322

Train the Data Collectors 322

Determine the Data Collection Timeline 322

Determine and Implement Instrument Disposition 322

Invite the Chosen Participants 322

Activate the Survey 323

schi18939_FM_i-xxxii.indd 29 2/15/18 2:09 PM

xxx >contents

Remind the Participants 323

Enter the Data 323

Enter the Data 323

Prepare the Data 325

Postcollection Coding of Data 325

Edit the Data 329

Complete 329

Accurate 331

Appropriately Coded 332

Examine the Data 332

Exploratory Data Analysis 332

Cross-Tabulation 341

The Importance of Percentages 342

Other Table-Based Analysis 343

>summary 345

>key terms 346

>discussion questions 347

>cases 348

Appendix: Better Tables 349

Appendix: Describing Data Statistically 352

>part V Analyze and Interpret Data 357

14 Stage 4: Hypothesis Testing 358 Introduction 359

Statistical Significance 359

The Logic of Hypothesis Testing 361

Type I Error 362

Type II Error 365

Statistical Testing Procedures 367

Probability Values (p Values) 367

Tests of Significance 368

Types of Tests 368

How to Select a Test 370

Selecting Tests Using the Choice Criteria 371

One-Sample Tests 372

Two-Independent-Samples Tests 375

Two-Related-Samples Tests 378

k-Independent-Samples Tests 382

k-Related-Samples Tests 389

>summary 391

>key terms 392

>discussion questions 393

>cases 394

15 Stage 4: Measures of Association 395

Introduction 396

Bivariate Correlation Analysis 396

Pearson’s Product Moment Coefficient r 396

Scatterplots for Exploring Relationships 398

The Assumptions of r 401

Computation and Testing of r 401

Simple Linear Regression 407

The Basic Model 408

Concept Application 408

Method of Least Squares 411

Predictions 413

Testing the Goodness of Fit 415

Nonparametric Measures of Association 418

Measures for Nominal Data 418

Measures for Ordinal Data 422

>summary 427

>key terms 428

>discussion questions 428

>cases 430

>part VI Stage 5: Report the Research 431

16 Stage 5: Research Reports: Supported Insights and Recommendations 432

Introduction 433

Emphasis on Insights 434

Shift in Reporting Structure 435

Audience-Centric Planning 436

schi18939_FM_i-xxxii.indd 30 2/15/18 2:09 PM

>contents xxxi

Audience Analysis 436

Audience Composition 437

Research Question Knowledge 437

Research Predispositions 438

Recommendation Effects 438

Audience Effect 438

Supplemental Purposes 439

Report Structure 439

Content and Style: Organize 440

Report Components 441

Traditional Frameworks for Organization 443

Content and Style: Visualize 444

Presenting Findings 449

Visualization Specifically for the Oral Report 456

Content and Style: Compile 457

Infographics 459

Content and Style: Practice 459

Content and Style: Deliver 460

Ethical Considerations in Reporting 460

>summary 461

>key terms 462

>discussion questions 463

>cases 464

Appendix: Better Reports 465

>part VII Research Project Overview 475

17 An Integrated Example 476 The Need for Research 477

Receipt of an RFP at Visionary Insights  477

The VI Proposal for BrainSavvy Research 479

The Proposal Review at BrainSavvy  479

The BrainSavvy Finalist Interview  479

Visionary Insights Captures the Research Contract 482

Stage 1: Clarify the Research Question 482

The Dallas Trip 482

VI Project Structure and Exploration 483

Interviews 484

Data File Analysis 484

Process Modeling 486

Focus Groups 486

Investigative and Research Questions 486

Stage 2: Research Design 488

Sampling Design 488

Data Collection Design 488

Stage 3: Data Collection, Preparation, and Examination  492

Activate the Survey 492

The Analysis Plan  492

Data Preparation and Examination 493

Stage 4: Analyze the Data 493

Conference Call with Parma the Following Tuesday 493

Stage 5: Report the Research 494

Oral Report 495

Post-Project Team Debriefing 495

The Continuation Contract 496

Three and a Half Months Later 496

>summary 506

>discussion questions 506

>case index 507

>appendices 517

a Business Research Proposals and RFPs (with Sample RFP) 518

Proposing Research 518

The Research Proposal 518

Proposal Structure 520

Executive Summary 520

Problem Statement 520

Research Objectives 522

Literature Review 522

Importance/Benefits of the Study 522

Research Design 522

Data Analysis 523

Nature and Form of Results 523

Qualifications of Researchers 523

Budget 524

Schedule 525

schi18939_FM_i-xxxii.indd 31 2/15/18 2:09 PM

xxxii >contents

Facilities and Special Resources 525

Project Management 526

Bibliography 526

Appendices 526

Creating a Request for Proposal 527

Proposal Administration 527

Summary Problem Statement 527

Technical Section 528

Management Section 529

Contracts and Licensing Section 529

Pricing Section 529

Format 530

Qualify a Researcher to Receive the RFP 530

Evaluate a Research Company’s Proposal 530

Evaluating an Unsolicited Research Proposal 532

Sample RFP: Covering Kids 534

Background 534

Contractual Needs 535

Evaluation Task 535

Anticipated Time Schedule 536

Proposal Instructions 536

Review Criteria 537

b Focus Group Discussion Guide 538 Background 538

c Nonparametric Significance Tests 540

One-Sample Test 540

Kolmogorov-Smirnov Test 540

Two-Samples Tests 541

k-Samples Tests 545

d Selected Statistical Tables 547

References 558

Glossary 577

Indexes

Name 592

Company 595

Subject 598

schi18939_FM_i-xxxii.indd 32 2/19/18 11:24 AM

Building the Foundation for Research

Chapter 1 Research Foundations and Fundamentals

Chapter 2 The Research Process: An Overview

Chapter 3 Stage 1: Clarify the Research Question

>part I

schi18939_ch01_001-025.indd 1 2/14/18 2:37 PM

LO1-1   How business research and data analytics complement each other.

LO1-2 The language of professional researchers.

>learningobjectives After reading this chapter, you should understand . . .

“As big data increases, we see a

parallel growth in the need for ‘small data’

to answer the questions it raises.”

William C. Pink,

senior partner

Creative Analytics

Research Foundations and Fundamentals

>chapter 1

schi18939_ch01_001-025.indd 2 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

3

>The Role and Process of Research Every manager in the course of his or her career, regardless of his or her field, will make thousands of decisions of various types: strategic, tactical, and procedural. Each decision starts with a problem or opportunity—a dilemma. A strategic decision determines the general approach; a tactical one, a method for executing the strategic decision; and a procedural one, the specifics for executing the tactical deci- sion. For example, brick-and-mortar retailers have been having significant difficulties as more people have embraced smartphones. Increasingly, people are bypassing stores to shop online (dilemma). Hobby Lobby, however, is thriving. This retailer chooses to serve the crafting market (strategic) and emphasizes supplying the needs of painters, photographers, jewelry makers, quilters, floral designers, and interior decorators (strategic). To serve photographers, it carries matting boards, tools for cutting such board, and frames, but not cameras (tactical). It might identify each new product for its photography area by reviewing an in-person, profit-based pitch (procedural). The company is family-owned (strategic). It makes its decisions guided by values emphasizing strengthening family (strategic). As a result, its brick-and-mortar stores are open fewer hours (tactical), and its full-time people are paid three times the minimum wage (tactical).1 Researchers also make these kinds of decisions, deciding, for example, to use a communication study (strategic) and choosing a mobile survey (tactical) with participants recruited by posting an invitation on a company’s Facebook page (procedural). Today, the pressure on managers to justify their decisions, in an effort to guarantee a return on the investment of the resources (people, money, time, equipment, facilities) that each decision requires, is enormous. The journey from dilemma to decision uses information as fuel.

Companies have always collected data. Each organization is not equally adept, however, at using those data to develop meaningful information and insights useful for making good decisions. Over the last decade, some organizations have used newly available tools (faster computing power, better data analytic software) to tap into data it has already collected, data that have been languishing in departmen- tal silos or company data warehouses. What once was a pool of data has become a veritable ocean of data. Some firms are drowning. Others are barely staying afloat. Still others are grasping the opportunity to use this ocean of data as a foundation for their strategic direction and gain competitive advantage.2 Those companies that have been successful have found fuel for their decisions. But fuel comes in differ- ent grades; think regular gasoline verses rocket fuel. Using only historic data to make a current decision is one approach to decision making; enriching that fuel mixture by collecting new data specific to a given dilemma, is another.

The field you are about to study is in the midst of upheaval and disruption.3 For almost a century, researchers have been viewed by managers as technical support. These specialists were brought in on projects when technical expertise in research methodology and data analysis were needed. But, within the last few years, that has been changing. New technology, new and better computing tools (artificial intelligence, virtual reality, better mobile equipment, Internet of Things), and even a new computing environment (the cloud) are adding to industry chaos. In the midst of this sea change in business, new pressures are being put on the researcher. It is no longer acceptable to merely add to the data pool; busi- ness managers need clearly communicated insights from each new data addition. Researchers are now expected to not only be technologically competent, but to have an understanding of how businesses and organizations work. And managers, who once delegated research projects to specialists, are expected to be conversant with research methodologies and tools. Welcome to the new world of research, where researchers are data storytellers and insight providers, critical to helping provide strategic and tactical direction.

Research versus Data Analytics Facing each new dilemma, it is the manager’s decision whether he or she has sufficient information— drawn from data previously collected, either internal or external to the firm—or needs more information to make an appropriate decision. Managers draw on data from existing internal data sources (called a decision support system) when engaging in data analytics. For example, Amazon, in an attempt to increase our order, mines its data to provide us with a list of products that others—who bought what we

schi18939_ch01_001-025.indd 3 2/14/18 2:37 PM

>part I Building the Foundation for Research

4

According to the latest report from MIT/Sloan Management

Review and SAS, data analytics is not living up to its hype. The

report classifies analytic users on three levels of maturity: ana-

lytical innovators (who apply analytics strategically), analytical

practitioners (who apply analytics operationally), and the ana-

lytically challenged (who rely more on management experience

than analytics to make decisions). Analytical innovators—those

benefiting the most from the application of analytics (both the

extraction of insights and their dissemination to affect organiza-

tional actions)—have, rather than growing in the last four years,

basically remained stagnant. They propose several reasons for

this, including a lack of senior management commitment and a

focus on operational rather than strategic use of data.

Some examples of analytical innovators, however, give us

role models for the practice. Bank of England (BoE) is an ana-

lytic innovator; to fulfill its regulatory role in the British economy,

it is aggregating datasets, both microeconomic and macroeco-

nomic, for the first time. BoE has “hired a chief data officer, cre-

ated a data lab, established an advanced analytics group and

formed a bank-wide data community.” General Electric, also an

analytic innovator, created a new business unit, as well as a huge

software division, to manage a cloud-based platform that aggre-

gates and analyzes sensor data from industrial machines. “GE’s

strategy for data and analytics has become tightly linked to its

corporate strategy, a tremendous corporate shift for what was

once a traditional manufacturing conglomerate.”

Analytics Under-delivers Compared to its Hype

>snapshot

are ordering—also bought. Such data is often referred to as big data due to the extensive size of many of these databases. Exhibit 1-1 provides some ideas for sources.

When existing data is mined, it may be used for a purpose other than that for which it was originally in- tended. In its mobility division, Siemens AG, the engineering powerhouse with almost 350,000 employees in 200 countries,4 builds systems into its trains. These systems generate more than 1 billion data points5 per train per year. They are used to track that train’s performance and maintenance activity and to learn from any train malfunctions or accidents. Collecting train data to understand and improve that train’s performance employs repetitive, ongoing observation research. However, drawing insights from data accumulated from European and U.S. trains to help design new rolling stock for China and Russia employs data analytics.

In another example, customer relationship management (CRM) software may initially be used to facilitate the sales process and improve the effectiveness of sales appeals to various customer groups.

State of Data Analytics

10%

49%

41%

Analytical Innovators

Senior management drives an analytics culture; strategic use of data Higher levels of data management & analytic skills

Focus on prediction and prescription

Analytical Practitioners

Working to become data driven; operational use Have ‘just good enough’ data; more of what they need. Use complex analytics with some prediction

Analytically Challenged

Rely more on management experience than data Lack appropriate data and analytic skills Simple analytics; mostly descriptive; focus on cost reduction

Source: Sam Ransbotham, David Kiron, and Pamela Kirk Prentice, “Beyond the Hype: The Hard Work Behind Analytic Success,” MIT/Sloan Management Review,

April 2016, downloaded April 29, 2016 (http://marketing.mitsmr.com/PDF/57381-MITSMR-SAS-Analytics2016.pdf?utm_source=WhatCounts%2c+Publicaster+Edition&

utm_medium=email&utm_campaign=darpt16&utm_content=Download+the+Report+%28PDF%29&cid=1).

schi18939_ch01_001-025.indd 4 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

5

Additionally, a call center is designed to answer questions, provide technical support, or funnel prospects into the sales process; many calls to a call center are recorded to improve performance. Using data analyt- ics, a firm might mine these two datasets to extract insights that help design a new customer landing page for the firm’s website. Businesses are getting better at data blending,6 combining data from separate data files (e.g., financial, human resources [HR], CRM, inventory management, and manufacturing) into a new composite data file, and then querying that composite data file to help make decisions. While the informa- tion that comes from data blending has an important role in decision making, it is not the same as research.

Assume for the moment that you are the manager of a full-service restaurant. You are experiencing significant turnover in your server pool, and some long-time customers have commented that the friendly at- mosphere, which has historically drawn them to your door, is changing. Where will you begin to try to solve this problem? Your business intelligence system is designed to provide ongoing information about events and trends in the technological, economic, political-legal, demographic, cultural/social, and competitive arenas (see Exhibit 1-2). It reveals that wait-staff turnover is high in your industry, regulations on restaurant opera- tions have become more stringent, and some area competitors are experimenting with increasing wait-staff wages while eliminating tips. You also review your firm’s financial records and HR records to determine pay, tips, pre-hire experience, and work hours of those who left and compare that information with those who stayed. Is this sufficient information, or is this a problem for which additional research should be used?

>Exhibit 1-1 Where Business Collects Information

Type of Data Where/How Data Source

Transactional Online and in-store purchases Customer Online, phone, in-store inquiries Potential customer, customer Warehouse and shipping manifests Logistic partners, employee Machine performance Machine data log

Observational Online Web visits and in-store shopping trips Customer, employee Competitor interactions Customer Click-through paths on Web Potential customer, customer In-store customer service interactions Customer, employee Stock price valuations Investors Biometric measures (e.g., neuromarketing, fMRI, PET, eye tracking)

Potential customer, customer, employee

Conversational (Touch points)

Surveys, online and in-store intercepts Potential customer, customer, employee Call center interactions Customer, employee In-store customer service interactions Customer, employee Web chat interactions Customer, employee In-store checkout Customer, employee Candidate interviews Potential employee Performance reviews Employee Exit interviews Employee Annual stockholder meetings Investor Financial performance presentations Financial analyst, institutional investor Listening tours Customer, supplier, logistic partner, employee,

decision influencer Twitter posts Customer, employee, competitor, trade

associations, distributor Facebook posts (company site) Customer, employee, trade associations, distributor Blog activity Customer, employee, competitor, trade

associations, distributor Other social media posts or discussions Customer, employee, competitor, trade

associations, distributor

Internet Analytics Keyword searches Potential customer, customer Click analysis Potential customer, customer Google+ Potential customer, customer

Our interviews and research for this edition revealed several sources of research data. This table is adapted from that research and author experience as well as from material by Cyn-

thia Clark, “5 Ways to Learn What Customers Aren’t Telling You,”1to1 Magazine, March 5, 2012, accessed March 8, 2012 (http://www.1to1media.com/view.aspx?docid=33464); and

“Harness the Conversation: Business in Today’s Social World,”Cvent, accessed March 8, 2012 (http://www.cvent.com/en/sem/business-in-todays-social-worldsurvey-ebook.shtml).

schi18939_ch01_001-025.indd 5 2/14/18 2:37 PM

>part I Building the Foundation for Research

6

>Exhibit 1-2 Some Sources of Business Intelligence

Website of agency or department

Speeches by elected o�cials

Recordings of public proceedings

Press releases or press events

Syndicated studies

Government reports

Business research

Government reports

Literature searches

Presentations at conferences

Literature searches

Syndicated industry studies

Press releases or press events

Website

Records of public proceedings

Clipping services

Business research

Patent filings

Syndicated industry studies

Literature search

Presentations at conferences

Press releases or press events

Clipping services

Websites

Syndicated studies

Public opinion organizations

Business research

Government reports

Government/ Regulatory

Demographic Business

Intelligence

Cultural & Social

Economic

Technological

Competitive

In his book, Small Data: The Tiny Clues That Uncover Huge

Trends, author Martin Lindstrom talks about the importance of

knowing why. Lindstrom isn’t an advocate of only using big data,

indicating big data lacks insight because it focuses on analysis

rather than emotional connection. His book focuses on what

he’s learned as he has visited or lived in more than 2,000 homes

throughout the world and how those ethnographic observations

paid big dividends. In a Knowledge@Wharton interview, Lind-

strom described financially troubled Danish toymaker Lego. In

2002, Lego had ventured away from its core small blocks, instead

emphasizing movies, theme parks, apparel, and large building

blocks (based on big data about Millennials) only to discover—via

interviews and ethnographic observations in homes across Eu-

rope—that it was taking away the major reason children play with

the toy: the sense of accomplishment. In explanation, Lindstrom

writes, “children attain social currency among their peers by play-

ing and achieving a level of mastery at their chosen skill.” Lego re-

focused on the small blocks based on its collection of small data.

These actions brought Lego back from near bankruptcy. “You

have to remember that Big Data is all about analyzing the past,

but it has nothing to do with the future. Small Data, . . . seemingly

insignificant observations you identify in consumers’ homes, is . . .

the emotional DNA we leave behind.”

Sources: Martin Lindstrom, Small Data: The Tiny Clues That Uncover Huge Trends,

St. Martin's Press (February 23, 2016), pp 1-2; and “Why Small Data Is the New

Big Data,” Knowledge@Wharton, March 24, 2016. Downloaded March 25, 2016

(http://adage.com/article/special-report-4as-conference/tipping-point-j/303268/?utm

_source=daily_email&utm_medium=newsletter&utm_campaign=adage&ttl

=1459463433).

Big versus Small Data

>snapshot

©Cr-Management GmbH & Co. KG/Getty Images

schi18939_ch01_001-025.indd 6 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

7

Perhaps you are the head of your state’s department of transportation, charged with determining which roads and bridges will be resurfaced or replaced in the next fiscal year. You have data on which roads and bridges handle the most traffic, as well as those roads/bridges representing the greatest eco- nomic disaster if closed. However, the state’s manager of public information has expressed concern about the potential for public outcry if work is once again directed to more affluent regions of the state. The manager suggests using new research to assist in making your decision because the decision is one with numerous operational, financial, and public relations ramifications. Should you authorize new research?

The Research Process A deep dive into historical data rarely illuminates the ‘why’ behind actions. And the whys change over time; what was true a year ago might not be true today. To fulfill this new role of insight provider, you’ll need an understanding of both the process and the tools used by a researcher. Business research is de- fined as a systematic inquiry that provides information to guide a specified managerial decision. More specifically, it is a set of processes that include planning, acquiring, analyzing, and reporting relevant data, information, and insights to decision makers in ways that mobilize the organization to take appro- priate actions. These actions are designed to maximize performance and help accomplish organizational goals. Typically, the overall process is divided into the following stages:

1. Clarify the research question. 2. Design the research. 3. Collect and prepare the data. 4. Analyze and interpret the data.

5. Report insights and recommendations.

Exhibit 1-3 provides a graphic of the process that we will develop in this text. At times, a manager may start his or her journey at the beginning and proceed stage-by-stage to its culmination. At other times, a

>Exhibit 1-3 The Research Process

Report Insights and Recommendations

ExplorationExploration

Analyze and Interpret the Data

Design the Research

Clarify the Research Question

Collect and Prepare Data

Data Collection Design

Sampling Design

Stage 1

Stage 2

Stage 3

Stage 4

Stage 5

schi18939_ch01_001-025.indd 7 2/14/18 2:37 PM

>part I Building the Foundation for Research

8

manager may need only a portion of the process, given information that is available from a variety of other sources. Research is often characterized by much smaller datasets than big data. Once the research is pre- sented, the manager has one very important decision: How shall he or she resolve the management problem?

Research and the Scientific Method

Intelligent, curious people who have a driving need to seek answers are at the heart of great research. The foundation of the business research process is the scientific method. The essential tenets of the scientific method are:

• Clearly defined concepts, constructs, variables, methods, and procedures.

• Empirically testable hypotheses: a way exists to gather evidence that directly supports/refutes any hypothesis.

How does research keep an organization secure from criminal

“bad actors” in the cyber security arena? It’s used to spot the

threat before it happens, to understand an organization’s vulner-

abilities, to spot the attack venues, and more.

Over the last decade, cyber attacks have become more fre-

quent, more sophisticated, more complex, and easier for the bad

actors, all at the same time. New digital tools make it possible

for these criminals to do tasks that just recently would require

sophisticated programming expertise. Today, they can purchase

whatever tools they need from criminal networks using Bitcoin,

a digital currency that makes tracking the purchase and find-

ing the criminal very difficult. As Richard Cassidy, cyber security

evangelist with Alert Logic, one of the nation’s leading managed

security providers, explains, “Companies are vulnerable from

three types of bad actors. Not all pose the same degree of dam-

age.” Hacktivists have a political or social agenda, and garnering

media attention is their goal; cyber criminals may also want the

media attention, but they seek monetary gain from the data they

capture; advanced persistent threats (APTs) are the most dan-

gerous and spend significant time, money, and resources prior

to crafting a target-specific attack and do so for significant mon-

etary gain and/or damage to the target.

From research, Alert Logic discovered that passwords (49.9

percent) and email addresses (45.5 percent) remain the prize

target of bad actors, along with usernames (37.7 percent) and

names (29.4 percent). Using this stolen information in combina-

tion with information readily available on company websites and

social media sites like LinkedIn, a bad actor can obtain access to

servers, databases, web domains, and more.

Attacks via phishing emails and software application

plug-ins are the chief avenues of access to deliver malware.

A phishing email is disguised to appear as though from a

trusted source—for example, another employee or your boss.

A plug-in is a software component that adds a specific feature

to an existing computer program. Employees today bring their

own mobile devices to work and often use software that hasn’t

been rigorously evaluated for inappropriate plug-ins. Through

these portals, malware (malicious software such as viruses,

worms, Trojan horses, and spyware) can be injected into an

organization’s system.

“Bad actors often try out their approach before launching

the real attack,” claimed Cassidy. So catching them in this pre-

liminary test should be the goal. The problem is that it takes an

organization, on average, 205 days to identify that it has been

compromised, and by then, the purpose of the attack has been

accomplished.

www.alertlogic.com

Source: Richard Cassidy, “Behind The Scenes: Cybercrime Threat Landscape,” Brightalk

webcast, April 27, 2016, downloaded May 26, 2016 (https://www.brighttalk.com/webcast/11

587/201299?autoclick=true&utm_medium=web&utm_source=brighttalk-promoted&utm

_campaign=player-page-feed&utm_content=promoted).

Research on Cyber Security

>snapshot

©solarseven/Shutterstock

schi18939_ch01_001-025.indd 8 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

9

• Direct observation of phenomena (facts).

• Conclusions drawn from statistical evidence rather than inferred justification (educated guesses).

• The self-correcting process: ability to replicate and reassess validity of conclusions.

>The Language of Research So where do we start to understand the preceding material? We start with the language of research. When we do research, we seek to know “what is” in order to understand, explain, or predict phenom- ena. We might want to answer the question “What will be the department employees’ reaction to a new flexible work schedule?” or “Why did the stock market price surge higher when all normal indicators suggested it would go down?” When dealing with such questions, we must agree on definitions. Which employees of the department: clerical or professional? What reaction? What are normal indicators? These questions require the use of concepts, constructs, operational definitions, and variables.

Concepts Concepts are used to understand and communicate information. The success of research hinges on (1) how clearly we conceptualize and (2) how well others understand the concepts we use. We design hypotheses using concepts. We devise measurement scales using concepts by which to test these hypoth- eses. We gather and analyze data using measurement concepts.

A concept is a generally accepted collection of meanings or characteristics associated with certain events, objects, conditions, situations, or behaviors:

• Concepts are created when we classify and categorize events, objects, conditions, situations, or behaviors—identifying common characteristics beyond any single observation.

• Concepts are acquired through personal experience or the experience of others.

• Concepts use words as labels to designate them; these words are derived from our experiences.

• Concepts have progressive levels of abstraction—that is, the degree to which the concept does or does not have something objective to refer to. At one extreme are objective concepts; at the other, abstractions. Table is an objective concept. We have images of tables in our mind. Personality is an abstract concept as it is much more difficult to visualize.

Think of a movie ticket as a concept. What comes to mind is not a single example, but your collected memories of all movie tickets from which you define a set of specific and definable characteristics (mate- rial, movie title use, perforation, multiple parts, screen location, etc.). For another example, assume you see a man passing and identify that he is running rather than walking, skipping, crawling, or hopping. Each movement represents a different concept. We also use concepts to identify that the moving object is an adult male rather than a truck or a horse.

Ordinary concepts make up the bulk of communication in research. Ordinary, however, does not mean unambiguous. We might, for example, ask research participants for an estimate of their family’s total income. Income may seem to be a simple, unambiguous concept, but we will receive varying an- swers and confusing data unless we restrict or narrow the concept by specifying:

• Time period, such as weekly, monthly, or annually.

• Before or after income taxes are deducted.

• For head of household only or for all household members.

• For salary and wages only or also including tips, bonuses, dividends, interest, and capital gains.

• To include or not include in-kind income (e.g., free rent, employee discounts, vacations, or food stamps).

We run into difficulty when trying to deal with less ordinary phenomena or advance new ideas. One way to handle this problem is to borrow a concept from another language or from another field. As- sume we are researching a brand logo’s design strength. We can borrow the term gestalt from German, which translates as form or shape and means an organized whole more than the sum of its parts.7 Or we

schi18939_ch01_001-025.indd 9 2/14/18 2:37 PM

>part I Building the Foundation for Research

10

might use the physics concept of gravitation to explain why people shop where they do or the geographic concept of distance to describe degree of variability between the attitudes of employees on a new work schedule.

Sometimes we need to make a word cover a different concept or develop new labels for a concept. When we adopt new meanings or develop new labels, we begin to develop a specialized language or jar- gon. While jargon contributes to efficiency of communication among specialists or a particular group, it excludes everyone else. Jargon is often avoided in business research for this reason unless the sample is very narrowly defined.

Constructs When research requires us to work with abstract concepts, we define one or more constructs. A construct is an abstract idea specifically invented for a given research and/or theory-building purpose. We build constructs by combining simpler, more concrete concepts, especially when the idea or image we intend to convey is not subject to direct observation. Consider this example: Heather is a human resource analyst at CadSoft, an architectural software company that employs technical writers to write product manuals, and she is analyzing task attributes of a job in need of redesign.

Exhibit 1-4 illustrates some of the concepts and constructs Heather is dealing with. The concepts at the bottom of the exhibit (format accuracy, manuscript errors, and keyboarding speed) define a con- struct that Heather calls “presentation quality.” Presentation quality is not directly observable. It is an invented construct, used to communicate the combination of meanings presented by the three objective, measurable concepts that Heather has discovered are related empirically. She is able to observe key- boarding speed, for example, by timing a person’s entry of a paragraph.

Concepts at the next higher level of abstraction in Exhibit 1-4 are vocabulary, syntax, and spelling. Heather also finds them to be related. They form a construct that she calls “language skill.” She has chosen this label because the three concepts together define the language requirement in the job descrip- tion. Language skill is placed at a higher level of abstraction in the exhibit because two of the concepts it comprises, vocabulary and syntax, are more difficult to observe and their measures are more complex.

>Exhibit 1-4 Constructs Composed of Concepts in a Job Redesign

“Job Interest Construct” (Components unknown by analyst)

“Language Skill Construct”

“Presentation Quality Construct”

Most abstract

Most concrete

L e

ve l o

f a b

st ra

ct io

n Vocabulary

Syntax

Spelling

Manuscript errors

Format accuracy

Keyboarding speed

schi18939_ch01_001-025.indd 10 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

11

Heather has not yet defined the last construct, “job interest.” It is the least observable and the most dif- ficult to measure. It will likely be composed of numerous concepts—many of which will be quite abstract. Highly abstract constructs can be inferred only from the data; these are presumed to exist but must await further testing and definition. Heather will have the beginning of a conceptual scheme if research shows the concepts and constructs in this example to be interrelated and if their connections can be supported. In graphic form, the conceptual scheme depicts the relationships among the knowledge and skill require- ments necessary to clarify the job redesign effort.

Operational Definitions Confusion about the meaning of constructs or concepts can destroy a research study’s value without the knowledge of the researcher or its sponsor. Definitions are one way to reduce this danger.

Researchers distinguish between dictionary definitions and operational definitions. In the more fa- miliar dictionary definition, a concept is defined with a synonym. For example, a customer is defined as a patron; a patron, in turn, is defined as a customer or client of an establishment; a client is defined as one who employs the services of any organization or a patron of any shop.8 These circular definitions may be adequate for general communication but not for research. In research, we measure concepts and constructs, and this requires more rigorous operational definitions.

An operational definition is a definition stated in terms of specific criteria for measurement or test- ing. We must be able to count, measure, or in some other way gather the information through our senses. Whether the object to be defined is physical (e.g., a can of soup) or highly abstract (e.g., achieve- ment motivation), the definition must specify the characteristics and how they are to be observed. The specifications and procedures must be so clear that any competent person using them would classify the object in the same way.

To do this, you need operational definitions. Operational definitions may vary, depending on your purpose and the way you choose to measure them. College undergraduates are grouped by class. No one has much trouble understanding such terms as senior, junior, sophomore, and so forth. But the task may not be that simple if you must determine which students comprise each class. Here are two different situations involving a survey among students where we want to classify their answers by their class level. Each uses a different definition of the same concept:

1. You ask them to report their class status and you record it. In this case, class is freshman, sopho- more, junior, or senior; you accept the answer each respondent gives as correct. The operational definition for class: how the student themselves classify their class.

2. You ask them to define their class by registrar guidelines. The operational definition for class: semester hours of credit completed by the end of the prior spring semester and recorded in each student’s record in the registrar’s office:

• Freshman Fewer than 30 hours’ credit • Sophomore 30 to 59 hours’ credit • Junior 60 to 89 hours’ credit • Senior 90 or more hours’ credit

These examples deal with relatively concrete concepts, but operational definitions are even more critical for constructs. Suppose one tries to measure a construct called “socialization.” Are we referring to someone’s level of activity on Facebook and other social media, whether they like to entertain, or whether they can converse with others? We would probably develop questions on access, posts, sharing, likes, etc., in the first instance; the number of times and the types of entertaining they do for the second; and the number, types of language used, and types of reactions received for the third as we grapple with the operational definition. We may use a measurement approach already developed and validated by other researchers or create our own. The measurement approach chosen operationally defines the construct.

We may need to provide operational definitions for only a few critical concepts or constructs, but these will almost always be the definitions used to develop the relationships found in hypotheses and theories.

schi18939_ch01_001-025.indd 11 2/14/18 2:37 PM

>part I Building the Foundation for Research

12

Variables A variable is a measurable symbol of an event, act, characteristic, trait, or attribute.9 In practice, one or more variables are used as a substitute for a concept or construct. As researchers are interested in rela- tionships among concepts and constructs, researchers are interested in relationships among variables. Variables come in various types: independent, dependent, moderating, and extraneous (including con- trol, confounding, and intervening). We’ll expand this list when we get into the concept of measurement.

Independent and Dependent Variables

The dependent variable (DV) is of primary interest to the researcher; it is measured, predicted, or other- wise monitored and is expected to be affected by manipulation of an independent variable (IV), another variable of primary interest. In each relationship, there is at least one independent variable (IV) and one dependent variable (DV). As one writer notes:

Researchers hypothesize relationships of independence and dependence: They invent them, and then they try by real- ity testing to see if the relationships actually work out that way.10

The assignment of the variable type (dependent vs. independent) depends on the assumed relation- ship the researcher is studying. If you were interested in studying the impact of the length of the working week on productivity, you would make the length of working week the IV. If you were focusing on the relationship between age of the worker and productivity, then age would be the IV. Exhibit 1-5 lists some terms that have become synonyms for independent variable and dependent variable. Although it is easy to establish whether an IV influences a DV, it is much harder to show that the relationship between an IV and DV is a causal relationship.

Exhibit 1-6 summarizes the many variable types, while Exhibit 1-7 graphically shows their relation- ship to each other using an example. Researchers recognize that there are often several independent variables that might be studied and that they are probably at least somewhat related and, therefore, not

When you read in the business press or attend a business

presentation, you are exposed to many constructs. If you

don’t struggle to define these, you miss out on a lot of what

the article is telling you. Let’s see how good you are at spot-

ting these.

During the most recent Ad Age Digital Conference, in a

panel discussion on major issues advertisers and marketers

know they have but aren’t ready to confront—the elephants

in the room—Lou Paskalis, senior vice president of enterprise

media at Bank of America shared, “We’re optimizing advertising

and people want storytelling. People are ad blocking because

they don’t like what we are doing.” What are the constructs in

this statement? What might be the operational definition of these

constructs? What is(are) the management problem(s) confronting

advertisers and their agencies?

The panel also revealed there are challenges beyond digital

advertising’s oft-cited viewability, ad blocking, and transparency

(which refers to kickbacks and rebates from media to advertising

agencies). “The [industry trade associations] can’t agree on what

transparency is and, clearly, they are not speaking the same

language. And that leads to a lack of trust,” shared Ron Amram,

vice president of media at Heineken USA. What are the constructs

and operational definitions for these constructs? What is the

management problem(s) confronting advertisers and advertising

media?

Paskalis also shared, “We need to go back to what [the ad-

vertising] business is about. We need to ask, ‘Is this compelling?’

Clients need to step up and pay full value for things that are

great, but we need to stop commoditizing everything.” What

are the constructs and the operational definitions for these con-

structs? What is the management problem revealed for advertis-

ers and advertising agencies?

Source: George Slefo, “Digital Conference: Marketers Chime in on Elephants in the

Room, Known and Unknown,” Ad Age, April 05, 2016, downloaded April 6, 2016

(http://adage.com/article/special-report-digital-conference/marketers-chime-elephants

- d i g i t a l / 3 0 3 4 1 3 / ? u t m _ s o u r c e = d a i l y _ e m a i l & u t m _ m e d i u m = n e w s l e t t e r & u t m

_campaign=adage&ttl=1460499287).

Identifying and Defining Constructs

>snapshot

schi18939_ch01_001-025.indd 12 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

13

independent among themselves. When we study simple relationships, all other variables are considered unimportant and are ignored by the researcher. But for more complex relationships, researchers must identify and measure those variables.

In Exhibit 1-7a, a causal relationship is illustrated by an arrow pointing from the independent variable to the dependent variable.

Moderating Variables

A moderating variable (MV) is a second independent variable believed to have a significant contributory effect on the original IV–DV relationship. For example, one might hypothesize that in an office situation:

The introduction of a four-day working week (IV) will lead to higher productivity (DV), especially among younger workers (MV).

In Exhibit 1-7a, the arrow pointing from the moderating variable to the arrow between the IV and DV shows the difference between an IV directly affecting the DV and an MV affecting the relationship between an IV and the DV. In this case, the researcher hypothesizes that a different relationship between the four-day week and productivity might result from age differences among the workers. Hence, after introduction of a four-day working week, the productivity gain for younger workers is compared with that for older workers. For example, let’s assume that the productivity of younger workers is 12 percentage points higher than that for older workers before the introduction of the four-day working week. Assume that the productivity of all workers having a four-day working week is six percentage points higher than for those of workers having a five-day working week. If the productivity of a younger worker having a four-day working week is only 18 percentage points higher than the productivity of an older worker, there is no moderating effect (12 + 6 = 18), because the 18 percentage points are the sum of the main effects and the moderating effect should show a surplus. However, if the productivity of younger workers was 25 percentage points higher, then the moderating effect of a worker’s age would be obvious.

Other Extraneous Variables

An almost infinite number of extraneous variables (EVs) exist that might conceivably affect a given relationship. Taking the example of the effect of the four-day working week again, one would normally

>Exhibit 1-5 Independent and Dependent Variables: Synonyms

Independent Variable Dependent Variable

Predictor Criterion

Presumed cause Presumed effect

Stimulus Response

Predicted from ... Predicted to ...

Antecedent Consequence

Manipulated Measured outcome

>Exhibit 1-6 A Summary of Variable Types

Variable Type Symbol Presumed Effect on IV-DV Relationship Action Needed

Dependent DV Concept/construct of interest Measure

Independent IV Primary variable believed to have significant effect on DV Manipulate

Moderating MV Alternative IV; possible significant contributory effect on IV-DV Measure

Control CV Might influence the IV-DV, but effect is not at the core of the problem studied

Ignore; effect is randomized

Confounding CFV Alternative IV; unknown effect on IV-DV Measure

Intervening IVV Theoretically might affect; effect can’t be determined Infer effect from IV and MV on DV

schi18939_ch01_001-025.indd 13 2/14/18 2:37 PM

>part I Building the Foundation for Research

14

think that weather conditions, the imposition of a local sales tax, the election of a new mayor, and thousands of similar events and conditions would have little effect on work-week length and productiv- ity. Most can safely be ignored because their impact occurs in such a random fashion as to have little effect. Others might influence the DV, but their effect is not at the core of the problem we investigate. Control variables (CV) are extraneous variables that we measure to determine whether they influence our results, as we want to make sure our results are not biased by excluding them. In Exhibit 1-7b, weather is shown as a CV; the broken line indicates that we included it in our research hypothesis because it might influence the DV, but we consider it irrelevant for the investigation of our research problem.

Extraneous variables can also be confounding variables (CFVs) to our hypothesized IV–DV relation- ship, similar to moderating variables. You may consider that the kind of work being done might have an effect on the impact of workweek length on productivity; productivity gains might not be universal across all types of work. This might lead you to introducing type of work as a confounding variable (CFV). In our example, we would study the effect of the four-day working week within groups (e.g., office workers vs. manufacturing workers vs. distribution plant workers). In Exhibit 1-7b, we included the type of work as a CFV, with a broken line.

>Exhibit 1-7 Relationships among Types of Variables

(a)

(b)

(c)

CFV: type work

CFV: type work

IVV: special project

MV: workers' age

MV: workers' age

DV: productivity

DV: productivity

MV: workers' age

DV: productivity

IV: four-day working

week

IV: four-day working

week

IVV: job

satisfaction

CV: weather

CV: weather

IV: four-day working

week

schi18939_ch01_001-025.indd 14 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

15

Intervening Variables

The variables mentioned with regard to causal relationships are concrete and clearly measurable—that is, they can be seen, counted, or observed in some way. Sometimes, however, one may not be com- pletely satisfied by the explanations they give. Thus, while we may recognize that a four-day workweek results in higher productivity, we might think that this is not the whole story—that workweek length affects some other variable that, in turn, results in higher productivity. The intervening variable (IVV) is a factor that theoretically affects the DV but cannot be observed or has not been measured; its ef- fect must be inferred from the effects of the independent and moderating variables on the observed phenomenon.11 In our example, one might view the intervening variable (IVV) to be job satisfaction, giving a hypothesis such as:

The introduction of a four-day working week (IV) will lead to higher productivity (DV) by increasing job satisfaction (IVV).

Here we assume that a four-day work week increases job satisfaction; similarly, we can assume that the introduction of special project work might influence productivity. Exhibit 1-7c illustrates how theoretical constructs, which are not directly observed, fit into our model.

When the first confirmed case of bovine spongiform encepha-

lopathy (BSE—known as “mad cow” disease) was discovered

in a Washington state dairy cow in December 2003, numerous

countries banned U.S. beef imports, bringing the $3.2 billion ex-

port industry to a standstill. That year, the U.S. Department of

Agriculture (USDA) performed random tests on approximately

0.03 percent of all slaughtered cattle, about 20,000 cows of the

nearly 40 million head of cattle slaughtered annually. In com-

parison, western European countries tested 10 million cows and

Japan tested each of its 1.2 million slaughtered cows.

Theories are essential to a researcher’s quest to explain

and predict phenomena while creating business opportunities

and informing public policy. One USDA theory is that the best

way to identify sources of cattle-born disease is to monitor a

cow from birth to slaughter. Thus, the USDA wanted a national

livestock database. After evaluating the options, the USDA

proposed another theory: Cows tagged with radio frequency

identification devices (RFID) would create the most accurate

database.

About the size of a quarter, the RFID tag is stapled to the

base of the animal’s ear. It is programmed with a numeric code

that is scanned by a stationary or handheld device when a cow

reaches a new location in the production process. As cows

move from farm to feeding lot to slaughterhouse, each animal’s

origin and location can be updated in the national database.

But RFID tags can be damaged, dislodged, or tampered

with. Slaughterhouses need additional safeguards to be

certain these devices don’t end up in the meat. “All you need

is one chip in someone’s burger and you’ve got a problem,”

says Brian Bolton, vice president of marketing for Optibrand.

This Colorado company offers a different theory for the best

identification and tracking: A camera that records the unique

vascular patterns in a cow’s retina at each stage of the beef

production chain is the most reliable. With retinal scanning,

Bolton says, “the tracking technology is contained in the hand-

held reader. It takes a tiny picture of a cow’s retina and then

links it to that animal’s computerized record.” Meatpacker Swift

& Co., the nation’s third-largest beef processor, has been using

Optibrand’s devices for several years. Retinal scan wands also

read RFID tags, access global positioning receivers, and stamp

each scan with a location record. However, retinal scanning is

not always practical because scans must be taken about an

inch from an animal’s eye.

In addition to RFID and retinal scanning, beef producers and

processors implement other tracking systems, thus implement-

ing their own theories. Some use implantable computer chips

and others use DNA matching systems. While still preferring RFID

technology, the USDA’s director of national animal identification,

John F. Wiemers, concedes, “We think there’s room for all these

technologies.”

Which tracking theory do you favor? What are the most im-

portant variables you would consider in justifying your decision? 

www.usda.gov; www.optibrand.com; www.jbsswift.com

Radio Chips versus Retinal Scans: Which Theory Offers the Best Protection?

>snapshot

schi18939_ch01_001-025.indd 15 2/14/18 2:37 PM

>part I Building the Foundation for Research

16

Hypotheses, Theories, and Models Hypotheses, theories, and models serve researchers in different ways but are related.

• A hypothesis is an unsubstantiated assumption about the relationship between concepts and con- structs; it drives the research.

• A theory is comprised of data-tested, supported hypotheses; it is derived from research.

• A model is a visualization of a theory; it is used for clarification and to enhance understanding.

Hypotheses

At the core of any good research, then, is a carefully constructed hypothesis. A hypothesis can be phrased as a declarative statement (descriptive) or a question about the relationship between two or more concepts or constructs that may be judged as true or false. It is always conjecture and formulated for empirical testing/measurement.

Descriptive format Question format

American cities are experiencing budget difficulties due to a decline in manufacturing.

Are American cities experiencing budget difficulties due to a decline in manufacturing?

Because if drives the research, crafting a hypothesis serves several important functions:

• It encourages researchers to think about the likely relationships to be found.

• It encourages researchers to think about the relevant facts (those needed to support or reject the hypothesis) and those facts that are not relevant.

• It suggests which research design is likely to be most appropriate.

• It is useful for testing statistical significance.

• It provides a framework for organizing the conclusions that result from the research.

To consider specifically the role of the hypothesis in determining the direction of the research, sup- pose we use this example:

Husbands and wives agree on their respective roles in vehicle purchase decisions.

The hypothesis specifies who shall be studied (married couples), in what context they shall be studied (their vehicle purchase decision making), and what shall be studied (their individual perceptions of their roles). This hypothesis suggests that the best research design is a communication-based study, either a survey or interview. We have at this time no other practical means to ascertain perceptions of people except to ask about them in one way or another. In addition, we are interested only in the roles that are assumed in the vehicle purchase decision. The study should not, therefore, seek information about roles husbands and wives might assume in other purchase decisions—say, electronics or furniture or movies. A study based on this hypothesis might reveal that husbands and wives disagree on their perceptions of roles, but the differences may be explained in terms of some factor(s) other than gender (i.e., mechanical knowledge, passion for cars, age, social class, religion, personality, etc.).

Types of Hypotheses There are numerous types of hypotheses. A descriptive hypothesis states the existence, size, form, or distribution of some concept/construct. For example, “The life expectancy of airplane Model 707 exceeds 16 years.”

A relational hypothesis describes a relationship between two or more concepts/constructs. Each re- lationship describes a correlational or causal relationship. With causal hypotheses one variable being studied is assumed to cause a specific effect on other variables studied. With correlational hypotheses the variables being studied occur together, but there is no assumption of causation. For example:

• Women older than 50 (concept) purchase (concept) more or less of our product (concept) than those younger than 25 (concept). [correlational (neutral) relationship]

• Students who attend class regularly (construct) earn higher grades (concept) than those who do not attend regularly (construct). [correlational (positive) relationship]

schi18939_ch01_001-025.indd 16 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

17

• U.S. brand cars (concept) are perceived by American consumers (construct) to be of lesser quality (construct) than foreign brand cars (concept). [correlational (negative) relationship]

In the first example we hypothesize that the two groups purchase different levels of our product. In the second, there may be other factors that influence grades; it might be the amount of studying. In the last example, we hypothesize that “country of origin” (concept) influences “perceived quality” (con- struct), but we don’t know what factors cause “perceived quality.” Correlational hypotheses are often made when we believe there are more basic causal forces that affect the concepts/constructs (e.g., maybe a lack of product recalls causes high perceived quality) or when we do not have enough evidence to claim a stronger linkage.

Causal hypotheses not only predict the cause (cause means roughly to “help make happen”) but also the effect. Here are three examples of causal hypotheses:

• An increase in the price of salvaged copper wire (concept/cause) leads to an increase in scaveng- ing in abandoned homes (construct/effect).

• Exposure to a company’s message concerning recent injuries (concept/cause) leads to more favor- able attitudes among employees toward safety (construct/effect).

• Loyalty to a particular grocery store (construct/cause) leads to purchasing that store’s private brands (concept/effect).

In proposing or interpreting causal hypotheses, the researcher must consider the direction of influ- ence. One would assume the price of copper wire influences scavenging rather than the reverse. Once would also assume that the measure of attitude follows the release of the information about on-the-job in- juries. Sometimes our ability to identify the direction of influence depends on the research design. Store loyalty and purchasing of store brands appear to be interdependent. Loyalty to a store may increase the probability of someone buying the store’s private brands, but satisfaction with the store’s private brand may also lead to greater store loyalty.

Reasoning and Hypotheses Every day we reason with varying degrees of success. Reasoning— gathering facts consistent with the problem, proposing and eliminating rival hypotheses, measuring outcomes, developing crucial empirical tests, and deriving the conclusion—is pivotal to much of a re- searcher’s success. Two types of reasoning are of great importance to research in forming and testing hypotheses: induction and deduction.

Induction Researchers use induction to craft hypotheses. In induction, you start by drawing a con- clusion from one or more particular facts or pieces of evidence. The conclusion explains the facts, and the facts support the conclusion. To illustrate, suppose your firm spends $10 million on a regional pro- motional campaign and sales do not increase; these are facts. Under such circumstances, we ask, “Why didn’t sales increase?”

One likely answer to this question is that the promotional campaign was poorly executed (conclu- sion). This conclusion is an induction because we know from experience that regional sales should go up during a promotional event. Also we know from experience that if the promotion is poorly executed, sales will not increase. The nature of induction, however, is that the conclusion is only a hypothesis. It is one explanation, but there are others that fit the facts just as well. For example, each of the following hypotheses might explain why sales did not increase:

• A strike by the employees of our trucking firm prevented stock from arriving at retailers; regional retailers did not have sufficient stock to fulfill customer orders during the promotional period.

• A competitor lowered its price during the promotional period; customers bought their brand rather than ours.

• A category-five hurricane closed all our retail locations in the region for the 10 days during the promotion.

In this example, we see the essential nature of inductive reasoning. The inductive conclusion is an inferential leap beyond the evidence presented—that is, although one conclusion explains the fact of no sales increase, other conclusions also might explain the fact. It may even be that none of the conclusions we advanced correctly explain the failure of sales to increase.

schi18939_ch01_001-025.indd 17 2/14/18 2:37 PM

>part I Building the Foundation for Research

18

For another example, let’s consider the situation of Tracy Nelson, a salesperson at the Square Box Company. Tracy has one of the poorest sales records in the company. Her unsatisfactory performance prompts us to ask the question, “Why is she performing so poorly?” From our knowledge of Tracy’s sales practices, the nature of box selling, and the market, we might conclude (hypothesize):

• Tracy makes too few sales calls per day to build a good sales record.

• Tracy’s territory does not have the market potential of other territories.

• Tracy’s sales-generating skills are so poorly developed that she is not able to close sales effectively.

• Tracy does not have authority to lower prices and her territory has been the scene of intense price- cutting by competitive manufacturers, causing her to lose many sales to competitors.

• Some people just cannot sell boxes, and Tracy is one of those people.

Each of the above hypotheses has some chance of being true, but we would probably have more confidence in some than in others. All require further confirmation before they gain our confidence. Confirmation comes with more evidence. The task of research is largely to (1) determine the nature of the evidence needed to confirm or reject hypotheses and (2) design methods by which to discover and measure this other evidence.

Deduction Researchers use deduction to plan research and draw insights from data that will test hy- potheses. Deduction is a form of reasoning that starts with one or more true premises and the conclusion flows from the premises given. For a deduction to be correct and sound, it must be both true and valid:

• Premises (reasons) given for the conclusion must agree with the real world (true).

• The conclusion must necessarily follow from the premises (valid).

A deduction is valid if it is impossible for the conclusion to be false if the premises are true. For ex- ample, consider the following simple deduction:

• All employees at BankChoice can be trusted to observe the ethical code. (Premise 1)

• Sara is an employee of BankChoice. (Premise 2)

• Sara can be trusted to observe the ethical code. (Conclusion)

If we believe that Sara can be trusted, we might think this is a sound deduction. But this conclusion cannot be a sound deduction unless the form of the argument is valid and the premises are true. In this case, the form is valid, and premise 2 can be confirmed easily. However, trillions of dollars each year

Researchers often use ob- servation when evaluating a customer's use of a product. Apply deductive reasoning to this image. Develop your own conclusions concerning what will happen next.

© E

ri k

Is ak

so n

/B le

n d

Im ag

es

schi18939_ch01_001-025.indd 18 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

19

in confirmed global employee theft12 will challenge premise 1. If one premise fails the acceptance test, then the conclusion is not a sound deduction. It should be apparent that a conclusion that results from deduction is, in a sense, already “contained in” its premises.13

Combining Induction and Deduction Induction and deduction are used together in research reasoning. John Dewey, psychologist and educational reformer, describes this process as the “double movement of reflective thought.”14 Induction occurs when we observe a fact and ask, “Why is this?” In answer to this question, we advance a tentative explanation (hypothesis). The hypothesis is plausible if it explains the fact (event or condition) that prompted the question. Deduction is the process by which we test whether the hypothesis is capable of explaining the fact. The process is illustrated in Exhibit 1-8:

1. You promote a product but sales don’t increase. (Fact 1)

2. You ask the question “Why didn’t sales increase?” (Induction)

3. You propose a hypothesis to answer the question: The promotion was poorly executed. (Hypothesis)

4. You use this hypothesis to conclude (deduce) that sales will not increase during a poorly executed promotion. You know from experience that ineffective promotion will not increase sales. (Deduction 1)

This example, an exercise in circular reasoning, points out that one must be able to deduce the ini- tiating fact from the hypothesis advanced to explain that fact. A second critical point is also illustrated in Exhibit 1-8. To test a hypothesis, one must be able to deduce from it other facts that can then be investigated. This is what research is all about. We must deduce other specific facts or events from the hypothesis and then gather information to see if the deductions are true. In this example:

5. We deduce that a well-executed promotion will result in increased sales. (Deduction 2)

6. We run an effective promotion, and sales increase. (Fact 2)

How would the double movement of reflective thought work when applied to the Tracy Nelson prob- lem? The process is illustrated in Exhibit 1-9. The initial observation (fact 1) leads to hypothesis 1 that Tracy is lazy. We deduce several other facts from the hypothesis. These are shown as fact 2 and fact 3. We use research to find out if fact 2 and fact 3 are true. If they are found to be true, they confirm our hypothesis. If they are found to be false, our hypothesis is not confirmed, and we must look for another explanation.

In most research, the process may be more complicated than these examples suggest. For instance, we often develop multiple hypotheses by which to explain the manager’s problem. Then we design a study to test all the hypotheses at once. Not only is this more efficient, but it is also a good way to reduce the attachment (and potential bias) of the researcher to any given hypothesis.

Fact 1: We promote a product but sales do not increase.

Induction

Deduction: Ine�ective promotion

will not increase sales.

Ded uctio

n: E �ec

tive prom

otio n

wil l inc

reas e sa

les.

Fact 2: We run an e�ective promotion and sales increase.

Hypothesis: The promotion was poorly executed.

Why?

>Exhibit 1-8 Why Didn’t Sales Increase

schi18939_ch01_001-025.indd 19 2/14/18 2:37 PM

>part I Building the Foundation for Research

20

The steps that follow represent one approach to assessing the validity of conclusions about observ- able events.15 These steps are particularly appropriate for business researchers whose conclusions result from empirical data. The researcher:

1. Encounters a curiosity, doubt, barrier, suspicion, or obstacle.

2. Struggles to state the problem—asks questions, contemplates existing knowledge, gathers facts and moves from an emotional to an intellectual confrontation with the problem.

3. Proposes a hypothesis (one plausible explanation) to explain the facts that are believed to be logi- cally related to the problem.

4. Deduces outcomes or consequences of that hypothesis—attempts to discover what happens if the results are in the opposite direction of that predicted or if the results support the expectations.

5. Formulates several rival hypotheses.

6. Devises and conducts a crucial empirical test with various possible outcomes, each of which se- lectively excludes one or more hypotheses.

7. Draws a conclusion (an inductive inference) based on acceptance or rejection of the hypotheses.

8. Feeds information back into the original problem, modifying it according to the strength of the evidence.

What Is a Strong Hypothesis? A strong hypothesis should fulfill three conditions:

• Adequate for its purpose.

• Testable.

• Better than its rivals.

The conditions for developing a strong hypothesis are detailed more fully in Exhibit 1-10.

Theories

We have many theories and use them continually to explain or predict what goes on around us. To the degree that our theories are sound (empirically supported) and fit the situation, we are successful in our explanations and predictions. For example, it is midday and you note that outside the natural light is dimming, dark clouds are moving rapidly in from the west, the breeze is freshening, the barometric pressure is falling, and the air temperature is cooling but above 50 degrees. Would your understanding

>Exhibit 1-9 Why Is Tracy Nelson’s Performance So Poor?

Fact 1: Tracy has a poor performance record.

Fact 2: Tracy is regularly late to work.

Induction

Deduction: Laziness results

in excessive tardiness.

Ded uctio

n: L azin

ess resu

lts in few

er

cus tom

er ca lls p

er d ay.

Fact 3: Tracy makes fewer calls per day than the average salesperson.

Hypothesis: Tracy is lazy.

Why?

schi18939_ch01_001-025.indd 20 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

21

of the relationship among these concepts/constructs (your weather theory) lead you to predict that it will rain?

A theory is an empirically supported description of the relationships among concepts, constructs, and hypotheses that are advanced to explain or predict phenomena. A theory, therefore, is comprised of data-tested, supported hypotheses; it is derived from research. Our ability to make rational decisions is based on our ability to develop theory. But a caution: No theory can ever be considered final because it is subject to challenge by new data.

In marketing, for example, the product life cycle theory describes the stages that a product category goes through in the marketplace.16 It was developed based on observing thousands of product introductions and their success path over time. The generalized product life cycle has four stages: introduction, growth, maturity, and decline. In each stage, many concepts, constructs, and hypotheses describe the influences that change revenue and profit. Definitions are used for communicating the claims of the theory. In the growth stage of this theory, for example, companies in the category spend heavily on promotion to create product awareness (construct). In the early period of this stage these expenditures may be made to fuel pri- mary demand (construct), improving product category awareness (construct) rather than brand awareness (construct). In this stage, sales (concept) increase rapidly because many customers (concept) are trying the product; and those who are satisfied and purchase again—repeat purchasers (concept)—are swelling the ranks. According to product lifecycle theory, if a given company is unable to attract repeat purchasers, this may mean death for its particular brand (hypothesis), even though the product category may endure. This theory has endured for more than 50 years.

Models

If you are an architect, a client might hire you to design their dream home. You might build them a physi- cal model so that they could better visualize your design. Such models, while time-consuming to create, often save time and money during the construction process by avoiding costly onsite design changes.

>Exhibit 1-10 Checklist for Developing a Strong Hypothesis

Criteria Interpretation

Adequate for Its Purpose 1. Does the hypothesis reveal the original problem condition?

2. Does the hypothesis clearly identify facts that are relevant and those that are not?

3. Does the hypothesis clearly state the condition, size, or distribution of some variable in terms of values meaningful to the research problem (descriptive)?

4. Does the hypothesis explain facts that gave rise to the need for explanation (explanatory)?

5. Does the hypothesis suggest which form of research design is likely to be most appropriate?

6. Does the hypothesis provide a framework for organizing the conclusions that result?

Testable 1. Does the hypothesis use acceptable techniques?

2. Does the hypothesis require an explanation that is plausible given known physical or psychological laws?

3. Does the hypothesis reveal consequences or derivatives that can be de- duced for testing purposes?

4. Is the hypothesis simple, requiring few conditions or assumptions?

Better Than Its Rivals 1. Does the hypothesis explain more facts than its rivals?

2. Does the hypothesis explain a greater variety or scope of facts than its rivals?

3. Is the hypothesis one that informed judges would accept as being the most likely?

schi18939_ch01_001-025.indd 21 2/14/18 2:37 PM

>part I Building the Foundation for Research

22

Marketing objective

Gain awareness

Stress di�erentiation

Maintain brand loyalty

Harvesting, deletion

0 +

–

Introduction Growth

Time

Maturity Decline

Total industry profit

Total industry sales revenue

S al

e s

re ve

n u

e o

r p

ro fit

Stage of the product life cycle

>Exhibit 1-11 Model of the Traditional Product Life Cycle Theory

Source: Adapted from Roger Kerin, Eric Berkowitz, Steven Hartley, and William Rudelius, Marketing, 7th ed. (Burr Ridge, IL: McGraw-Hill, 2003), p. 295.

LO1-1       Managers make thousands of decisions (strategic, tac- tical, and procedural) that use data as fuel. Business research and data analytics complement each other, but they are not synonymous. Data analytics needs data, some of which are provided by research. Managers often draw on data from existing internal data sources (called a decision support system) when engaging in

data analytics. When these data are mined, they may be used for a purpose other than that for which they were originally collected. Collecting data to understand and improve performance employs ongoing observation or communication research. However, using such data for a different purpose employs data analytics. Businesses are getting better at data blending—combining data from

>summary

A model, therefore, is a representation of a theory or system that is constructed to study some aspect of that sys- tem or the system as a whole. A research model’s purpose is to increase our understanding, prediction, and control of the complexities of the environment. While theory’s role is explanation or prediction, a model’s role is representa- tion. Models allow researchers and managers to character- ize present or future conditions: the effect of a raise on employee engagement, the effect of higher interest on bond purchase rates, the effect of advertising on purchase, the effect of scheduled maintenance on manufacturing defects. Exhibit 1-11 provides the model for the product life-cycle theory. Models are an important means of advancing theo- ries and aiding decision makers. ©

Jo h

n L

un d/

M ar

c R

om an

el li/

B le

n d

Im ag

es L

L C

schi18939_ch01_001-025.indd 22 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

23

separate datasets into a new actionable dataset and then querying that composite data to help make decisions. While the information that comes from data blending has a role in decision making, it is not the same as research.

Business research is a systematic inquiry that pro- vides information to guide decisions. More specifically, it is a process of determining, acquiring, analyzing and synthesizing, and disseminating relevant data, infor- mation, and insights to decision makers in ways that mobilize the organization to take appropriate actions that, in turn, maximize performance and help achieve organizational goals. The research process is a model for the development and interpretation of research studies. Several subprocesses or stages are included in the process, including (1) clarify the research, (2) design the research, (3) collect and prepare the data, (4) analyze and interpret the data, and (5) report insights and recommendations. Despite variations, the idea of a sequence is useful for developing a project and for keeping the project orderly as it unfolds. The research process is based on the tenets of the scientific method: clearly defined concepts, constructs, methods, and procedures; empirically testable hypoth- eses; direct observation of phenomena; conclusions drawn from statistical evidence rather than inferred justification (educated guesses); and the self-correcting process (ability to replicate and reassess validity of conclusions). Although the scientific method consists of neither sequential nor independent stages, the prob- lem-solving process that it reveals provides insight into the way research is conducted.

LO1-2                   Understanding the language of professional research- ers is crucial to doing great research. Researchers use concepts, constructs, operational definitions, and variables to form hypotheses. A concept is a gener- ally accepted collection of meanings or characters associated with an event, object, condition, situa- tion, or behavior. When we work with more abstract

phenomena, we invent constructs. When we are trying to understand the relationships between con- cepts and constructs, we often form a conceptual scheme or map. Each concept or construct needs an operational definition, its specific criteria for measure- ment. Variables serve as measurable substitutes for concepts and constructs in research. There are many different types: independent, dependent, moderating, and extraneous (including control, confounding, and intervening).

Hypotheses are the core of good research and are crafted as descriptive statements or questions. Each hypothesis is an unsubstantiated assumption about a relationship between two or more concepts or constructs; it drives the research study. Hypotheses are either descriptive or relational. Relational hypotheses are either correlational or causal. The research process is grounded in reasoning. The reasoning process is used for the development and testing of various hypoth- eses largely through the double movement of reflective thinking. Reflective thinking consists of sequencing induction and deduction in order to explain inductively (by hypothesis) a puzzling condition. In turn, the hypoth- esis is used in a deduction of further facts that can be sought to confirm or deny the truth of the hypothesis. Hypotheses are often crafted using inductive reasoning and then tested using deductive reasoning. Induction starts by drawing a conclusion from one or more facts or pieces of evidence. Deduction starts by proposing one or more true premises and draws a conclusion based on those premises. A theory is a data-supported description of the relationships between concepts and constructs in hypotheses; it is derived from research. Its role is prediction. A model is the visualization (representation) of a theory; it is used for clarification and to enhance understanding. Models allow researchers and managers to characterize present or future conditions.

business research 7

concept 9

conceptual scheme 11

construct 10

data blending 5

deduction 18

hypothesis 16

causal hypothesis 16

correlational hypothesis 16

descriptive hypothesis 16

relational hypothesis 16

induction 17

model 16

operational definition 11

reasoning 17

scientific method 8

theory 16

variable 12

control variable (CV) 14

confounding variable (CFV) 14

dependent variable (DV) (criterion variable) 12

extraneous variable (EV) 13

independent variable (IV) (predictor variable) 12

intervening variable (IVV) 15

moderating variable (MV) 13

>keyterms

schi18939_ch01_001-025.indd 23 2/14/18 2:37 PM

>part I Building the Foundation for Research

24

Terms in Review

1 What is business research?

2 Distinguish among the following sets of items, and suggest the significance of each in a research context:

a Concept and construct.

b Deduction and induction.

c Operational definition and dictionary definition.

d Concept and variable.

e Hypothesis and theory.

3 Describe the characteristics of the scientific method.

4 Below are some terms commonly found in a management setting. Are they concepts or constructs? Give two different operational definitions for each.

a First-line supervisor.

b Employee morale.

c Assembly line.

d Overdue account.

e Leadership.

f Union democracy.

g Ethical standards.

5 In your company’s management development program, there was a heated discussion between some people who claimed, “Theory is impractical and thus no good,” and others who claimed, “Good theory is the most practical ap- proach to problems.” What position would you take and why?

6 An automobile manufacturer observes the demand for its brand increasing as per capita income increases. Sales increases also follow low interest rates, which ease credit conditions. Buyer purchase behavior is seen to be depen- dent on age and gender. Other factors influencing sales appear to fluctuate almost randomly (competitor advertising, competitor dealer discounts, introductions of new competi- tive models).

a If sales and per capita income are positively related, clas- sify all variables as dependent, independent, moderating, extraneous, or intervening.

b Comment on the utility of a model based on the hypothesis.

Making Research Decisions

7 You observe the following condition: “Our female sales rep- resentatives have lower customer defections than do our male sales representatives.”

a Propose the concepts and constructs you might use to study this phenomenon.

b How might any of these concepts and/or constructs be related to explanatory hypotheses?

8 You are the office manager of a large firm. Your company prides itself on its high-quality customer service. Lately, complaints have surfaced that an increased number of incoming calls are being misrouted or dropped. Yesterday, when passing by the main reception area, you noticed the receptionist fiddling with his hearing aid. In the process, a call came in and would have gone unanswered if not for your intervention. This particular receptionist had earned an unsatisfactory review three months earlier for tardiness. Your inclination is to urge this 20-year employee to retire or to fire him, if retirement is rejected, but you know the indi- vidual is well liked and seen as a fixture in the company.

a Pose several hypotheses that might account for dropped or misrouted incoming calls.

b Using the double movement of reflective thought, show how you would test these hypotheses.

From Concept to Practice

9 Using Exhibits 1-8 an 1-9 as your guides, graph the induc- tions and deductions in the following statements. If there are gaps, supply what is needed to make them complete arguments.

a Repeated studies indicate that economic conditions vary with—and lag 6 to 12 months behind—the changes in the national money supply. Therefore, we may conclude the money supply is the basic economic variable.

b Research studies show that heavy smokers have a higher rate of lung cancer than do nonsmokers; therefore, heavy smoking causes lung cancer.

c Show me a person who goes to church regularly, and I will show you a reliable worker.

From the Headlines

10 Chipotle Mexican Grill continues to suffer from perception issues after a string of outbreaks, including E.coli, worried customers about the safety of eating at the fast casual chain. Chipotle’s strategy for getting customers back into its restaurants was to give away free tacos, burritos, and chips. And while its customer survey scores are improving, Chipo- tle is still operating at a loss. What concepts, constructs, and operational definitions should any future research deal with?

>discussionquestions

schi18939_ch01_001-025.indd 24 2/14/18 2:37 PM

>chapter 1 Research Foundations and Fundamentals

25

>cases*

Campbell-Ewald: R-E-S-P-E-C-T Spells Loyalty

Open Doors: Extending Hospitality to Travelers with Disabilities

HeroBuilders.com

*You will find a description of each case in the Case Abstracts section of this textbook. Check the Case Index to determine whether a case provides data, the research instrument, video, or other supplementary material. Cases and case supplements are available in Connect.

schi18939_ch01_001-025.indd 25 2/14/18 2:37 PM

LO2-1     The standards of good research.

LO2-2  What is happening within each stage of the research process.

LO2-3 Research process pitfalls and their solutions.

LO2-4 Ethical issues at each stage in the process.

>learningobjectives After reading this chapter, you should understand . . .

“Today individuals increasingly keep their

emails and documents on remote servers in data centers—

in short, in the cloud. But the transition to the cloud does not alter people’s

expectations of privacy.” Brad Smith,

President and Chief Legal Officer, Microsoft

The Research Process: An Overview

>chapter 2

schi18939_ch02_026-044.indd 26 2/19/18 11:26 AM

>chapter 2 The Research Process: An Overview

27

>What Is Good Research? Business research has inherent value only to the extent that it helps a manager make a better deci- sion to achieve organizational goals than a decision made with the information he already has. Interesting information about consumers, employees, competitors, equipment, facilities, locations, or the environment might be pleasant to have, but it has limited value if the information cannot be applied to making the best decision in a critical situation. As we explore the research process, we need quality standards that will distinguish good research from poor research and that provide a manager limited insight. These six standards are summarized here and linked to researcher respon- sibilities in Exhibit 2-1.

1. Purpose clearly defined. The purpose of business research—the problem involved or the decision to be made—should be clearly defined in writing. This statement of the decision problem should include its scope, limitations, and the precise meanings of all concepts, and variables significant to the research.

2. Research design thoroughly planned and executed. The procedural design of the research, and its choice among competing designs, should be clearly described and justified. Procedures should be detailed and carefully followed to yield results that are as objective as possible. Researchers must avoid personal bias in selecting research and sampling designs and collecting and recording data.

3. High ethical standards applied. Researchers often have significant latitude in designing and executing projects. They face ethical dilemmas at each stage of the process and must be prepared to address these issues. A research design must include safeguards against causing mental or physical harm, exploitation, invasion of privacy, and/or loss of dignity to participants, as well as ensuring the adequate training and safety of data collectors. Procedures for ensuring data integrity are critical.

4. Adequate analysis for decision maker’s needs. The analysis of the data should use appropriate techniques, be classified in ways that assist the researcher in reaching pertinent insights and conclusions, and clearly reveal the findings and procedures that have led to those conclu- sions. Insights should be limited to those for which the data provide an adequate basis. When statistical methods are used, the probability of error should be estimated, and the criteria of statistical significance applied; these should be revealed. The validity and reliability of data should be checked carefully.

5. Limitations frankly revealed. While few research designs are perfect, the researcher should report, with complete frankness, flaws in research design or design execution and estimate their effect on the findings. Some of the imperfections may have little effect on the validity and reliability of the data; others may invalidate them entirely. As a decision maker, you should question the value of research about which no limitations are reported.

6. Findings reported unambiguously; insights and conclusions justified. Except when confidentiality is imposed, research reports should reveal candidly the sources of data and the means by which data were obtained. Researchers should report procedural details to make it possible to estimate the validity and reliability of the data and increase confidence in the research itself, as well as any recommendations based on the research. Good researchers always specify the conditions under which their conclusions seem to be valid. Presentation of insights and conclusions should be comprehensive, easily understood by the audience, and organized so that the relevant decision maker can readily locate critical information.

>The Research Process Organizations usually treat the research process as a sequential process involving several clearly defined stages: clarify the research question, design the research, collect and prepare data, analyze and interpret data, and report insights and recommendations. Recycling, circumventing, and skipping occur among

schi18939_ch02_026-044.indd 27 2/19/18 11:26 AM

>part I Building the Foundation for Research

28

>Stage 1: Clarify the Research Question The purpose of this stage in the research process is to determine whether research should be conducted. During this stage, managers must answer the following questions:

1. What is/are the management dilemma(s) and which should be the focus?

2. What do we know that might help define and solve the dilemma?

3. What information do we need to make a decision?

4. Would knowing this information improve the decision(s) we must make to resolve the dilemma?

5. Where will the funding come from for any proposed research?

>Exhibit 2-1 Actions That Deliver Effective Research

Category Option

Purpose clearly defined ● Researcher distinguishes among symptom of organization’s problem, the manager’s perception of the problem, and the research problem.

● Researcher includes the scope and limitations within the statement of the research problem.

● Researcher operationally defines all concepts, constructs, and variables.

Research design thoroughly planned and executed

● Researcher provides a justification for the choice of a research design and the rejection of alternative designs.

● Researcher details procedures within the chosen design and carefully executes them.

● Researcher avoids personal bias in selecting research designs.

● Researcher defines exploratory procedures.

● Researcher defines target population and describes the sample unit along with the procedures for the selection.

● Researcher defines and follows procedures for data collection.

High ethical standards applied

● Researcher designs safeguards to protect study participants, organizations, clients, and researchers.

● Researcher recommendations do not exceed the scope of the study.

● Researcher ensures the study’s methodology and limitations sections reflect researcher’s restraint and concern for accuracy.

● Research protects confidentiality of subjects and sources, as necessary.

Adequate analysis for decision maker’s needs

● Researcher follows analytical procedures acceptable for the data collected and the decision maker’s needs.

● Researcher connects findings to the research instrument.

● Researcher connects insights and conclusions to the data collected, not exceeding the scope of the data.

● Researcher checks for data validity and reliability.

Limitations frankly revealed

● Researcher reveals the study’s limitations by comparing desired with actual procedure.

● Researcher reveals the sample’s limitations by comparing its characteristics with target population.

● Researcher reveals the impact of limitations on findings, insights, and conclusions.

Findings reported unambiguously; insights and conclusions justified

● Researcher clearly presents findings in words, tables, and graphs.

● Researcher logically organizes findings, insights, and conclusions to help the manager reach a decision.

● Researcher summarizes conclusions.

● Researcher provides a detailed table of contents in any written report to facilitate quick access to desired information.

steps. Some steps are begun out of sequence, some are carried out simultaneously, and some are omitted. Despite these variations, the idea of a sequence is useful for developing a project and for keeping the project orderly as it unfolds. Exhibit 2-2 models the research process we will develop in this text.

schi18939_ch02_026-044.indd 28 2/19/18 11:26 AM

>chapter 2 The Research Process: An Overview

29

Identify and Prioritize Dilemmas Managers don’t have unlimited resources, so they have to focus these where they believe they will accom- plish the most. This stage begins at the most general level with the management dilemma(s), a symptom of an actual problem, such as:

• Rising costs.

• Increasing tenant move-outs from an apartment complex.

• Declining sales.

• Increasing employee turnover in a restaurant.

• A larger number of product defects during the manufacture of an automobile.

• An increasing number of letters and phone complaints about post-purchase delivery service.

The management dilemma can also be triggered by an early signal of an opportunity or growing evidence that a fad may be gaining staying power, such as:

• The discovery of an expensive chemical compound that would increase the efficacy of a drug.

• The shift in job hunting to digital job boards.

• The rise in penetration of smart phone usage.

• Increasing numbers of firms mining data to develop digital dashboards.

>Exhibit 2-2 The Research Process, with Ethical Responsibilities

Report Insights & Recommendations

ExplorationExploration

Analyze & Interpret the Data

Design the Research

Clarify the Research Question

Collect & Prepare Data

Data Collection Design

Sampling Design

• Sponsor’s right to dilemma, options, and research question nondisclosure • Sponsor’s right to quality research • Researcher’s right to absence of sponsor deception and coercion

S ta

g e

1

• Participant’s right of informed consent • Participant’s right of confidentiality • Sponsor’s right to research nondisclosure • Sponsor’s right to quality research • Researcher’s right to absence of coercion

S ta

g e

2

• Participant’s right to safety • Participant’s right of confidentiality • Sponsor’s right to quality research • Researcher’s right to absence of participant deception • Researcher’s right to participant completion

S ta

g e

3 • Sponsor’s right to quality research • Sponsor’s right to findings nondisclosure • Researcher’s right to absence of sponsor coercion

S ta

g e

4

• Researcher’s right to adequate compensation for results

S ta

g e

5

schi18939_ch02_026-044.indd 29 2/19/18 11:26 AM

>part I Building the Foundation for Research

30

Identifying problem-oriented management dilemmas is rarely difficult, unless the organization fails to track its performance against prior performance and industry standards. These metrics are called key performance indicators (KPIs). KPIs exit at the company, business unit, and individual employee lev els. Company KPIs are driven by stockholder demands (sales, profits, growth, reputation), government regu- lations (safety, overtime, on-time performance), even employee unions (training, advancement). At the business unit, companies might track a variety of KPIs; for example, a production line might track manu- facturing output, production defects, time frame of manufacturer, supply fulfillment, etc. For the office of human resources, KPIs might be employee turnover, employee engagement, applicants per open posi- tion, online training participation, safety training compliance scores, performance review ratings, etc. Even individual employees have KPIs, although we might not call them this; these include pay increases,

Doing research with mothers of infants and small children

or expectant mothers can be problematic. These women be-

tween 21–40 have lots of distractions and their time availabil-

ity is often at odd hours. Marcus Thomas LLC (it describes itself

as an idea agency) and QualVu (a FocusVision company that

helps its clients find truth) collaborated to help a durable baby

goods producer better understand a mother’s choices.

Specifically, the study objectives were to “gain insight into

the mother’s perceptions of specific products, uncover their

path to purchase, [understand] purchase motivations, and [share

their] experiences at key retailers (Target, Walmart, and Babies

‘R’ Us),” explained Jennifer Hirt-Marchand, associate partner,

strategic insights executive with Marcus Thomas. Qualitative re-

search would allow Marcus Thomas a deeper understanding of

their subjects, so they opted for video diaries, and QualVU of-

fered the platform.

Recruited participants needed to have a baby or be preg-

nant and have access to a computer and high-speed Internet

access. Those who didn’t already have a video camera were

provided one (that they could keep at the end of the study).

“We knew from secondary research provided by Forrester

Research that women in the age group were active on social

networks and accustomed to sharing their personal stories

online, using both text and video,” shared Edwige Winans, as-

sociate research director at Marcus Thomas. As moderator,

Winans prerecorded “discussion points” and posted them to

the QualVu platform, where participants accessed them. “What

we tried to achieve was a semblance of a relationship with the

participant, using the video.” The participants answered various

questions by uploading video diaries to the site over a four-

day period. The participants actually have “conversations with

themselves,” explained Winans. “Their diaries provide really

candid responses.” Participants answer different discussion

topics each day; the methodology allows participants to answer

the questions on their own schedule. Videos are monitored

while the study is ongoing, and the moderator uses follow-up

emails to prompt or redirect participants for further clarification,

if necessary.

Over the course of the study, 21 compensated participants

responded to eight topics with uploaded diaries on each topic

(a typical video diary study is 7 to 10 topics). QualVu research-

ers watched the videos, identifying trends in comments, and

provided a video report. “The client was especially interested in

comments on retailer in-store merchandising elements,” shared

Winans. “But our most important task was to keep an open

mind.” Some of the most valuable insights came from secondary

cues, like the participant’s home environment, props the partici-

pant used in her diaries, and the product demonstrations she did

in stores or in her home.

“A key consideration in recommending this methodology to

our client was to ensure that technology was intuitive and natural

to the participant and allow them to comfortably and confidently

share their thoughts and perspectives,” stated Hirt-Marchand.

The client got just what it needed.

www.marcusthomasllc.com; www.focusvision.com

New Moms and Moms-to-be offer Video Insights

>snapshot

©Hero/Corbis/Glow Images

schi18939_ch02_026-044.indd 30 2/19/18 11:26 AM

>chapter 2 The Research Process: An Overview

31

promotions, training experiences, project assignments, project leadership, new customers captured, and so on. Sophisticated businesses track these KPIs through digitized, perpetually maintained dashboards— a data visualization tool that shows current and a period of prior status on each metric, usually on one screen. Even small businesses can dashboard KPIs create with the software tools available.

Identifying opportunity-based management dilemmas is more time consuming and difficult. It requires monitoring obscure developments in a variety of industries, as well as emerging trends in your own industry. Those companies looking for opportunities are trying to be first movers; they want to capitalize on an environmental trend that others haven’t recognized. Jeff Bezos, chief executive officer (CEO) of Amazon, is always seeking these kinds of dilemmas: customers getting frustrated at not finding the solutions they sought at stores, companies needing a more reliable delivery option, readers wanting to take more books with them when they travel than luggage space permits, etc. The search for opportunity-based management dilemmas is full of risk, but when the approach pays off, it often pays big.

However, choosing one dilemma on which to focus may be difficult. Solving different dilemmas offers different rewards. One approach is to estimate the payoff of solving a dilemma, using this value to prioritize. This is more easily done for problems (e.g., estimating the cost of reducing customer returns) than for opportunities. Ultimately, to choose incorrectly puts a company or business on an unproductive path. As a manager, only practice makes you proficient at this task. For new managers, or for established managers facing new responsibilities, developing several management dilemma-to- research question hierarchies, each starting with a different dilemma, will assist in the choice process. To develop these, much exploratory research is used, tapping into published secondary sources and mining company data.

Exploration Seeking existing information is often used not only to identify dilemmas (e.g., identify industry standard to compare to company performance), but also to ask the right questions and better understand deci- sion options. Historic company data are increasingly being used in exploration as better tools to tap into digital data warehouses have become available.1 Much of this information may not be in search able databases; it may be in written reports, where accessibility is an issue. At this stage, it always pays the

We are attracted to experi- ments at an early age due to our unlimited curiosity. A great researcher fosters curiosity as an important skill.

© B

le n

d/ Im

ag e

S ou

rc e

schi18939_ch02_026-044.indd 31 2/19/18 11:26 AM

>part I Building the Foundation for Research

32

researcher to search aggressively because the cost of exploration is far less than that of conducting new research. However, historic information and published information always have an inherent problem: it is a reflection of the past, not the present, and the situation or environment reflected in that past rarely matches the current situation.

Exploration is itself a type of research; it is usually highly unstructured. If you have a paper to write, you might start with a Google search using key words to search on a topic of interest. You are likely to get thousands of links but many won’t directly relate to the information you seek. So you keep refining your search terms. And you try a different search engine, like Bing, to see if it gives you different results. Researchers might use Google or another search engine, but they are also likely to seek information in proprietary databases that general search engines can’t access. The most important proprietary informa- tion source to a researcher might be company records, both written and digital.

During the process of exploration, the manager has, hopefully, defined plausible management ac- tions (options) to address the dilemmas. Each of these options may be put into two groups: those that are feasible given company resources, and those that aren’t currently feasible. Besides estimated payoff, managers use feasibility to prioritize management dilemmas and focus their attention. It is equally likely that we have raised numerous additional questions without answers, and thus identified what we don’t know.

To move forward, the manager needs to frame a research question—the hypothesis that best states the objective of the research—and often specifies a choice between action options. The answer to this ques- tion must provide the manager with the desired information necessary to make a decision with respect to the management dilemma. Arriving at the research question usually involves additional exploration.

Value and Budget the Research While identifying dilemmas and prioritizing them, exploration helped the manager determine the re- search question and what they know and don’t know. Capturing the information you don’t know (con- ducting a research project) always has a cost. So before a manager moves forward, he or she needs to estimate whether acquiring that information is worthwhile. Managers have prior experiences that frequently lead them to decisions without research. If that experience, when facing the newly chosen dilemma, would lead the manager to the same decision as new research, then the research has no value. However, if a decision is unlikely to be made without new research, then the research has value. The only way this estimate of value can be made is to consider plausible methods by which the information could be collected and from whom or what and to then determine the cost. Using this cost estimate, the man- ager determines a budget. If research has value, then the manager obtains the resources budgeted for the project. In most companies of any size, this is a competitive and political activity, pitting one manager’s research proposal against another’s. If the process is fair, the manager with the best justification will earn the budget he or she needs.

>Stage 2: Design the Research Project The research design is the blueprint for collecting data that fulfills objectives and answers questions. This stage of the planning answers two questions:

• Data collection design. What data need to be collected, how, and in what format?

• Sampling design. From whom or what do the data need to be collected, and from how many?

Data Data are collective units of information (e.g., production defects, purchases or returns, a person’s atti- tudes, behaviors, motivations, attributes, photos, recorded comments, etc.) from a subject or case (peo- ple, events, machines, personnel records, etc.) measured by a data collector (person or a sensing device or machine, digital or mechanical) following consistent procedures. Data are of two classifications:

schi18939_ch02_026-044.indd 32 2/19/18 11:26 AM

>chapter 2 The Research Process: An Overview

33

primary or secondary. Primary data are recorded directly; they are raw or unprocessed. For example, a researcher conducts a personal interview with a subject and digitally records the answers. Primary data are the focus of most business research studies. Secondary data contain at least one level of interpreta- tion. For example, a researcher discovers a research report or article and uses the summarized graphical representation of the data contained within that report to extract data for his own research. Secondary data are often used during exploration and are helpful in the interpretation of new primary data on similar subjects.

We also describe data by its properties and group them into four types based on the type of measure- ment scale used to collect it: nominal, ordinal, interval, or ratio. Nominal data provide classification but no order, equal distance, or natural origin. When we ask people whether they prefer online training or classroom training, we group their data based on a classification system (online/classroom). In the classification, no group is more than or less than another. Ordinal data provide classification and order but no equal distance or natural origin. We can collect employees’ home addresses then group them by whether they are within a two-mile radius of the business, two to five miles away from the plant, or beyond five miles; in the groups, we have classification and order (based on distance) but the distance represented by each group is not equal. Interval data reveal classification, order, equal distance, but no natural origin. If we give a person an attitude question and ask them to use a five-point scale of agree- ment (1 = strongly agree; 2 = agree; 3 = neither agree nor disagree; 4 = disagree; 5 = strongly disagree), we collect interval data. We assume that those who agree with a statement feel less strongly than some- one else who chose “strongly agree.” Because we anchored the words with numbers, we assume equal distance between word descriptors, but there is no natural origin. Finally, ratio data offer classification, order, equal distance, and natural origin. If we measure the number of defects produced by a machine in a 24-hour period, we get an actual number. That number can’t be less than 0, its natural origin. We also know a machine that produces 10 defects a day produces a greater number of defects than a machine that produces only three defects a day; thus we also have classification and order. One defect is exactly one defect away from two defects; thus we have equal distance. Knowing statistically what we want to do with data often determines which of the four types we collect.

Category Characteristics of Data Basic Empirical Operation Example

Nominal Classification but no order, equal distance, or natural origin 

Determination of equality Gender (male, female)

Ordinal Classification and order but no equal distance or natural origin

Determination of greater or lesser than

Doneness of meat (well, me- dium well, medium rare, rare)

Interval Classification, order, equal distance, but no natural origin 

Determination of equality of inter- vals or differences

Temperature in degrees

Ratio Classification, order, equal distance, and natural origin

Determination of equality of ratios Age in years

One writer suggests data also may be characterized by their abstractness, verifiability, elusiveness, and truthfulness (closeness to the phenomenon).2

• Abstractness. If what we measure is objective (e.g., someone’s height) it is on one extreme of the abstractness continuum. We may not be able to observe and record some concepts and constructs directly, so we rely on variable substitutions. Someone’s attitude about job satisfaction is at the op- posite end of the abstract continuum.

• Verifiability. Data are processed by our senses. When sensory experiences consistently produce the same result, our data are said to be trustworthy because they have been verified.

• Elusiveness. Some data of interest occur at a speed that may make measurement during occurrence difficult. In these instances, we rely on video or audio recordings of events or people’s memories. In other instances, the effects of an observable event may be quickly gone; in those instances, we also rely on people’s memories. To capture elusive data, we need to measure as near to the event as possible.

• Truthfulness. If the data collected accurately reflect the phenomena being measured, they are con- sidered truthful. Primary data collected at the time of the event being measured are considered the closest to truth.

schi18939_ch02_026-044.indd 33 2/19/18 11:26 AM

>part I Building the Foundation for Research

34

At the American Advertising Federation’s D2 Digital Dialogue

conference Scott Franzer, associate director with OMG and re-

sponsible for P&G’s programmatic digital ad buys, shared that

P&G was experiencing “three to five times greater ROI through

programmatic buying than . . . through traditional [digital ad

buys].” Just a few years ago “the processes of ad buying and

placement [of digital media ads] were handled by humans, and

often conducted over the phone,” explained Michael Benisch,

director of artificial intelligence, Rocket Fuel Inc. But today’s

digital ads are bought via computer programmatically through

an “astonishing range of complex calculations” that no human

could do in the time available.

How does programmatic buying work? The consumer clicks

on a content link. Behind the scenes, the publisher’s computer

profiles the consumer using previous contact history to deter-

mine which ads will appeal most to that consumer and earn the

publisher the greatest advertising revenue. The publisher’s com-

puter sends out a bid, with information about the consumer’s

profile, through various ad exchanges to numerous advertisers.

Each advertiser determines if it wants to bid on that consumer.

“Machine learning algorithms can decide exactly how much they

want to bid for [each consumer’s] attention—and in some cases

they can even create the right ad and serve it on the fly,” shared

Benisch. All this happens in about 100 milliseconds.

Advertising research demonstrates that each consumer

needs “to see an ad multiple times before the message

[penetrates] to generate a response. And, for big-ticket, long

purchase-cycle items like cars, it takes even more exposures

over a longer time period. This is . . . why getting greater tar-

geting capability through programmatic is important—to reach

[the consumer] just often enough to make the sale, but not so

often that [the seller annoys them and wastes money].” So, great

results—a higher return on an advertising investment—is all about

the algorithm and the data points it considers. “[The] more data

points an algorithm [evaluates], . . . the more capable it is of learn-

ing from each bidding opportunity . . . , the smarter the system.”

What role might research play in identifying powerful data

points to create the smarter algorithm? Or is it all about following

the click data?

www.pg.com

Research and Programmatic Algorithms

>snapshot

Consistency in collection produces verifiable data, but that does not guarantee its truthfulness. The Census Bureau, which manages the largest survey in the U.S.every 10 years, conducts a sample of follow-up interviews to de - termine that households have been counted and that the data are truthful— that they accurately reflect the household being measured.

© U

.S . C

en su

s B

ur ea

u, P

ub lic

I n

fo rm

at io

n O

ffi ce

( P

IO )

schi18939_ch02_026-044.indd 34 2/19/18 11:26 AM

>chapter 2 The Research Process: An Overview

35

Data Collection Design  Selecting a data collection design is complicated by the availability of a large variety of methods, tech- niques, procedures, and protocols. For example, a researcher may decide on a survey, but she must choose it from a long list of research methodologies; other methodologies might be equally viable and the choice—of which will be used and which are rejected—must be justified. Once the method is chosen, the researcher faces other questions. Should the survey be administered via the Internet, via mobile de- vice, or through a digitally recorded personal interview? Should all relevant data be collected at one time, or would multiple measurements over a period of time be a better choice? How will the questionnaire or interview guide be structured? What instructions will be provided? How will questions be worded? Should the responses be scaled or open-ended? How will reliability and validity be achieved? Will char- acteristics of the interviewer influence responses to the measurement questions? What kind of training should the data collectors receive? These questions represent only a few of the decisions that have to be made and justified when just one method is chosen.

Although selecting an appropriate design may be complicated by the widening range of options, the creative researcher actually benefits from this array. The numerous combinations spawned by the abundance of tools may be used to construct alternative perspectives on the same problem. By creating a design using diverse methodologies, researchers are able to achieve greater insight than if they followed the most frequently used method or the method receiving the most media attention. Although pursu- ing research on a single research problem from a multimethod, multistudy strategy is not currently the norm, such designs are getting increasing attention from researchers and winning industry awards for effectiveness.3 One of the major trends in business is remote workforce collaboration; this is increasingly evident in the research process.4 Collaborating to identify several plausible competing designs should be considered before settling on a final data collection design

According to the latest Greenbook Research Industry Trends (GRIT) report, the top four emerging techniques, among both research buyers and providers all involve Internet use. Mobile surveys continue to gain increasing acceptance by both research providers (77 percent) and managers (65 percent). The use of online communities made a big move over the last four years as a major source of samples. The gaining acceptance of social media analytics and better tools for text analytics explain their respective number 3 and number 4 positions. The interest in agile research has influenced a rise in the use of micro-surveys. As Gregg Archibald, managing partner with Gen2 Advisers, notes, “Big data, Automation, and AI are all seen as game-changers” for the research industry. What’s next? “It’s not just about the tools and processes of providing insights, it is about integrating disparate sources of knowledge for gaining a more complete understanding of people and issues.” Read the report for yourself. It comes out every year free. www.greenbook.org

>picprofile 100

Mobile Surveys

Emerging Research Techniques

Online Communities

Social Media Analytics

Text Analytics

Webcam-based Interviews

Mobile Qualitative

Mobile Ethnography

Eye Tracking

Micro Surveys

Research Gamification

Facial Analysis

Prediction Markets

30 4020 60 7050 80

Research client (n=322) Research provider (n=1261)

schi18939_ch02_026-044.indd 35 2/19/18 11:26 AM

>part I Building the Foundation for Research

36

Sampling Design This subprocess answers the question: From whom or what (target population) does the data need to be collected, how, and from how many (cases)? The steps in this subprocess include:

• Define the target population.

• Define a case.

• Define the number of cases needed.

• Define the procedure for how each case will be recruited or selected.

The first step in planning the sampling design is to define the target population, those people, events, or records that possess the desired information to answer the research question. If a study’s objective is to examine the attitudes of U.S. automobile assemblers about quality improvements, the population may be defined as “the entire adult population of auto assemblers employed by the auto industry in the United States.” Using operational definitions for adult and assembler and the relevant job descriptions included under “assembly” and “auto industry in the United States” may further limit the population under study. The researcher may also restrict the research to specific companies or brands, vehicle types, or assembly processes.

Each unit in that target population is a case. Each case possesses the information on one or more variables that need to be measured. For example, using the automobile assembler study, each case is an employee. Each employee can be defined by his or her years on the job, his or her performance ratings, his or her skill set, the number of companies he or she has worked for, etc.

The number of cases we need depends on the size, variability, and accessibility of the target population. During sampling design we determine whether a sample or a census is desired. Taking a census requires that the researcher examines or counts every case in the target population. Using a sample requires examining a portion of the cases in the target population. When a target population is small, a census might be necessary; when it is large, a census might not be feasible or desired. For most business research, we use samples. We use statistical formulas for determining ideal sample sizes and types.

When researchers use a sample, they must determine a procedure for selecting which specific cases are most likely, collectively, to be representative. If our sample is drawn from records or events or machine performance data, statistical procedures to assure random selection are often employed. With people, however, accessibility may be a problem. Recruiting individuals to participate in research studies has become more difficult (higher decline rates) and more costly (higher incentives needed). As a result, researchers (1) employ sampling specialists to recruit par- ticipants or (2) use established communities of individuals who are more willing to participate in studies.5 We still must employ consistent procedures for recruiting and keeping participants engaged; the mobile devices most individuals carry today have given researchers more options to accomplish these tasks.

>Stage 3: Collect and Prepare the Data This stage contains two subprocesses: actual collection of the data and preparation of the data for analysis.

Collect the Data Questionnaires, standardized tests, observation checklists, event notes, video or audio recordings, photo- graphs, personal diaries, and instrument calibration logs are among the devices used to record primary data. Experiments, observation studies, online surveys, focus groups, photo ethnographic studies, and even research gaming are just a few of the methods used today. Increasingly, researchers combine more than one method to answer the research question(s)—for example, using a survey followed by sending a subject into a designated store to take video of what he or she is experiencing when making a choice

schi18939_ch02_026-044.indd 36 2/19/18 11:26 AM

>chapter 2 The Research Process: An Overview

37

from a product array. Each research method used in a project requires varying degrees of scientific rigor, but each must follow consistent procedures to ensure the quality of the data. It is at this stage that the methods and the procedures specified during data collection design are executed.

Prepare the Data Most data must be edited prior to analysis to ensure consistency across respondents, objects, or events and to locate and address errors and omissions. Data preparation processes address recording errors, improve legibility, and clarify unclear or unexpected responses. As data are processed by computer, the alphanumeric codes assigned to each data unit reduce responses to more manageable categories for storage and processing. These codes also assist with sorting, tabulating, and analyzing. Computers have made it possible to enter, merge, edit, and code data using fewer steps and at a faster speed, shortening the time frame of many research projects.

Processes for data preparation include more than code assignment. For example, a researcher may include early summarization of data to locate out-of-range answers and determine if coding is correct. The survey question requested Yes (code 1) or No (code 2) answers, but during data entry, some 3s were entered. Will you review the original mail survey for the correct answer or discard the data for this person on this question? What will you do if this question is so fundamental to the survey that, without it, you cannot correctly interpret other responses? Will you discard all of this person’s answers? These types of issues are addressed during this stage.

>Stage 4: Analyze and Interpret the Data Managers need information, not raw data. Researchers generate information by analyzing data. Data analysis involves reducing accumulated data by developing summaries (descriptive statistics of the vari- ables), looking for patterns by looking for relationships among variables, and applying statistical tech- niques. For example, assume a researcher polls 2,000 people from prospects for a new generation of mobile phone (target population). Each respondent responds to four questions:

• Do you prefer the convenience of Pocket-Phone over existing mobile phones?

• Are there transmission problems with Pocket-Phone?

While intercept surveys are ex- pensive due to the involvement of an interviewer and travel issues, if your target population is congregated—like at a con- cert or mall—it is sometimes the most efficient method to collect the data.

© w

ds to

ck /G

et ty

I m

ag es

schi18939_ch02_026-044.indd 37 2/19/18 11:26 AM

>part I Building the Foundation for Research

38

• Is Pocket-Phone better suited to worldwide transmission than your existing mobile phone?

• Would cost alone persuade you to purchase Pocket-Phone?

Each of these questions elicits a yes/no response, meaning their answers produce 8,000 pieces of raw data. Assume the researcher asked each respondent six demographic questions (age, gender, education, etc.); the raw data have more than tripled. If the researcher used a four-point scale rather than a yes/no response, the raw data increase again. Besides summarizing the data, data analysis also reduces the data. Reducing the original 8,000 data points produces eight statistics (for each question, the percentage of “yes” answers and percentage of “no” answers). If the researcher cross-tabulates the four answers by gen- der to look for patterns, he now has 16 statistics to interpret, as well as appropriate tests of significance to evaluate. Imagine if we ask one open-ended question (Describe an undesirable situation with your current phone?). A researcher would have to identify keywords, as well as sentiments, further expanding the data.

At the end of the data analysis process a researcher will often have reams of statistical information, but he or she must whittle this down to key findings (key patterns in the data). When the researcher interprets the findings in light of the manager’s research questions, he or she develops insights. Insights determine whether the hypotheses and theories are supported or refuted. Increasingly, managers ask researchers to make recommendations (actions the manager should take that are empirically supported) based on their interpretations of the data.

>Stage 5: Report Insights and Recommendations In this stage of the process, the researcher reports and transmits the findings, insights, and any recom- mendations to the manager. The researcher adjusts the type of reporting—and any style and organization of the report—to the target audience, the occasion, the method of presentation, and the purpose of the research. The results may be communicated via formal oral presentation, conference call, letter, detailed written report, top-line summary, infographic, or some combination of any or all of these methods. Re- ports are developed from the manager’s or information user’s perspective and from whether and how the information is to be shared within or outside the organization. The manager’s foremost concern, solving the management dilemma, drives this stage.

A manager might receive a report but not take action. Inferior communication of results might be a primary reason. Managers don’t always have the same statistical knowledge as a researcher, so careful choice of words, and providing explanations as needed, is important. Inaction after a project might also be caused by environmental factors beyond the researcher’s control—a shift in business strategy, a change in a competitor’s strategy, the disappearance of a dilemma, etc. For example, members of the Associa- tion of American Publishers wanted an ad campaign to encourage people to read more books. One find- ing from the $125,000 study, however, found that only 13 percent of Americans bought general-interest books in stores. When the time came to commit the $14 million to the ad campaign to raise book sales, member interest had faded and the project died.6

>The Research Project Time Frame At its core, research is a time-based project, and project management tools can be used to depict se- quential and simultaneous steps and estimate scheduling and timetables for each activity or phase of the research. Exhibit 2-3 shows the use of the critical path method (CPM) for a research project. The pathway from start to end that takes the longest time to complete is called the critical path. Any delay in an activity along this path will delay the end of the entire project. Researchers also use spreadsheets and Gantt charts (see Exhibit 2-4) to manage the various research activities. The most important lessons in using any planning tool are to make sure all tasks are included and that each task is tracked in terms of time so that time lost in one task is recaptured during another. Tracking project task timing is valuable in planning future research projects.

schi18939_ch02_026-044.indd 38 2/19/18 11:26 AM

>chapter 2 The Research Process: An Overview

39

>Exhibit 2-3 CPM Schedule of Research Design

Complete questionnaire

Field interviewing

Code data and analyze

Write report

Meet client

Client interviews

Conduct exploratory interviews

Develop proposal

Literature review

1

2

35

6 10

3 4 7 8 3 8 8

6

4 1

5

Complete executive questionnaire

Arrange executive appointments

Interview executives

3 2

3

Final check of background records

9

Milestones:

3 Proposal approval 7 Interviews completed 9 Final report completed

EndStart

Critical Path: S–1–3–4–7–8–9–E

Time to Completion: 40 working days

>Exhibit 2-4 Gantt Chart of a Research Project Day 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

Client Interviews Literature Review Exploratory Interviews Justify Research Design Submit Proposal for Design Receive Approval for Design Develop Executive Interview Guide Arrange Executive Interviews Conduct Executive Interviews Executive Interviews Completed Complete Questionnaire Survey in Progress Survey Completed Prepare Data Analyze Data Prepare Report(s) Report Completed Deliver Oral Report

>Research Process Pitfalls and Solutions Although it is desirable for research to be thoroughly grounded in management decision priorities, stud- ies can wander off target or be less effective than they should be. Here are some pitfalls to watch for and how each can be addressed.

Ill-Defined Management Problems Some research questions are so complex, value laden, and bound by constraints that they don’t lend themselves to traditional forms of analysis. These questions or problems have characteristics that are vir- tually the opposite of well-defined problems. They have too many interrelated facets for measurement to handle with accuracy.7 Or methods do not presently exist, or if the methods were to be invented, they still might not provide the data necessary to solve the problem.8 One author describes the differences like this:

To the extent that a problem situation evokes a high level of agreement over a specified community of problem solvers regarding the referents of the attributes in which it is given, the operations that are permitted, and the consequences of those operations, it may be termed unambiguous or well defined with respect to that community. On the other hand, to the extent that a problem evokes a highly variable set of responses concerning referents of attributes, permissible operations, and their consequences, it may be considered ill-defined or ambiguous with respect to that community.9

schi18939_ch02_026-044.indd 39 2/19/18 11:26 AM

>part I Building the Foundation for Research

40

Spending sufficient time on exploration and problem definition may address this issue, but some ill- defined research questions are really unresearchable ones.

Unresearchable Questions Not all questions are researchable. To be researchable, a question must (1) be one for which data col- lection and analysis can provide the answer, (2) be answerable with current data collection techniques, and (3) allow for some units within target population the to be accessible and willing to share the data. And some questions don’t meet these criteria. There is little a researcher can do if a question can’t be addressed with data. However, if he overcomes this first issue, then the researcher would need to develop new research designs and data collection procedures to address the last two issues.

Overembracing Data Analytics If you have a researchable problem, the existence of big data can seduce a manager, seemingly eliminat- ing the need for dilemma-specific research. Just because substantial data is available does not mean the modern management information system provides substantial knowledge. Each field in a database was originally created for a specific reason and drawn from a specific source. These decisions may or may not be compatible with the management dilemma facing the organization today. Data are also often found in unconnected silos, making extracting desired information difficult. While accessing such data is often a starting point in decision-based research, rarely will such activity answer all questions related to a par- ticular research question. Researchers can avoid this problem by pointing out the lack of compatibility to the current dilemma of the source, time frame, and operational definitions of data variables on which the manager (or his or her superior) wants to rely.

The Manager’s Hidden Agenda(s) It is important to remember that all of a manager’s motivations for seeking research are not always obvious to a researcher. Managers are often rewarded (higher pay, better office location, more status or prestige) for building empires of authority. In the ideal scenario for quality research, a manager ex- presses a genuine need for specific information on which to base a decision. Sometimes, research may be authorized even though a decision has been made; it’s seen as a measure of protection for a manager in case he or she is criticized later. At other times, a research study is focused on a less-attractive action option; while not justified, such a study may win approval because it supports a manager’s pet idea. Researchers facing hidden agendas often experience pressure to pursue research questions or specific research designs that, with analysis, seem inappropriate. Researchers avoid hidden agendas by asking the manager a lot of questions when clarifying the research question; the more questions asked, the more true motivations are revealed.

Favored-Technique Syndrome Not all researchers are comfortable with every plausible methodology, so they stay within their comfort zone. These researchers recast any research question so that it is amenable to their favored methodology— a survey, for example. Avoiding this syndrome is as simple as demanding that the researcher justify his or her selection of a research design and rejection of all alternative designs.

Researcher Inexperience Finally, lack of experience can create numerous pitfalls. Researchers may be tempted to rely too heav- ily on exploratory data to help define the research question or interpret the data in a new study. In

schi18939_ch02_026-044.indd 40 2/19/18 11:26 AM

>chapter 2 The Research Process: An Overview

41

the absence of sufficient data, researchers sometimes substitute personal experiences to form their insights, thereby misdirecting conclusions. Researchers may be tempted to project the experiences from a limited sample to the population as a whole, thereby overstating the importance of a finding. Assigning a research project to a researcher with sufficient experience—or assigning a more seasoned researcher to oversee more novice researchers—goes a long way to avoiding the problems

Offshoring is defined as the movement of a process done

at a company in one country to the same or another com-

pany in a different country. These processes in research

often include IT (information technology), business, or knowl-

edge processes. Researchers use offshoring to reduce

costs, speed a research time frame, access hard-to-recruit

research participants, or obtain specialized or proprietary

methodologies.

There are significant risks associated with offshoring re-

search services, as Gordon Morris, global insights manager

for Sony Ericsson (London, UK), discovered. During a global

research project for the Experia X10 Android phone, sensitive

business plans were shared with the contracted researchers.

As a result of offshoring of some research services, information

about the phone launch leaked several months early. “We esti-

mate the potential damage caused by the leak at approximately

£100 milion,” shared Morris. Leaks may be more likely to occur

when offshoring is used because intellectual property standards,

safe computing standards, and contract laws vary from country

to country. High employee turnover in some developing coun-

tries can also add to this risk.

A survey fielded by the Foundation for Transparency in

Offshoring (FTO) among 850 U.S. and international research

buyers and providers indicated that clients were more likely to

think their projects did not involve offshoring than was actually

true. Also, clients were much more likely than research services

suppliers to think clients should be told about offshoring. “Very

few buyers have sufficient information to assess the relative

strengths and risks associated with offshoring,” said Tom H. C.

Anderson, FTO founder and chairman and managing partner of

the research consultancy Anderson Analytics.

As Sonia Baldia, partner at Mayer Brown LLP and a legal ex-

pert on offshoring, explains, “Clients absolutely need to know

about any offshore subcontracting and [its] location in order to

risks and protect themselves.”

www.offshoringtransparency.org

Is Your Research Leaving the Country?

>snapshot

Clients

P e

rc e

n t

o f

‘Y e

s’ R

e sp

o n

se s

92

69 62

74

100

Q. Regardless of what your stance is on o�shoring, do you believe research suppliers have an obligation to tell their clients that they o�shore? (n=850)

90

80

70

60

50

40

30

20

10

0 All

Suppliers O�shoring Suppliers

Non-o�shoring suppliers

Source: ©Foundation for Transparency in Offshoring

schi18939_ch02_026-044.indd 41 2/19/18 11:26 AM

>part I Building the Foundation for Research

42

of inexperience. Managers today are pressured to do more of their own research; they need to be forthcoming when their knowledge and skill is insufficient to conduct a needed research project and recruit or hire the necessary expertise.

> Research Process Ethical Issues and Responsibilities

Research invades workers’ and people’s lives at some level, so appropriate behavior is always a concern. It also accesses proprietary and often sensitive records. Exhibit 2-2 links basic ethical rights to their appropriate stage(s) in the process.

Managers, researchers, and participants all have the responsibility to behave ethically. The ethical issues focus on deception, privacy, quality, notice, choice, access, security, respect, and bias. For the participant, his or her primary ethical responsibilities are truthfulness (researcher’s right to absence of participant deception) and completion of research tasks (researcher’s right to quality research).

For the manager, their ethical responsibilities are record accuracy (researcher’s right to quality re- search), purpose transparency and truthfulness (researcher’s right to absence of sponsor deception), discouraging falsification of research results (researcher’s right to absence of sponsor coercion), and fulfilling contractual obligations (researcher’s right to adequate compensation for results).

The researcher has the largest number of responsibilities because they must serve the needs of the participant as well as the manager. For the participant, these responsibilities include obtaining informed consent as required (participant’s right of informed consent), accepting the subject’s choice on whether to participate (participant’s right to privacy), following appropriate safety procedures (participant’s right to safety), and keeping participant identity confidential (participant’s right to confidentiality). For the manager, these responsibilities include protecting manager confidentiality (sponsor’s right of purpose nondisclosure and sponsor’s right of sponsorship nondisclosure); following acceptable industry practices and procedures (sponsor’s right to quality research); and protecting data, findings, insights, and recommendations (sponsor’s right to findings nondisclosure).

LO2-1       What characterizes good research? Generally, one expects good research to be purposeful with a clearly defined focus and plausible goals; with defensible, ethical, and repeatable procedures; and with evidence of objectivity. The reporting of procedures—their strengths and weaknesses—should be complete and honest. Appropriate analytical techniques should be used; insights and conclusions drawn should be limited to those clearly justified by the findings; and reporting of findings, insights, and conclusions should be clearly pre- sented and professional in tone, language, and appear- ance. Managers should always choose a researcher who has an established reputation for meeting these criteria.

LO2-2 The research process is a sequential process involving clearly defined stages: clarify the research question, design the research, collect and prepare data, analyze and interpret data, and report insights and recommen- dations. Stage 1: Clarify the Research Question starts

with a management dilemma, either a problem or an opportunity. Exploration reveals what managers know and don’t know. Managers must prioritize dilemmas, choosing to focus on the one that is most important. It is also during this stage that managers justify that research will yield a better decision and develop a budget for a more formal study. Stage 2: Design the Research Project determines the sampling design and the data collection design. Sampling design determines the parameters of interest in a target population, whether a census or sample will be conducted, the procedures for selecting and recruiting each case, and the number of cases. Data collection design determines the type of data to be collected, either primary or secondary or some combi- nation, the method, and the specific procedures for col- lecting it. Stage 3: Collect and Prepare the Data involves the actual data collection (including measurement instrument use and any training involved with using it),

>summary

schi18939_ch02_026-044.indd 42 2/19/18 11:26 AM

>chapter 2 The Research Process: An Overview

43

case 36

census 36

dashboard 31

data 32

data analysis 37

findings 38

insights 38

interval data 33

key performance indicator (KPI) 30

management dilemma 29

nominal data 33

ordinal data 33

primary data 33

ratio data 33

recommendations 38

research design 32

research process 27

research question(s) 32

sample 36

secondary data 33

target population 36

>keyterms

as well as the preparation of data (to include dealing with out-of-range and missing data). Stage 4: Analyze and Interpret Data involves data summarization; looking for patterns among variables; and applying statistical techniques with the purpose of discovering findings and insights and, ultimately, making recommendations in relation to the research question and management di- lemma. Stage 5: Report the Results involves determining the audience(s), method, timing, and location to report the findings, insights, and recommendations.

LO2-3 Studies can wander off target or be less effective if the researcher experiences any of these pitfalls: ill-defined research question(s), unresearchable questions, overe- mbracing data analytics, a manager’s hidden agenda(s), application of favored-technique syndrome, or researcher inexperience. The ways to address these include a thor- ough exploration phase in Stage 1, avoiding questions where data aren’t helpful, pointing out the lack of compatibility of stored data with the current dilemma,

thorough interviewing of the manager for his or her motivations, seeking a justification for any research de- sign that includes why alternative designs were rejected, and seeking the researcher with sufficient experience.

LO2-4         Ethical research process issues focus on deception, privacy, quality, notice, choice, access, security, respect, and bias. A participant’s primary ethical responsibilities are truthfulness and completion. A manager’s responsibilities are accuracy, transparency, truthfulness, discouraging the falsification of results, and fulfilling contractual obligations. The researcher has the most ethical responsibilities,and these center around both participants and managers. For participants, a researcher is responsible for participants’ safety and absence of harm, consent, privacy and confidentiality. For man- agers, a researcher is responsible for confidentiality, following acceptable industry standards for quality research, and protecting data, as well as its findings, insights, and recommendations. 

Terms in Review 1 Distinguish between the following and identify the research

use of each.

a primary and secondary data

b finding and insight

c census and sample

d management dilemma and research question

e finding and insight

2 Kathy Lee Berggren, a professor of oral communication at Cornell University, indicates “a lot of my students really [only] scratch the surface with the type of research they’re doing.” According to Andy Guess, at Inside Higher Ed, “Just because students walk in the door as ‘digital natives,’ doesn’t mean they’re equipped to handle the heavy lifting of digital databases and proprietary search engines that

comprise the bulk of modern, online research techniques. Students erroneously think a Google search is research.” As you evaluate the factors that guarantee good research, what actions do you propose to narrow the gap between students’ research competence and what is required of a modern college graduate about to become a manager with research responsibilities?

Making Research Decisions 3 SupplyCo is a supplier to a number of firms in an indus-

try. Several companies are involved in the manufacturing process—from processed parts to creation of the final product—with each firm adding some value to the product. SupplyCo provides materials and services to every part of the chain. By carefully study its customer data, SupplyCo discovers a plausible new model for manufacturing and dis- tributing industry products that would increase the overall

>discussionquestions

schi18939_ch02_026-044.indd 43 2/19/18 11:26 AM

>part I Building the Foundation for Research

44

efficiency of the industry, reduce costs of production, and result in greater sales and profits for some of the industry’s participants (SupplyCo’s customers). Implementing the model might hurt the sales and profits of some SupplyCo customers that did not have the manpower, plant or equip- ment, or financial resources to participate in the change. These companies might potentially go out of business, decreasing SupplyCo’s sales and profits in the short term. SupplyCo is considering research to further study the new model. SupplyCo. had the data within its warehouse only because of its relationship with its customers.

a Is it a betrayal to use the data in a manner that would cause some customers harm? Or does SupplyCo have a more powerful obligation to its stockholders and employ- ees to aggressively research the new model?

b What are the implications of your answers even if there is no violation of law, industry regulation, or company policy?

4 Based on an analysis of the last six months’ sales, your boss notices that sales of beef products are declining in your chain’s restaurants. As beef entrée sales decline, so do profits. Fearing beef sales have declined due to several newspaper stories reporting E. coli contamination discov- ered at area grocery stores, he suggests a survey of area restaurants to see if the situation is pervasive.

a What do you think of this research suggestion? 

b How, if at all, could you improve on your boss’s formula- tion of the research question?

5 You have received a business research report done by a con- sultant for your firm, a life insurance company. The study is a survey of customer satisfaction based on a sample of 600. You are asked to comment on its quality. What will you look for?

6 As area sales manager for a company manufacturing and marketing outboard engines, you have been assigned the

responsibility of conducting a research study to estimate the sales potential of your products in the domestic (U.S. or Canadian) market. Discuss key issues and concerns arising from the fact that you, the manager, are also the researcher.

From Concept to Practice 7 Using Exhibit 2-2 and case examples from some research

company’s website, discover if favored-technique ap- proaches to research design dominate a firms’ activities.

8 Using Exhibit 2-2, find a case study of a research example in which a clear statement of the management dilemma leads to a precise and actionable research. (Hint: Visit research company websites; use a search engine to find examples.)

From the Headlines 9 Hearsay Social monitors how workers at large companies

interact with outsiders on Facebook, LinkedIn, and other social media sites. Companies pay $100,000 or more for a digital dashboard that alerts supervisory managers if employees are violating privacy policies, regulatory compli- ance rules, or other company policies. If you were a large client, like Farmers Insurance, what information would you want to know?

10 Researchers recently reported in Cell Metabolism that a natural ingredient found in milk can protect mice against obesity even while they enjoyed diets that were high in fat. Mice given high doses of nicotinamide riboside along with their high-fat meals burned more fat. Their muscles worked better, providing them better endurance and making them better runners. If you worked for the National Milk Produc- ers Federation, could you justify a research project to de- termine the value of a national ad campaign to promote the qualities of milk in a balanced diet and for avoiding obesity based on this information found during exploration? If not, develop an argument to justify a research study.

>cases*

Akron Children’s Hospital

Calling Up Attendance

Covering Kids with Health Care

Donatos: Finding the New Pizza

Goodyear’s Aquatred

HeroBuilders.com

Inquiring Minds Want to Know—NOW!

Lexus SC 430

Mastering Teacher Leadership

NCRCC: Teeing Up a New Strategic Direction

Ohio Lottery: Innovative Research Design Drives Winning

Ramada Demonstrates Its Personal Best™

State Farm: Dangerous Intersections

USTA: Come Out Swinging

*You will find a description of each case in the Case Abstracts section of this textbook. Check the Case Index to determine whether a case provides data, the research instrument, video, or other supplementary material. Cases and case supplements are available in Connect.

schi18939_ch02_026-044.indd 44 2/19/18 11:26 AM

LO3-1 The question hierarchy.

LO3-2 The purposes and process of exploration.

LO3-3 How internal and external exploration differ.

LO3-4 The process and goal of research valuation.

LO3-5 The process and justification needed to budget for research.

LO3-6 Ethical issues at this stage of the process.

>learningobjectives After reading this chapter, you should understand . . .

“A beautiful question is an ambitious yet

actionable question that can begin to shift the way we perceive

or think about something—and that might serve as a catalyst to

bring about change.” Warren Berger,

consultant and author, A More Beautiful Question

Stage 1: Clarify the Research Question

>chapter 3

schi18939_ch03_045-068.indd 45 2/19/18 11:28 AM

>part I Building the Foundation for Research

46

>Stage 1 in the Research Process

>Exhibit 3-1 Tasks Associated with Stage 1 of Research Process

Define Exploration Strategy

Define Management-Research Question Hierarchy

Value the Research

Justify & Budget for Research

Stage 1 uses exploratory research to accomplish the following tasks as depicted in Exhibit 3-1:

1. Define the management–research question hierarchy.

2. Define and execute the exploration strategy.

3. Value the research.

4. Justify and budget for research.

>The Question Hierarchy A useful exercise to start Stage 1 of the research process (Exhibit 3-2) is to state the basic management dilemma (the problem or opportunity that requires a decision), then try to develop other questions by progressively breaking down the original question into more specific ones. You can think of the outcome of this process as the management–research question hierarchy.

The process takes the manager and his or her research collaborator through various brainstorming and exploratory research exercises to define the following:

• Management question—This is a restatement of the manager’s dilemma(s) in question form. Think of it like a query about what should be done about the dilemma. Applying exploratory research to this question often reveals several options and information about whether an option might address the management dilemma.

• Research questions—These questions explore the various options and forms one or more hypoth- eses that best states the objective of the research; these questions focus the researcher’s attention. Exploratory research helps identify the options and the hypotheses.

• Investigative questions—These are questions the researcher must answer to satisfactorily answer the research question; what the manager feels he or she needs to know to arrive at a conclusion about the management dilemma. Exploratory research will identify what information is known and what is unknown.

The final question in the hierarchy isn’t addressed until Stage 2: Research Design.

• Measurement questions—These are questions that participants in research are asked or what specifically is observed in a research study.

This multistep process is presented in Exhibit 3-2 and in the example in Exhibit 3-3. The role of exploration in this process is depicted in Exhibit 3-4.

schi18939_ch03_045-068.indd 46 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

47

>Exhibit 3-3 SalePro’s Management–Research Question Hierarchy Declining sales is one of the most common symptoms serving as a stimulus for a research project, especially a continuing pattern that is unexplained. SalePro, a large manufacturer of industrial goods, faces this situation. Exploration (1) reveals that sales, in fact, should not be declining in the South and Northeast. Environmental factors there are as favorable as in the growing regions. Subsequent exploration (2, 3) leads management to believe that the problem is in one of three areas: salesperson compensation, product formulation, or trade advertising. Further exploration (4) has SalePro management narrowing the focus of its research to alternative ways to alter the sales compensation system, which (5) leads to a survey of all sales personnel in the affected regions.

Why are our sales declining in the South and Northeast, while sales are booming in the Southwest?

How can we improve sales in the South and Northeast?

Should we introduce a 2 percent incentive commission-based compensation system on all sales over quota for salespeople in the South and Northeast or a 5-percent-of-profit regional bonus to the region that increases sales by 10 percent over quota (to be shared proportionately among the salespeople in the region)? Should we modify the product formula for distribution in the South and Northeast? Should we increase the level of advertising via trade publications in South and Northeast editions?

Please rate your level of concern for each of the following outcomes if management were to change your compensation to a commission-based system compared to the current salary system. For each outcome, indicate a number between 1 and 7 where 7 = extreme concern, 4 = neither concerned nor unconcerned, and 1 = no concern at all.

______ Lack of predictability of monthly pay. ______ Increased internal competition for sales prospects. ______ Reduced time for postsale servicing of customer needs. ______ Reduced incentive for postsale servicing of customer needs.

4

3

2 1

What is the likelihood that we will lose excellent salespeople in the South and Northeast if we implement the compensation change? What is the likelihood that current customer satisfaction in these regions will decrease? What is the likelihood that future sales to existing customers will be lost?

>Exhibit 3-2 Management–Research Question Hierarchy

What is the recommended course of action, given the research insights?

What should be asked or observed to obtain the information the manager needs?

What does the manager need to know to choose the best alternative from the available courses of action?

What symptoms cause management concern? What environmental stimuli raise management interest?

Management Dilemma

What plausible courses of action are available to management to correct the problem or take advantage of the opportunity, and which should be considered?

How can management eliminate the negative symptoms? How can management fully capitalize on an opportunity?

Management Question 1

Investigative Questions 3

Measurement Questions 4

Management Decision

Research Question(s) 2

schi18939_ch03_045-068.indd 47 2/19/18 11:28 AM

>part I Building the Foundation for Research

48

The Management Question The management question restates the management dilemma in question form; the four types are de- tailed in Exhibit 3-5. Here’s an example. A bank’s president is concerned about low deposit growth, possibly the result of changes in their competitive environment. While lowered deposits directly affect profits, another part of the profit weakness is revealed in increasing customer complaints. The qualified researcher knows that the management question as originally stated is too broad to guide a definitive business research project. As a starting point, the broadly worded question “What can we do to increase profits?” might be fine, but the bank will want to refine its management question into these more specific subquestions:

• How can we improve deposits?

• How can we improve internal operations that currently result in customer complaints?

Discover Management

Dilemma

1 Define

Management Question

2 Define

Research Question(s)

This may be either a problem or an opportunity. At this stage you may even

have identified symptoms rather than problems or opportunities.

Using collected exploratory information, you word the dilemma or the correction of the symptom in question form, usually

starting with “How can the organization . . .?”

Several research questions may be formulated at this stage. Each question is

an alternative action that management might take to solve the management dilemma. Usually the most plausible

action, or the one that o�ers the greatest gain using the fewest resources, is

researched first.

Exploration At this stage you review published sources and

interview information gatekeepers to understand the true management dilemma, not just its symptoms.

1 Exploration The purpose of this stage is to clarify the possible

management actions that might be taken to solve the management dilemma. This stage usually involves interviews

with information gatekeepers, brainstorming with experts, and other qualitative research techniques.

>Exhibit 3-4 Formulating the Research Question, the Process

schi18939_ch03_045-068.indd 48 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

49

This separation of the management question into two subquestions may not have occurred without a discussion between the researcher and the manager.

Assume the bank has done no formal business research in the past. It has little specific information about competitors or customers and has not analyzed its internal operations. To move forward in the management–research question hierarchy and define the research question, the client needs to collect some exploratory information:

• What factors are contributing to the bank’s failure to achieve a stronger growth rate in deposits?

• How well is the bank doing regarding customer satisfaction and financial condition compared to industry norms and competitors?

Small-group discussions are conducted among employees and managers, and trade association data are acquired to compare financial and operating statistics from company annual reports and end-of-year division reports. From the results of these two exploratory activities, it is obvious that the bank’s opera- tions are not as progressive as its competitors’, but it has its costs well in line. So the revised management question becomes “What should be done to make the bank more competitive?” The process of explora- tion will be critical in helping the bank identify its options.

In addition to solving problems, managers are likely to be looking for opportunities. So let’s look at another case, TechByte. This company is interested in enhancing its position in a given technology that appears to hold potential for future growth. This interest or need might quickly elicit a number of questions:

• How fast might this technology develop?

• What are the likely applications of this technology?

• What companies now possess this technology, and which ones are likely to make a major effort to obtain the technology?

• How much will the new technology absorb in resources?

• What are the likely payoffs?

In the preceding exploration of opportunities, researchers would probably begin with specific litera- ture, looking only for certain aspects, such as recent developments, predictions by informed individuals about the prospects of the technology, identification of those involved in the area, and accounts of suc- cessful ventures or failures by others in the field. After becoming familiar with the literature, researchers might seek interviews with scientists, engineers, and product developers who are well known in the field.

>Exhibit 3-5 Types of Management Questions

Categories General Question Sample Management Questions

Evaluation of solutions (choices between concrete actions to solve problems or take advantage of opportunities)

● How can we achieve the objectives we have set?

● Should we reposition brand X as a thera- peutic product from its current cosmetic positioning?

● How can we improve our program for prod- uct repairs and servicing?

Choice of purpose (choice of objectives) 

● What do we want to achieve? ● What goals should XYZ try to achieve in its next round of union negotiations?

● What goals should we set for sales and profits in the next 5 years?

● What goals should XYZ set for its return on its investment portfolio?

Troubleshooting (diagnosing ways an organization is failing to meet its goals)

● Why is our (blank) program not meeting its goals?

● Why does our department have the lowest sales-to–web page visit ratio?

● Why does our recruiting program generate the lowest hire-to-offer ratio in the industry?

Control (monitoring or diagnosing ways an organization is failing to meet its goals)

● How well is our (blank) pro- gram meeting its goals?

● What is our product line’s sales-to- promotion cost ratio?

schi18939_ch03_045-068.indd 49 2/19/18 11:28 AM

>part I Building the Foundation for Research

50

They would give special attention to those who represent the two extremes of opinion in regard to the prospects of the technology. If possible, they would talk with persons having information on particularly thorny problems in development and application. Of course, much of the information will be confiden- tial and competitive. However, skillful investigation can uncover many useful indicators.

An unstructured exploration allows the researcher to develop and revise the management question and determine what is needed to secure answers to the proposed question.

The Research Question One task of a researcher is to assist the manager in formulating a research question that fits the management dilemma. A research question best states the objective of the business research study. It is a more specific question from the manager’s perspective that must be answered. It may be more than one question or just one. Incorrectly defining the research question is the fundamental weakness in the business research process. Time and money can be wasted studying an option that won’t help the manager rectify the original dilemma.

Assume the bank’s president has agreed to have the business research be guided by the following research question: Should the bank position itself as a modern, progressive institution (with appro- priate changes in services and policies) or maintain its image as the oldest, most reliable institution in town?

Glenn Kelman, the CEO of Redfin, recently wrote a blog post

around the provocative question, “What will America housing

be like when all the cars leave?” He was alluding to several

technical changes and changes in the behavior of Millennials.

One research study shows that younger generations have not

embraced the car-dependent, large-square-foot house in the

suburbs that their parents and grandparents enjoy. Instead,

these younger homebuyers are seeking smaller, urban spaces

nearer to restaurants and entertainment options that are closer

to work—or as Redfin’s company suggests, houses with a great

Walk Score. The success of such television shows as HGTV’s

Small Space, Big Style and Property Virgins also is testament to

that conclusion.

A study by the Transportation Research Center at the Uni-

versity of California estimated that there are approximately eight

parking spaces for every car—that inventory of parking not only

includes where the owner resides, but also where he or she

works, shops, eats, etc. That’s approximately 2 billion spaces. Ad-

ditionally, we have the technology for self-driving cars, and such

cars will be constantly moving—not taking space in a garage or

parking lot. According to Tesla CEO Elon Musk and Google, the

reality of self-driving cars may be only two to five years away. Re-

member the battle between NYC and Uber? A research study of

taxis and Uber rides showed that Uber serves the boroughs out-

side of Manhattan better than taxis (22 percent of Uber pickups

vs. 14 percent of taxi pickups), and many of Uber’s strongholds

were far from subways or a long transit trip away from Midtown.

So, in an environment of fewer or no cars per housing unit,

of self-driving cars, of transportation seen as a service rather

than an asset, will we need garages? If Millennials are eating out

more, will they need elaborate kitchens? How large—or small—

will houses need to be to satisfy this group? If you were a home

builder serving any metro area, what would your management–

research question hierarchy look like?

Housing and Millennials

>snapshot

©Image Source/Getty Images

schi18939_ch03_045-068.indd 50 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

51

Fine-Tuning the Research Question

The term fine-tuning might seem an odd choice for research, but it creates an image that most research- ers come to recognize. Fine-tuning the question is precisely what a skillful practitioner must do after the exploration is complete. At this point, a clearer picture of the management and research questions begins to emerge. After the researcher does a preliminary review of the literature, a brief exploratory study, or both, the project begins to crystallize in one of two ways:

1. It is apparent that the question has been answered and the process is finished (no further research is needed).

2. A question different from the one originally addressed has appeared.

The research question does not have to be materially different, but it will have evolved in some fash- ion. This is not cause for discouragement. The refined research question(s) will have better focus and will move the business research forward with more clarity than the initially formulated question(s).

In addition to fine-tuning the original question, the researcher should address other research question–related activities in this phase to enhance the direction of the project:

1. Examine the variables to be studied. Are they satisfactorily defined? Have operational definitions been used where appropriate?

2. Review the research questions with the intent of breaking them down into specific second- and third-level questions.

3. If hypotheses (tentative explanations) are used, make sure they are appropriately stated.

4. Determine what evidence must be collected to answer the various questions and hypotheses.

5. Set the scope of the study by stating what is not a part of the research question. This will establish a boundary to separate contiguous problems from the primary objective.

Investigative Questions Investigative questions represent the information that the decision maker needs to know; they are the questions the researcher must answer to satisfactorily arrive at a conclusion about the research question. To study the market, the researcher working on the bank project develops two major investigative ques- tions. Each question has several subquestions. These questions provide insight into the lack of deposit growth:

1. What is the public’s position regarding financial services and their use? a. What specific financial services are used? b. How attractive are various services? c. What bank-specific and environmental factors influence a person’s use of a particular service?

2. What is the bank’s competitive position? a. What are the geographic patterns of our customers and of our competitors’ customers? b. What demographic differences are revealed among our customers and those of our

competitors? c. What descriptive words or phrases does the public (both customers and noncustomers)

associate with our bank? With our competitors? d. How aware is the public of our bank’s promotional efforts? e. What opinion does the public hold of our bank and its competitors? f. How does growth in services compare among competing institutions?

Measurement Questions Measurement questions are the actual questions that researchers use to collect data in a study. They could become questions on a survey or elements on an observation checklist. Measurement questions may or may not be defined during Stage 2: Research Design, but topics for such questions and even actual sample questions may be discovered during exploration.

schi18939_ch03_045-068.indd 51 2/19/18 11:28 AM

>part I Building the Foundation for Research

52

To assist in the formation of the management–research question hierarchy, and later in the tasks related to valuing research and budgeting for research, an exploration strategy must answer these questions:

• What information sources should be used? These sources are designed to reveal information to help the manager understand the dilemma, identify courses of action (options), and clarify concepts, constructs, operational definitions, and variables needed to frame the research question and beyond

• What methods will be used to extract information from these sources? The methods detail how manager will communicate with or locate a desired source.

Exploration is loosely structured research but with a defined purpose. It involves some or all of the following: studying company records; tapping into company databases; exploring professional, scientific, and academic literature; and consulting experts, as well as those involved with the problem or its solu- tion. Through exploration, researchers:

• Understand the management dilemma, management problem, and research question.

• Discover how others have addressed and/or solved problems similar to the current management dilemma or management question.

• Establish priorities for dilemmas and problems to be addressed.

• Identify action options and develop hypotheses.

• Gather background information on the topics in order to refine the research question.

• Identify information that should be gathered to formulate investigative questions.

A Generation X software engineer wanted a last family vaca-

tion before his oldest child headed off to college. The family

chose Europe; first a quick stop to visit friends in London, then

on to Paris. Among their criteria for Parisian lodging: a central

location, access to laundry and kitchen facilities, and space for

four for four nights. They rejected established hotels as too

stuffy or too sleek. Their choice for booking accommodations:

Airbnb.

Airbnb is an app and website that connects people seek-

ing lodging with renters who have listed their personal houses,

apartments, and guest rooms on either platform. The company

boasts more than 1.5 million listings in 34,000 cities and more than

190 countries, as well as more than 40 million total guests. The

idea for Airbnb came to two guys who didn’t have jobs but did

have an apartment for which they needed rent money. During

a convention period in their hometown of San Francisco, when

standard hotel rooms were at a premium or totally unavailable,

they rented out the floor space and air mattresses to three des-

perate conventioneers. Some investors later, Airbnb was born.

Warren Berger, in his book A More Beautiful Question:

The Power of Inquiry to Spark Breakthrough Ideas, notes

that innovative ideas are the result of asking questions that

challenge assumptions. In the case of Airbnb, that assump-

tion would be that visitors to San Francisco would not want

to stay in people’s homes or apartments but would prefer to

stay in recognized, established hotels. “It’s one thing to see a

problem and to question why the problem exists, and maybe

even wonder whether there might be a better alternative,”

explains Berger. “It’s another to keep asking those questions,

even after experts have told you, in effect, ‘You can’t change

this situation; there are good reasons why things are the way

they are.’”

Berger suggests any question creates an unresolved tension

in our minds, demonstrates what we don’t know, organizes our

thinking, and enables us to act in the face of uncertainty. But a

beautiful question  is phrased in such a way that it (1) reveals a

problem or opportunity, (2) challenges existing assumptions,

(3) shifts people’s thinking, and (4) serves as a catalyst for

change/action.

How might the creators of RedBull, Zipcar, Uber, the mobile

phone, or the suitcase with wheels have phrased their beautiful

question?

The Art of Asking the Right Questions

>snapshot

>The Exploration Strategy

schi18939_ch03_045-068.indd 52 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

53

• Develop operational definitions for concepts, constructs, and variables.

• Provide the evidence for justifying and budgeting research.

• Set the foundation for the final research design.

• Exploration can also • Identify sources for and actual questions that might be used as measurement questions in any

subsequent research. • Identify sources for actual sample frames that might be used in sampling design.

A thorough exploration strategy comprises subtasks (Exhibit 3-6): (1) examination of published sources, (2) examination of digital sources, and (3) interviews with knowledgeable people. The first two are discussed in extensive detail in the remainder of the chapter. The third is explored more thoroughly here.

Information Sources In identifying sources, a good researcher will look at several classifications of information and use both internal and external sources within those classifications (see Exhibit 3-7). Some information has more value than others. Information sources are generally categorized into three levels (1) primary sources, (2) secondary sources, and (3) tertiary sources. As the source levels indicate, primary sources have more value than secondary sources, and secondary sources have more value than tertiary sources.

Primary sources are original works of research or raw data without interpretation or pronouncements that represent an official opinion or position. Included among the primary sources are memos; letters; complete interviews or speeches (in audio, video, or written transcript formats); laws; regulations; court decisions or standards; and most government data, including census, economic, and labor data. Other internal sources of primary data are inventory records, personnel records, purchasing requisition forms, statistical process control charts, and reports of earlier research. A researcher or manager seeking pri- mary data can look internally or externally (e.g., the U.S. government offers extensive primary research useful for business).

Secondary sources are interpretations of primary data. Encyclopedias, textbooks, handbooks, maga- zine and newspaper articles, and most newscasts are considered secondary information sources. Indeed, many of the secondary reference materials businesses use fall into this category. Internally, dashboards of key performance indicators (KPIs) and investor annual reports would be examples of secondary

>Exhibit 3-6 Exploration Strategy Methods

Interviews (Individual or

Group)

Exploration Strategy

Examination of Digital Sources

Examination of Published Sources

schi18939_ch03_045-068.indd 53 2/19/18 11:28 AM

>part I Building the Foundation for Research

54

sources because they are compiled from a variety of primary sources. While to an outsider the annual report is viewed as a primary source, because it represents the official position of the corporation, to a researcher it represents secondary data. A firm searching for secondary sources can search either inter- nally or externally.

Tertiary sources may be interpretations of a secondary source but generally are represented by in- dexes, bibliographies, and other finding aids (e.g., Internet search engines). These finding aids can be exceptionally valuable in providing direction for secondary searchers, as well as identifying people and events to explore.

When looking for sources, don’t forget to look beyond a firm’s immediate industry. Some of a re- searcher’s best sources will be environmental, government, legal, and social sources, not only ones on business topics.

People as Sources

A strong search of published and digital sources will reveal the identity of people who can add valuable insight. Think of these individuals as information gatekeepers. They fall into four groups: (1) institu- tional memory guardians or company historians, (2) industry or problem experts, (3) company employ- ees involved directly in the problem, and (4) company employees likely to be affected by any considered solution.

As the management dilemma is explored and the research question begins to formalize, those with long institutional memories are invaluable. These may or may not be current employees and may include those who have retired or left the company to pursue their careers elsewhere. What they offer is not only information about past decisions and effects of those decisions, but decision-making context. Very large and enduring organizations understand the value of historical context and may have an employee desig- nated as a company historian. These individuals are responsible for company archives and are invaluable

>Exhibit 3-7 Exploration Strategy Sources

Exploration Strategy

Primary Sources

Published Sources

Secondary/Tertiary Sources

Digital Sources

Company Historians External Sources Internal Sources

Industry Experts on Problem or Solution

Employees involved with the Problem

Employees a�ected by the Solution

Published Sources

Digital Sources

Finding Aids

Actual Source

schi18939_ch03_045-068.indd 54 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

55

in helping a researcher because they may know others who should be interviewed. One study found that the person with the most historical context on a dilemma was a member of the janitorial staff, so keep an open mind about who might be valuable. If a name is mentioned repeatedly, that’s a person who should be interviewed. Individual, face-to-face interviews are most productive here.

A second group of individuals to tap are industry or problem experts. These individuals can provide substantiation for trends, the identity of firms who have faced similar problems and identified solu- tions, background on management dilemmas in the broad sense, and more. For businesses involved in scientific breakthroughs, experts can provide valuable interpretations for the researcher or manager. Thanks to the Internet, professional communities often provide a rich source of industry experts. An- other source is financial analysts. You might have noticed in the business press that reporters often reach out to experts to comment on breaking stories; you can do the same. Often industry/problem experts publish, so they are likely to be identified by some published sources or have their own Internet blogs. Keep your eyes open for names that repeat. These individuals might be inside or outside a company. Individual interviews are most productive here, but phone, video chat, or email might have to replace face-to-face interviews.

The third group comprises company employees directly involved in the current dilemma. Interview them about current conditions, likely contributing factors, and their perspective on action options. The employees you are looking for are those who hold extreme positions because you want breadth of ideas. When interviewing employees, individual or group interviews can be used, using informal or more for- mal techniques, such as focus groups, brainstorming sessions, online or mobile bulletin boards, etc.

The last group, although not necessary the least important, is company employees who would likely be directly affected by possible action decisions being considered. While at this stage of the research a manager might not want the solutions being considered to be fully known, most internal employee grape- vines make it difficult to keep such considerations under wraps. Such individuals will likely have valuable perceptions of how likely actions would affect the work environment, work collaborations and relaton- ships, productivity, morale, etc. Again you are looking for employees who hold extreme perspectives because you want breadth of ideas. You can use individual or group interviews, informal or more formal techniques, such as focus groups, brainstorming sessions, online and mobile bulletin boards here as well.

Published Sources 

In most cases, the external exploration strategy will begin with an examination of published sources, called a literature search—a review of books, as well as articles in journals or professional literature that relate to your management dilemma. A literature search requires the use of the library’s online catalog and one or more bibliographic databases or indexes. This review should include high-quality, web-published materials.

In general, this literature search starts with the management dilemma or question and has four steps:

1. Consult encyclopedias, dictionaries, handbooks, and textbooks to identify key terms, people, companies, or events relevant to your management dilemma or management question.

2. Apply these key terms, names of people or companies, and events in searching indexes, bibliog- raphies, and the web to identify specific secondary sources.

3. Locate and review specific secondary sources for relevance to the management dilemma.

4. Evaluate the value of each source and its content.

While many published sources now can be accessed digitally, many others cannot. A strong explora- tion of available published sources doesn’t ignore nondigital sources just because they are more difficult to access.

External Sources

There are dozens of types of information sources, each with a special function. In this section we de- scribe the information types used most by researchers in the exploratory phase, many of which help them find other valuable sources.

schi18939_ch03_045-068.indd 55 2/19/18 11:28 AM

>part I Building the Foundation for Research

56

Raw and Summarized Data In the United States, businesses are blessed with detailed, government-collected data from numerous agencies. Data.gov is a portal to downloadable data from numerous government agencies, including data on agriculture, climate, education, energy, finance, health, manufacturing, science and industry, consumer expenditures, and much more. Much of these data come from surveys, but some from censuses of various populations. Every 10 years, the U.S. govern- ment conducts the decennial census, one of the largest surveys conducted anywhere. It collects volumes of data on everything from household demographics to housing type. The 2010 Census contained the fewest questions in history and was also the most expensive.1 Periodically, the Census Bureau updates population estimates with sample studies, as well as conducting numerous other studies on current topics of interest, including retail and food services, housing, capital expenditures, time use, wholesale trade, business dynamics, research and development, county business patterns, exports and imports, and more (more than 130 surveys each year). The Census Bureau also conducts the Economic Census, the “official 5-year measure of American business and the economy,” a key data source for many business decisions.2 Most of the summarized data are available through Census.gov. The Bureau of Labor

Exploratory research is unstructured. You don’t know exactly

what to ask or what you’ll discover, only the focus of the discov-

ery. Sometimes what you discover is enough to make a decision,

and sometimes you discover that you need more information to

move forward.

Tire brands offer lots of information to consumers to differ-

entiate their brand from their competitors. But you and I, as tire

purchasers, often need an interpreter to make sense of this in-

formation. We turn to our friendly, knowledgeable salesperson

at our local tire store. What you might not know is that the sales-

person is being offered competitive financial incentives to en-

courage them to recommend one brand versus another. These

incentives might be in the form of cash or cash alternatives,

points redeemable for merchandise, or travel experiences. One

big problem with these incentives is that they are all pretty much

the same; thus they can’t accomplish the job of making one tire

brand more attractive to the salesperson to sell.

One company wanted to identify the critical issues for

salespersons—besides the amount of the incentive—when

choosing one incentive offer over another and thereby rec-

ommending one brand. It chose a readily available tool, their

marketing staff, to find out. The company gave the marketing

team one week to visit tire stores and talk to salespeople and

any other people who processed the paperwork for incentives.

Their charge: determine how to solve the question of incentive

program nondifferentiation. No formal interview guide was cre-

ated. No interview training was provided.

Through their exploratory research, they learned the

following:

● Incentives needed to be in cash, not merchandise or

travel experiences.

● Incentive-claims paperwork needed to be easy to

process, not take hours or days to complete. ● Incentive claims needed to pay out quickly, even multiple

times per week. ● Problematic incentive claims needed to be flagged quickly,

the sales rep needed to be notified immediately, and the

issue needed to be resolvable with one email or phone call. ● Superior performers—those that sell more of your brand

than others—should be offered higher-level incentives.

The most popular program was a competitor’s offer that in-

volved immediately loading a reward card with incentive cash

for each sale; it could be used by the salesperson like a reload-

able debit card. As one salesperson told them, “People come to

work to earn a living, not an inflatable boat.” 

Using Interviews to Refine the Management Question

>snapshot

©Tetra Images/Alamy

schi18939_ch03_045-068.indd 56 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

57

Statistics (www.bls.gov) provides data access and publishes numerous summaries in the area of inflation and prices, pay and benefits, spending and time use, unemployment, employment, and productivity.

Indexes and Bibliographies Indexes and bibliographies are the mainstay of any search strategy because they help you identify and locate a single book or journal article from among the millions pub- lished. The single most important bibliography in any library is its online catalog. As with all other in- formation types, there are many specialized indexes and bibliographies unique to business topics. These can be very useful in a literature search to find authors and titles of prior works on the topic of interest.

Skill in searching bibliographic databases is essential for the business researcher. For the novice or less skilled, we provide two how-to documents through Connect. The first, “Bibliographic Data- base Searches,” reviews the process of searching. The second, “Advanced Searches,” reveals the more advanced techniques of skilled searchers.

Dictionaries In business, as in every field, there are many specialized dictionaries that define words, terms, or jargon unique to a discipline. Many of these specialized dictionaries include information on people, events, or organizations that shape the discipline. They are also an excellent place to find acronyms. A growing number of dictionaries and glossaries (terms in a specialized field, area, or topic plus their definitions) are now available on the web. Information from dictionaries and glossaries may be used to identify key terms for a search of an online or printed database.

Encyclopedias Researchers use an encyclopedia to find background or historical information on a topic or to find names or terms that can enhance search results in other sources. For example, you might use an encyclopedia to find the date that Microsoft introduced Windows and then use that date to draw more information from an index to the time period. Encyclopedias are also helpful in identifying the experts in a field and the key writings on any topic. One example of an encyclopedia is the Online TDM Encyclopedia published by the Victoria Transportation Policy Institute.

The Online TDM Encyclopedia is a comprehensive source of information about innovative management solutions to transportation problems. The Encyclopedia provides detailed information on dozens of Transportation Demand Man- agement (TDM) strategies, plus chapters on their planning, evaluation and implementation. It can help you view trans- portation problems from a new perspective, and expand the range of possible solutions to apply.3

The U.S. government is the world’s largest source of data and informa- tion used by managers of all business disciplines and of all types of organi- zations. Learning how to navigate this government portal is a critical skill for a researcher.

S ou

rc e:

w w

w .u

sa .g

ov

schi18939_ch03_045-068.indd 57 2/19/18 11:28 AM

>part I Building the Foundation for Research

58

Another example drawn from the area of finance is the Encyclopedia of Private Equity and Venture Capital, published by VC Experts Inc., a provider of expertise and opportunity in private equity and venture capital.4

Handbooks A handbook is a collection of facts unique to a topic. Handbooks often includes statis- tics, directory information, a glossary of terms, and other data such as laws and regulations essential to a field. The best handbooks include source references for the facts they present. The Statistical Abstract of the United States is a valuable and frequently used handbook. It contains an extensive variety of facts, an excellent and detailed index, and a gateway to even more in-depth data for every table included. One handbook with which students and managers alike are familiar is the Occupational Outlook Handbook published by the U.S. Bureau of Labor Statistics. In it you can find details about many business occupa- tions and whether each is expanding or declining.5 Many handbooks are quite specialized, such as the one published by the Potato Association of America. It reveals not only consumption patterns, but also potato-growing and -processing statistics.6

One of the most important handbooks, especially in the business-to-business arena, is the North American Industry Classification System, United States (NAICS). Jointly designed with Canada and Mexico to provide comparability in business statistics throughout North America, especially as new businesses and new business sectors develop, this classification system of all businesses replaced the Standard Industrial Classification in 1997.7

Directories Directories are used for finding names and addresses as well as other data. Although many are available and useful in printed format, directories in digitized format that can be searched by certain characteristics or sorted and then downloaded are far more useful. Many are available free through the web, but the most comprehensive directories are proprietary (i.e., must be purchased). An especially useful directory is the Encyclopedia of Associations (called Associations Unlimited on the web), which provides a list of public and professional organizations plus their locations and contact numbers.8 Several directories are available for firms seeking research services providers, including GreenBook: A Guide for Buyers of Marketing Research Services, and BlueBook, the member directory of Insights Association.9

Articles, Reports, Blogs, and Books Media-published articles, whether in the professional, public, or academic press, are valuable sources for understanding the dilemma. We access these by searching various data-finding aids, such as those mentioned earlier, and through recommendations from our professional colleagues. Books are often maligned as an exploratory source, but they are valuable for putting a management dilemma in context, especially when they provide a detailed bib- liography. In this era of self-publication, professional topic blogs are ubiquitous. Some of the things a researcher searches for in this category are prior research studies; reviews of studies, documents, or records; industry trend reports; and articles on problems and their solutions, as well as notable people and events. Some articles and reports are only available on the unindexed web, which means you need the resources of your local librarian for navigation or experience with portals that reveal such sites.

Presentations, Videos, and Webinars Business professionals join organizations as a means of keeping current in their field and to develop professionally. As a result, these organizations often sponsor conferences at which noted professionals speak. The content of these presentations is often critical for providing evidence of possible dilemmas. Business professionals spend increasing amounts of time watching videos on business topics as a method of professional development, but also to familiarize themselves with what their counterparts are doing. These videos are the byproduct of professional presentations at conferences or webinars or produced as promotional materials (e.g., product demonstrations). Each has the ability to offer valuable information, so don’t forget to search YouTube, SlideShare.net,and LinkedIn during exploration.

schi18939_ch03_045-068.indd 58 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

59

Two key trends in secondary data exploratory research are

the incorporation of social media data and automation. Social

media have been shown to provide significant influence on

decision making in both the consumer and industrial spaces.

Increasingly, researchers have spotted early signs of emerging

trends using social media, before more formal research con-

firms such changes. Automation has to do with simplifying the

capture and aggregation of all types of secondary data, includ-

ing (but not exclusively) social media data; analyzing these data;

and reporting them to critical decision makers through push

emails, key information alerts, and sharable reports. Automation

offers the researcher more efficient monitoring of secondary

data, increased productivity by doing each task more quickly,

and reductions in human error and omission, thereby increasing

data quality.

Imagine you need information on automation in the work-

place. You want to monitor what’s being said and written on the

topic of jobs being modified or eliminated due to automation of

processes in preparation for preparing a research project for the

retail industry. You find one relevant article on the topic in Mc­

Kinsey Quarterly. It claims, “Very few occupations will be auto-

mated in their entirety in the near or medium term. Rather, certain

activities are more likely to be automated, requiring entire busi-

ness processes to be transformed, and jobs performed by people

to be redefined, much like the bank teller’s job was redefined with

the advent of ATMs.” Luckily, that article contains a bibliography,

but the search for each article is time consuming and not alto-

gether successful. And you need newer articles as technology

advances are making automation easier. You also want to know

what’s being said in social media on the topic—not just in retail,

but in other industries as well—including such venues as LinkedIn,

Facebook, Twitter, and YouTube. You could manually search each

of these sites, using keywords that relate to your topic, but it’s

really time consuming and you have a relatively short deadline.

What if you could set up an automated process in minutes

that would collect everything said or written on your topic—

every article, post, blog, forum, tweet, video—and you could add

to that everything in the major retail and business, as well as the

academic, press? And what if it updated every time something

new was written or posted? And, what if it was collected in one

digital place, with embedded links for easy access? Founded

in 2005, Netvibes pioneered the first personalized dashboard

publishing platform for the web. Today, Netvibes powers cus-

tomized dashboards for more than 1,000 of the world’s leading

brands, interactive agencies, government organizations, non-

profits and enterprises. Each dashboard aggregates information

to a single computer screen. The owner of the dashboard speci-

fies sources (media, people, etc.), as well as the keywords or

tags, that guide the search widgets. As the owner reviews the

aggregated information, he or she can add additional tags—for

example, sentiments (positive, negative, etc.)—that can be tallied

and graphed for use in reporting. The owner can also set profiles

for alerts to be automatically sent to him by email when new

information joins the dashboard. “All you really need to know is

what data you want to listen to,” shared Kim Terca, director UX

and PR at Netvibes.

www.netvibes.com; www.gartner.com

Secondary Data Sources Go Real Time

>snapshot

Source: Gartner Research

schi18939_ch03_045-068.indd 59 2/19/18 11:28 AM

>part I Building the Foundation for Research

60

>Exhibit 3-8 Evaluating Websites as Information Sources

Evaluation Factor Questions to Answer

Purpose ● Why does the site exist?

● How evident is the purpose it is trying to convey?

● Does it achieve its purpose?

● How does its purpose affect the type and bias of information presented?

Authority ● What are the credentials of the author or institution or organization sponsoring the site?

● Does the site give you a means of contacting anyone for further information?

● Who links to this site?

● If facts are supplied, where do they come from?

Scope ● How old is the information?

● How often is it updated?

● How much information is available?

● Is it selective or comprehensive?

● What are the criteria for inclusion?

● If applicable, what geographic area or time period or language does it cover?

● How does the information presented compare with that on similar sites?

● Is it a series of links only (a metasite), or is there added value?

● What is the nature of the added value?

● What information did you expect to find that was missing?

● Is the site self-contained, or does it link to other websites?

Audience ● Whom does the site cater to?

● What level of knowledge or experience is assumed?

● How does this intended audience affect the type and bias of the information?

Format ● How quickly can you find needed information?

● How easy is the site to use? Is it intuitive?

● Does it load quickly?

● Is the design appealing?

● Are there navigation buttons?

● Is there a site map or search button?

● Is there an easily identifiable Help button?

● Is Help helpful?

● Are pages in ASCII or graphic format?

● Is the information downloadable into a spreadsheet or word processing program, if desired?

Evaluating External Sources

A researcher should conduct a source evaluation on any secondary source to evaluate the information’s credibility. This is especially true of web sources where anyone can publish anything. Researchers should evaluate and select information sources based on five factors in Exhibit 3-8:

• Purpose—the explicit or hidden agenda of the information source.

• Scope—the breadth and depth of topic coverage, including time period, geographic limitations, and the criteria for information inclusion.

• Authority—the level of the data (primary, secondary, tertiary) and the credentials of the source author(s).

• Audience—the characteristics and background of the people or groups for whom the source was created.

• Format—how the information is presented and the degree of ease of locating specific information within the source.

schi18939_ch03_045-068.indd 60 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

61

Internal Sources  Not all published sources are external to the firm; many are internal and include department or busi- ness unit reports, annual reports, research reports, memos, letters, email, etc. An organization’s own internal historical data are often an underutilized source of information in the exploratory phase. Due to employee turnover, the manager may lack knowledge that such historical sources exist, or based on time or budget constraints and the lack of an organized archive, the researcher may choose to ignore such sources. Although digging through data archives can be as simplistic as sorting through a file con- taining employee service records or inventory shipping manifests, or rereading company reports and management-authored memos that have grown dusty with age, if these sources relate to the management dilemma or management question, they should never be totally ignored. Increasingly, managers and researchers are more likely to mine internal digital databases than paper ones during exploration, due to time pressures.

As computing power has increased and large-volume data storage become more common, accessing more recent data archives has become more likely, although not necessarily easier. A data warehouse is an electronic repository for data that organizes large volumes of data into categories or folders to facili- tate retrieval, interpretation, and sorting by end users. The data warehouse provides an accessible ar- chive to support dynamic organizational intelligence applications. The keywords here are accessible and dynamic. Data in a data warehouse must be continually updated (dynamic) to ensure that managers have access to data appropriate for real-time decisions. Today’s cloud-based data warehouses have eliminated data silos for some businesses by housing data in a central repository where standard architecture and consistent data definitions are applied. These data are available to decision makers or cross-functional teams for direct analysis or through intermediate storage facilities, regardless of the manager’s depart- ment or geographic location. However, such a system only works if it is constructed for integration and compatibility.

The more accessible the databases that comprise the data warehouse, the more likely a researcher will use such databases during exploration. Some data in a data warehouse were once primary data, col- lected for a specific purpose. Any patterns revealed will be used for purposes other than those originally intended. For example, in an archive of employee performance reports, we have a wealth of data about different classifications of workers, training, how often reviews are conducted, metrics used to evaluate performance, standards, etc. Initially the company generated these reports to determine pay scale, apply termination rules, determine if training were needed, etc.

Numerous companies build large consumer purchase be- havior databases by collect- ing transaction data made via store-owned credit programs or frequent purchase loyalty identification programs not linked directly with payment plans. Studying such data can reveal the likely success of a new product introduction or the sales lift effect of a price incentive.

© st

oc kn

sh ar

es /G

et ty

I m

ag es

schi18939_ch03_045-068.indd 61 2/19/18 11:28 AM

>part I Building the Foundation for Research

62

When a researcher mines performance report archives, the search is for patterns (e.g., promotion, training, termination, etc.). Traditional database queries are often unidimensional and historical—for example, “How many IT employees were trained?” In contrast, advanced data mining uses artificial intelligence to discover patterns and trends in the data and to infer rules from these patterns. With the rules discovered from the data mining, a manager is able to support, review, and/or examine alternative courses of action for solving a management dilemma, alternatives that may later be studied further in the collection of new primary data.

Data-mining technology provides two capabilities for the researcher or manager seeking better understanding of the management dilemma: pattern discovery and prediction. Data-mining tools can be programmed to sweep regularly through databases and identify previously hidden patterns. An example of pattern discovery is the detection of stolen credit cards based on analysis of credit card transaction records. This pattern could be used to strengthen security protocols. Master- Card processes millions of transactions daily and uses pattern discovery to detect fraud in real time.10 Other uses of pattern discovery include defining retail purchase patterns (used for inven- tory management and planning promotional efforts), identifying call center volume f luctuations (used to identify staff training needs and product defect detection), and locating abnormal data that could represent data-entry errors (used to evaluate the need for training, employee evaluation, or security).

During exploration, managers or researchers use data mining’s predictive capabilities to identify pos- sible action options to solve management dilemmas. Bank of America and Mellon Bank both use data mining’s predictive capabilities to pinpoint marketing programs that attract high-margin, low-risk cus- tomers. Bank of America focuses on credit lines and retail lending; Mellon Bank focuses on its home equity line of credit.11

The process of digitally extracting information from databases has been done in some industries for years. It was State Farm Insurance’s ability to mine its extensive nationwide database of acci- dent locations and conditions at intersections that allowed it to identify high-risk intersections and then plan a primary data study to determine alternatives to modify such intersections. Functional areas of management and select industries are currently driving data-mining projects: marketing, customer service, administrative/financial analysis, sales, manual distribution, insurance, fraud de- tection, and network management.12 However, two problems have limited the effectiveness of this method of exploration: Getting the data has been both difficult and expensive, and processing it into information has taken time—making it historical rather than predictive. While data warehouses offer great promise, businesses have a long way to go to achieve fully accessible and integrated data warehouses.

Evaluating Internal Sources Internal sources should be evaluated on the following

• Compatibility with the current management dilemma, problem, and research question (the more compatible the data, the more valuable the data).

• Timelessness (the more the data hasn’t been affected by time, the more valuable the data).

• Recentness (the more current the data, the more valuable the data).

• Quality of original research process (the more that standards of good research were applied in collecting the original data, the more valuable the data).

• Author(s) experience of the individual(s) compiling the source (the more experienced the individuals, the more valuable the data).

You should always try to identify the political and managerial situations in existence during the cre- ation of internal sources, as these add valuable context for evaluating their value. For many businesses, the most valuable information will be embedded in the firm’s current digital dashboards, as these data will tie directly to the firms current KPIs.

schi18939_ch03_045-068.indd 62 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

63

>Value the Research By the time a researcher or manager arrives at this task, they know the research question. In a business situation, investing in research should increase the desired effect on KPIs in much the same way as any other investment of resources. These effects might include increased revenue, profits, applications, em- ployee morale, and supplier loyalty or decreased expenses, injuries, defects, or employee turnover. One source suggests that the value of research be defined as “the difference between the result of decisions made with [new] information and the result that would be made without it.”13 While such a definition is simple, its actual application is more difficult.

Ideally, a manager needs to determine a research project’s value prior to conducting the research. Most business research is designed to help choose between known alternative options. In such decisions we might use option analysis to determine the value of conducting research. We need several pieces of information, including two drawn from decision theory:

• Two or more options; these are actions a manager might take to address a management dilemma (e.g., replace a training program, change a product’s design, realign the flow of an assembly line, contract with a different distribution company, etc.).

• One or more decision variables; these are projected quantifiable effects of any option being con- sidered (e.g., the dollars projected for increased sales or profits from a new product introduction, the reduction of recruiting costs if employee turnover could be reduced, or a reduction in the num- ber of on-the-job injuries.)

• A decision rule; this is the criteria the manager will use to make an unbiased decision among the options using the decision variable(s). For example: Choose the option that results in the fewest (criterion) on-the-job injuries (decision variable).

• The estimated budget needed for the research design to assess each option.

Plausible options, decision variables, and decision rules are often revealed during exploration. For example, low college student retention is plaguing colleges (management dilemma). Colleges are try- ing to increase their student retention rate (decision variable). They could identify different alternative programs that are used at other colleges that positively affected retention—peer mentoring, alumni men- toring, introductory course option for each major, freshman success program, etc. (options). Each of these programs has various elements—manpower, space, materials, etc.—and each element has a cost. A college could project various retention rate scenarios—best case, likely case, worst case scenarios—with each option. Their choice between options would be based on choosing the option that resulted in the highest student retention from one year to the next (decision rule).

If we can project the costs and outcomes of each option, how valuable would it be to know with cer- tainty which one or combination of options would actually deliver the desired results (e.g., 10 percent higher student retention)? Because each management dilemma is unique, the resulting research ques- tion that will drive the research design will be unique. We can estimate the cost of a research project by first identifying alternative research designs—there is never just one way to conduct research. While any particular research design has unique properties, it also shares common properties with other similar designs. We use these commonalities to develop an estimate. Two approaches work: (1) consult research experts within or outside the firm, or (2) consult company records for actual costs of prior projects em- ploying similar research designs. A seasoned researcher or a manager in a larger company that conducts hundreds of research projects each year might choose the latter approach because the information on costs for various research designs would be readily available.

If the best decision among available options could have been made without the benefit of research infor- mation, then the research has negative value (the cost of the research project, plus the opportunity cost of the time and dollars spent on that project). On the other hand, if the correct decision could not have been made without new research information, it has positive value. If a manager’s assessment projects positive research value, he or she will recommend budgeting for research. Exhibit 3-9 models this subprocess.

Of course, it may not be possible to estimate a research's true value before conducting that research. In that scenario, we determine its value after the decision is made, an ex post facto evaluation. This type

schi18939_ch03_045-068.indd 63 2/19/18 11:28 AM

>part I Building the Foundation for Research

64

of evaluation comes too late to help a manager decide whether a current research decision should be made. The ex post facto evaluation process, however, adds to a manager's ability to make future research decisions.

>Exhibit 3-9 Value the Research

Research Question

Option Analysis Management

Decision without

Research

Execute Research Design

Justify & Budget

Research

Negative Research Value

Positive Research Value

Approved

Rejected

>Justify and Budget for Research Every decision in business—the purchase of a new machine, the hire of a new employee, the launch of an ad campaign, the issue of stock—needs an advocate. Organizations don’t have unlimited resources. A manager who seeks to conduct research must be an advocate—not only to justify the costs to be spent on research, but also to identify the source of funds. A budget to conduct research is usually one of two types:

• Task budget. Dollars are drawn from the discretionary reserves of an organization or the budget of an in-house research operation to fund a research project. A discretionary reserve might be based on a fixed percentage of projected or prior year’s sales. The process is a competitive process among managers; the best crafted argument should, but doesn’t always, win approval, often due to internal politics.

• Functional area budget. Dollars drawn from a portion of the manager’s business unit’s operational funds and allocated to research activities. The manager has the authority to spend budget dollars as he or she desires. This process is also competitive because, internally, the manager has numer- ous goals and priorities. Government agencies, not-for-profits, and businesses alike frequently manage research budgeting this way.

Depending on the level of competition, firms may require a formal written proposal. If at this stage the manager may recognize that he has neither the skills or the resources to conduct research himself, a proposal for research is prepared to send to a sample of outside research organizations. You’ll find this process and a sample proposal in Appendix A.

Without a persuasive argument for a sufficient budget, a research project might never be approved. A manager will be forced to make a decision with existing information.

schi18939_ch03_045-068.indd 64 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

65

>Stage 1 Ethical Issues and Their Solutions While a manager might have assigned a subordinate to conduct the exploration strategy, in many cases the person doing all Stage 1 tasks is the manager herself. Regardless of the organizational structure, the man- ager shares the heaviest weight in executing Stage 1 ethically. The primary issue involved is the manager’s/ researcher’s right to quality research. Some actions that can guarantee these rights include full disclosure of facts (leading to the management dilemma), unfettered access to necessary internal sources (both people and records), clear identification of exploration’s scope (to avoid wasting the researcher’s time), absence of restrictions on secondary sources within that scope, encouragement to share all discoveries, and the absence of hidden agendas on discovered options (to avoid misstating the research question).

Obtaining a budget for research is a competitive process. When competing for a portion of a finite pool of dollars, every research project needs an advocate. The stronger the argument of direct positive value, the more likely the manager will obtain necessary funds.

LO3-1     The manager starts the process of research question formation by focusing on the management dilemma. This problem/opportunity is translated into a management question and then into a research question—the major objective of any research study. In turn, the research question is further expanded into investigative ques- tions. These questions represent what the manager needs to know in order to address the management question. The question hierarchy ends with measure- ment questions that are answered by respondents in a survey or answered about each case in an obser- vational study; the topics may be revealed during this stage but are not refined until Stage 3: Design the Research.

LO3-2 Refining the research question is the ultimate objective of exploration. This phase of the process uses

information to expand understanding of the management dilemma, look for ways others have addressed and/or solved problems similar to the management dilemma or management question, and gather background information on the topic to refine the research question. Exploration involves familiarization with the available literature, delving into company records, interviews (individuals and group) with experts and those involved with the problem or a possible solution. Exploration as a process uses a two-pronged approach, pursuing internal and external sources (often simultaneously) and pursuing people’s perceptions of the problem or its solutions.

LO3-3         External sources are published works of three types: primary, secondary, and tertiary. Primary sources are original works of research or raw data without interpreta- tion. Secondary sources are another’s interpretation of

>summary

© F

ra n

ci sc

o C

ru z/

P ur

es to

ck /S

up er

st oc

k

schi18939_ch03_045-068.indd 65 2/19/18 11:28 AM

>part I Building the Foundation for Research

66

primary research. Tertiary sources may be interpretations of secondary sources or, more commonly, finding aids. Researchers use several tools to identify information sources, including indexes and bibliographies, dictionaries, encyclopedias, handbooks, and direc- tories. One of the hardest tasks in using secondary/ tertiary published sources is evaluating the quality of the information. Five factors to consider when evalu- ating the quality of the source are purpose, scope, authority, audience, and format. External sources can also involve interviewing, through individual or group interviews, individuals who have knowledge of the problem or its solutions, or knowledge of research methodology. Extracting internal information often involves delving into company records or company databases. Some of these may be searchable, but others might not be digital, thus hampering easy searchability. The process of searching digital internal sources is referred to as data mining; its purpose is pattern discovery and predicting trends and behaviors.

LO3-4          The value of research can be defined as “the difference between the result of decisions made with [new] informa- tion and the result that would be made without it.” In order to value research before it is conducted, researchers can use option analysis. This process needs four pieces of information: two or more options (actions the manager might take), one or more decision variables (quantifiable effects of any option), a decision rule (the criteria the manager will use to make an unbiased decision among options), and the estimated budget for the research design to assess each option. Plausible options, deci- sion variables, and decision rules are often revealed during exploration. Estimating budgets for research

designs relies on prior experience or the consultation with experts. If the right decision can be made without new research, then research has no value. If, however, the correct decision cannot be made without new research, the research must be justified and budgeted.

LO3-5 Managers need to be advocates for necessary research. This means they need to identify a source of funds. Budgets for research generally fall into two categories: task and functional area. Task budgets draw funds from discretionary reserves or the operating budget of an in-house research operation. With a functional area budget, the manager uses a portion of the business unit’s operating funds to cover the cost of the research. Both budgeting processes are competitive and, sometimes, political. The greater the level of competition for funds in task budgeting, the greater the likelihood that a formal written research proposal will be required. Without a persuasive argument of a sufficient budget, a research project might never be approved.

LO3-6 The primary ethical issue during this stage of the research process is the right to quality research. Actions that guarantee quality include full disclosure of facts (leading to the management dilemma), un- fettered access to necessary internal sources (both people and records), clear identification of explora- tion’s scope (to avoid wasting the researcher’s time), absence of restrictions on secondary sources within that scope, encouragement to share all discoveries, and the absence of hidden agendas on discovered options (to avoid misstating the research question.

bibliography 57

data warehouse 61

decision rule 63

decision variable 63

dictionary 57

directory 58

encyclopedia 57

exploration 52

exploration strategy 52

handbook 58

index 57

investigative questions 46

literature search 55

management dilemma 46

management question 46

management–research question hierarchy 46

measurement questions 46

option analysis 63

primary sources 53

research question(s) 46

secondary sources 53

source evaluation 60

tertiary sources 54

>keyterms

schi18939_ch03_045-068.indd 66 2/19/18 11:28 AM

>chapter 3 Stage 1: Clarify the Research Question

67

Terms in Review

1 Explain how each of the five evaluation factors for a secondary source influences its management decision- making value.

a Purpose

b Scope

c Authority

d Audience

e Format

2 Define the distinctions among primary, secondary, and tertiary sources in a secondary search.

3 What problems of secondary data quality might researchers face? How can they deal with them?

Making Research Decisions 4 TJX Co., the parent company of T.J.Maxx and other

retailers, announced in a Securities and Exchange Commission filing that more than 45 million credit and debit card numbers had been stolen from its IT systems. The company had taken some measures over a period of a few years to protect customer data through obfuscation and encryption. But TJX didn’t apply these policies uniformly across its IT systems. As a result, it still had no idea of the extent of the damage caused by the data breach. If you were TJX, what internal sources could you use to evaluate the safety of your customer’s personal data?

5 Confronted by low sales, the president of Oaks International Inc. asks a research company to study the activities of the customer relations department in the corporation. What are some of the important reasons that this research project may fail to make an adequate contribution to the solution of management problems?

6 You have been approached by the editor of Gentlemen’s Magazine to carry out a research study. The magazine has been unsuccessful in attracting shoe manufacturers as advertisers because men’s clothing stores are a small and dying segment of their business. Gentlemen’s Magazine is distributed chiefly through men’s clothing stores; the manufacturers reasoned that it was, therefore, not a good vehicle for their advertising. The editor believes that a survey of men’s clothing stores in the United States will probably show that these stores are important outlets for

men’s shoes and are not declining in importance as shoe outlets. Develop the management–research question hierarchy that will help you to develop a specific research proposal.

7 Develop the management–research question hierarchy for a management dilemma you face at work or with an organi- zation to which you volunteer.

8 How might you use exploration using internal sources if you were a human resources officer or a supervis- ing manager trying to increase compliance with safety policies?

9 When Oreo cookies turned 100, Nabisco celebrated with a limited-edition-flavor release called Birthday Cake Oreos, with the taste of Funfetti cake and rainbow sprinkles within the filling center. According to Vanity Fair, this Oreo is a very sweet smell and taste experience. Oreos are the number- one packaged cookie brand, selling more than $778.8 million each year. Oreo comes in more than 30 flavor versions, including original, Golden, Double Stuff, and Green Tea. If you were deciding whether to introduce a new Oreo as part of the celebration, how would you frame the question hierarchy? 

From Concept to Practice 10 Develop the management-research question hierarchy

(Exhibits 3-2, 3-3) citing the management dilemma, man- agement question, and research question for each of the following:

a The president of a home health care services firm.

b The vice president of investor relations for an auto manufacturer.

c The retail advertising manager of a major metropolitan newspaper.

d The chief of police of a major city.

From the Headlines 11 At a BMW staff meeting, employees were told “We’re in

the midst of an electric assault,” referring to electric cars introduced by Mercedes, Porsche, Jaguar, and most notably, Tesla. A fund manager at Union Investment parroted this sentiment, “BMW is falling behind in electrics.” Develop an exhibit similar to Exhibit 3-4 for this management dilemma.

>discussionquestions

schi18939_ch03_045-068.indd 67 2/19/18 11:28 AM

>part I Building the Foundation for Research

68

>additionalcontent

You’ll find the following appendices available from Connect,

Appendix: Bibliographic Database Searches

Appendix: Advanced Searches

>cases*

A Gem of a Study

Akron Children’s Hospital

Calling Up Attendance

Donatos: Finding the New Pizza

HeroBuilders.com

Inquiring Minds Want to Know—NOW!

Mastering Teacher Leadership

NCRCC: Teeing Up and New Strategic Direction

Ohio Lottery: Innovative Research Design Drives Winning

Ramada Demonstrates Its Personal Best™

State Farm: Dangerous Intersections

USTA: Come Out Swinging

*You will find a description of each case in the Case Abstracts section of this textbook. Check the Case Index to determine whether a case provides data, the research instrument, video, or other supplementary material. Cases and case supplements are available in Connect.

schi18939_ch03_045-068.indd 68 2/19/18 11:28 AM

Chapter 4 Stage 2: Research Design, An Overview

Chapter 5 Stage 2: Sampling Design

Chapter 6 Stage 2: Data Collection Design: Qualitative Research

Chapter 7 Stage 2: Data Collection Design: Observation Research

Chapter 8 Stage 2: Data Collection Design: Experiments

Chapter 9 Stage 2: Data Collection Design: Survey Research

>part IIII

The Design of Business Research

schi18939_ch04_069-084.indd 69 2/14/18 2:53 PM

LO4-1    The tasks that comprise research design.

LO4-2 The types of variable relationships.

LO4-3 The major descriptors of research designs.

>learningobjectives After reading this chapter, you should understand . . .

“Good research is about using the right

platform for each piece of a study, for each activity

or interaction.”

Nick Drew,

VP Insights and Strategy,

Fresh Intelligence

Stage 2: Research Design, An Overview

>chapter 4

schi18939_ch04_069-084.indd 70 2/14/18 2:53 PM

>chapter 4 Stage 2: Research Design, An Overview

71

>What Is Research Design? At its most simplistic, research design is a blueprint for fulfilling research objectives. Regardless of the numerous definitions of research design,1 they all have common properties. A design:

• Is a time-based, procedural plan for every research activity.

• Is always focused on the research question.

• Guides selection of sources of information (cases).

• Provides the framework for specifying the relationships among the study’s variables.

The tasks covered by the term research design are wide-ranging, as depicted in Exhibit 4-1. Our objec- tive here is not for you to acquire the details of research design in one reading, but for you to understand its scope and to get a glimpse of the available options for tailoring a design to an organization’s particular research needs.

We finished Stage 1 of the research process by determining whether research would be justified given when we didn’t know—our investigative questions. In detailing the design, a researcher must develop:

• The sampling design.

• The data collection design.

• The measurement instrument(s).

Some researchers choose to start the process of planning these three tasks with data collection de- sign. Others prefer to start with sampling design. Either approach is appropriate.

>Exhibit 4-1 Research Design in the Research Process

Investigative Questions

Design the Research

Collect and Prepare the Data

Data Collection Design

Sampling Design

Develop Measurement Instrument

Stage 2

Stage 1

Stage 3

Measurement Questions

schi18939_ch04_069-084.indd 71 2/14/18 2:53 PM

>part II The Design of Business Research

72

In sampling design, the task is to identify the sources of any information you did not find during explora- tion. The investigative questions you developed in Stage 1 are the focus of this process. Whatever sam- pling design you choose must answer two questions:

• Who or what should be measured?

• How do we access or recruit that source?

While in business research, people don’t al- ways represent the target population of inter- est, we’ll use an example here in which they would. Let’s say that a firm is experiencing declining worker productivity that appears to parallel a decline in worker morale. The man- agement question is, “What can be done to increase productivity?” During our Stage 1 ex- ploration of secondary sources, we learn that worker morale often affects productivity and is itself affected by (1) a worker’s perception of job uncertainty; and (2) on-the-job stress caused by perceptions (a) of the unfairness of a superior’s expectations, (b) that there are insufficient resources to perform the job well, and (c) that worker efforts are unappreciated. We also find out that workers often feel they don’t know what is going on in the firm outside their own responsibilities; this makes them feel isolated. We determine a list of investigative questions built around the research question of

whether to implement an easily accessible, company-wide KPI dashboard to increase transparency of com- pany performance with hopes of improving morale. Many of the investigative questions deal with what information items workers deem important; others deal with what causes worker stress, and how workers perceive the supportiveness of their work environment. Our exploratory study of performance evaluations revealed these records would give us little insight on the problem. Thus the workers themselves become our target population. If we had a small firm, we might ask all workers to participate in the research proj- ect. But if we have thousands of workers in numerous locations speaking a variety of languages, we might choose instead to conduct research using a presentative sample. Because these are our workers, we can encourage the chosen workers to participate by incentivizing them with paid time off work.

Once we determine a sample is our choice, we must choose a sample type based on what we plan to do with the information. It’s costly to implement a digital dashboard, so we want to make sure it will have the desired effect before spending the money. Ideally, we want to predict all worker response with the evidence measured from just a sample of the workers, so we want a higher-order sample—a probabil- ity sample rather than a nonprobability sample. Depending on the variability of jobs within our worker structure, we might choose a specific type of probability sample—a stratified sample—to be sure to mea- sure those attitudes among different classifications of workers. Using an employee roster, divided by job classification, to draw our sample will ensure its representativeness.

So knowing we will measure worker attitudes, perceptions, and feelings, in our sampling design we have made several sampling design decisions:

• We will draw a sample of workers from a company roster divided by job classifications.

• We will apply probability sampling procedures to that roster to extract a stratified sample based on job classification, identifying specific workers.

• We will recruit those specific workers to participate, compensating them for time off the job to encourage their participation.

>Sampling Design

© E

ri k

Is ak

so n

/G et

ty I

m ag

es

schi18939_ch04_069-084.indd 72 2/14/18 2:53 PM

>chapter 4 Stage 2: Research Design, An Overview

73

>Data Collection Design Data collection design involves a large number of decisions related to how, when, how often, and where data will be collected. In research we measure variables of interest. During data collection design we make numerous choices in order to accurately measure these variables and any relationships between these variables.

An Example We’ll use an example to discuss data collection design dimensions. For a bank, let’s assume that it is considering building a new branch in an area of town where it is not yet established. The bank wants to stimulate savings based on the results of a formal study of existing sav- ings activity at its current branches. If the manager hypothesized that savings activity is related to nearness of the branch bank to one’s home, he might choose a location with a sizable population of homes. But buying land and constructing a branch is a significant investment of resources. The bank’s board of directors will want to be sure that such an in- vestment is empirically supported.

Assume the research question might be, “Is the distance from the account owner’s residence to the branch related to saving ac- tivity?” Or using the hypothesis format, “Sixty percent of current savers live within a two- mile radius of the branch.” In the first phase of the study, we use savings records to cross- tabulate the distance from the account owner’s residence to his or her branch and account activity. It shows that differential rates of activity are related to an account owner’s location. A cross-tabulation of home-to-branch distance to account size and gender of account owner also shows a relationship. Our task is to determine if the variables are independent (unrelated) or, if they are not, to determine the strength or magnitude of the relationships. Neither procedure tells us which variable is the cause. For example, we might be able to conclude that gender and distance are related to savings activity but not that gender is a causal factor in choice of branch.

Let’s say the examination of account records reveals a concentration of near-to-their-branch savers. Their accounts were typically larger and more active than those of account holders who live at a dis- tance. So we choose to add a second phase to our study: a survey of savers to provide information on stage in the family life cycle, attitudes toward savings, family income levels, and motivation for establish- ing the account. Correlation of this information with known savings data shows that women own larger accounts and that these women with larger accounts are non-working-widowed or working-single women who are older than the average account holder.

The correlation between nearness to the branch and the probability of having an account at the branch might suggest the question, “Why would people who live far from the branch have an account there?” The researcher might hypothesize that:

1. Distant savers (operationally defined as those with addresses more than two miles from the branch) have accounts at the branch because they once lived near the branch; they were “near” when the account decision was made.

2. Distant savers actually live near the branch, but the address on the account is outside the two- mile radius; they are “near,” but the records are inaccurate.

© Y

in Y

an g/

G et

ty I

m ag

es

schi18939_ch04_069-084.indd 73 2/14/18 2:53 PM

>part II The Design of Business Research

74

3. Distant savers work near the branch; they are “near” by virtue of their work location.

4. Distant savers are not normally near the branch but responded to a promotion that encouraged savers to bank via computer; they are “near” by virtue of electronic banking.

When these hypotheses are tested, the researcher learns that a substantial portion of the distant sav- ers can be accounted for by hypotheses 1 and 3. The conclusion: Location is closely related to saving at the branch.

Here are the data collection design decisions that were revealed:

• All design decisions are consistent with the hypothesis that savings activity is related to “nearness” of the branch to the account holder’s home.

• We operationally defined the variable of interest (nearness) as “a home address within two miles of the branch where their account was opened.”

• The board of directors will want statistical data to support a recommendation to build a new branch.

• Variables of interest include home location, a measure of nearness, the branch where participants opened their account, branch(es) where they do their saving, types of account activity (deposits, withdrawals, CDs, IRAs), frequency of account activity, attitudes toward saving, how account holders access their accounts (branch visits from home, work, or via Internet), and account holder demographics (age, gender, etc.).

• We cannot manipulate our primary variable of interest (nearness).

• While we would love to measure an causal relationship between nearness and saving activity, we will seek evidence of a strong correlational relationship.

• The research will be conducted in two phases: (1) an examination of savings records, followed by (2) a survey of savers. • The examination of savings records allows the researcher to track savings activity over the time

the account has been open and link that activity to initial home address and any changes in home location during the life of the account.

• A survey allows the researcher to collect more demographic data on the saver who opened the account, as well as attitudinal information about saving, motivational information about why the account was opened, and behavioral information about employment and location of employment.

Design Dimensions All these decisions address dimensions important in data collection design. Exhibit 4-2 groups these di- mensions.2 A brief discussion of each dimension illustrates its nature and contribution to a researcher’s design choices.

• Objective of the study.

• Researcher’s ability to manipulate variables to be studied.

• The topical scope.

• Measurement emphasis of research techniques.

• Complexity of design.

• Method of data collection.

• Research environment.

• Time dimension.

• Participant’s research awareness.

Objective of the Study

This is the most important dimension of any data collection design because it influences choices in all others. A reporting study provides a summation of data, often recasting data to achieve a deeper

schi18939_ch04_069-084.indd 74 2/14/18 2:53 PM

>chapter 4 Stage 2: Research Design, An Overview

75

>Exhibit 4-2 Dimensions of Data Collection Design

Dimensions Data Collection Design Options

Objective of the study ● Reporting (who, what, when, where, how—in the past)

● Descriptive (who, what, when, where, how—now)

● Causal

● Explanatory (why, how—now)

● Predictive (what, why, in future)

Researcher’s ability to manipulate variables being studied

● Experimental (ability to manipulate)

● Ex post facto (no ability to manipulate)

The topical scope of the study ● Statisticalstudy (significant breadth, some depth)

● Case study (little breadth, significant depth)

The measurement emphasis of research techniques

● Qualitative (emphasis on measures of meaning)

● Quantitative (emphasis on measures of frequency)

Complexity of design ● Single methodology

● Multiple methodologies

The method of data collection ● Monitoring

● Communication study

The research environment ● Field setting

● Laboratory research

● Simulation

The time dimension ● Cross-sectional (one measurement at one point in time)

● Longitudinal (many measurements in specified period of time)

The participants’ research awareness ● No deviations perceived

● Deviations perceived, not researcher related

● Deviations perceived, researcher induced

understanding or to generate statistics for comparison. In a study of crime, for example, such a study might tally the number of employee thefts that take place in shopping malls versus free-standing stores. Businesses use reporting studies, but because they rely on existing data, they may not cover all the vari- ables a researcher needs to measure. Phase 1 of the bank research is a reporting study.

A descriptive study attempts to determine who, what, where, when, or how much. Descriptive research on employee theft would measure the types of theft committed (clothing vs. electronics vs. housewares vs. money), how often, when (time of year, time of day, day of week), where (receiving dock, stockroom, sales floor, financial office), and by whom (gender, age, years of service, departmental assignment). Descriptive studies collect new data, which may be combined with existing data; such a study can be simple or complex. Phase 2 of our bank example is a descriptive study. These studies serve a variety of research objectives:

• Descriptions of phenomena or characteristics associated with a target population.

• Estimates of proportions of a population that have these characteristics.

• Discovery of associations among variables of interest.

A causal-explanatory study attempts to explain relationships among variables—to determine how one variable produces changes in another. For instance, such a study might try to explain why the crime rate is higher in mall A than in mall B or why male employees steal more than female employees. A causal-predictive study attempts to predict an effect on a dependent variable by manipulating another variable (independent) while holding all other variables constant. For example, researchers conducting a causal-predictive study might test whether installation of video surveillance cameras on the receiving

schi18939_ch04_069-084.indd 75 2/14/18 2:53 PM

>part II The Design of Business Research

76

dock and in stockrooms would reduce employee theft in mall stores. While causal inferences may be neither permanent nor universal, over time these inferences allow us to build knowledge of presumed causes. Such empirical conclusions provide us with successive approximations to the truth and the foun- dation of theories.

Researcher Ability to Manipulate Variables

With an ex post facto design, researchers have no ability to manipulate variables. Instead, the re- searcher studies participants who have been exposed to the independent variable and those who have not been so exposed and compare the results. In our bank research, we study savings account holders who live “near” to a branch and those who don’t using an ex post facto design. It is important that the researchers using this design not influence the variables; to do so would introduce bias. The researcher is limited to holding factors constant by judicious selection of participants according to strict sampling procedures and by statistical manipulation of findings. Businesses use ex post facto data collection designs extensively.

In an experiment, the researcher must be able to control and/or manipulate the variables in the study. It is enough that the researcher can cause variables to be changed or held constant in keeping with the research objectives. Results from experiments provide the most powerful support possible for a hypoth- esis of causation. Banks can use experiments; for example, extending (or decreasing) the hours of ATM access to determine if ATM use would increase (decrease) compared to prior use.

The Topical Scope

Statistical studies attempt to capture a population’s characteristics by making inferences from a sample’s characteristics. They are designed for breadth rather than depth, extracting the same data from multiple cases. Hypotheses are tested quantitatively. Generalizations about findings are presented based on the representativeness of the sample and the validity of the design. In our bank research, our records study is statistical; our survey of savings account holders is statistical. Much of business research involves statistical studies.

Case studies place more emphasis on a full contextual analysis of one or a few events or conditions and their interrelationships. An emphasis on contextual detail provides valuable insight for problem solving, evaluation, and strategic or tactical change. This detail may be secured from multiple sources of information, allowing evidence to be verified and avoiding missing data. A single, well-designed case study can provide a major challenge to a hypothesis and provide a source of new hypotheses and constructs simultaneously.3 Before the use of surveillance cameras, a business might have made as- sumptions about how employees followed or didn’t follow safety procedures based on the infrequent observations of a manager; the eight-hour footage from a single camera (case) could challenge those assumptions. Businesses often use case studies. One often-used example is to study a firm that has achieved a solution to a similar problem as the one a manager faces. Such a case might be described as a best-practice case.

Measurement Emphasis

Business researchers use both qualitative and quantitative research, sometimes in the same study. Qualitative research use interpretive techniques to describe, decide, translate, and otherwise come to terms with the meaning of certain phenomenon. Qualitative research is often used during exploration, but it may also follow a statistical study to further understand what was measured. With the increasing use of mobile devices to conduct qualitative research involving larger samples, such studies may be suf- ficient on their own to guide a business decision. Qualitative research is dependent on the researcher for its rigor.

Quantitative research is interested in the quantity, frequency, or magnitude of a phenomenon; it is the basis of a statistical study. Quantitative research is dependent on the quality of the measurement instru- ments used for observation, surveys, and experimental tests. In our bank research, it is likely that our board of directors will be much happier with both phases being conducted quantitatively.

schi18939_ch04_069-084.indd 76 2/14/18 2:53 PM

>chapter 4 Stage 2: Research Design, An Overview

77

Complexity of Design

A single-methodology design would employ only one method—a survey, for example. However, business research is often complex, including more than one data collection methodology. A multiple-methodology design might include an array of methods—for example, a series of personal interviews, followed by a series of focus groups, followed by a survey. One of the most common multiple-methodology studies is

Following the financial crisis, AIG found itself under in creasing

scrutiny. It had an urgent need to demonstrate to share holders

that it could differentiate performance. When Julia Smith

joined the HR team at AIG, she was told the performance

management system was in peril. It was cumbersome, lacked

manager ownership, and focused only on what employees

were doing—without the effects of those actions—and

provided little direction to employees on how to improve or

enhance their performance.

AIG undertook a massive research project to determine what

should replace its current system. It partnered with ThinkTank.net

to conducted virtual focus groups, drawing participants from its

more than 60,000 employees worldwide. AIG learned the ideal

system needed to inspire employees to do their best, offer real-

time feedback and constructive coaching, and reveal transfor-

mative development opportunities.

AIG recruited the expertise of noted expert in behavioral

economics, as well as cognitive and social psychology, Danny

Kahneman. And it interviewed hundreds of HR professionals

about their performance appraisal systems to find best prac-

tices. The program that developed would evaluate 18 specific

behaviors (e.g., able to anticipate problems, collaborative and

helpful, open to new ways of doing things) using an online

evaluation system in which both managers and employees

have input; each behavior is needed to reinforce AIG values.

The research taught AIG to frame the behaviors positively and

to use everyday language rather than jargon. It also learned

that the appraisal entry needed to be online, use simplistic

scales, and be easy to complete by both manager and em-

ployee. It also needed to generate a consistent, computer-

generated report on each employee. This report needed to

show the employee not only his or her behavior achievement

in comparison with others with the same type of job, but also

how he or she compares with employees throughout the com-

pany regardless of job type or level. The research revealed

the need to share the report with employees before any man-

ager discussion and to prepare a discussion guide for man-

agers and employees to use when they discussed the report.

AIG put its best ideas into a new program focused on

two primary goals—reward collaboration and teamwork—while

discouraging inter-employee competition. It then launched a pilot

with 3,000 employees to test the new program and surveyed

participants after the pilot. After tweaking the program, it went com-

pany-wide in 2016. Employees find the new program fair (96 per-

cent), accurate (92 percent), and helpful to improving performance

(92 percent), with 86 percent of employees indicating their rela-

tionship with their manager improving. Volunteer employees are

featured in video clips focused on each of the 18 behaviors to en-

courage their fellow employees. That’s what buy-in looks like.

AIG.com; ThinkTank.net

AIG AND Research Design

>snapshot

©John Lund/Drew Kelly/Blend Images LLC

schi18939_ch04_069-084.indd 77 2/14/18 2:53 PM

>part II The Design of Business Research

78

the two-stage design, which employs qualitative research techniques to further understand the manage- ment problem and frame the research question, followed by a quantitative statistical study to collect data related to the measurement questions. Our bank study is using a multiple-methodology design employing observation of records and a survey of account holders.

For a business, the data collection design complexity choice is often influenced by the action option(s) the manager is considering and how much each would cost to implement; the more expensive the op- tions being considered, the more likely that business might want to know with some statistical certainty that any hypothesis it formed is correct. This might lead the business to invest more in research. A busi- ness considering a major shift in its overall business strategy, for example, could lead to a more complex and more expensive data collection design. Our bank considering building a new branch is in just such a situation.

Method of Data Collection

In a monitoring study the researcher inspects the activities of a subject or the nature of some material or activity and records the information from observations; the researcher does not attempt to elicit responses from anyone. Traffic counts at an intersection, license plates recorded in a restaurant park- ing lot, a search of the library collection, an observation of the actions of a group of decision makers, the study of records of accidents by State Farm to identify dangerous intersections—all are examples of monitoring. These observations can be made in person or through digital or mechanical means. Our bank research employs monitoring in phase 1.

In a communication study, the researcher questions the participants and collects their responses by personal or impersonal means. The collected data may result from (1) interview, telephone, or web con- versations; (2) self-administered or self-reported instruments sent through the mail, left in a convenient location (e.g., hotel room), or transmitted electronically or by other means; or (3) instruments presented before and/or after a treatment or stimulus condition in an experiment. We use the term communication because collecting data by questioning encompasses more than the survey method. Our bank research employs a communication study in phase 2.

Research Environment

Designs also differ as to whether measurement occurs under staged or manipulated conditions (laboratory conditions) or under actual environmental conditions (field conditions). Laboratory condi- tions might be used in highly sensitive situations where you want the research to remain confidential (e.g., a taste test of a new menu item for Wendy’s that, if known by competitors, would reveal too soon a pos- sible new strategic direction). It is assumed when you collect data using field conditions that participants are not as likely to alter their actual behavior, behaviors of which they may not be fully aware. But it is also likely that you will not observe as many participants using field conditions. We are using visual eth- nography more in business research, where, for example, we ask participants to take video of a shopping trip using their smartphone, then upload the video to a researcher. Setting up a fake store (laboratory), and asking participants to “shop” that store while we video them, is an alternative design, but it would be extremely costly. Some field conditions might put a participant or researcher at risk (e.g., observing how a participant handles a contrived, unsafe work environment or sending an interviewer into a crime- infested area to do interviews); in such cases, an alternative design would be explored. When we intercept customers in a mall or store to conduct a survey, we are using field conditions. The second phase of the bank study will likely communicate with the savings account holders in their homes, a field study.

Both field and laboratory conditions may use simulations—a replica of a system or process. As computer programming advances, some businesses are using mathematical models to simulate char- acteristics under various conditions and relationships (e.g., performance of an airplane engine de- sign on its ability to deflect birds); others are using virtual reality (e.g., to determine the strength of a package design4). Role-playing and other behavioral activities may also be viewed as simulations. One behavioral role-play simulation in retail involves mystery shoppers where a researcher poses as a potential customer to record and evaluate the sales person’s knowledge, customer service approach, and procedural skills.

schi18939_ch04_069-084.indd 78 2/14/18 2:53 PM

>chapter 4 Stage 2: Research Design, An Overview

79

It doesn’t take much effort to find a business press

article that encourages the elimination of rating-

based performance management (PM) systems.

TIAA is a Fortune  100 financial institution with

12,800 employees that manages investment

accounts for more than 5 million people employed

in 16,000 institutions. It decided to approach an-

ecdotal employee dissatisfaction with their PM

processes more deliberately, designing a multi-

stage, multi-year research project.

“Asking the right question to start the research

process was critical,” shared Michelle Shail, manager of talent

management at TIAA. “Should we eliminate formal reviews and/

or ratings?” was the wrong question, even though it was, in part,

the annual review ratings that caused the dissatisfaction among

employees. TIAA discovered that the right question was actually

two questions: “How can we develop a PM process that will drive

desired business outcomes?” and “What PM process will best fit

our culture?”

Their first step was to conduct quantitative and qualitative-

analyses of employee survey data from the prior five years.

It partnered with i4CP to develop insights, connections, and

guidance. Based on employee survey feedback, TIAA discov-

ered that many employees found that the current PM process

didn’t help them improve their performance—one of its key

goals. TIAA tallied time spent in the PM process and dis-

covered the 71,800 hours was mostly spent on compliance

documentation, not conversations between manager and em-

ployee. Additionally, employees were concerned about elimi-

nating the rating system because it was integrated into their

compensation system.

TIAA conducted direct benchmarking interviews with human

resources (HR) professionals in more than 35 companies to help

design a possible new system. Based on the HR interviews, TIAA

learned that any great PM system contained three elements:

(1) individual goal setting aligned with organizational goals, (2)

coaching and feedback from both managers and peers, and

(3) reviews that truly differentiated performance and transpar-

ently linked compensation with performance. “Every organiza-

tional culture is different. Ours is very focused on developing

detailed documentation. Yet we discovered that employees and

their managers wanted PM information condensed and put in

snippets.”

TIAA designed a new PM system that focused on increas-

ing coaching and feedback; incorporated accountability not only

for business results, but also individual behaviors; simplified

documentation; and increased management capability. To test-

components of the new program, TIAA launched three different

pilots (research studies). The first pilot tested the “Get Feed-

back” tool for peer evaluations—research had revealed that em-

ployees wanted to control who evaluated them and who had

access to that evaluation.

The second pilot evaluated the critical behaviors list. It

learned that nine key competencies or behaviors were critical to

achieving company goals. They grouped the desired behaviors

into three categories quality, people, and character.

The third pilot evaluated replacing the formal mid-year review

with two brief check-in conversations, which were flexible in

terms of timing but focused on employee goals and behaviors

to accomplish those goals. They used focus groups to deter-

mine if the check-in interviews of the new process were taking

place between employees and their managers and whether

these provided the desired personal development feedback.

They used discussions with 49 volunteer employee ambassa-

dors, who were both passionate about the new process as well

as strong performers, for additional feedback and to get critical

buy-in among employees.

While it’s early yet to determine if this new PM program will be

successful, one thing it is doing is saving time—one to three hours

per year for every employee and 5 to 18 hours for every manager.

www.tiaa.org; www.talentmgt.com; i4cp.com

TIAA Performance Management Overhaul

>snapshot

Quality

• Provide exceptional service • Establish strategic priorities • Drive continuous Improvement

People

• Communicate Openly • Foster collaboration • Develop self and others

Character

• Build trust • Take ownership • Make sound decisions

Time Dimension

Longitudinal studies are repeated measures of the same variable over a specified period. The advantage of a longitudinal study is that it can track changes over time or involve participants in a variety of activities. Longitudinal studies can study the same people over time (e.g., via panel) or use different participants for each subsequent measurement (e.g., via cohort groups: the financial service industry might study

schi18939_ch04_069-084.indd 79 2/14/18 2:53 PM

>part II The Design of Business Research

80

the needs of aging baby boomers by sampling 40- to 45-year-olds in 1995, 50- to 55-year-olds in 2005, and 60-to 65-year-olds in 2015). Some types of information once collected cannot be collected a second time from the same person without the risk of injecting bias because participant awareness has been enhanced by introduction to a stimulus. The longitudinal study of public awareness of an advertising campaign over a six-month period would require different samples for each measurement. While longitu- dinal research is important, the constraints of budget and shortened decision-making time often impose the need for cross-sectional studies. Phase 1 of the bank research is a longitudinal study.

Cross-sectional studies are carried out once and represent a snapshot at one point in time. The ad- vantages of a cross-sectional study are that it can be accomplished more quickly, at a lower cost, and error sources caused by changes over time are minimized. Discovering past attitudes, history, and future expectations can be discovered through a cross-sectional study; however, interpreting such data requires care. In the bank study, the survey is a cross-sectional study.

Participants’ Research Awareness

The usefulness of a design may be reduced when a participant perceives that research is being con- ducted. This awareness influences the outcomes of the research in subtle ways or more dramatically as we learned from the pivotal Hawthorne studies of the late 1920s.5 Although there is no widespread evidence of attempts by participants or respondents to please researchers through successful hypothesis guessing or evidence of the prevalence of sabotage, when participants believe that something out of the ordinary is happening, they may behave less naturally. There are three levels of perception:6

• Participants perceive no deviations from everyday routines.

• Participants perceive deviations, but as unrelated to the researcher.

• Participants perceive deviations as researcher-induced.

In mystery shopper research, if a retail sales associate knows she is being observed and evaluated—with consequences in future compensation, scheduling, or work assignment—she is likely to change her perfor- mance. In all research environments and control situations, researchers need to be vigilant to effects that may alter their conclusions. Participants’ awareness concerns serve as a reminder of the value of classifying one’s study, to examine validation strengths and weaknesses, and to be prepared to qualify results accordingly. In phase 1 of the bank research, the account holders would be unaware. In phase 2, they would be aware.

>Design the Measurement Instrument When you get to this task, the sample design and data collection design choices are completed and the researcher knows the variables to be measured. The activities included in this task are

• Determine the appropriate measurement approach for each data collection task in the design.

• Craft the specific measurement questions or elements of an observation checklist.

• Determine the order of questions or elements.

• Determine the necessity of introductions, definitions, instructions, and conclusions for partici- pants; develop these non-question elements.

• Craft the measurement instrument(s).

• Determine the necessity of training for data collectors.

• Craft any training guides and activities.

A measurement instrument is a tool for collecting data on a study’s variables. There are many different types: survey questionnaires, interview guides, group discussion guides, and observation checklists, to name a few. Both qualitative and quantitative studies need measurement instruments. Also during this task, indi- vidual questions and whole instruments are tested, either with a sample of participants or their surrogates, with other researchers, or both. Tests determine wording problems, question order problems, the need for clarifying definitions or instructions, and more. These tests are always conducted as the instrument will actually be used, face-to-face or phone or personal interview, Internet chat, mobile survey, field observation, etc. In the bank study, we’ll need a record’s checklist for the observation study and a survey for bank savers.

schi18939_ch04_069-084.indd 80 2/14/18 2:53 PM

>chapter 4 Stage 2: Research Design, An Overview

81

Founded in 2010, rapidly growing San Diego-based organic

juice company Suja offers a line of non-GMO, cold-pressured,

preservative- and gluten-free juices and smoothies through

14,000 natural, grocery, box, and mass-market retailers. “One

recent study showed while 4% of households actually consume

organic foods,” shared Suja’s brand manager Bella Tumini,

“many more are interested. Suja is trying to democratize organic

for the 44% who are trying to incorporate organic foods into

their diet.”

In order to stay competitive and stand out to consumers, com-

panies that offer organic and/or juice products must innovate con-

stantly. Without a research budget, Suja considered their research

approach as “scrappy,” using secondary research to identify trends

and turning to their own head of innovation—a Culinary Institute-

trained chef—to develop juices to match these trends. Innovation

ideas were vetted using mostly the opinions of employees, as well

as those of their family and friends. Additionally, the team used do-it-

yourself (DIY) surveys, drawing participants from their social media

followers, leaving room for biased results. The team lacked the time

and research expertise necessary for survey creation, data analy-

sis, and report building. “Having the DIY tools without the people

wasn’t helpful because we’re such a small team,” explained Tumini.

By using this “gut” process, Suja was able to innovate-

quickly—often taking just six weeks from idea to distribution.

“That speed allowed us to stay ahead of the competition and

establish Suja as a leader in the category,” explained Tumini. And

while the Suja team has launched well over 100 products since

inception, almost half of those have been discontinued. In order

to pivot quickly based on product performance, the marketing

team employs a flexible approach when it comes to testing and

launching different products, especially being a team of three

people with a tight budget.

In just four years, the company grew from 300 to 14,000

stores in which Suja products are sold. Yet many brands have

entered the category, making it harder on consumers to make

a choice at the shelf. As such, the marketing team sought to de-

crease the amount of time spent at the shelf by differentiating

Suja’s product labels from competitors. Before moving forward

with a redesign of the leading product line’s packaging, the team

had to be sure the update would enhance the brand’s perfor-

mance while appealing to both of their consumer segments.

The Suja team decided to partner with GutCheck, a provider

of online agile market research, in order to ensure a packag-

ing and brand update would be worth the risk. GutCheck offers

several online agile research solutions that are fast, adaptable,

and iterative (giving clients the option for more waves or stages

of research, if needed). “By automating certain steps in the re-

search process that don’t add value, research strategists are

able to focus on what matters: research design, data collection,

analysis, and delivering reports within days, instead of months,”

shared Matt Marta, CEO, GutCheck. The Suja team was attracted

by the speed and iterative nature of GutCheck’s research for the

redesign of its Suja Essentials package.

Together, the Suja marketing team and GutCheck research

team designed the multiphase research to answer the key ques-

tion, “Are consumers more or less likely to buy the new packaging

versus the current one?” Phase I (five days from recruit to report)

employed an Instant Research Group. An Instant Research Group is

an online qualitative discussion where respondents—recruited from

global panel providers—answer open-ended questions and follow-

up inquiries posted by a trained moderator over the course of sev-

eral days. This methodology allows for quick consumer feedback

and gives respondents the ability to interact with others in the group.

To avoid bias, all participants first answer each question themselves.

Then, on the next screen, they see all participants’ responses and

can react to and comment on each other’s responses. This encour-

ages conversations and sparks ideas among respondents. This

phase was exploratory and included 30 participants who evaluated

four package designs based on key metrics: likes, dislikes, overall

appeal, purchase intent, believability, and uniqueness.

Primary findings showed that the product claims were believ-

able based on the presence of certification labels, which added

How Agile Research Helped Prove the Value in a Packaging Redesign

>closeup

Courtesy of GutCheck and Suja

schi18939_ch04_069-084.indd 81 2/14/18 2:53 PM

>part II The Design of Business Research

82

LO4-1           Research design is a time-based plan for the execution of all research activities. It has three tasks: sampling design, data collection design, and development of the measurement instrument. Each of these stages is integrated, although they are tackled separately. Some researchers choose to start with sampling design, while others choose to start with data collection design; either starting point is appropriate. Sampling design answers who or what is the source of measurement and how we access or recruit that source.

LO4-2 Data collection design involves the measurement of variables and the search for relationships between those variables. Data collection design involves a large number of decisions related to how, when, how often, and where data will be collected. Design

of the measurement instrument involves translating investigative questions into measurement questions that are appropriate for the chosen data collection design. Researchers not only craft instruments, but also test them to reveal problems for both the participant and any data gatherers.

LO4-3 A number of different design dimensions influence data collection design: objective of the study, the re- searcher’s ability to manipulate variables in the study, the topical scope, the measurement emphasis, the complexity of design, the method of data collection, the research environment, the time dimension, and whether participants are research aware. Each of these dimensions represent multiple options to the researcher.

>summary

trust in the brand. Respondents also liked that the bottle color

matched the juice color. Verbatims indicated the new designs

were perceived as clean, stylish, modern, and unique—all what

the Suja team wanted.

These qualitative findings were used to further optimize the-

new package concept. Phase 2, the test of the refined design

against the existing design, utilized an Agile Screen (five days

from recruit to report), which is an online quantitative survey

designed to provide quick consumer feedback based on a set

of key metrics, including purchase intent, uniqueness, believ-

ability, quality, and product function for this particular project.

This methodology helps clients reduce the ideas or concepts

that merit further attention. For this phase, a sample size of 200

respondents was used. Sample size is an important component

of statistical significance. “In order for our results to be consid-

ered representative and projected to the broader population,

we suggest a minimum sample size of 200. As the sample size

drops below 200, the margin of error widens and the results in

the survey have a higher risk of not encapsulating the true mean

of the population,” explained Laura Jett, research manager at

GutCheck. Participants for both research phases were screened

based on demographic and attitudinal/behavioral requirements,

including household shopping responsibility and type of stores

frequently visited; the sample needed to include shoppers from

both natural and conventional consumer segments to ensure

neither segment would be alienated by the packaging update.

With only 45 days to implement a design prior to the start of

Suja’s primary selling season, the short five-day-recruit-to-report

time frame was critical. Within days, it was clear the new packag-

ing outperformed the current (at time of study) design by nearly

a two-to-one ratio, giving the Suja team confidence to move for-

ward with the redesign. After taking into account all consumer

feedback, the Suja team made updates to the packaging, which

led to the final design, now available in stores.

Most importantly, the research results proved to both the-

marketing team and the key stakeholders that the packaging

redesign was not only worth the investment, but that the new

packaging would likely increase purchases. Having the results

so fast allowed the team to launch the new product line in time

for their peak buying season.

“The Suja Essentials product line is our number-one revenue-

driver and makes up the largest percentage of our portfolio; be-

cause of the risk involved in updating its packaging, this was one of

the most important studies we’ve done as a company,” said Tumini.

www.sujajuice.com; www.gutcheckit.com

>closeupcont’d

schi18939_ch04_069-084.indd 82 2/14/18 2:53 PM

>chapter 4 Stage 2: Research Design, An Overview

83

case study 76

causal-explanatory study 75

causal-predictive study 75

communication study 78

cross-sectional study 80

data collection design 73

descriptive study 75

ex post facto design 76

experiment 76

field conditions 78

laboratory conditions 78

longitudinal study 79

measurement instrument 80

monitoring study 78

multiple-methodology design 77

qualitative research 76

quantitative research 76

reporting study 74

research design 71

sampling design 72

single-methodology design 77

statistical study 76

>keyterms

Terms in Review

1 Distinguish between the following:

a. Exploratory and formal studies.

b. Experimental and ex post facto research designs.

c. Descriptive and causal studies.

Making Research Decisions

2 You have been asked to determine how hospitals prepare and train volunteers. Because you know relatively little about this subject, how will you find out? Be as specific as possible.

3 You are the administrative assistant for a division chief in a large holding company that owns several hotels and theme parks. You and the division chief have just come from the CEO’s office, where you were informed that the guest com- plaints related to housekeeping and employee attitude are increasing. Your onsite managers have mentioned some tension among the workers but have not considered it un- usual. The CEO and your division chief instruct you to inves- tigate. Suggest at least three different types of research that might be appropriate in this situation.

4 Kraft split off its global snack business, naming it Mondelez International. The name came from a massive name sug- gestion program from Kraft employees, involving more than 1, 700 names by more than 1, 000 Kraft employees over a five-month process. The name plays on the connotation of “delicious world.” How would you design this process to en- sure that viable names were produced?

5 P&G recently moved its IAMS pet food division from Vandalia (OH) to Mason (OH). The move affected 240 workers, who were offered positions to join the 2, 000 workers already occupying the pet food division headquarters in Mason. A company statement indicated, “We’re doing this to increase productivity, collaboration, and access to P&G’s resources/expertise.” P&G also told employees that it was beginning a separate, multi-month study on how to increase collaboration and efficiencies with the 250 employees still working in its pet food research and development (R&D) complex located in Lewisburg (OH). What research might be included in the multi-month study to determine the future of the Lewisburg R&D facility and its employees?

From Concept to Practice

6 Using the eight design descriptors (Exhibit 4-2), describe a study to determine why enrollment in your major at your uni- versity is increasing (decreasing).

7 Use the eight design descriptors in Exhibit 4-2 to profile the research described in the chapter Snapshots.

From the Headlines

8 Uber is facing a large number of lawsuits, among them a charge of technology theft from Google, installation of soft- ware that blanks out the availability of Uber rides near gov- ernment buildings or when government-employed riders call, discrimination in the workplace, etc. Using an Internet search on this topic, design a study to determine the effect on ridership of the charge you examined.

>discussionquestions

schi18939_ch04_069-084.indd 83 2/14/18 2:54 PM

>part II The Design of Business Research

84

>cases*

A Gem of a Study

Calling Up Attendance

Campbell-Ewald Pumps Awareness into the American Heart Association

Covering Kids with Health Care

Donatos: Finding the New Pizza

Goodyear’s Aquatred

Inquiring Minds Want to Know—NOW!

Ohio Lottery: Innovative Research Design Drives Winning

Open Doors: Extending Hospitality to Travelers with Disabilities

Proofpoint: Capitalizing on a Reporter’s Love of Statistics

Ramada Demonstrates Its Personal Best™

Starbucks, Bank One, and Visa Launch

Starbucks Card Duetto Visa

State Farm: Dangerous Intersections

Volkswagen’s Beetle

* You will find a description of each case in the Case Abstracts section of this textbook. Check the Case Index to determine whether a case provides data, the research instrument, video, or other supplementary material. Cases and case supplements are available in Connect.

schi18939_ch04_069-084.indd 84 2/14/18 2:54 PM

LO5-1 The six tasks that comprise sampling design.

LO5-2 The premises on which sampling theory is based.

LO5-3 The characteristics of accuracy and precision for measuring sample validity.

LO5-4 The two categories of sampling methods and the variety of sampling techniques within each category.

LO5-5 The various sampling techniques and when each is used.

LO5-6 The ethical issues related to sampling design.

>learningobjectives After reading this chapter, you should understand . . .

“When you sample something, you're using

the crutch of borrowing chords and melodies from a song that's

already great, that's already stood the test of time, that's already special.”

Gerald Earl Gillum (G-Eazy),

American rapper and record producer

Stage 2: Sampling Design

>chapter 5

schi18939_ch05_085-121.indd 85 2/14/18 3:08 PM

>part II The Design of Business Research

86

>Sampling Design

>Exhibit 5-1 Sampling Design in the Research Process

Sampling Design

RejectAccept

Investigative Questions

Choose CensusChoose Sample

Identify Existing Sample Frames

Define Number of Cases

Define Population Parameters

Define Target Population & Case

Define & Evaluate Sample Frames

Define Sampling Method

Nonprobability Probability

Define Selection & Recruiting Protocols

Draw Cases

Evaluate Existing Sample Frames

Modify or Construct Sample Frames

Select Sample Frames

This subprocess of research design answers the question: From whom or what (target population) does the data need to be collected and how and from how many (cases)? The steps in this subprocess (Exhibit 5-1) include:

1. Define the target population and a case (describe those entities—collectively and individually— that possess the desired information about the chosen variables and its parameters).

2. Define the population parameters of interest (summary descriptors—proportion, mean, variance— of study variables) in the population.

3. Identify and evaluate the sample frame (list of cases within the target population) or create one.

4. Define the number of cases needed (choose between a census or sample; choose the size of any sample).

5. Define the appropriate sampling method (the type of sample to be used).

6. Define the sampling selection and recruitment protocols (choose standardized procedures or custom-design ones).

schi18939_ch05_085-121.indd 86 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

87

If a study’s objective is to understand data security breaches, then the target population might be “any event where data were accessed by an unauthorized source.” Using operational definitions for “accessed” and “unauthorized source” would be critical. IT experts indicate exploratory data breaches by hackers are often so well hidden that they may not be discovered for months, if at all, until a major breach occurs. Would these exploratory breaches also be part of the target population? For each data breach (case), the population parameters of interest might be whether data were accessed but not modified, what modifications occurred, whether data were permanently lost, the method of access, etc. The sample frame would be the list of every such data breach within a specified period of time (e.g., five years). The number of cases we need depends on the size, variability, and accessibility of the target population. Depending on the number of breaches a company experienced, using a census (evaluating all instances of unauthorized access) might be feasible. If the number of breaches is very large, and exploration showed that one method of access and type of breach was common (limited variability), then using a sample (examining only a portion of unauthorized access records) might be chosen. In this instance, a researcher would need special forensic computing skills, thus using a sample might also be more desirable due to time constraints on a limited pool of researchers.

>Define the Target Population and Case In business, a target population can be any of the following (Exhibit 5-2), with a case being a single ele- ment drawn from that target population:

• People (individuals or groups: e.g., employees, customers, suppliers).

• Organizations or institutions (companies, trade associations, professional online communities, unions).

• Events and happenings (e.g., trade association meetings, presentations to financial analysts, industry conventions, employee picnics).

• Objects or artifacts (e.g., products, machines, production waste or byproducts, tools, maps, process models, ads).

• Settings and environments (e.g., warehouses, stores, factories, distribution facilities).

• Texts (e.g., annual reports, productivity records, social media posts, emails, memos, reports).

>Exhibit 5-2 Common Types of Target Populations in Business Research

Organizations

Events

Objects

Settings

Texts & Records

People

Target Population

schi18939_ch05_085-121.indd 87 2/14/18 3:08 PM

>part II The Design of Business Research

88

The definition of the target population may be apparent from the management problem or the re- search question(s), as it was in our data breach study, but often it is not so obvious. Sometimes there is more than one option for the target population. The researcher will choose one or more options that will provide him or her with the most answers to the investigative questions.

In the discussion that follows, we will use a dining study on a college campus: The researchers at Metro University (Metro U) are exploring the feasibility of creating a dining club whose facilities would be available on a membership basis. To launch this venture, they will need to make a substan- tial investment. Research will allow them to reduce many risks. Thus, the research question is: Would a membership dining club be a viable enterprise? Some investigative questions that flow from the research question include:

1. Who would patronize the club, and on what basis?

2. How many would join the club under various membership and fee arrangements?

3. How much would the average member spend per month?

4. What days would be most popular?

5. What menu and service formats would be most desirable?

6. What lunch times would be most popular?

7. Given the proposed price levels, how often per month would each member have lunch or dinner?

8. What percent of the people in the population say they would join the club, based on the pro- jected rates and services?

Is the target population for the dining club study at Metro University defined as “full-time day students on the main campus of Metro U”? Or should the population include “all persons employed at Metro U”? Or should townspeople who live in the neighborhood be included? Without knowing

In the midst of the financial crisis in the automobile industry,

Ford’s James Farley decided his research was excluding a very

important target population: dealers. With dealers controlling 75

percent of advertising expenditures for the auto giant, Farley

thought excluding them as research participants was suicidal.

So he recruited 30 of the most influential dealers to fly to Detroit

to provide information and critique the creative proposals of the

Ford ad agency, Team Detroit.

Farmington Hills (MI) full-service research firm Morpace put

the dealers through an intensive group interview experience.

The dealers were soon challenged with questions. “Which

incentives work and which don’t?” “What does the Ford brand

mean to you?” “What is wrong with Ford’s advertising?” In

subsequent sessions, the dealers were asked to critique ad

slogans and branding strategies and recommend those that

best captured the Ford experience. The dealers left the 72-hour

marathon session enthusiastic about the direction Ford was tak-

ing and with significant buy-in for the next ad campaign. Farley’s

actions gave voice to its dealers with its altered research sam-

pling design.

www.ford.com; www.morpace.com; www.teamdetroit.com

Ford Reenergizes by Changing Its Sampling Design

>snapshot

©Joe Raedle/Staff/Getty Images

schi18939_ch05_085-121.indd 88 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

89

the likely patron for the new venture, it is not obvi- ous which of these is the appropriate target popula- tion. Assume the Metro University Dining Club is to be solely for the students and employees on the main campus. The researchers might define the population as “currently-enrolled students and employees (full- and part-time) of Metro U, main campus, and their families.” Thus, any single student or employee would be a likely case.

> Define the Population Parameters

Population parameters are summary descriptors (e.g., incidence proportion, mean, variance, etc.) of vari- ables of interest in the population. Sample statistics are descriptors of those same relevant variables computed from sample data. Sample statistics are used as estima- tors of population parameters. The sample statistics are the basis of our inferences about the population. Ex- hibit 5-3 indicates population parameters for the Metro U dining study.

Depending on how measurement questions are phrased, each will collect a different level of data. Each different level of data also generates different sample statistics. Thus, choosing the parameters of interest will actually dictate your sample type and its size. Data have different properties depending on how they were collected. Exhibit 5-4 reviews the data types and these properties.

When the variables of interest in the study are measured on interval or ratio scales, we use the sample mean to estimate the population mean and the sample standard deviation to estimate the population standard deviation. When the variables of interest are measured on nominal or ordinal scales, we use the sample proportion of incidence (p) to estimate the population proportion and the pq to estimate the population variance where q = (1 − p). The population proportion of incidence “is equal to the number of cases in the population belonging to the category of interest, divided by the total number of cases in the population.”1 Proportion measures are necessary for nominal data and are widely used for other measures as well. The most frequent proportion measure is the percentage. In the Metro U study, examples of nominal data are the proportion of a population that expresses interest in joining the club (e.g., 30 percent; therefore p is equal to 0.3 and q, those not interested,

>Exhibit 5-3 Example Population Parameters in the Metro U Dining Study

Population Parameter of Interest Data Level and Measurement Scale

Frequency of eating on or near campus at a restaurant within the last 30 days

● Ratio data (actual number of eating experiences)

● Ordinal data (less than 5 times per month, greater than 5 but fewer than 10 times per month, greater than 10 times per month)

Proportion of student/employees expressing interest in the dining club

● Nominal data (interested, not interested)

Proportion of students/employees spending money per person per visit

● Interval data ($5–9.99, $10–14.99, $15–19.99, $20–24.99, $25–29.99)

© W

av eb

re ak

M ed

ia /G

et ty

I m

ag es

schi18939_ch05_085-121.indd 89 2/14/18 3:08 PM

>part II The Design of Business Research

90

equals 0.7) or the proportion of married students who report they now eat in restaurants at least five times a month.

There may also be important subgroups in the population about whom we would like to make esti- mates. For example, we might want to draw conclusions about the extent of dining club use that could be expected from married faculty versus single students, residential students versus commuter students, and so forth. Such questions have a strong impact on the nature of the sampling frame we accept (we would want the list organized by these subgroups, or within the list each characteristic of each case would need to be noted), the design of the sample, and its size.

>Exhibit 5-4 Data Types and Characteristics

Data Type Data Characteristics Example

Nominal Classification Respondent type (faculty, staff, student)

Ordinal Classification and Order Preferred doneness of steak (well done, medium well, medium rare, rare)

Interval Classification, order, & distance

How rated last restaurant experience (scale of 1-10; l=very poor, 10=exceptional)

Ratio Classification, order, distance & natural origin

Average $ amount spent per person for last dinner in restaurant.

>Define the Sampling Frame The sampling frame is the list of cases in the target population from which the sample is actually drawn. Ideally, it is a complete and correct list of population members only. As a practical matter, however, the sampling frame often differs from the desired population. For the dining club study, the Metro U direc- tory would be the logical first choice as a sampling frame. Published directories are usually accurate when published in the fall, but suppose the study is being done in the spring. The directory may contain errors and omissions because some people will have withdrawn or left since the directory was published, while others will have enrolled or been hired. Usually university directories don’t mention the families of students or employees. Just how much inaccuracy one can tolerate in choosing a sampling frame is a

With the growing number of people in cell-phone-only households, the printed phone directory has become obsolete as a sample frame for house- hold research. Specialized business directories are still viable options for business-to- business research.

© O

lly y/

S h

ut te

rs to

ck

schi18939_ch05_085-121.indd 90 2/19/18 11:34 AM

>chapter 5 Stage 2: Sampling Design

91

matter of judgment. You might use the directory anyway, ignoring the fact that it is not a fully accurate list. However, if the directory is a year old, the amount of error might be unacceptable. One way to make the sampling frame for the Metro U study more representative of the population would be to secure a supplemental list of the new students and employees as well as a list of the withdrawals and terminations from Metro U’s registrar and human resources databases. You could then craft your own sample frame by adding and deleting information from the original directory. Or, if their privacy policies permit, you might just request a current listing from each of these offices and combine these lists to create your sampling frame.

A greater distortion would be introduced if a branch campus population were included in the Metro U directory. This would be an example of a too-inclusive frame—that is, a frame that includes many cases other than the ones in which we are interested. A university directory that includes faculty and staff retirees is another example of a too-inclusive sampling frame.

Often you have to accept a sampling frame that includes people or cases beyond those in whom you are interested. You may have to use a telephone directory to draw a sample of business telephone numbers. Fortunately, this is easily resolved. You draw a sample from the larger population and then use a screening procedure to eliminate those who are not members of the group you wish to study.

The Metro U dining club survey is an example of a sampling frame problem that is readily solved. Often one finds this task much more of a challenge. Suppose you need to sample the members of an ethnic group, say, Asians residing in Little Rock, Arkansas. There is probably no list of this population. Although you may use the general city directory, sampling from this too-inclusive frame would be costly and inefficient because Asians represent only a small fraction of Little Rock’s population. The screen- ing task would be monumental. Because ethnic groups frequently cluster in certain neighborhoods, you might identify these areas of concentration and then use a city directory, which is organized by street address, to draw the sample.

It is not until we begin talking about sampling frames and sampling methods that international re- search starts to deviate. International researchers often face far more difficulty in locating or building sample frames. Countries differ in how each defines its population; this affects census and relevant population counts.2 Some countries purposefully over sample to facilitate the analysis of issues of par- ticular national interest; this means we need to be cautious in interpreting published aggregate national figures.3 These distinctions and difficulties may lead the researcher to choose nonprobability techniques or different probability techniques than they would choose if doing such research in the United States or other developed countries. In a study that is fielded in numerous countries at the same time, researchers may use different sampling methodologies, resulting in hybrid studies that will need care to be combined. It is common practice to weight sample data in cross-national studies to develop sample data that are representative.4 Choice of sampling methods is often dictated by culture as much as by communication and technology infrastructure. Just as all advertising campaigns would not be appropriate in all parts of the world, all sampling techniques would not be appropriate in all subcultures. The discussion in this text focuses more on domestic than international research. It is easier to learn the principles of research in an environment that you know versus one in which many students can only speculate. Ethnic and cultural sensitivity should influence every decision of researchers, whether they do research domestically or internationally.

>Define the Number of Cases The ultimate test of a sampling design is how well any cases we measure represent the characteristics of the target population the design purports to represent.

Sample versus Census  Most people intuitively understand the idea of why drawing a sample works. One taste from a drink tells us whether it is sweet or sour; we don’t need to drink the whole glass. If we select a few ads from a maga- zine, we assume our selection reflects the characteristics of the full set. If some members of our staff favor

schi18939_ch05_085-121.indd 91 2/14/18 3:08 PM

>part II The Design of Business Research

92

Early in panel development, panelists were offered $100 to

$150 to join and participate in qualitative studies. Historically,

these participants participated in longer engagements (on-

line qualitative, ethnography studies or face to face in-depth

interviews). As random digit dialing became less produc-

tive for recruiting respondents and the Internet became

more widely used, researchers started using panels—also

called communities—to recruit participants for quantitative

research. Quantitative respondents who are engaged for

much shorter periods are more likely to be paid $1 to $10 per

survey completed.

The professional respondent—one who takes repeated

surveys—was once considered a deterrent to quality research.

“As researchers, ” explained Jessica Broome, PhD, principal of

Jessica Broome Research, “we wanted to keep the ‘cheaters’ and

‘repeaters’ out of our studies, believing they biased results.”

As decision cycles shorten, the demand for better and more

timely information means attracting and retaining qualified par-

ticipants. During the last three decades, increasingly this means

researchers turn to panels, and by design, these participants

are asked to participate in numerous studies.

“We wanted to know ‘Who are these people willing to take

repeated surveys?’” explained Broome. “And given that some sur-

veys are overly long and others poorly designed, ‘Why do they

do this?’” Broome teamed with Kerry Hecht, Director of Research

Services, Recollective, a division of Ramius Corporation, to find

out what motivates panel respondents and if their motivations are

likely to reduce the quality of the information they provide.

The Qualitative Study

Broome and Hecht designed a multistage study that drew par-

ticipants from multiple panel providers, including Critical Mix,

Schlesinger Associates, and Swagbucks. “We started with a

5-day online qualitative community study with 20 people to

explore what got them started as a survey panelist and what

kept them going,” explained Hecht. “While money is a motivator

in keeping panelists engaged, they also shared the influences

of intrinsic motivators like fun, feeling useful, contributing to

important decisions, and participation being more interesting

than time spent on social media.”

For any particular study, panelists are often screened exclu-

sively on demographics. “Because participants derive intrinsic

benefits, panelists will sometimes fudge on screening informa-

tion in an attempt to be included in a study, ” shared Broome.

When a panelist doesn’t meet the desired demographic param-

eters, they are told “You don’t qualify” but are rarely told why.

“But once they are included, ” explained Broome, “participants

claim honesty drives their responses; they think lying on survey

questions would undermine the study.”

“Often panelists expressed feeling abused, misled, and dis-

respected. For example, when they are told about survey length,

they often felt deceived when a promised 15-minute survey took

45 minutes, or when the survey was not only long, but boring,”

explained Broome. Increasingly, panelists in quantitative stud-

ies are tech-savvy. “They understand what current technology

should permit a survey company to do—like eliminate the need

to ask demographic questions repeatedly in the same survey

process or use earlier answers to filter later questions, ” claimed

Hecht. “They basically think researchers can make the experi-

ence so much better.” “The research industry needs their in-

sights, but can treat panelists with disdain," claimed Broome.

The Quantitative Study

Broome and Hecht followed their qualitative exploration study

with Phase 2, a mobile-optimized quantitative study of 1,499

participants, also drawn from various panel providers and fielded

by Propeller Insights. Each panelist took a topical survey that

included a creativity assessment. Additionally, half the group

(750) took a VARK assessment. VARK assesses visual, aural/

audio, read/write, and kinesthetic learning preferences through

a series of learning scenario questions. Creativity was assessed

through a battery of 38 statements requiring agreement or dis-

agreement, as well as a checklist of 54 descriptors. “We dis-

covered that participants didn’t favor any one of the learning

approaches, nor were they outliers on the creativity assessment,”

shared Broome.

Who’s Really Taking Your Surveys?

>closeup

schi18939_ch05_085-121.indd 92 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

93

a particular software training strategy, we infer that others will also. The basic idea of taking a sample is that by selecting some cases in a population, we may draw conclusions about the entire target population.

There are several compelling reasons for using a sample (a subset of the target population) rather than a census (all cases within a population), including (1) lower cost, (2) greater speed of data collection, (3) availability of population cases, and (4) greater accuracy of results. The advantages of taking a sample over a census are less compelling when two conditions exist: (1) a census is feasible due to a small target population and (2) a census is necessary when the cases are quite different from each other.5 When the population is small and variable, any sample we draw may not be representative of the popula- tion from which it is drawn. The resulting values we calculate from the sample are incorrect as estimates of the population values.

Lower Cost The economic advantages of taking a sample rather than a census are massive. Consider the cost of taking a census. The 2020 U.S. Census of the Population is expected to cost as much as $30 billion, barring any natural disasters and using 2010 costs as a barometer; the Census Bureau’s own estimate is $22 billion.6 By any reasonable standard, continuing to take a census of the population every 10 years is unsustainable. Is it any wonder that researchers in all types of organizations ask, “Why should we spend thousands of dollars interviewing thousands of employees in our company if we can find out what we need to know by asking only a few hundred?”

Greater Speed of Data Collection  Due to the smaller number of cases in a sample, using a sample drawn from a target population will always take less time than conducting a census.

Broome and Hecht’s Phase 2 research revealed factors es-

sential for reducing participant discontinuation: Don’t ask screen-

ing questions within a survey once a participant has answered

a detailed screener to qualify; merge the data. Participants don’t

care that different companies do different aspects of the research;

they want researchers to avoid duplication. Don’t ask essentially

the same question in different ways; once is enough. Use pre-

vious question responses to determine later questions asked;

panelists are willing to share their thoughts, behaviors, and lives

but don’t want their time wasted. Participants have different pref-

erences and styles; let them respond as they wish—with text or

video, exclusively on a mobile device or a computer. “They shared

lots of ideas for making surveys more engaging and interesting,”

claimed Hecht, “including, adding music or video, getting rid of

grid questions, and reducing survey time to 15 minutes or less.”

Using information extracted from an open question and ana-

lyzed with OdinText, Phase 2 research also revealed that “most

participants were a member of one or two panels but struggled

to remain engaged with their panels,” shared Hecht. “Participants

appreciate feeling like a valued part of a full process, not just

an interchangeable cog in a wheel. They like to participate

in studies that are interesting to them and about products or

services that are relevant to their daily lives—knowledge is a

powerful motivator.” Many also cited meeting interesting people

and hearing different viewpoints as motivation. “They also love

learning the results of studies they participate in and under-

standing why we as researchers do what we do or ask what we

ask,” shared Broome.

Recently, a SurveyMonkey study also found panelists gave

thoughtful, consistent answers over time. It released results of a

1,000-person international panel assessment study, using three

surveys with the same respondents, one in each of three se-

quential months, which checked for quality-reducing behaviors

like straight lining (repeatedly choosing the same answer choice

in matrix questions), poor open response validity (responding

with nonhelpful, gibberish answers), and whether they were un-

focused or not paying attention. Of the panelists, 97 percent,

97 percent, and 94 percent or more passed each of these tests,

respectively, with no difference between men and women.

And in terms of response reliability, over the three waves of

surveys, in 23 indicators SurveyMonkey tracked, only three

items showed significantly significant change among U.S. par-

ticipants: time-to-complete, choice of the “other” response, and

attitude about “moral acceptability of alcohol use.”

jessicabroomeresearch.com; recollective.com; surveymonkey.com

>closeupcont’d

schi18939_ch05_085-121.indd 93 2/14/18 3:08 PM

>part II The Design of Business Research

94

Availability of Population Cases Some situations require using a sample. Safety is a compelling appeal for most vehicles. Yet we must have evidence to make safety claims. Therefore, we crash-test cars to evaluate bumper strength or efficiency of airbags to prevent injury. In testing for such evidence, we destroy the cars we test. A census would require complete destruction of all cars manufactured. Drawing a sample is also the only process possible if the population is infinite.

Better Quality Results W. Edwards Deming, who developed the sampling techniques used by the U.S. Census Bureau and the Bureau of Labor Statistics, argues that the quality of a study is often better with a sample than with a census. He suggests samples possess “the possibility of better interviewing (testing), more thorough in- vestigation of missing, wrong, or suspicious information, better supervision, and better processing than is possible with complete coverage.”7 Error related to research comes from two sources: errors related to the sample itself (sampling error, or estimates of a variable drawn from a sample differ from true value of a population parameter) and error not related to the sample but to all other decisions made in the research design (nonsampling error). As Exhibit 5-5 depicts, and research supports, nonsampling error often exceeds sampling error.8 For a sample to be valid, it must offer both accuracy and precision. The U.S. Bureau of the Census, while mandated to take a census of the population every 10 years, shows its confidence in sampling by taking sample surveys to check the accuracy of its census. The U.S. Bureau of the Census knows that in a census, segments of the population are seriously undercounted.

Accuracy Accuracy is the degree to which bias is absent from the sample. When the sample is drawn properly, the measure of behavior, attitudes, or knowledge (the measurement variables) of some cases will be less than (thus, underestimate) the measure of those same variables drawn from the population. Also, the measure of the behavior, attitudes, or knowledge of other cases will be more than the population values (thus, overestimate them). Variations in these sample values offset each other, resulting in a sample value that is close to the population value. Thus an accurate (unbiased) sample is one in which the underes- timators offset the overestimators. For these offsetting effects to occur, however, there must be enough cases in the sample, and they must be drawn in a way that favors neither overestimation nor underestima- tion. Only when the population is small, accessible, and highly variable is accuracy likely to be greater with a census than a sample.

For example, assume in the Metro U study you measured likelihood to join the dining club. Hypo- thetically, you could measure via sample or census. In a hypothetical census, 32 percent of the popula- tion (currently enrolled students and current employees) said they would join. However, 52 percent of a sample made that choice. With both results for comparison, you would know that your sample was not representative because it significantly overestimated the population value of 32 percent. Without the

>Exhibit 5-5 Sources of Error

Sampling Error

What

Why

Non-sampling Error

Sample estimate is

not the population parameter

Inappropriate Sampling Method

Errors in Measurement

Instrument

Behavioral E�ects

schi18939_ch05_085-121.indd 94 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

95

census information, you might not know that you have significant error. Unfortunately, in most studies taking a census is not feasible, so we need an estimate of the amount of error.9

Systematic variance has been defined as “the variation in measures due to some known or unknown influences that ‘cause’ the scores to lean in one direction more than another.”10 Homes on the corner of the block, for example, are often larger and more valuable than those within the block. Thus, a sample that selects only corner homes will cause us to overestimate home values in the area.

Increasing the sample size can reduce systematic variance as a cause of error. However, even the larger size won’t reduce error if the sample frame from which you draw your cases is biased. The clas- sic example of a sample with systematic variance was the Literary Digest presidential election poll in 1936, in which more than 2 million people participated. The poll predicted Alfred Landon would defeat Franklin Roosevelt for the presidency of the United States. Your memory is correct; we’ve never had a president named Alfred Landon. The poll drew its sample from telephone owners, who were in the middle and upper classes—at the time, the bastion of the Republican Party—while Roosevelt appealed to the much larger working class, whose members could not afford to own phones and typically voted for the Democratic Party candidate.

Precision A second criterion of a valid sample is precision of estimate. Researchers accept that no sample will fully represent its population in all respects. However, to interpret the findings of research, we need a measure of how closely the sample estimate represents the population parameter on any variable of inter- est. The numerical descriptors that describe samples may be expected to differ from those that describe populations because of random fluctuations inherent in the sampling process. Sampling error reflects the influence of chance in drawing the sample cases. Sampling error is what is left after all known sources of systematic variance have been accounted for. In theory, sampling error consists of random fluctuations only, although some unknown systematic variance may be included when too many or too few cases possess a particular characteristic.

Precision is measured by the standard error of estimate, a type of standard deviation measurement; the smaller the standard error of estimate, the higher is the precision of the sample. The ideal sample produces a small standard error of estimate. However, not all types of sample design provide estimates of precision, and samples of the same size can produce different amounts of error.

Sample Size So assume you have chosen a sample rather than a census. Much folklore surrounds the question: How many cases should comprise your sample. The most pervasive myths are (1) a sample must be large or it is not representative and (2) a sample should bear some proportional relationship to the size of the population from which it is drawn. With nonprobability samples, researchers use subgroups, rules of thumb, and budget considerations to settle on a sample size. In probability sampling, how large a sample should be is a function of the variation in the population parameters under study and the estimating precision needed by the researcher. Some principles that influence sample size include:

• The greater the dispersion or variance within the population, the larger the sample must be to pro- vide estimation precision.

• The greater the desired precision of the estimate, the larger the sample must be.

• The narrower or smaller the error range, the larger the sample must be.

• The higher the desired confidence level in the estimate, the larger the sample must be.

• The greater the number of subgroups of interest within a sample, the greater the sample size must be, as each subgroup must meet minimum sample size requirements.

Cost considerations influence decisions about the size and type of sample and the data collection methods. Probability sample surveys incur list costs for sample frames, callback costs, and a variety of other costs that are unnecessary when nonprobability samples are used. As research has budgetary

schi18939_ch05_085-121.indd 95 2/14/18 3:08 PM

>part II The Design of Business Research

96

constraints, this may encourage a researcher to use a nonprobability sample. When the data collection method is changed, the amount and type of data that can be obtained also change. Note the effect of a $2,000 budget on hypothetical sampling considerations:11

• Simple random sampling: $25 per interview; 80 completed interviews.

• Telephone interviews: $10 per participant; 200 completed interviews.

• Self-administered questionnaire: $12 per participant; 167 completed instruments.

• Geographic cluster sampling: $20 per interview; 100 completed interviews.

The investment required to open the dining club at Metro U justifies the more careful probability ap- proach which means the researcher needs a quality sample frame. See the chapter appendix, Calculate the Sample Size.

If a census is chosen, then the researcher need not determine sample method or size. For most busi- ness research, however, we use samples. The remainder of the chapter will focus on sampling designs using samples.

>Define the Sampling Method The researcher faces a basic choice in sampling method: a probability sample or nonprobability sample. Any discussion of the relative merits of probability versus nonprobability sampling clearly shows the technical superiority of the former. Yet businesses often use nonprobability methods.

Key to the difference between nonprobability and probability samples is the term random. In the dictionary, random is defined as “without pattern”or as “haphazard.” In sampling, random means some- thing else entirely. Probability sampling is based on the concept of random selection—a controlled proce- dure that assures that each case is given a known nonzero chance of selection. This procedure is never haphazard. Only probability samples provide estimates of precision. When a researcher is making a deci- sion that will influence the expenditure of thousands, if not millions, of dollars, an estimate of precision is critical. Under such conditions, we can have substantial confidence that the sample is representative of the population from which it is drawn. In addition, with probability sample designs, we can estimate an error range within which the population parameter is expected to fall. Thus, we can reduce not only the chance for sampling error, but also estimate the range of probable sampling error present. Also, only probability samples offer the opportunity to generalize the findings to the population of interest from the sample population. Although exploratory research does not necessarily demand this, explanatory, descriptive, and causal studies do.

Alternatively, nonprobability sampling is arbitrary and subjective; when we choose subjectively, we usually do so with a pattern or scheme in mind (e.g., only talking with department heads or only talking with women). Each member of the target population does not have a known nonzero chance of being included. Allowing data collectors to use their judgment in drawing records or choosing par- ticipants is arbitrary. Early Internet samples had all the drawbacks of nonprobability samples. Those individuals who frequented the Internet were not representative of most target populations, because far more young, technically-savvy men frequented the Internet than did any other demographic group. As Internet access has reached saturation, with some estimates near 88 percent in the United States, this concern is diminished.12 Such samples now closely approximate non-Internet samples. Of increas- ing concern, however, is what the Bureau of the Census labels the “great digital divide”—low-income and ethnic subgroups’ underrepresentation in their use of technology compared to the general popu- lation.13 Additionally, many Internet samples were, and still are, drawn substantially from panels or communities. These are composed of individuals who have self-selected to become part of a pool of individuals interested in participating in online research. There is much discussion among profes- sional researchers about whether Internet samples should be treated as probability or nonprobability samples. Some admit that any sample drawn from a panel is more appropriately treated as a nonprob- ability sample; others vehemently disagree, citing the success of such well-known panels as Nielsen Media’s People Meter panels for radio audience assessment.14 As you study the differences here, you should draw your own conclusion.

schi18939_ch05_085-121.indd 96 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

97

>Exhibit 5-6 Types of Sampling Designs

Representation Basis

Case Selection Probability Nonprobability

Unrestricted Simple random Convenience

Restricted Complex random

● Systematic

● Cluster

● Stratified

● Double

Purposive

● Judgment

● Quota

● Snowball

With a probability sample, a researcher can make probability-based confidence estimates of various parameters that cannot be made with nonprobability samples. Choosing a probability sampling tech- nique has several consequences. A researcher must follow appropriate procedures so that:

• Interviewers or others cannot modify the selections made.

• Only the selected cases from the original sampling frame are included.

• Substitutions are excluded except as clearly specified and controlled according to predetermined protocols.

Despite all due care, the actual sample drawn will not match perfectly the probability sample that is originally planned. Some people will refuse to participate, and others will be difficult, if not impossible, to find. Thus, no matter how careful we are in replacing those who refuse or are never located, sampling error is likely to rise.

With personnel records available at Metro U and a population that is geographically concentrated, a probability sampling method is possible in the dining club study. University directories are generally available, and the costs of using a simple random sample would not be great here. Then, too, because the researchers are thinking of a major investment in the dining club, they would like to be highly confident they have a representative sample.

The researcher makes several decisions when choosing a design that uses a sample. These are repre- sented in Exhibit 5-6. The sampling decisions flow from the management question, the research ques- tion, and the specific investigative questions that evolve from the research question. These decisions are influenced by requirements of the project and its objectives, level of risk the researcher can tolerate, budget, time, available resources, and culture.

>Probability Sampling Simple Random Sampling The unrestricted, simple random sample is the purest form of probability sampling. Because all prob- ability samples must provide a known nonzero probability of selection for each population element, the simple random sample is considered a special case in which each population element has a known and equal chance of selection.

Probability of selection = Sample size

___________ Population size

The Metro U dining club study has a population of 20,000. If the sample size is 300, the probability of selection is 1.5 percent (300/20,000 = 0.015). In this section, we use the simple random sample to build a foundation for understanding sampling procedures and choosing probability samples. The simple random sample is easy to implement with automatic dialing (random dialing) and with computerized voice response systems. However, it requires a list of population elements, can be time-consuming and expensive, and can require larger sample sizes than other probability methods. Exhibit 5-7 provides an overview of the steps involved in choosing a random sample.

schi18939_ch05_085-121.indd 97 2/14/18 3:08 PM

>part II The Design of Business Research

98

Complex Probability Sampling The limitations of simple random sampling have led to the development of alternative designs that are superior to the simple random design in statistical and/or economic efficiency.

A more efficient sample in a statistical sense is one that provides a given precision (standard error of the mean or proportion) with a smaller sample size. A sample that is economically more efficient is one that provides a desired precision at a lower dollar cost. We achieve this with designs that enable us to lower the costs of data collecting, usually through reduced travel expense and interviewer time.

In the discussion that follows, four alternative probability sampling approaches are considered: (1) systematic sampling, (2) stratified random sampling, (3) cluster sampling, and (4) double sampling.

Systematic Sampling

A versatile form of probability sampling is systematic sampling. In this approach, every kth element in the population is sampled, beginning with a random start of any case in the range of 1 to k. The kth case, or skip interval, is determined by dividing the sample size into the population size to obtain the skip pattern applied to the sampling frame. This assumes that the sample frame is an accurate list of the population; if not, the number of cases in the sample frame is substituted for population size.

k = Skip interval = Population size

___________ Sample size

The major advantage of systematic sampling is its simplicity and flexibility. It is easier to instruct field interviewers to choose the dwelling unit listed on every kth line of a listing sheet than it is to use a random numbers table. With systematic sampling, there is no need to number the entries in a large personnel file before drawing a sample. To draw a systematic sample, do the following:

• Identify, list, and number the cases in the population.

• Identify the skip interval (k).

• Identify the random start.

• Draw a sample by choosing every kth entry.

>Exhibit 5-7 How to Choose a Random Sample

Selecting a random sample is accomplished with the aid of computer software, a table of random numbers, or a calculator with a random number generator. Drawing slips out of a hat or Ping-Pong balls from a drum serves as an alternative if every case in the sampling frame has an equal chance of selection. Mixing the slips (or balls) and returning them between every selection ensures that every case is just as likely to be selected as any other.

A table of random numbers (such as Appendix D, Exhibit D-10) is a practical solution when no software program is available. Random number tables contain digits that have no systematic organization. Whether you look at rows, col- umns, or diagonals, you will find neither sequence nor order. Exhibit C-1 in Appendix C is arranged into 10 columns of five-digit strings, but this is solely for readability.

Assume the researchers want a sample of 10 from a population of 95 cases. How will the researcher begin?

Other approaches to selecting digits are endless: horizontally right to left, bottom to top, diagonally across columns, and so forth. Computer selection of a simple random sample will be more efficient for larger projects.

1 Assign each case within the sampling frame a unique number from 01 to 95.

2 Identify a random start from the random number table. Drop a pencil point-first onto the table with closed eyes. Let’s say the pencil dot lands on the eighth column from the left and 10 numbers down from the top of Exhibit C-1, marking the five digits 05067.

3 Determine how the digits in the random number table will be assigned to the sampling frame to choose the specified sample size (researchers agree to read the first two digits in this column downward until 10 are selected).

4 Select the sample cases from the sampling frame (05, 27, 69, 94, 18, 61, 36, 85, 71, and 83) using the above process. (The digit 94 appeared twice and the second instance was omitted; 00 was omitted because the sampling frame started with 01.)

schi18939_ch05_085-121.indd 98 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

99

Invoices or customer accounts can be sampled by using the last digit or a combination of digits of an invoice or customer account number. Time sampling is also easily accomplished.

Systematic sampling can introduce subtle biases. A concern with systematic sampling is the possible periodicity in the population that parallels the sampling ratio. In sampling restaurant sales of dessert by drawing days of the year, a skip interval of 7 would bias results, no matter which day provides the random start. A less obvious case might involve a survey in an area of apartment buildings where the typical pattern is eight apartments per building. A skip interval of 8 could easily over sample some types of apartments and under sample others.

Another difficulty may arise when there is a monotonic trend in the population cases. That is, the population list varies from the smallest to the largest case or vice versa. Even a chronological list may have this effect if a measure has trended in one direction over time. Whether a systematic sample drawn under these conditions provides a biased estimate of the population mean or proportion depends on the initial random draw. Assume that a list of 2,000 commercial banks is created, arrayed from the largest to the smallest, from which a sample of 50 must be drawn for analysis. A skip interval of 40 (2,000 divided by 50) beginning with a random start at 16 would exclude the 15 largest banks and give a small-size bias to the findings.

The only protection against these subtle biases is constant vigilance by the researcher. Some ways to avoid such bias include:

• Randomize the target population before drawing the sample (e.g., order the banks by name rather than size).

• Change the random start several times in the sampling process.

• Replicate a selection of different samples.

Although systematic sampling has some theoretical problems, from a practical point of view it is usually treated as equivalent to a simple random sample. When similar cases are grouped within the sampling frame, systematic sampling is statistically more efficient than a simple random sample. This might occur if the listed cases are ordered chronologically, by size, by class, and so on. Under these conditions, the sample approaches a proportionate stratified sample. The effect of this ordering is more pronounced on the results of cluster samples than for other samples and may call for a proportionate stratified sampling formula.15

Stratified Random Sampling

Most populations can be segregated into several mutually exclusive subpopulations, or strata. The pro- cess by which the sample is constrained to include cases from each of the segments is called stratified random sampling. University students can be divided by their class level, school or major, gender, and so forth. After a population is divided into the appropriate strata, a simple random sample can be taken within each stratum. The results from the study can then be weighted (based on the proportion of the strata to the population) and combined into appropriate population estimates.

There are three reasons a researcher chooses a stratified random sample: (1) to increase a sample’s statistical efficiency, (2) to provide adequate data for analyzing the various subpopulations or strata, and (3) to enable different research methods and procedures to be used in different strata.16

Stratification is usually more efficient statistically than simple random sampling and at worst it is equal to it. With the ideal stratification, each stratum is homogeneous internally (cases are similar) and heterogeneous with other strata (cases of one stratum are not similar to cases within another stratum). This might occur in a sample that includes members of several distinct ethnic groups. In this instance, stratification makes a pronounced improvement in statistical efficiency.

Stratification is also useful when the researcher wants to study the characteristics of certain popula- tion subgroups. Thus, if one wishes to draw some conclusions about activities in the different classes of a student body, stratified sampling would be used. Similarly, if a restaurant were interested in testing menu changes to attract younger patrons while retaining its older, loyal customers, stratified sampling using age and prior patronage as descriptors would be appropriate. Stratification is also called for when different methods of data collection are applied in different parts of the population, a research design that is becoming increasingly common. This might occur when we survey company employees at the

schi18939_ch05_085-121.indd 99 2/14/18 3:08 PM

>part II The Design of Business Research

100

Twice yearly Keynote Systems evaluates the performance of five

search engines, including market leader Google, AOL Search,

Yahoo! Search, Ask.com, and MSN Search. Keynote, a “worldwide

leader in services that improve online business performance and

communications technologies,” uses an online panel to perform

“interactive Web site tests to assess user experience,” profiling

not only how people use search engines, but why they search

as they do. Keynote allocates participants and experimental

treatments as in Exhibit CU-1: 2,000 people are randomly drawn

from more than 160,000 panel members and invited to partici-

pate via e-mail. They are assigned randomly to five groups of

400; each group is assigned a particular search engine. Whether

participants have any experience with that particular engine is

not a criterion for assignment. Each group is assigned a series

of search tasks, starting with a general task—Think about any-

thing you would like to search for; go and search that—to more

specific tasks—find a local establishment, a product, an image,

and a news item. Each search engine–allocated group essen-

tially performs the same series of tasks. From their activities, Key-

note generates 250,000 metrics (including time involved in the

search, whether the search was successful, etc.). It matches these

metrics to survey data used to measure satisfaction, perceived

difficulty, and specific frustrations. From this combined data it de-

velops several indices.

“One of the things we noted from a series of such tests was

that Google repeatedly received rave reviews, even in instances

where performance measures told a different story,” shares se-

nior research consultant Lance Jones. With almost 60 percent

market share, Google has strong recognition and tends to set

the bar in search site design. Is its brand that powerful that it

can influence attitudes even in the face of conflicting perfor-

mance experience? If the brand is not a factor, which search

engine would produce the most satisfying and useful results,

the best sponsored results, and the best presentation and de-

sign? Keynote wanted to design an experiment that would show

the power of the search engine brand. To do that, it needed to

remove brand identity from the search results. Its solution was

to design a generic-appearing search engine website and re-

sults format page, feeding actual search results into its generic

format.

For the brand power test (Exhibit CU-2) 2,000 participants

were again divided into five groups and assigned one search

engine. This time, however, half the participants were assigned

to a branded group (n = 200) and would see the results with

a text line “Results brought to you by Yahoo/Google/Ask, etc.”;

the other half (n = 200) would see the same results but without

the brand notation line (n = 200). All five search engines were

tested using the tasks performed in the standard twice-annual

test, but all the results seen by participants were actually gener-

ated using the assigned search engine, then fed into the ge-

neric results presentation. “The results pages were delivered

live and participants would have perceived no difference in

elapsed time, as the results were delivered within milliseconds

of what the standard search would have delivered,” explained

Jones. The test produced 1, 600 queries that generated 12 dis-

tinct metrics.

Keynote Systems Tests the Power of Search

>closeup

>Exhibit CU-1 Participant Allocation in Search Engine Test

Grocery (n = 200)

Restaurant (n = 200)

Movie (n = 200)

Self-Select (n = 200)

General Task (n = 200)

Local Task (n = 200)

Product Task (n = 200)

Image Task (n = 200)

New Item Task (n = 200)

Experiment Participants (n = 2000)

Search Engine A Group (n = 400)

Search Engine B Group (n = 400)

Search Engine C Group (n = 400)

Search Engine D Group (n = 400)

Search Engine E Group (n = 400)

Source: Lance Jones, senior research consultant.

schi18939_ch05_085-121.indd 100 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

101

home office with one method but must use a different approach with employees scattered throughout the country or the world.

If data are available on which to base a stratification decision, how shall we go about it?17 The ideal stratification would be based on the primary variable under study. If the major concern were to learn how often per month patrons would use the Metro U dining club, then the researcher would stratify on the expected number of use occasions. The only difficulty with this idea is that if we knew this informa- tion, we would not need to conduct the study. We must, therefore, pick a variable for stratifying that we believe will correlate with the frequency of club use per month, something like days at work or class schedule as an indication of when a sample case might be near campus at mealtimes.

Researchers often have several important variables about which they want to draw conclusions. A reasonable approach is to seek some basis for stratification that correlates well with the major variables. It might be a single variable (class level), or it might be a compound variable (class by gender). In any event, we will have done a good stratifying job if the stratification base maximizes the difference among strata means and minimizes the within-stratum variances for the variables of major concern.

The more strata used, the closer you come to maximizing inter-strata differences (differences between strata) and minimizing intrastratum variances (differences within a given stratum). You must base the decision partially on the number of subpopulation groups about which you wish to draw separate conclu- sions. Costs of stratification also enter the decision. The more strata you have, the higher the cost of the

Is a brand powerful? Here are some sample results for Google;

keep in mind that the branded group and the unbranded group

saw the exact same results pages. On the unbranded group,

the calculated Google results satisfaction score was 732 (on a

1,000-point scale), while the branded group delivered an 800;

Google’s sponsored results satisfaction was 763 (unbranded) com-

pared to 809 (branded); full design satisfaction was 753 (unbranded)

compared to 806 (branded). Evaluate the design of this sample.

www.keynote.com

>closeupcont’d >Exhibit CU-2 Participant Allocation in Brand Power Test

Experiment Participants (n = 2000)

Search Engine A Group (n = 400)

Search Engine B Group (n = 400)

Search Engine C Group (n = 400)

Search Engine D Group (n = 400)

Search Engine E Group (n = 400)

Branded Group (n = 200)

General Task (n = 200)

Local Task (n = 200)

Product Task (n = 200)

Image Task (n = 200)

New Item Task (n = 200)

Grocery (n = 200)

Restaurant (n = 200)

Movie (n = 200)

Self-Select (n = 200)

Unbranded Group (n = 200)

General Task (n = 200)

Local Task (n = 200)

Product Task (n = 200)

Image Task (n = 200)

New Item Task (n = 200)

Grocery (n = 200)

Restaurant (n = 200)

Movie (n = 200)

Self-Select (n = 200)

schi18939_ch05_085-121.indd 101 2/14/18 3:08 PM

>part II The Design of Business Research

102

research project due to the cost associated with more detailed sampling. There is little to be gained in estimating population values when the number of strata exceeds six.18

The size of the strata samples is calculated with two pieces of information: (1) how large the total sample should be and (2) how the total sample should be allocated among strata. In deciding how to allocate a total sample among various strata, there are proportionate and disproportionate options.

Proportionate versus Disproportionate Sampling In proportionate stratified sampling, each stratum is proportionate to the stratum’s share of the total population. This approach is more popular than any of the other stratified sampling procedures. Some reasons for this include:

• It has higher statistical efficiency than a simple random sample.

• It is much easier to carry out than other stratifying methods.

• It provides a self-weighting sample; the population mean or proportion can be estimated simply by calculating the mean or proportion of all sample cases, eliminating the weighting of responses.

On the other hand, proportionate stratified samples often gain little in statistical efficiency if the strata measures and their variances are similar for the major variables under study.

Any stratification that departs from the proportionate relationship is disproportionate stratified sampling. There are several disproportionate allocation schemes. One type is a judgmentally determined disproportion based on the idea that each stratum is large enough to secure adequate confidence levels and error range estimates for individual strata. The following table shows the relationship between pro- portionate and disproportionate stratified sampling.

Stratum Population Proportionate Sample Disproportionate Sample

Male 45% 45% 35%

Female 55 55 65

A researcher makes decisions regarding disproportionate sampling, however, by considering how a sample will be allocated among strata. One author states,

In a given stratum, take a larger sample if the stratum is larger than other strata; the stratum is more variable internally; and sampling is cheaper in the stratum.19

If one uses these suggestions as a guide, it is possible to develop an optimal stratification scheme. When there is no difference in intra-stratum variances and when the costs of sampling among strata are equal, the optimal design is a proportionate stratified sample.

While disproportionate stratified sampling is theoretically superior, there is some question as to whether it has wide applicability in a practical sense. If the differences in sampling costs or variances among strata are large, then disproportionate sampling is desirable. It has been suggested that “differ- ences of several-fold are required to make disproportionate sampling worthwhile.20

The process for drawing a stratified sample is:

• Determine the variables to use for stratification.

• Determine the proportions of the stratification variables in the population.

• Select proportionate or disproportionate stratification based on project information needs and risks.

• Divide the sampling frame into separate frames for each stratum.

• Randomize the case listing within each stratum’s sampling frame.

• Follow random or systematic procedures to draw the sample from each stratum.

Cluster Sampling

In a simple random sample, each population element is selected individually. The population can also be divided into groups of elements with some groups randomly selected for study. This is cluster sampling. Cluster sampling differs from stratified sampling in several ways, as indicated in Exhibit 5-8.

schi18939_ch05_085-121.indd 102 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

103

Two conditions foster the use of cluster sampling: (1) the need for more economic efficiency than can be provided by simple random sampling and (2) the frequent unavailability of a practical sampling frame listing individual cases.

Statistical efficiency for cluster samples is usually lower than for simple random samples chiefly be- cause clusters often don’t meet the need for heterogeneity and, instead, are homogeneous. For example, families in the same block (a typical cluster) are often similar in social class, income level, ethnic origin, and so forth. Although statistical efficiency in most cluster sampling may be low, economic efficiency is often great enough to overcome this weakness. The criterion, then, is the net relative efficiency resulting from the trade-off between economic and statistical factors. It may take 690 interviews with a cluster design to give the same precision as 424 simple random interviews. But if it costs only $5 per interview in the cluster situation and $10 in the simple random case, the cluster sample is more attractive ($3,450 versus $4,240).

Area Sampling Much research involves populations that can be identified with some geographic area. When this occurs, it is possible to use area sampling, the most important form of cluster sampling. This method overcomes the problems of both high sampling cost and the unavailability of a practical

>Exhibit 5-8 Comparison of Stratified and Cluster Sampling

Stratified Sampling Cluster Sampling

1. We divide the population into a few subgroups.

● Each subgroup has many cases in it.

● Subgroups are selected according to some crite- rion that is related to the variables under study.

2. We try to secure homogeneity within subgroups.

3. We try to secure heterogeneity between subgroups.

4. We randomly choose cases from within each subgroup.

1. We divide the population into many subgroups.

● Each subgroup has few cases in it.

● Subgroups are selected according to some crite- rion of ease or availability in data collection.

2. We try to secure heterogeneity within subgroups.

3. We try to secure homogeneity between subgroups.

4. We randomly choose several subgroups that we then typically study in depth.

A low-cost, frequently used method, the area cluster sample may use geographic sample units (e.g., city blocks).

© Te

tr a

Im ag

es /A

la m

y

schi18939_ch05_085-121.indd 103 2/14/18 3:08 PM

>part II The Design of Business Research

104

sampling frame for individual cases. Area sampling methods have been applied to national populations, county populations, and even smaller areas where there are well-defined political or natural boundaries.

Suppose you want to survey the adult residents of a city. You would seldom be able to secure a listing of such individuals. It would be simple, however, to get a detailed city map that shows the blocks of the city. If you take a sample of these blocks, you are also taking a sample of the adult residents of the city.

Design In designing cluster samples, including area samples, we must answer several questions:

1. How homogeneous are the resulting clusters?

2. Shall we seek equal-size or unequal-size clusters?

3. How large a cluster shall we take?

4. Shall we use a single-stage or multistage cluster?

5. How large a sample is needed?1

1. When clusters are homogeneous, this contributes to low statistical efficiency. Sometimes one can improve this efficiency by constructing clusters to increase intracluster variance. In the dining club study, researchers might have chosen a course as a cluster, choosing to sample all students in that course if it enrolled students of all four class years. Or maybe they could choose a departmental office that had faculty, staff, and administrative positions as well as student workers. In area sampling to increase intra- cluster variance, researchers could combine into a single cluster adjoining blocks that contain different income groups or social classes.

2. A cluster sample may be composed of clusters of equal or unequal size. The theory of clustering is that the means of sample clusters are unbiased estimates of the population mean. This is more often true when clusters are naturally equal, such as households in city blocks. While one can deal with clusters of unequal size, it may be desirable to reduce or counteract the effects of unequal size. There are several approaches to this:

• Combine small clusters and split large clusters until each approximates an average size. • Stratify clusters by size and choose clusters from each stratum. • Stratify clusters by size and then subsample, using varying sampling fractions to secure an overall

sampling ratio.21

3. There is no a priori answer to the ideal cluster size question. Comparing the efficiency of differing cluster sizes requires that we discover the different costs for each size and estimate the different variances of the cluster means. Even with single-stage clusters (where the researchers interview or observe every element within a cluster), it is not clear which size (say, 5, 20, or 50) is superior. Some have found that in studies using single-stage area clusters, the optimal cluster size is no larger than the typical city block.22

4. Concerning single-stage or multistage cluster designs, for most large-scale area sampling, the tendency is to use multistage designs. Several situations justify drawing a sample within a cluster, in preference to the direct creation of smaller clusters and taking a census of that cluster using one-stage cluster sampling:23

• Natural clusters may exist as convenient sampling units yet, for economic reasons, may be larger than the desired size.

• We can avoid the cost of creating smaller clusters in the entire population and confine subsam- pling to only those large natural clusters.

• The sampling of naturally compact clusters may present practical difficulties. For example, inde- pendent interviewing of all members of a household may be impractical.

5. The answer to how many subjects must be interviewed or observed depends heavily on the specific cluster design, and the details can be complicated. Unequal clusters and multistage samples are the chief complications, and their statistical treatment is beyond the scope of this book.24 Here we will treat only single-stage sampling with equal-size clusters (called simple cluster sampling). It is analogous to simple random sampling. We can think of a population as consisting of 20,000 clusters of one student each, or 2,000 clusters of 10 students each, and so on. Assuming the same specifications for precision and confidence, we should expect that the calculation of a probability sample size would be the same for both clusters.

schi18939_ch05_085-121.indd 104 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

105

>Exhibit 5-9 Comparison of Probability Sampling Designs

Type Description Advantages Disadvantages

Simple Random

Cost: High

Use: Moderate

Each population case has an equal chance of being se- lected into the sample.

Sample drawn using random number table/generator.

Easy to implement with auto- matic dialing (random-digit dialing) and with computerized voice response systems.

Requires a listing of popula- tion cases.

Takes more time to implement.

Uses larger sample sizes.

Produces larger errors.

Systematic

Cost: Moderate

Use: Moderate

Using a random start, selects a population case and follow- ing the sampling skip interval selects every kth case.

Simple to design.

Easier to use than the simple random.

Easy to determine sampling dis- tribution of mean or proportion.

Periodicity within the popu- lation may skew the sample and results.

If the population list has a monotonic trend, a biased estimate will result based on the start point.

Stratified

Cost: High

Use: Moderate

Divides population into subpopulations or strata and draws a simple random sample from each stratum. Results may be weighted and combined.

Researcher controls sample size within strata.

Increased statistical efficiency.

Provides data to represent and analyze subgroups.

Enables use of different meth- ods in strata.

Increased error will result if subgroups are selected at different rates.

Especially expensive if population strata must be created.

Cluster

Cost: Moderate

Use: High

Population is divided into internally heterogeneous sub- groups. Some subgroups are randomly selected for further study.

Provides an unbiased estimate of population parameters if properly done.

Economically more efficient than simple random.

Lowest cost per sample, espe- cially with geographic clusters.

Easy to do without a popula- tion list.

Often lower statistical efficiency (more error) due to subgroups being homogeneous rather than heterogeneous.

Double

(Sequential or     multiphase)

Cost: Moderate

Use: Moderate

Process includes collecting data from any type sample. Based on the information found, a subsample is selected for further study.

May reduce costs if first stage results in enough data to stratify or cluster the population.

Increased costs if indis- criminately used.

Double Sampling

It may be more convenient or economical to collect some information by sample and then use this infor- mation as the basis for selecting a subsample for further study. This procedure is called double sampling, (also called sequential sampling or multiphase sampling) It is usually found with stratified and/or cluster designs. The calculation procedures are described in more advanced texts.

Double sampling can be illustrated by the dining club example. You might use an email survey or another inexpensive survey method to discover who would be interested in joining such a club and the degree of their interest. You might then stratify the interested respondents by degree of interest and sub- sample among them for intensive interviewing on expected consumption patterns, reactions to various services, and so on. Whether it is more desirable to gather such information by one-stage or two-stage sampling depends largely on the relative costs of the two methods.

Because of the wide range of sampling designs available, it is often difficult to select an approach that meets the needs of the research question and helps to contain the costs of the project. To help with these choices, Exhibit 5-9 may be used to compare the various advantages and disadvantages of probability

schi18939_ch05_085-121.indd 105 2/14/18 3:08 PM

>part II The Design of Business Research

106

sampling. Nonprobability sampling techniques are covered in the next section. They are used frequently and offer the researcher the benefit of low cost. However, they are not based on a theoretical framework and do not operate from statistical theory; consequently, they produce selection bias and nonrepresenta- tive samples. Despite these weaknesses, their widespread use demands their mention here.

>Nonprobability Sampling With a subjective approach like nonprobability sampling, the probability of selecting population cases is unknown. There are a variety of ways to choose persons or cases to include in the sample. Often we allow the choice of subjects to be made by field workers on the scene. When this occurs, there is greater op- portunity for bias to enter the sample selection procedure and to distort the findings of the study. Also, we cannot estimate any range within which to expect the population parameter. Given the technical advantages of probability sampling over nonprobability sampling, why would anyone choose the latter? There are some practical reasons for using the less precise methods.

We may use nonprobability sampling procedures because they satisfactorily meet the sampling ob- jectives. Although a random sample will give us a true cross section of the population, this may not be the objective of the research. If there is no desire or need to generalize to a population parameter, then there is much less concern about whether the sample fully reflects the population. Often researchers have more limited objectives. They may be looking only for the range of conditions or for examples of dramatic variations. This is especially true in exploratory research in which one may wish to contact only certain persons or cases that are clearly atypical.

Additional reasons for choosing nonprobability over probability sampling are cost and time. Prob- ability sampling clearly calls for more planning and repeated callbacks to ensure that each selected case is contacted. These activities are expensive. Carefully controlled nonprobability sampling often seems to give acceptable results, so the investigator may not even consider probability sampling. While prob- ability sampling may be superior in theory, there are breakdowns in its application. Even carefully stated random sampling procedures may be subject to careless application by the people involved. Thus, the ideal probability sampling may be only partially achieved because of the human element.

It is also possible that nonprobability sampling may be the only feasible alternative. The total popula- tion may not be available for study in certain cases. At the scene of a major event, it may be infeasible to attempt to construct a probability sample. A study of past correspondence between two companies must use an arbitrary sample because the full correspondence is normally not available.

In another sense, those who are included in a sample may select themselves. In mail surveys, those who respond may not represent a true cross section of those who receive the questionnaire. The receivers of the questionnaire decide for themselves whether they will participate. In web-based surveys those who volunteer don't always represent the appropriate cross section—that’s why screening questions are used before admitting a participant to the sample. There is, however, some of this self-selection in almost all surveys because every respondent chooses whether to be interviewed.

Convenience Nonprobability samples that are unrestricted are called convenience samples. They are the least reliable design but normally the cheapest and easiest sample to draw. Researchers or field workers have the free- dom to choose whomever they find: thus, the name “convenience.” Examples include informal pools of friends and neighbors, people responding to a newspaper’s invitation for readers to state their positions on some public issue, a TV reporter’s “person-on-the-street” intercept interviews, or the use of employees to evaluate the taste of a new snack food.

Although a convenience sample has no controls to ensure precision, it may still be a useful procedure. Often you will take such a sample to test ideas or even to gain ideas about a subject of interest. In the early stages of exploratory research, when you are seeking guidance, you might use this approach. The results may present evidence that is so overwhelming that a more sophisticated sampling procedure is unnecessary. In an interview with students concerning some issue of campus concern, you might talk to 25 students selected sequentially. You might discover that the responses are so overwhelmingly one- sided that there is no incentive to interview further.

schi18939_ch05_085-121.indd 106 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

107

Purposive Sampling A nonprobability sample that uses certain criteria to select cases is called purposive sampling. There are two major types—judgment sampling and quota sampling.

Judgment sampling occurs when a researcher selects sample members to conform to some criterion. In a study of labor problems, you may want to talk only with those who have experienced on-the-job discrimination. Another example of judgment sampling occurs when election results are predicted from only a few selected precincts that have been chosen because of their predictive record in past elections.

When used in the early stages of an exploratory study, a judgment sample is appropriate. When one wishes to select a biased group for screening purposes, this sampling method is also a good choice. Com- panies often try out new product ideas on their employees. The rationale is that one would expect the firm’s employees to be more favorably disposed toward a new product idea than the public. If the product does not pass this group, it does not have prospects for success in the general market.

Quota sampling is the second type of purposive sampling. We use it to improve representativeness. The logic behind quota sampling is that certain relevant characteristics describe the dimensions of the population. If a nonprobability sample has the same distribution on these characteristics, then it is likely to be representative of the population regarding other variables on which we have no control. Suppose the student body of Metro U is 55 percent female and 45 percent male. The sampling quota would call for sampling students at a 55 to 45 percent ratio.

In most quota samples, researchers specify more than one control dimension. Each should meet two tests: it should (1) have a distribution in the population that we can estimate and (2) be pertinent to the topic studied. We may believe that responses to a question should vary depending on the gender of the respondent. If so, we should seek proportional responses from both men and women. We may also feel that undergraduates differ from graduate students, so this would be a dimension. Other dimensions, such as the student’s academic discipline, ethnic group, religious affiliation, and social group affiliation, also may be chosen. Only a few of these controls can be used. To illustrate, suppose we consider the following:

• Gender: Two categories—male, female.

• Class level: Two categories—graduate, undergraduate.

• College: Six categories—arts and science, agriculture, architecture, business, engineering, other.

• Religion: Four categories—Protestant, Catholic, Jewish, other.

• Fraternal affiliation: Two categories—member, nonmember.

• Family social-economic class: Three categories—upper, middle, lower.

In an extreme case, we might ask an interviewer to find a male undergraduate business student who is Catholic, a fraternity member, and from an upper-class home. All combinations of these six factors would call for 288 such cells to consider. This type of control is known as precision control. It gives greater assurance that a sample will be representative of the population. However, it is costly and too difficult to carry out with more than three variables.

When we wish to use more than three control dimensions, we should depend on frequency control. With this form of control, the overall percentage of those with each characteristic in the sample should match the percentage holding the same characteristic in the population. No attempt is made to find a combination of specific characteristics in a single person. In frequency control, we would probably find that the following sample array is an adequate reflection of the population:

Population Sample

Male 65% 67%

Married 15 14

Undergraduate 70 72

Campus resident 30 28

Greek member 25 23

Protestant 39 42

schi18939_ch05_085-121.indd 107 2/14/18 3:08 PM

>part II The Design of Business Research

108

Quota sampling has several weaknesses. First, the idea that quotas on some variables assume a representativeness on others is argument by analogy. It gives no assurance that the sample is repre- sentative of the variables being studied. Often, the data used to provide controls might be outdated or inaccurate. There is also a practical limit on the number of simultaneous controls that can be applied to ensure precision. Finally, the choice of subjects is left to field workers to make on a judg- mental basis. They may choose only friendly looking people, people who are convenient to them, and so forth.

Despite the problems with quota sampling, it is widely used by opinion pollsters and business researchers. Probability sampling is usually much more costly and time-consuming. Advocates of quota sampling argue that although there is some danger of systematic bias, the risks are usually not that great. Where predictive validity has been checked (e.g., in election polls), quota sampling has been generally satisfactory.

Snowball This design has found a niche in recent years in applications where respondents are difficult to identify and are best located through referral networks. It is also especially appropriate for some qualitative stud- ies. In the initial stage of snowball sampling, individuals are discovered and may or may not be selected through probability methods. This group is then used to refer the researcher to others who possess similar characteristics and who, in turn, identify others. Similar to a reverse search for bibliographic sources, the “snowball” gathers subjects as it rolls along. Various techniques are available for selecting a nonprobability snowball with provisions for error identification and statistical testing. Let’s consider a brief example.

The high end of the U.S. audio market is composed of several small firms that produce ultra- expensive components used in recording and playback of live performances. A risky new technology for improving digital signal processing is being contemplated by one firm. Through its contacts with a select group of recording engineers and electronics designers, the first-stage sample may be identified for interviewing. Subsequent interviewees are likely to reveal critical information for product develop- ment and marketing.

Variations on snowball sampling have been used to study drug cultures, teenage gang activities, power elites, community relations, insider trading, and other applications where respondents are difficult to identify and contact.

> Define the Selection and Recruiting Protocols Whether the researcher chooses a census or sample, selection and recruiting answers the following questions:

• How will the researcher select cases (from a sample frame or without a sample frame)?

• How will each desired case/gatekeeper be contacted?

• How will the case/gatekeeper be convinced to participate?

• What follow-up procedures will be used to guarantee the case/gatekeeper completes the research?

The sampling methodology often has specific protocols for drawing cases (e.g., the simple random sample, the stratified sample, etc.); we’ve given you these in the earlier section. If the chosen method doesn’t have such a protocol—as with nonprobability samples—you must create one. Will the data

schi18939_ch05_085-121.indd 108 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

109

Mixed-access sampling means that multiple methods are used to invite participants to a research study—phone, email, mobile or wireless, addressed based/mail, etc. According to the CDC, approximately 97% of possible participants are reachable by phone, while on 87% are reachable online. Mixed-access sampling reduces non-coverage error and nonresponse error. Once a participant is recruited, regardless of the means, he or she may complete the study by a different mode (e.g., recruited by phone but take a survey online). Sample recruitment is increasingly done by mixed access.

www.surveysampling.com, www.pewresarch.org, www.cdc.gov

>picprofile

Percent of U.S. Households Accessible by Phone Method of Sampling Invitation

Landline and Cell Phone Cell Phone Only

50.4%39.4%

Neither Cell nor LandlineLandline Phone Only

3.3%6.5%

collector have the latitude to choose anyone to participate, or will they need to follow specific guide- lines? In the Metro U study, if we use registrar and personnel lists for accuracy, systematic or stratified sampling protocols would be appropriate.

Research shows that participants are research weary. This means you may face difficulties gaining compliance and cooperation. Most researchers will tell you participants need to feel that they are part of something bigger than themselves in order to give of their time and expertise. They also may need to be offered an incentive for sharing their ideas and experiences. In the Metro U study, we could compel faculty to participate due to their employment, but it might backfire. It would be better to encourage par- ticipation as a means for the university to address financial issues (as they are plaguing many colleges). For students, we will need some other means of enticement.

Some individuals actually enjoy participating in research, feeling they have much to contribute when they are knowledgeable about the topic. These individuals often volunteer for research panels or join communities of like-minded individuals. They have volunteered, so recruiting them is not dif- ficult; keeping them engaged so that they follow through during a multi-activity research study is more problematic. To keep research participants engaged, researchers need to have procedures in place (e.g, follow-up emails or phone calls) to keep in touch with participants. For this group, an inducement to concluding a research project is the promise to share some or all of the results; this often works when the participant/gatekeeper is a business manager or professional. Finally, a charitable gift in the participant’s name is gaining favor as an incentive; this might work for faculty in the Metro U study. Food incentives often work for students. And both students and faculty would likely respond to money as an incentive.

We often use email or phone to contact potential cases or gatekeepers. The PicProfile in this chapter offers an example done by email. When your target population is a person, they will often inquire about why they were chosen. If the research sponsor is willing, researchers might divulge the sample frame, or at least its essence to the potential case as a means to recruit them (e.g., “You have posted on our Facebook in the last two weeks” or “You have a Plenti loyalty card”). In the Metro U study, email con- tacts would be appropriate for both students and faculty.

schi18939_ch05_085-121.indd 109 2/14/18 3:08 PM

>part II The Design of Business Research

110

If you could feed a hungry child or prevent the euthanasia of a

dog by taking a survey, would you?

Researchers have been using online panels to draw samples

for Web, email, and mobile surveys for the last decade. Most of

these panels are developed using three sources: advertising net-

works (sample providers advertise via the web, e-mail, and other

media to attract individuals who are willing to participate in sur-

veys), loyalty programs (a sponsor company uses its own list of

individuals who are part of its loyalty efforts and recruit them to

take surveys), and social media (sample providers use Facebook,

Twitter, and numerous other social media to recruit participants).

Sampling firms under the current model employing these sources

often incentivize their participants with money, Internet currency

or points, or prizes. Unfortunately, one of the drawbacks is a small

pool of individuals—even though in the millions—from which thou-

sands of firms are drawing their participants.

The founding partners of Research for Good (RFG) were con-

cerned that people receiving a personal incentive to give their

opinion tend to only represent a particular segment of the pop-

ulation. Also, there was the ongoing industry concern about the

development of “professional respondents”—when volunteer

participants are recruited by hundreds of sampling companies in

roughly the same three ways to meet an ever-growing demand

for survey participants. RFG was also concerned about social re-

sponsibility. So, it developed an incentivized sampling model that

uses charitable donations to attract uniquely different respondents.

Its SaySo for Good panel draws on the 90 percent of U.S. and

Canadian adults who support at least one charity. When joining the

panel, each participant chooses a charity to receive their survey

incentive. The RFG database of charities includes every government-

registered charity in both the U.S. and Canada. RFG delivers to the

specified charity $1.00 or 25 percent of the budgeted cost of a per

completed survey for each participant (i.e., if a completed survey is

budgeted at $8.00 per participant, then the charity receives $2.00).

RFG has discovered that participants attracted by charity

incentives are different—in both behavior and attitude—than

those attracted to a research panel by other means. These sur-

vey takers tend to be cause-minded, are financial supporters or

volunteers to charities, and are not likely to be motivated by typi-

cal cash or prize incentives. Their response rates are higher, as

well as their completion rates. And they are infrequent survey

takers, reducing the concern about “professional respondents.”

RFG now counts thousands of charity-incentivized members in

its panel. “By attracting new participants, we will generate better

data quality while we serve the greater good, ” emphasized co-

founder Sean Case. As a result, RFG now composes its samples

by including this charity-incentive model.

www.researchforgood.com; www.saysoforgood.com

Research for Good: Using Charity as an Incentive

>snapshot

Research for Good Sample Composition

Donation only

Donation + points/cash Donation + virtual currency

>Ethical Issues and Their Solutions Ethical issues in relation to sampling fall into three broad categories: deception, incentives, and quality. Many studies in business require that the sponsor not be revealed (sponsor nondisclosure) and that the purpose be protected (purpose nondisclosure); participant deception is often the method to ensure the first, while a consent agreement is a means to the second. If the sponsor permits, after-research debriefing is one solution available for deception. Otherwise, getting informed consent to participate from each participant/gatekeeper is critical.

Incentives may be offered to participants, so a researcher must keep those promises as soon as possi- ble. Quality is the responsibility of both the participant and the researcher. For the participant, truthfully answering screening questions to ensure they are the correct participant is critical and keeping their par- ticipation confidential (if that was a condition of participation). Quality is also dependent on researchers (1) using the best sample frame available and (2) conscientiously following selection protocols, whether these are industry standards or project specific.

schi18939_ch05_085-121.indd 110 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

111

LO5-1 Sampling design comprises six primary tasks:

a. Define the target population and a case.

b. Define the population parameters of interest. 

c. Identify and evaluate the sample frame or create one.

d. Define the number of cases needed.

e. Define the appropriate sampling method.

f. Define the sampling selection and recruitment protocols.

LO5-2 An effective use of a sample versus a census is based on two premises: (1) enough similarity exists among the cases in a target population that a few of these cases will adequately represent the characteristics of the target population and (2) that although some elements in a sample underestimate a population pa- rameter, others overestimate this value.  The result of these tendencies is that a sample statistic is generally a good estimate of a population parameter.

LO5-3 A quality sample has both accuracy and precision. An accurate sample is one in which there is little or no bias or systematic variance. A sample with adequate precision is one that has a sampling error that is within acceptable limits for the study’s purpose.

LO5-4 Many sampling techniques are available. They may be classified by their representation basis and element selection techniques.

Representation Basis

Element Selection

Probability Nonprobability

Unrestricted Simple random Convenience

Restricted Complex random

● Systematic

● Cluster

● Stratified

● Double

Purposive

● Judgment

● Quota

● Snowball

Probability sampling is based on random selection—a con- trolled procedure that ensures that each population case is given a known nonzero chance of selection. The sim- plest type of probability approach is simple random sam- pling. In this design, each case within the target population has an equal chance of being included in a sample. In con- trast, nonprobability selection is arbitrary. In unrestricted sampling, each sample case is drawn from the target population without defining criteria. In restricted sampling, the selection process follows more complex rules.

Complex sampling is used when conditions make simple random samples impractical or uneconomical. The four major types of complex random sampling

discussed in this chapter are systematic, stratified, clus- ter, and double sampling. Systematic sampling involves the selection of every kth element in the population, be- ginning with a random start between elements from 1 to k. Its simplicity in certain cases is its greatest value.

Stratified random sampling is based on dividing a population into subpopulations and then randomly sampling from each of these strata. This method usu- ally results in a smaller total sample size than would a simple random design. Stratified samples may be proportionate or disproportionate.

In cluster sampling, we divide the population into convenient groups and then randomly choose the groups to study. It is typically less efficient from a statistical viewpoint than the simple random because of the high degree of homogeneity within the clusters. Its great advantage is its savings in cost—if the popula- tion is dispersed geographically—or in time. The most widely used form of clustering is area sampling, in which geographic areas are the selection elements.

At times, it may be more convenient or economi- cal to collect some information by sample and then use it as a basis for selecting a subsample for further study. This procedure is called double sampling.

Nonprobability sampling also has some compelling practical advantages that account for its widespread use. Often probability sampling is not feasible because the population is not available. Then, too, frequent break- downs in the application of probability sampling discount its technical advantages. You may find also that a true cross section is often not the aim of the researcher. Here the goal may be the discovery of the range or extent of conditions. Finally, nonprobability sampling is usually less expensive to conduct than is probability sampling.

Convenience samples are the simplest and least reliable forms of nonprobability sampling. Their primary virtue is low cost. One purposive sample is the judg- mental sample, in which one is interested in studying only selected types of subjects. The other purposive sample is the quota sample. Subjects are selected to conform to certain predesignated control measures that secure a representative cross section of the popu- lation. Snowball sampling uses a referral approach to reach particularly hard-to-find respondents.

LO5-5 Ethical issues in sampling design relate to decep- tion, compensation, and quality. These are addressed through informed consent; compensation fulfillment (including cash, charitable donations, food, etc); and quality procedures, including adequate participant screening, using the best sample frame available, and conscientiously following selection protocols, whether these are industry standards or project specific.

>summary

schi18939_ch05_085-121.indd 111 2/14/18 3:08 PM

>part II The Design of Business Research

112

area sampling 103

case 86

census 93

cluster sampling 102

convenience samples 106

disproportionate stratified sampling 102

double sampling 105

judgment sampling 107

nonprobability sampling 96

nonresponse error 109

nonsampling error 94

population parameters 86

population proportion of incidence 89

probability sampling 96

proportionate stratified sampling 102

quota sampling 107

sample 93

sample frame 86

sample statistics 89

sampling error 94

sampling method 96

simple random sample 97

skip interval 98

snowball sampling 108

stratified random sampling 99

systematic sampling 98

systematic variance 95

target population 86

>keyterms

Terms in Review

1 Distinguish between:

a Statistic and parameter.

b Sample frame and target population.

c Case and target population.

d Simple random and complex random sampling.

e Convenience and purposive sampling.

f Sample precision and sample accuracy.

g Nonsampling and sampling error.

h Proportionate and disproportionate samples.

2 Under what conditions would you recommend:

a A probability sample? a nonprobability sample?

b A simple random sample? a cluster sample? a stratified sample?

c A disproportionate stratified probability sample?

3 You plan to conduct a study using unrestricted sampling. What subjective decisions must you make?

4 Describe the differences between a probability sample and a nonprobability sample.

5 Why would a researcher use a quota purposive sample?

Making Research Decisions

6 Your task is to interview a representative sample of attend- ees for the large concert venue where you work. The new season schedule includes 200 live concerts featuring all types of musicians and musical groups. Because neither the number of attendees nor their descriptive characteris- tics are known in advance, you decide on nonprobability

sampling. Based on past seating configurations, you can calculate the number of tickets that will be available for each of the 200 concerts. Thus, collectively, you will know the number of possible attendees for each type of music. From attendance research conducted at concerts held by the Glacier Symphony during the previous two years, you can obtain gender data on attendees by type of music. How would you conduct a reasonably reliable nonprob- ability sample?

7 Your large firm is about to change to a customer-centered organization structure, in which employees who have rarely had customer contact will now likely significantly influence customer satisfaction and retention. As part of the transition, your superior wants an accurate evaluation of the morale of the firm’s large number of computer technicians. What type of sample would you draw if it was to be an unrestricted sample?

8 When Nike introduced its glow-in-the-dark Foamposite One Galaxy sneakers, fanatics lined up at distributors around the country. As crowds became restless, jockeying for position at the front of increasingly long lines for the limited-supply shoes, Footlocker canceled some events. It’s been sug- gested that Nike should sell its limited-release introductions online rather than in stores to avoid putting its customers’ safety in jeopardy. What sample group would you suggest Nike use to assess this suggestion?

From Concept to Practice

9 Using Exhibit 5-9 as your guide, for each sampling tech- nique, describe the sample frame for a study of employers’ skill needs in new hires using the industry in which you are currently working or wish to work.

>discussionquestions

schi18939_ch05_085-121.indd 112 2/14/18 3:08 PM

>chapter 5 Stage 2: Sampling Design

113

>cases*

Akron Children’s Hospital

Calling Up Attendance

Campbell-Ewald Pumps Awareness into the American Heart Association

Campbell-Ewald: R-E-S-P-E-C-T Spells Loyalty

Can Research Rescue the Red Cross?

Goodyear’s Aquatred

Inquiring Minds Want to Know—NOW!

Marcus Thomas LLC Tests Hypothesis for Troy-Bilt Creative Development

Ohio Lottery: Innovative Research Design Drives Winning

Pebble Beach Co.

Starbucks, Bank One, and Visa Launch Starbucks Card Duetto Visa

State Farm: Dangerous Intersections

The Catalyst for Women in Financial Services

USTA: Come Out Swinging

Volkswagen’s Beetle

From the Headlines

10 As populations age and young workers reject work requir- ing repetitive operations, factories around the world are turning to robots. It is estimated that the United States has 176 robots for every 10,000 workers, with South Korea (531) and Germany (301) leading the charge and China, with its investment in being a leading maker of robots, expected to

advance quickly (90,000 robots installed by 2016, one-third of the world’s total). You are a worldwide manufacturer of industrial equipment with 10 plants and more than 10,000 manufacturing employees and supervisors. Some of your plants are represented by unions. Create the sampling design to study robotics as a solution to a manufacturing worker shortage.

* You will find a description of each case in the Case Abstracts section of this textbook. Check the Case Index to determine whether a case provides data, the research instrument, video, or other supplementary material. Cases and case supplements are available in Connect.

schi18939_ch05_085-121.indd 113 2/14/18 3:08 PM

>Exhibit Ca-1 Random Samples of Preferred Lunch Times

1 2 3 4 5

f

2,000 4,000 7,000 4,000 3,000

N = 20,000

11:00 a.m. 11:30 a.m.

12:00 p.m. 12:30 p.m. 1:00 p.m.

Y = Time

Y

1 2 3 4 5

f

1 2 4 2 1

Samples

n1

Y

1 2 3 4 5

f

1 2 5 2 0

n3 n4n2

Y

1 2 3 4 5

f

0 1 5 1 3

Y

1 2 3 4 5

f

1 1 3 4 1

Population of preferred lunch times

n1 = 10 X1 = 3.0 s = 1.15

n2 = 10 X2 = 2.8 s = 0.92

n3 = 10 X3 = 3.6 s = 1.07

n4 = 10 X4 = 3.3 s = 1.16

μ = 3.1 or 12:03 p.m. σ = 0.74 or 22.2 minutes

114

>appendix Calculate the Sample Size

Basic Concepts for Sampling

In the Metro University Dining Club study, the university is exploring whether to develop a membership-only dining facility on campus. One set of measured questions deals with lunch behaviors for five different groups (faculty, stu- dents, staff, alumni, “friends” of the university. We explore probability sampling and the various concepts used to de- sign the sampling process using this example.

Exhibit Ca-1 shows the Metro U dining club study popu- lation (N = 20,000) consisting of five subgroups based on their preferred lunch times. The values 1 through 5 represent the preferred lunch times of 11 a.m., 11:30 a.m., 12 noon, 12:30 p.m., and 1 p.m. The frequency of response (f) in

the population distribution, shown beside the population subgroup, is what would be found if a census of the elements was taken. Normally, population data are unavailable or are too costly to obtain. We are pretending omniscience for the sake of the example.

Point Estimates Now assume we sample 10 elements from this population without knowledge of the popula- tion’s characteristics. We use a sampling procedure from a statistical software program, a random number generator, or a table of random numbers. Our first sample (n

1 = 10)

provides us with the frequencies shown below sample n 1 in

Exhibit Ca-1. We also calculate a mean score, X 1 = 3.0, for

this sample. This mean would place the average preferred

schi18939_ch05_085-121.indd 114 2/14/18 3:08 PM

115

lunch time at 12 noon. The mean is a point estimate and our best predictor of the unknown population mean, µ (the arithmetic average of the population). Assume further that we return the first sample to the population and draw a second, third, and fourth sample by the same procedure. The frequencies, means, and standard deviations are as shown in the exhibit. As the data suggest, each sample shares some similarities with the population, but none is a perfect duplication because no sample perfectly replicates its population.

Interval Estimates We cannot judge which estimate is the true mean (accurately reflects the population mean). However, we can estimate the interval in which the true µ will fall by using any of the samples. This is accomplished by using a formula that computes the standard error of the mean:

σ _ X =

σ __ √

__ n

where

σ _ x = standard error of the mean or the standard devia-

tion of all possible _

X s  σ = population standard deviation  n = sample size

The standard error of the mean measures the stan- dard deviation of the distribution of sample means. It varies directly with the standard deviation of the popu- lation from which it is drawn (see Exhibit Ca-2): If the standard deviation is reduced by 50 percent, the stan- dard error will also be reduced by 50 percent. It also varies inversely with the square root of the sample size. If the square root of the sample size is doubled, the stan- dard error is cut by one-half, provided the standard de- viation remains constant.

Let’s now examine what happens when we apply sample data (n

1 ) from Exhibit Ca-1 to the formula. The

sample standard deviation from sample n will be used as an unbiased estimator of the population standard deviation:

σ _ X =

s ____ √ __

n

where

s = standard deviation of the sample, n 1

n 1 = 10

_

X 1 = 3.0

s 1 = 1.15

Substituting into the equation:

σ _ X =

s __ √ __

n = 1.15

____ √

___ 10 = 0.36

Estimating the Population Mean

How does this improve our prediction of µ from X? The standard error creates the interval range that brackets the point estimate. In this example, μ is predicted to be 3.0 or 12 noon (the mean of n

1 ) ±0.36. This range may

be visualized on a continuum (see diagram at bottom of next page).

We would expect to find the true μ between 2.64 and 3.36—between 11:49 a.m. and 12:11 p.m. (if 2 = 11:30 a.m. and 0.64 = (30 minutes) = 19.2 minutes, then 2.64 = 11:30 a.m. + 19.2 minutes, or 11:49 a.m.). Since we assume omniscience for this illustration, we know the popula- tion average value is 3.1. Further, because standard errors have characteristics like other standard scores, we have 68 percent confidence in this estimate—that is, one standard error encompasses ±1 Z or 68 percent of the area under the

>Exhibit Ca-2 Effects on Standard Error of Mean of Increasing Precision

Reducing the Standard Deviation by 50% Quadrupling the Sample

σ _ x = s ___ √

__ n σ _ x =

0.74 ___

√ __

10 = 0.234 σ _

x = 0.8 ___

√ __

25 = 0.16

σ _ x = 0.37 ___

√ __

10 = 0.117 σ _

x = 0.8 ___

√ __

100 = 0.08

where

σ _ x = standard error of the mean

σ _ x = standard deviation of the sample

n = sample size

Note: A 400 percent increase in sample size (from 25 to 100) would yield only a 200 percent increase in precision

(from 0.16 to 0.08). Researchers are often asked to increase precision, but the question should be, at what cost? Each

of those additional sample elements adds both time and cost to the study.

>appendix Calculate the Sample Size

schi18939_ch05_085-121.indd 115 2/14/18 3:08 PM

>part II The Design of Business Research

116

normal curve (see Exhibit Ca-3). Recall that the area under the curve also represents the confidence estimates that we make about our results. The combination of the interval range and the degree of confidence creates the confidence interval. To improve confidence to 95 percent, multiply the standard error of 0.36 by ± 1.96 (Z), because 1.96 Z covers 95 percent of the area under the curve (see Exhibit Ca-4). Now, with 95 percent confidence, the interval in which we would find the true mean increases to ± 0.70 (from 2.3 to 3.7 or from 11:39 a.m. to 12:21 p.m.).

Parenthetically, if we compute the standard deviation of the distribution of sample means in Exhibit Ca-1, [3.0, 2.8, 3.6, 3.3], we will discover it to be 0.35. Compare this to the standard error from the original calculation (0.36). The result is consistent with the second definition of the standard error: the standard deviation of the distribution of sample means (n

1 , n

2 , n

3 , and n

4 ). Now let’s return to the

dining club example and apply some of these concepts to the researchers’ problem.

If the researchers were to interview all the students and employees in the defined population, asking them, “How many times per month would you eat at the club?” they would get a distribution something like that shown in part A of Exhibit Ca-5. The responses would range from zero to as many as 30 lunches per month with a μ and σ.

However, they cannot take a census, so μ and σ remain unknown. By sampling, the researchers find the mean to be 10.0 and the standard deviation to be 4.1 eating experi- ences (how often they would eat at the club per month). In part C of Exhibit Ca-5, three observations about this sample distribution are consistent with our earlier illus- tration. First, it is shown as a histogram; it represents a frequency distribution of empirical data, while the smooth curve of part A is a theoretical distribution. Second, the sample distribution (part C) is similar in appearance but is not a perfect duplication of the population distribution (part A). Third, the mean of the sample differs from the mean of the population.

>Exhibit Ca-3 Confidence Levels and the Normal Curve

+1.96 σx _–1.96 σx

_ +1 σx _0–1 σx

_

68%

95%

>Exhibit Ca-4 Standard Errors Associated with Areas under the Normal Curve Standard Error (Z) Percent of Area* Approximate Degree of Confidence

1.00 68.27 68%

1.65 90.10 90

1.96 95.00 95

3.00 99.73 99

*Includes both tails in a normal distribution.

True mean 3.1

63.300.346.2

11:49 a.m. X 12:11 p.m.

=

schi18939_ch05_085-121.indd 116 2/14/18 3:08 PM

117

If the researchers could draw repeated samples as we did earlier, they could plot the mean of each sample to se- cure the solid line distribution found in part B. According to the central limit theorem, for sufficiently large samples (n = 30), the sample means will be distributed around the population mean approximately in a normal distribution. Even if the population is not normally distributed, the dis- tribution of sample means will be normal if there is a large enough set of samples.

Estimating the Interval for the Metro U Dining Club Sample Any sample mean will fall within the range of the distribution extremes shown in part B of Exhibit Ca-5. We also know that about 68 percent of the sample means in this distribution will fall between x

3 and

x 4 and 95 percent will fall between x

1 and x

2 .

If we project points x 1 and x

2 up to the population dis-

tribution (part A of Exhibit Ca-5) at points x′1 and x′2, we see the interval where any given mean of a random sample

>Exhibit Ca-5 A Comparison of Population Distribution, Sample Distribution, and Distribution of Sample Means of Metro U Dining Club Study

Part A Population distribution

Part C Sample distribution

Part B Distribution of means from repeated samples of a fixed size (n = 64)

68% of the area

– σ

x'1

x1

s

x3 x x4 x2

x'2

– σ

µ

– σ + σ

X

s s

10.0

s

Note: The distributions in these figures are not to scale, but this fact is not critical to our understanding of the dispersion relationship depicted.

>appendix Calculate the Sample Size

schi18939_ch05_085-121.indd 117 2/14/18 3:08 PM

>part II The Design of Business Research

118

of 64 is likely to fall 95 percent of the time. Because we will not know the population mean from which to measure the standard error, we infer that there is also a 95 percent chance that the population mean is within two standard errors of the sample mean (10.0). This inference enables us to find the sample mean, mark off an interval around it, and state a confidence likelihood that the population mean is within this bracket.

Because the researchers are considering an investment in this project, they would want some assurance that the population mean is close to the figure reported in any sam- ple they take. To find out how close the population mean is to the sample mean, they must calculate the standard error of the mean and estimate an interval range within which the population mean is likely to be.

Given a sample size of 64, they still need a value for the standard error. Almost never will one have the value for the standard deviation of the population (σ), so we must use a proxy figure. The best proxy for σ is the standard deviation of the sample (s). Here the standard deviation (s = 4.1) was obtained from a pilot sample:

σ _ X =

s ____ √ __

n = 4.1

____ √

__ 64 = 0.51

If one standard error of the mean is equal to 0.51 visit, then 1.96 standard errors (95 percent) are equal to 1.0 visit. The students can estimate with 95 percent confidence that the population mean of expected visits is within 10.0 ± 1.0 visit, or from 9.0 to 11.0 mean visits per month.

Changing Confidence Intervals The preceding es- timate may not be satisfactory in two ways. First, it may not represent the degree of confidence the researchers want in the interval estimate, considering their financial risk. They might want a higher degree of confidence than the 95 percent level used here. By referring to a table of areas under the normal curve, they can find various other com- binations of probability. Exhibit Ca-6 summarizes some of those more commonly used. Thus, if the students want a greater confidence in the probability of including the popu- lation mean in the interval range, they can move to a higher

standard error, say, X ± 3σ X . Now the population mean lies

somewhere between 10.0 ± 3 (0.51) or from 8.47 to 11.53. With 99.73 percent confidence, we can say this interval will include the population mean.

We might wish to have an estimate that will hold for a much smaller range, for example, 10.0 ± 0.2. To secure this smaller interval range, we must either (1) accept a lower level of confidence in the results or (2) take a sample large enough to provide this smaller interval with the higher de- sired confidence level.

If one standard error is equal to 0.51 visit, then 0.2 visit would be equal to 0.39 standard error (0.2/0.51 = 0.39). Referring to a table of areas under the normal curve (book Appendix C, Exhibit C-1), we find that there is a 30.3 percent chance that the true population mean lies within ±0.39 standard error of 10.0. With a sample of 64, the sample mean would be subject to so much error variance that only 30 percent of the time could the researchers expect to find the population mean between 9.8 and 10.2. This is such a low level of confidence that the researchers would normally move to the second alternative; they would increase the sample size until they could secure the desired interval estimate and degree of confidence.

Calculating the Sample Size for Questions In- volving Means Before we compute the desired sample size for the Metro U dining club study, let’s review the in- formation we will need:

1. The precision desired and how to quantify it: a. The confidence level we want with our estimate. b. The size of the interval estimate.

2. The expected dispersion in the population for the in- vestigative question used.

3. Whether a finite population adjustment is needed.

The researchers have selected two investigative question constructs as critical—“frequency of patronage” and “inter- est in joining”—because they believe both to be crucial to making the correct decision on the Metro U dining club opportunity. The first requires a point estimate, the second

>Exhibit Ca-6 Estimates Associated with Various Confidence Levels in the Metro U Dining Club Study

Approximate Degree of Confidence Interval Range of Dining Visits per Month

68% µ is between 9.48 and 10.52 visits

90% µ is between 9.14 and 10.86 visits

95% µ is between 8.98 and 11.02 visits

99% µ is between 8.44 and 11.56 visits

schi18939_ch05_085-121.indd 118 2/14/18 3:08 PM

119

a proportion. By way of review, decisions needed and de- cisions made by Metro U researchers are summarized in Exhibit Ca-7.

Precision With reference to precision, the 95 percent confidence level is often used, but more or less confidence may be needed in light of the risks of any given project. Similarly, the size of the interval estimate for predicting the population parameter from the sample data should be de- cided. When a smaller interval is selected, the researcher is saying that precision is vital, largely because inherent risks are high. For example, on a 5-point measurement scale, one-tenth of a point is a very high degree of precision in comparison to a 1-point interval. Given that a patron could eat up to 30 meals per month at the dining club (30 days times 1 meal per day), anything less than one meal per day would be asking for a high degree of precision in the Metro U study. The high risk of the Metro U study warrants the 0.5 meal precision selected.

Population Dispersion The next factor that affects the size of the sample for a given level of precision is the population dispersion. The smaller the possible dis- persion, the smaller will be the sample needed to give a representative picture of population members. If the pop- ulation’s number of meals ranges from 18 to 25, a smaller sample will give us an accurate estimate of the popula- tion’s average meal consumption. However, with a popu- lation dispersion ranging from 0 to 30 meals consumed, a larger sample is needed for the same degree of confidence in the estimates. Since the true population dispersion of estimated meals per month eaten at Metro U dining club is unknowable, the standard deviation of the sample is used as a proxy figure. Typically, this figure is based on any of the following:

• Previous research on the topic.

• A pilot test or pretest of the data instrument among a sample drawn from the population.

>Exhibit Ca-7 Metro U Sampling Design Decision on “Meal Frequency” and “Joining” Constructs

Metro U Decisions

Sampling Issues “Meal Frequency” (interval, ratio data)

“Joining” (nominal, ordinal data)

1. The precision desired and how to quantify it:

● The confidence researcher wants in the esti- mate (selected based on risk)

● The size of the interval estimate the researcher will accept (based on risk)

95% confidence (Z = 1.96)

±0.5 meal per month

95% confidence (Z = 1.96)

±0.10 (10 percent)

2. The expected range in the population for the question used to measure precision:

Measure of Central Tendency

● Sample mean

● Sample proportion of population with the given attribute being measured

0 to 30 meals

10

0 to 100%

30%

Measure of Dispersion

● Standard deviation

● Measure of sample dispersion

4.1 pq = 0.30(0.70) = 0.21

3. Whether a finite population adjustment should be used

No No

4. Estimate of standard deviation of population:

● Standard error of mean

● Standard error of the proportion

0.5/1.96 = 0.255 0.10/1.96 = 0.051

5. Sample size calculation See formula (p. 120) See formula (p. 121)

6. Calculated sample size n = 259* n = 81

*Because both investigative questions were of interest, the researcher would use the larger of the two sample sizes calculated, n = 259 for the study.

>appendix Calculate the Sample Size

schi18939_ch05_085-121.indd 119 2/14/18 3:08 PM

>part II The Design of Business Research

120

• A rule of thumb (one-sixth of the range based on six standard deviations within 99.73 percent confidence).

If the range is from 0 to 30 meals, the rule-of-thumb method produces a standard deviation of 5 meals. The researchers want more precision than the rule-of-thumb method pro- vides, so they take a pilot sample of 25 and find the stan- dard deviation to be 4.1 meals.

Population Size A final factor affecting the size of a random sample is the size of the population. When the size of the sample exceeds 5 percent of the population, the finite limits of the population constrain the sample size needed. A correction factor is available in that event.

The sample size is computed for the first construct, meal frequency, as follows:

σ _ X =

s ____ √ __

n

√ __

n = s ____ σ _

X

n = s2 ____ σ _

X

n = (4.1)2

_______ (0.255)2

n = 258.5 or 259 where

σ x = 0.255 (0.51/1.96)

If the researchers are willing to accept a larger interval range (±1 meal), and thus a larger amount of risk, then they can reduce the sample size to n = 65.

Calculating the Sample Size for Ques- tions Involving Proportions

The second key question concerning the dining club study was, what percentage of the population says it would join the dining club, based on the projected rates and services? In business, we often deal with proportion data. An ex- ample is a CNN poll that projects the percentage of people who expect to vote for or against a proposition or a can- didate. This is usually reported with a margin of error of ±5 percent.

In the Metro U study, a pretest answers this question using the same general procedure as before. But instead of the arithmetic mean, with proportions, it is p (the pro- portion of the population that has a given attribute)1—in this case, interest in joining the dining club. And instead of the standard deviation, dispersion is measured in terms of p × q (in which q is the proportion of the population not having the attribute), and q = (1 − p). The measure of

dispersion of the sample statistic also changes from the standard error of the mean to the standard error of the proportion σ

p .

We calculate a sample size based on these data by making the same two subjective decisions—deciding on an acceptable interval estimate and the degree of confi- dence. Assume that from a pilot test, 30 percent of the students and employees say they will join the dining club. We decide to estimate the true proportion in the population within 10 percentage points of this figure ( p = 0.30 ±0.10). Assume further that we want to be 95 per- cent confident that the population parameter is within ±0.10 of the sample proportion. The calculation of the sample size proceeds as before:

±0.10 = desired interval range within which the population proportion is expected (subjective decision)

1.96σ p = 95 percent confidence level for estimating

the interval within which to expect the population proportion (subjective decision)

σ p = 0.051 = standard error of the proportion

(0.10/1.96)

pq = measure of sample dispersion (used here as an estimate of the population dispersion)

n = sample size

σ p = √

___

pq

___ n

n =

pq __

σ p 2

n = 0.3 × 0.7

______ ( 0.051 ) 2

n = 81

The sample size of 81 persons is based on an infinite population assumption. If the sample size is less than 5 percent of the population, there is little to be gained by using a finite population adjustment. The students interpreted the data found with a sample of 81 chosen randomly from the population as: We can be 95 percent confident that 30 percent of the respondents would say they would join the dining club with a margin of error of ±10 percent.

Previously, the researchers used pilot testing to generate the variance estimate for the calculation. Suppose this is not an option. Proportions data have a feature concerning the variance that is not found with interval or ratio data. The pq ratio can never exceed 0.25. For example, if p = 0.5, then q = 0.5, and their product is 0.25. If either p or q is greater than 0.5, then their product is smaller than 0.25 (0.4 × 0.6 = 0.24, and so on). When we have no information

schi18939_ch05_085-121.indd 120 2/14/18 3:08 PM

121

regarding the probable p value, we can assume that p = 0.5 and solve for the sample size:

n = pq

__ σ

p 2

n = (0.50)(0.50)

___________ (0.51)2

n = 0.25

___________ (0.51)2

n = 96 where

pq = measure of dispersion n = sample size σp = standard error of the proportion

If we use this maximum variance estimate in the dining club example, we find the sample size needs to be 96 per- sons in order to have an adequate sample for the question about joining the club.

When there are several investigative questions of strong interest, researchers calculate the sample size for each such variable—as we did in the Metro U study for “meal frequency” and “joining.” The researcher then chooses the calculation that generates the largest sample. This ensures that all data will be collected with the necessary level of precision.

>appendix Calculate the Sample Size

schi18939_ch05_085-121.indd 121 2/14/18 3:08 PM

LO6-1 The nature of qualitative research and its distinctions from quantitative research.

LO6-2 The types of business decisions that use qualitative methods.

LO6-3 The variety of qualitative research methods.

LO6-4 The importance and responsibility of the interviewer.

LO6-5 Ethical issues related to qualitative research.

>learningobjectives After reading this chapter, you should understand . . .

“The only way to capture a deeply personal

insight, which will help you evoke that emotion in

consumers, is through qualitative.” Gia Calhoun,

global insights manager

Burt's Bees

Stage 2: Data Collection Design: Qualitative Research

>chapter 6

schi18939_ch06_122-147.indd 122 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

123

Qualitative research methodologies have roots in a variety of disciplines, including anthropology, so- ciology, psychology, linguistics, communication, economics, and semiotics. Possibly because of these origins, qualitative methods have not enjoyed the unqualified endorsement of upper management. Some senior managers maintain that qualitative data are too subjective and susceptible to human error and bias in data collection and interpretation. The fact that results cannot be generalized from a qualitative study to a larger target population is considered a fundamental weakness. Thus some senior managers believe a qualitative study provides an unstable foundation for expensive and critical business decisions. Increasingly, however, managers are turning to these techniques because quantitative techniques take too long and fall short of providing true insights needed to make those ever-more-expensive business decisions.1

Historically, qualitative methodologies have been available much longer—some as early as the 19th century—than the quantitative tools managers rely on so heavily. While qualitative research currently accounts for about 20 percent of research expenditures by businesses, new technology is helping to rapidly expand the insights drawn from such methods [e.g., social media analysis (netnography), eth- nography (observation of human cultures), crowdsourcing (tasking a group of people with a problem or task, usually via the Internet, the availability of online communities and virtual groups, and most notably the smartphone].2 Some believe that the use of qualitative research by business is on the verge of massive expansion.3 Even the most recent GreenBook Research Industry Trends report cites several qualitative methods in its top 12 emerging research techniques (Exhibit 6-1).

Researchers deal with the issue of management’s concern over the trustworthiness of qualitative data by employing exacting protocols when using qualitative methodology:4

• Carefully using literature searches to build probing questions.

• Thoroughly justifying the methodology or combination of methodologies chosen.

• Executing the chosen methodology in its natural setting (field study) rather than a highly controlled setting (laboratory).

• Choosing sample participants for relevance to the breadth of the issue rather than how well they represent the target population.

• Developing and including questions that reveal the exceptions to a rule or theory.

• Carefully structuring the data analysis.

• Comparing data across multiple sources and different contexts.

• Conducting peer-researcher debriefing on results for added clarity, additional insights, and reduced bias.

>Exhibit 6-1 Emerging Qualitative Research Techniques

0 10 20 30 40 50 60 70

Online communities

Social media analytics

Text analytics

Webcam-based interviews

Mobile qualitative

Mobile ethnography

Research client (n = 322) Research provider (n = 1261)

Source: This graph was developed from data presented in Leonard F. Murphy, “GRIT Report,” Greenbook, June 2016, p. 5, downloaded June 30, 2016 (https://www

.greenbook.org/images/GRIT/2016_grit/GRIT_Report_2016_Edition.pdf).

schi18939_ch06_122-147.indd 123 2/14/18 3:12 PM

>part II The Design of Business Research

124

>What Is Qualitative Research? Qualitative research includes an “array of interpretive techniques which seek to describe, decode, trans- late, and otherwise come to terms with the meaning, not the frequency, of certain more or less naturally occurring phenomena in the social world.”5 These techniques include individual depth interviews (IDIs) and group interviews, as well as multiple-method approaches involving interviews and observation tech- niques (e.g., case studies, ethnography, grounded theory, action research). Qualitative research can be conducted asynchronously (from multiple participants at different times) or synchronously (from all par- ticipants at the same time). Qualitative research can be conducted as a cross-sectional study (e.g., one interaction per participant during the research period) or as a longitudinal study (e.g., multiple interac- tions per participant during the research period).

If the manager needs to know only what happened, or how often things happened, quantitative research methodologies would serve the purpose. But to understand the different meanings that people place on their experiences often requires research techniques that delve more deeply into people’s hid- den interpretations, understandings, and motivations. Qualitative research is designed to tell the re- searcher how (process) and why (meaning) things happen as they do. Judith Langer, a noted qualitative researcher, indicates that qualitative research is ideal if you want to extract feelings, emotions, motiva- tions, perceptions, people’s “language,” or self-described behavior.6 Exhibit 6-2 offers some examples of appropriate uses of qualitative research in business.

Qualitative research is based on “researcher immersion in the phenomenon to be studied, data which provide a detailed description of events, situations and interaction between people and things,

>Exhibit 6-2 Some Business Uses for Qualitative Research

Decision Arena Questions to Be Answered

Job analysis ● Does the current assignment of tasks generate the most productivity?

● Does the advancement through different job levels incorporate the necessary training to foster the strongest performance?

Advertising concept development ● What images should we use to connect with our target customers’ motivations?

Productivity enhancement ● What actions could we take to boost worker productivity without gen- erating worker discontent?

New product development ● What would our current market think of a proposed product idea?

● We need new products, but what should they be to take advantage of our existing customer-perceived strengths?

● Which products will create the greatest synergy with our existing prod- ucts in terms of ROI and distribution partner growth?

Benefits management ● Should our compensation plan be more flexible and customizable?

● How do employees perceive wellness-prevention programs as com- pared to corrective health programs in terms of value?

Retail design ● How do consumers prefer to shop in our store? Do they shop with a defined purpose, or are they affected by other motives?

Process understanding ● What steps are involved in cleaning a wood floor? How is our product perceived or involved in this process?

Market segmentation ● Why does one demographic or lifestyle group use our product more than another?

● Who are our customers and how do they use our product to support their lifestyle?

● What is the influence of culture on product choice?

Union representation ● How do various departments perceive the current effort to unionize our plant? Where and what are the elements of discontent?

Sales analysis ● Why have once-loyal customers stopped buying our service?

schi18939_ch06_122-147.indd 124 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

125

[thus] providing depth and detail.7 Qualitative research—sometimes labeled interpretive research be- cause it seeks to develop understanding through detailed description—contributes to building theory but rarely tests it.

Qualitative data were once exclusively about texts. Today, detailed descriptions of events, situations, and interactions—verbal, visual, and audio—constitute the data. Data may be contained within transcrip- tions of interviews or video focus groups, as well as in notes taken during those interactions. By defini- tion, qualitative techniques generate reams of words and images that need to be coded and analyzed by humans for meaning. While computer software is increasingly used for the coding process in qualitative research, at the heart of the qualitative process is the researcher—and his or her experience—framing and interpreting the data.8

In qualitative research, the researcher and manager (research sponsor) often have significant involve- ment in collecting and interpreting the data. The researcher may serve as a participant or a catalyst, as a participant observer, or as an interviewer. The manager may observe (in some cases via webcast of inter- views directly to the sponsor’s desktop computer), influence interview questions, and add interpretations

Businesses in the midst of digital transformation use digital ad-

vances, as well as improve their use of traditional technologies,

to change customer relationships, internal processes, and busi-

ness models. The MIT Center for Digital Business and Capgemini

Consulting set out to investigate whether and how large, tradi-

tional companies ($1 billion or more in annual sales) around the

world—as opposed to high-tech companies or digital startups—

are managing and benefiting from this digital transformation.

They conducted a multiyear study using in-depth interviews

with 157 executives (half business leaders and half IT leaders) in

50 companies in 15 countries. Interviews focused on how four

digital technologies—analytics, mobile, social media, and em-

bedded devices—are being used in the enterprise.

The study revealed companies follow three different

strategies to address three key areas of transformation (cus-

tomer experience, operational processes and business mod-

els): substitution (e.g., using smartphones rather than PCs for

email), extension (a significant improvement—e.g., using tab-

lets rather than paper for processing orders), or breakthrough

(fundamentally redefining a process or function—e.g., using

embedded devices and analytics to create fully automated

factories). Substitution dominates across the technologies ex-

cept in the use of social media. Embedded devices are still an

emerging technology; less than one-fourth of companies are

using these.

The interviews revealed that “very few businesses are gain-

ing the full value of technologically-enabled change, even with

technologies they already own.” Some have improved process

efficiency and improved financial data for strategic decisions,

but not learned how to use the information in negotiating with

suppliers. Others have gathered detailed information on buyer

behavior, but lack the analytical capabilities to tap the power

of the data they have collected. Still others haven’t begun the

transformation process.

Slow progress in digital transformation is caused by lack of

impetus at the top; lack of a clear business case for some tech-

nologies within some industries; regulatory, privacy, and security

concerns; missing skills, especially in analytics; organizational

culture related to changes in jobs due to automation, informa-

tion empowerment, and labor relations; or a poor information

technology base. Digital transformation successes reveal where

the IT team has a deep understanding of underlying business

processes and acts as internal process consultants; they ener-

gize the digital transformation.

Why did the researchers choose to use IDIs to collect this

data?

ebusiness.mit.edu; capgemini.com

Digital Transformation Revealed Using IDIs

>snapshot

©Mutlu Kurtbas/Getty Images

schi18939_ch06_122-147.indd 125 2/14/18 3:12 PM

>part II The Design of Business Research

126

and insights during the debriefing process. The qualitative researcher uses content analysis of written or recorded materials drawn from personal expressions by participants, behavioral observations, and debriefing of data collectors, as well as the study of artifacts and trace evidence from the physical envi- ronment of the participant interaction.

Qualitative studies with their smaller sample sizes offer an opportunity for faster turnaround of find- ings. While speed should never be the primary reason for choosing a methodology, qualitative data may be especially useful to support low-risk decisions that must be made quickly. Both group and individual interviewing, the mainstay techniques of qualitative research, can be conducted in highly secure environ- ments. In this era of data security, protecting research from the eyes of competitors is critically impor- tant. Qualitative techniques are suitable to research in all business disciplines.

>Exhibit 6-3 Qualitative Research Design within the Research Process

Design the Research

Develop Interview/ Discussion Guide

Investigative Questions

Measurement Questions

Data Collection Design

Recruitment Screener

Sampling Design

Assign Pretasks

Schedule Activities

Develop Activities (Including Stimuli)

ASYNCSYNC

Collect and Prepare the Data Debrief Data

Collectors/Sponsors

Stage 1

Stage 2

Stage 3

>Qualitative Research Design Exhibit 6-3 reveals several key distinctions that affect the research design when qualitative research is used:

• The level of question development in the management–research question hierarchy prior to the com- mencing of qualitative research and how it affects development of the measurement instrument. In qualitative studies, the researcher starts with broader investigative questions and has the ability to modify the measurement questions asked or the observations made—while the study is in progress—to gain a deeper understanding.

schi18939_ch06_122-147.indd 126 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

127

• The preparation of the participant prior to the research experience. In qualitative studies participants involved in an interview-based design are not only thoroughly prescreened but often prepared with pre-study exercises before the research starts (pretasking), sensitizing them to the research topic if not its objective. Participants are recruited in large part due to their ability to communicate their innermost thoughts and feelings; this requires some forethought on the part of the participant. The integration of data and insights that come from data collectors and sponsors as part of the data collection and preparation processes (debriefing).

• The integration of data and insights from the data collectors and manager-sponsors during data collec- tion and preparation. In a qualitative study, multiple sources add to the insights, especially debrief- ing of data collectors, manager-sponsors, and others observing participant interviews.

Key to the preceding is a fundamental difference in qualitative research: The researcher is the center; he or she is responsible for guiding the process to extract the desired insights. In quantitative research, the researcher is mostly invisible to the participant. In any quantitative study, any and all of these dis- tinctions would inject unacceptable bias, reducing the value of the data. But in qualitative research, they are the essence of discovering the insights that are sought. Exhibit 6-4 reveals the distinctions between qualitative and quantitative designs.

>Qualitative Sampling Design Unlike in quantitative design, the qualitative researcher, while identifying the desired characteristics of a sample case, may not predetermine the project’s sample size. The researcher seeks data saturation— the point at which no new information is forthcoming and no new insights seem feasible.9 That said, sample sizes for qualitative research vary by technique but are generally far smaller than quantitative statistical studies. Of course, there are always exceptions to the rule of smaller sample sizes. AT&T conducted a qualitative study to develop its 800 Reasons ad campaign for using its long-distance service, using thousands of structured interviews with business professionals in dozens of cities over several weeks. These interviews provided numerous reasons why businesses used the AT&T 800 ser- vice, and each of these “reasons” became the focus of a television and/or magazine ad in the multi-ad campaign.10

Qualitative research employs nonprobability sampling. In the arbitrary and subjective procedure for selecting sample cases, not only is little attempt made to generate a representative sample, the researcher often seeks a nonrepresentative one in order to extract the most, and most significant insights. Several types of nonprobability sampling are common to qualitative research:

• Purposive sampling. Researchers choose participants for their unique characteristics or their expe- riences, attitudes, or perceptions; as conceptual or theoretical categories of participants develop during the interviewing process, researchers seek new participants to challenge emerging patterns. Today, this often includes drawing participants from existing communities or panels.

• Snowball sampling. Participants refer researchers to others who have characteristics, experiences, or attitudes similar to or different from their own.

• Convenience sampling. Researchers select any readily available individuals as participants.

Recruiting top-quality participants is one of the most important jobs of the qualitative researcher. Without the right people, they won’t achieve deep meaning. Researchers want people who:11

• Share common characteristics, so they can be comfortable with each other.

• Have a range of life experiences.

• Hold different perspectives on issues.

• Can articulate their ideas, experiences, feelings, attitudes, and beliefs.

• Want to share their ideas/experiences.

• Want to hear the ideas/experiences of others.

• Like to collaborate to solve problems.

schi18939_ch06_122-147.indd 127 2/14/18 3:12 PM

>part II The Design of Business Research

128

>Exhibit 6-4 Qualitative versus Quantitative Research

Qualitative Quantitative

Focus of research ● Understand and interpret ● Describe, explain, and predict

Researcher involvement ● High—researcher is participant or catalyst ● Limited; controlled to prevent bias

Research purpose ● In-depth understanding; theory building ● Describe or predict; build and test theory

Sampling design ● Nonprobability; purposive, convenience, snowball

● Probability

Sample size ● Small to medium ● Medium to large

Research design ● May evolve or adjust during the course of the project

● Often uses multiple methods simultaneously or sequentially

● Consistency is not expected ● Involves either cross-sectional or longitudinal

approach

● Determined before commencing the project ● Uses single method or mixed methods ● Consistency is critical ● Involves either a cross-sectional or a longitudi-

nal approach

Engagement exercises, including projective techniques

● Frequently used ● Rarely used

Participant preparation ● Pretasking is common ● No preparation desired to avoid biasing the participant

Data type and preparation ● Verbal, video, or pictorial descriptions ● Reduced to verbal codes

● Verbal descriptions ● Reduced to numerical codes for computerized

analysis

Data analysis ● Human analysis following computer or human coding; primarily nonquantitative

● Analysis is ongoing during the project ● Distinctions between facts and judgments less

clear—researcher to seek the contextual frame- work of the phenomenon being measured

● Involves debriefing of data collectors/moderators/ research sponsors/observers for greater insights

● Computerized analysis—statistical and math- ematical methods dominate

● Analysis may be ongoing during the project ● Maintains clear distinction between facts and

judgments

Insights and meaning ● Determined by type and quantity of free- response questions, involvement activities; adjustable during the process

● Probing is expected and used ● Insights are formed and tested during the

process

● Limited by the quality of the measurement instrument

● Limited opportunity to probe participants during data collection, with limited ability to re-interview participants

● Insights follow data collection and data entry, with limited ability to re-interview participants

Research sponsor involvement

● May participate by observing research in real time or via taped interviews, or conducting inter- views or observations

● Rarely has either direct or indirect contact with participant

Feedback turnaround ● Smaller sample sizes make some data collection faster for shorter possible turnaround

● Insights are developed as the research pro- gresses, shortening data analysis

● Larger sample sizes lengthen data collection; Internet methodologies can shorten turnaround but inappropriate for some studies

● Insight development follows data collection and entry, lengthening research time frame; software permits some tallying of responses as data collection progresses

Data security ● More absolute given use of restricted access facilities and smaller sample sizes

● Act of research in progress is often known by competitors, employees, etc. 

● Insights may be gleaned by competitors/ employees for some visible, field-based studies

Source: This exhibit was developed from material extracted from Judith Langer, The Mirrored Window: Focus Groups from a Moderator’s Point of View (Ithaca, NY: Paramount Market Publishing, 2001); Hy Mari-

ampolski, Qualitative Market Research: A Comprehensive Guide (Thousand Oaks, CA: Sage Publications, 2001); David Carson, Audrey Gilmore, Chad Perry, and Kjell Gronhaug, Qualitative Marketing Research

(Thousand Oaks, CA: Sage Publications, 2001); and Norman K. Denzin and Ynonna S. Lincoln, editors, The SAGE Handbook Qualitative Research, 4th ed. (Thousand Oaks, CA: Sage Publications, April 27, 2011).

schi18939_ch06_122-147.indd 128 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

129

Increasingly, we are using online communities and panels to find participants because both the time and effort we need from them is significant. But we also use organization members when the research topic is general or relevant to a particular group. For example, we might recruit from the members of a church to interview prospects on laundry product purchases (general topic), or recruit those who belong to a social club to evaluate restaurants (topic relevant to a particular group). When the desired participants are difficult to access (e.g., doctors, lawyers, scientists, etc.), researchers are likely to use the services of a sample recruiter who will call and email prospects until a sample group is formed. Recruit- ing can sometimes take as long as the actual qualitative study.

Much of qualitative research involves the deliberate preparation of the participant called pretasking. Finding people who are willing to participate at this level of engagement is critical. This step is important due to the desire to extract detail and meaning from the participant. A variety of creative and mental ex- ercises draw participants’ understanding of their own thought processes and ideas to the surface. Some pretasking exercises include:

• In-home product use (with instructions to use the product or medium—e.g., a magazine—repeatedly over the preparation period before the interview).

• Collecting and supplying visual stimuli (e.g., family photos of areas or rooms in their homes that they hate to clean or have trouble decorating, or a favorite item of clothing).

• Preparing a visual collage (e.g., take pictures over several weeks, with a one-time-use camera, of their children’s favorite outfits for different purposes or situations, or cutting pictures out of magazines that reflect how they feel when dealing with a particular supervisor).

• Preparing detailed diaries of behavior and perceptions (e.g., a record of their step-by-step experience preparing a travel proposal).

• Drawing a picture of an experience (e.g., what they felt like when they last shopped in a particular store).

• Writing a dialog of a hypothetical experience (e.g., how a conversation between the participant and a sales associate would progress when a complaint was not resolved).12

There is a growing awareness among participants that their information has value, and that they should be compensated in some way for sharing information about themselves.13 The more time in- volved, the more exercises required, the more rare the participant, the more follow-up contacts desired— all increase the participation incentive. Thus, researchers have to be prepared to incentivize individuals for participating in qualitative research; this might not always be in the form of cash, but could also be in terms of discounts, customer perks, time off from the job, etc. Artem Patakov of CEO of Noom Inc., the behavioral modification organization designed to improve heath care, says firms increasingly need a privacy philosophy that includes not just a commitment to protect participant privacy, control of data, and participant choice to participate in data collection, but also a policy to fairly compensate partici- pants for sharing information.14

>Qualitative Data Collection Design Interviews The interview is the primary data collection technique for gathering data in qualitative methodologies. Interview techniques vary based on the number of participants involved in the interview, the level of structure, the proximity of the interviewer to the participant, and the number of interviews conducted during the research time frame. An interview can be conducted individually or in groups. Exhibit 6-5 compares the individual and the group interview as a research methodology. Both have a distinct place in qualitative research.

Interviewing requires a trained interviewer (called a moderator for group interviews) or the skills gained from experience. These skills include making respondents comfortable, probing for detail with- out making the respondent feel harassed, remaining neutral while encouraging the participant to talk openly, listening carefully, following a participant’s train of thought, and extracting insights from hours

schi18939_ch06_122-147.indd 129 2/14/18 3:12 PM

>part II The Design of Business Research

130

of detailed descriptive dialogue. Skilled interviewers learn to use their personal similarities with or dif- ferences from their interviewee to mine for information; similarities are used to convey sympathy and understanding, while differences are used to demonstrate eagerness to understand and empathize. Un- like quantitative research, where we are interested in the data collector following a prescribed procedure, in qualitative research the interviewer needs flexibility to extract meaning. This usually means the inter- viewer has a fuller understanding of the dilemma and how the insights will be used. The skilled inter- viewer must be a “quick-study,” someone who can grasp an understanding of an issue without necessarily having prior experience with the topic or being a technical expert.

The researcher chooses either an unstructured interview (no specific questions or order of topics to be discussed, with each interview customized to each participant; generally starts with a participant narrative—their story related to the topic of the research) or a semistructured interview (generally starts with a few specific questions and then follows the individual’s tangents of thought with interviewer probes) or a structured interview (often uses a detailed interview guide similar to a questionnaire to guide the question order and the specific way the questions are asked, but the questions generally remain open- ended). In mobile or Internet-based interviews, the moderator’s questions may be pretaped, especially if the interviews are asynchronous. Structured interviews permit more direct comparability of responses; question variability has been eliminated and thus answer variability is assumed to be real. Structured in- terviews are the choice when activities (such as video, photos, collaging, etc.) are used. In the structured interview, the interviewer’s neutrality is enhanced.

The unstructured and semistructured interviews used in qualitative research are distinct from the structured interview in several ways. They:

• Rely on developing a dialog between interviewer and participant.

• Require more interviewer creativity.

• Use the skill of the interviewer to extract more and a greater variety of data.

• Use interviewer experience and skill to achieve greater clarity and elaboration of answers.

Interviews conducted face-to-face in specialized facilities offer the obvious benefit of being able to observe and record nonverbal as well as verbal behavior. An interview, however, can be conducted

>Exhibit 6-5 A Comparison of Individual Depth Interviews and Group Interviews

Individual Interview Group Interview

Research Objective

● Explore life of individual in depth

● Create case histories through repeated interviews over time

● Test a survey

● Orient the researcher to a field of inquiry and the language of the field

● Explore a range of attitudes, opinions, and behaviors

● Observe a process of consensus and disagreement

● Add contextual detail to quantitative findings

Topic concerns

● Detailed individual experiences, choices, biographies

● Sensitive issues that might provoke anxiety if discussed in a group

● Issues of public interest or common concern

● Issues where little is known or of a hypothetical nature

Participants

● Time-pressed participants or those difficult to recruit (e.g., doctors)

● Participants with sufficient lan- guage skills (e.g., those older than seven)

● Participants whose distinctions would inhibit participation

● Participants whose backgrounds are similar or not so dissimilar as to generate conflict or discomfort

● Participants who can articulate their ideas

● Participants who offer a range of positions on issues

schi18939_ch06_122-147.indd 130 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

131

by mobile phone or via the Internet. In these instances, the participant must willingly use his or her own device’s camera to allow the interviewer such observation. Phone and online interviews offer the opportunity to conduct more interviews within the same time frame and draw partici- pants from a wider geographic area. These approaches also save the travel expenses of moving trained interviewers to participants, as well as the travel fees associated with bringing participants to a neutral site. Using interviewers who are fresher and more comfortable in conducting an interview—often from their home or office—should increase the quality of the interview. Also, de- pending on the group from which participants are drawn, there may be insufficient numbers to conduct group interviews in any one location, forcing the use of mobile or online techniques to supplement the sample.

Interviewers as Consultants As data collectors, interviewers have unique responsibilities in qualitative research due to their special- ized skill set. The interviewer, in his or her role as a collaborative consultant:15

• Recommends method and level of structure.

• Proposes the criteria for drawing the sample participants.

• Writes the recruitment screener and may recruit participants.

• Controls the interview, but also plans—and may manage—the locations and facilities for the study.

• Develops the various pretasking exercises.

• Prepares the interview or discussion guide (recommends topics, question wording, question order).

• Prepares instructions.

• Prepares any research stimuli (e.g., activities, tools used within activities) to be used during the interview.

• Debriefs the manager-sponsor and others observing the research.

• Supervises the transcription process.

• Helps analyze the data and draw insights.

• Prepares or directs the preparation of the client report, including extracting video or audio clips for the oral report.

Within this list of responsibilities, two deserve special attention: the recruitment screener and the interview/discussion guide. The interviewer is often responsible for generating the screening questions used to recruit participants for the qualitative research. This preinterview uses a device similar to a questionnaire, called a recruitment screener. Exhibit 6-6 provides the various elements necessary for a comprehensive recruitment screener. Each question is designed to reassure the researcher that the person who has the necessary information and experiences, as well as the social and language skills to relate the desired information, is invited to participate. Data gathered during the recruitment process are incorporated into the data analysis phase of the research, as recruitment data provide additional context for participants’ expressions.

The interviewer needs to be able to extract information from a willing participant who often is not consciously aware that he or she possesses the information desired. Thus the actual interviewer gener- ates the interview or discussion guide, the list of topics to be discussed (unstructured interview) or the questions to be asked (semistructured) and in what order (structured), as well as any preparatory pre- tasks or during-research activities. In building this interview/discussion guide, many interviewers employ a hierarchical questioning funnel, depicted in Exhibit 6-7. Broader questions start the interview, designed to put participants at ease and give them a sense that they have a lot to contribute, followed by increas- ingly more specific questions to draw out detail.

schi18939_ch06_122-147.indd 131 2/14/18 3:12 PM

>part II The Design of Business Research

132

>Exhibit 6-6 What Is Included in the Recruitment Screener? 

For best effect, qualitative research takes creative, articulate, expressive individuals. Finding appropriate participants is the task of the researcher. Here are some common elements addressed at this phase of the research.

Category Option

Heading Include project name, date of interviews, identity of screener. 

Screening requirements Specify conditions that must be met to extend a prospect an offer to partici- pate; may include quotas for various demographic, lifestyle, attitudinal, or usage questions. 

identity information Include name of prospect, address, phone, email.

Introduction Describe purpose of study in a motivational way. Completely “blind” studies do not motivate participation.

Security questions Reveal possible participant overparticipation or conflicts of interest; similar infor- mation on spouse or immediate family members.

Demographic questions Determine match for age, gender, ethnicity or race, income, geography, employ- ment status, or occupation.

Product/brand usage/ purchase questions

Establish frequency of use, purchase, loyalty, etc. 

Lifestyle questions Establish the participant’s daily life experiences, as well as those of the person with whom the participant shares his or her life.

Attitudinal and knowledge questions

Look for breadth in perceptions, attitudes, opinions, knowledge.

Articulation and creative questions

Seek evidence that participant can articulate his or her ideas and form and express opinions; scenarios might include problem–solution questions or ask participant to confront an unusual challenge. (“What could you do with a brick?”)

Offer or termination Invite participation, discuss compensation and pretasking, set up interview, or indicate that the person is not right for the current study but may be right for future studies.

>Exhibit 6-7 Interview Question Funnel

Broad Issue

What do participants consider entertaining?

Narrow to one topic:

What type of movie do participants consider most entertaining?

Narrow the topic:

What do participants find most entertaining about action movies?

Narrow to the client’s specific interest:

Within action movies, how do the computer-enhanced or computer-generated action sequences contribute to the movie’s entertainment experience?

Source: This graphic was adapted from one developed by Judith Langer and was published in The Mirrored Window: Focus Group from a Moderator’s Point of View (Ithaca, NY: Paramount

Market Publishing, 2001). www.paramountbooks.com.

schi18939_ch06_122-147.indd 132 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

133

Creative Exercises Creative exercises or activities are common within a qualitative interview. Because researchers are often looking for hidden or suppressed meanings, projective techniques can be used. Projective techniques are “exercises or activities that are purposely ambiguous and unstructured in nature so that beliefs, feelings, attitudes, and motivations, which may otherwise be hard for consumers to articulate, can be uncov- ered.”16 These techniques allow the participant to project their feelings and ideas onto another—often hypothetical—individual. Some examples of these activities include:17

• Word or picture association Participants are asked to match images, experiences, emotions, prod- ucts and services, even people and places, to whatever is being studied. “Tell me what you think of when you think of Kellogg’s Special K cereal.”

• Completion/fill in the blank Participants are given an incomplete sentence, story, argument, or con- versation and asked to finish it. “Complete this sentence: People who buy over the Internet . . .”

• Collage Participants are asked to collect images (e.g., from personal photos, magazines, or the In- ternet) and use these to express their feelings or the feelings of others like them on a topic.

• Cartoons or empty balloons Participants are asked to write the dialog for a cartoonlike picture. “What will the employee say when she sees her production supervisor approach her?”

• Thematic Apperception Test Participants are confronted with a picture (usually a photograph or drawing) and asked to describe how the person in the picture feels and thinks.

• Component sorts Participants are presented with flash cards containing component features and asked to create new combinations.

• Sensory sorts Participants are presented with scents, textures, and sounds, usually verbalized on cards, and asked to arrange them by one or more criteria.

• Laddering or benefit chain Participants are asked to link functional features to their physical and psychological benefits, both real and ideal.

• Write a letter Participants are asked to write a letter to a person empowered to make things hap- pen (e.g., CEO, HR director, parent, etc.).

• Storytelling A scenario is described and participants are asked to tell a story related to the scenario.

• Projective drawing Participants are asked to draw a concept or create a model, and then asked to explain it.

• Third-party projection Participants are asked to describe what others feel, believe, think, or do.

• Role-playing Participants are asked to assume a role within the organization and respond to a situa- tion as they think that individual would. • Creative innovation role-play Participants assume an innovation position (e.g., scientist), create

something (e.g., new product, package), and then explain it.

• Imagination exercises Participants are asked to relate the properties of one thing/person/brand to another. “If Crest toothpaste were a college, what type of college would it be?” • Imaginary universe Participants are asked to assume that the brand and its users populate an

entire universe; they then describe the features of this new world. • Visitor from another planet Participants are asked to assume that they are aliens and are con-

fronting the situation for the first time; they then describe their reactions, questions, and atti- tudes about the situation.

• Personification Participants are asked to imagine inanimate objects with the traits, character- istics and features, and personalities of humans. “If brand X were a person, what type of person would brand X be?”

• Authority figure Participants are asked to imagine that the brand or product is an authority fig- ure and to describe the attributes of the figure.

• Ambiguities and paradoxes Participants are asked to imagine a brand as something else (e.g., Tide dog food or Marlboro cereal), describing its attributes and position in the category.

schi18939_ch06_122-147.indd 133 2/14/18 3:12 PM

>part II The Design of Business Research

134

Exercises often draw out less verbal members of a group. Projective techniques can dissipate tension caused by sensitive topics or can be useful when a change of focus in the interview is im- minent. A well-trained interviewer is required if the research demands that one or more of these techniques be included within an individual depth interview or group interview. These techniques are also time-consuming to apply, lengthening the time frame of the individual or group interview. With online interviews, participants are asked to develop these techniques on their own time, to shorten the time spent with the interviewer. These techniques lengthen data analysis time, but en- rich the data obtained.

Individual Depth Interviews An individual depth interview (IDI) is an interaction between an individual interviewer and a single par- ticipant. Individual depth interviews generally take between 20 minutes (telephone interviews) and two hours (prescheduled, face-to-face interviews) to complete, depending on the issues or topics of interest and the contact method used. Some techniques such as life histories may take as long as five hours. Par- ticipants are usually paid to share their insights and ideas; $1 per minute is the budgeting rule of thumb for general consumers, but much higher rates are demanded by participants representing highly skilled professionals.18

Interviewees are often provided with advance materials via mail, fax, or the Internet. Recently, ad- vances in technology have encouraged the use of detailed visual and auditory aids during interviews, creating the methodology known as computer-assisted personal interviews (CAPIs). CAPIs often use a structured or semistructured individual depth interview.

Several unstructured individual depth interviews are common in business research, including oral histories, cultural interviews, life histories, critical incident technique, and sequential (or chronologic) interviewing. Exhibit 6-8 describes these techniques and provides examples.

When Anderson Analytics wanted to help its research participants delve into their deepest thoughts on a research topic, they used projective techniques that employed a colorful cast of characters. Participants choose one from the cast that most closely represents them or another they are trying to describe. They then attributed attitudes, feelings, emotions, and perceptions to the character. “The added bonus of the characters is that some of the characters elicit similar responses [across research projects], so we have some idea on benchmarking,” shared Anderson Analytics managing partner Tom Anderson. “For instance the heavier male character with the guitar is often thought of as an outsider, ‘poser’ wanting desperately to fit in, generally not well liked.” www.odintext.com

>picprofile ©T om

H . C

. A n

de rs

on (

O di

nT ex

t In

c. )

schi18939_ch06_122-147.indd 134 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

135

>Exhibit 6-8 Qualitative Methods That Use IDIs

Types How Research Is Conducted How Research Is Used

Oral history (narrative)

Ask participants to relate their personal experi- ences and feelings related to historical events or past behavior.

To develop products [e.g., books: September 11, 2001: Stories from 55 Broad Street by Eddie T. Deerfield and Thomas T. Noland Jr. (editors); An Album of Memories: Personal Histories from the Greatest Generation by Tom Brokaw].

Cultural interviews Ask a participant to relate his or her experiences with a culture or subculture, including the knowl- edge passed on by prior generations and the knowledge participants have or plan to pass on to future generations.

To determine product positioning or advertising creation

(e.g., how people use baking soda leads to positioning the product as not just a baking ingredient but also a deodorizer, toothpaste substitute, etc.).

Life histories Extract from a single participant memories and experiences from childhood to the present day regarding a product or service category, brand, or firm. Participants are encouraged to share how the significant people in their lives talked about or were involved with the organi- zation, how their attitudes or preferences have changed over their lives with respect to the organization, and how their perceptions and preferences have been altered by their various life experiences.

To determine positioning for com- pany prior to an identity or name change

(e.g., Frosted Flakes and Tony the Tiger—ad spots where adults feel they must appear in disguise because they eat a “child’s cereal”).

Critical incident technique

The participant describes:

● What led up to the incident.

● Exactly what he or she did or did not do that was especially effective or ineffective.

● The outcome or result of this action and why this action was effective or what more effec- tive action might have been expected.

To evaluate manufacturing pro- cesses, personal sales and telemar- keting sales programs, compensation or incentive programs, or other management-related incidents.

Convergent interviewing (convergent and divergent interviewing)

Experts serve as participants in a sequential series of IDIs; researcher refines the questions with each interview in order to converge on the central issues or themes in a topic area.

To develop appropriate questions for all types of research (in exploratory research).

Sequential interviewing (chronologic interviewing)

Approach the participant with questions formed around an anticipated series of activities that did or might have happened, in order to have the partici- pant recall the detail of his or her own experience.

To determine store design, adver- tising development, and product design; it is used to extract details related to shopping behavior, advertising consumption behavior, and product use behavior.

Ethnography Interviewer and participant collaborate in a field-setting participant observation and unstructured interview.

To determine product redesign, advertising development, positioning, distribution selection; to discover reactions and attitudes of striking employees.

Grounded theory Using a structured interview, each subsequent interview is adjusted based on the findings and interpretations from each previous inter- view, with the purpose to develop general concepts or theories with which to analyze the data.

To determine product design or redesign and advertising and promo- tion development.

Source: This exhibit was developed from Hy Mariampolski, Qualitative Market Research: A Comprehensive Guide (Thousand Oaks, CA: Sage Publications, 2001), p. 53;

David Carson, Audrey Gilmore, Chad Perry, and Kjell Gronhaug, Qualitative Marketing Research (Thousand Oaks, CA: Sage Publications, 2001), pp. 84–89 and 152–157;

Anselm Strauss and Julia Corbin, Basics of Qualitative Research: Techniques and Procedure for Producing Grounded Theory (Thousand Oaks, CA: Sage Publications, 1998);

and Norman K. Denzin and Ynonna S. Lincoln, editors, The SAGE Handbook of Qualitative Research, 4th ed. (Thousand Oaks, CA: Sage Publications, April 27, 2011).

schi18939_ch06_122-147.indd 135 2/14/18 3:12 PM

>part II The Design of Business Research

136

For 50 years, Maritz Travel Company (MTC) has produced

meetings, events, and incentive travel experiences. But it is re-

ally in the motivation business, helping its business clients edu-

cate, motivate, and inspire their employees to achieve business

objectives. With the recent acquisition of Experient plus service

expansions and aggressive growth goals, MTC hired Marcus

Thomas LLC to answer the questions, “How should Maritz

Travel Company be structured in order to build equity in the

entire portfolio of brands while enhancing long-term value?”

and “How can Maritz Travel Company optimize positioning and

messaging for each of its audiences?” The complex nature of

this business demanded a comprehensive research approach,

outlining every detail from sample recruitment of some rare

and hard-to-reach audiences to a detailed plan for analysis and

deliverables.

After three weeks of extensive recruiting using client lists

and target lists developed from trade publications, the research

included 22 in-depth interviews with key purchase decision

makers (meeting planners and association and trade show orga-

nizers with event budgets from $500,000 to $1,000,000) as well

as influencers for each audience within key markets and indus-

tries. All participants were screened for prior awareness of Maritz

Travel or Experient; this screening included competitors to keep

participants blind to the research purpose until it was revealed

during the interview. Interviews were conducted via in-person

focus groups (90 minutes) or WebEx IDIs (45–60 minutes). “The

online WebEx interview platform allows us to show imagery and

other assets by sharing our screen, as well as view participants’

non-verbal reactions on camera,” explained Edwige Winans, as-

sociate research director at Marcus Thomas. Brand perceptions

were captured using projective techniques, and buying journey

experiences were captured using a pretask storytelling exer-

cise. All interviews were recorded.

The comprehensive interview guide was designed to elicit

the following:

1. Perceptions of Maritz Travel Company, Maritz Travel,

Experient, and key competitors.

2. The equity of each brand.

3. The perceived value of three brand structure scenarios: ● P&G model (sub-brands stand above the company

brand).

● Marriott model (the company brand and the sub-brands

are balanced). ● Harley Davidson model (the master brand is leveraged

at the expense of sub-brands).

4. Buying journey experiences.

“[Once MTC was revealed as the research sponsor], to test

the different brand structures, specific hypothetical scenarios

were worked into the discussions designed to capture respon-

dent gut reactions. A change in brand structure clearly became

imperative, as an alternate structure demonstrated the greatest

potential for impact,” revealed Stefanie Riediger, strategic plan-

ner at Marcus Thomas.

Upon completion of this research, a complete analysis of the

gathered primary research in combination with provided com-

pany intelligence and secondary research, helped shape key

conclusions and implications for MTC. Marcus Thomas analyzed

responses according to their TRU approach: what was Truthful,

Relevant, and Unique about each sub-brand within the house of

brands. A final written and oral report outlined detailed findings

and summarized conclusions, with strategic implications, recom-

mendations, and next steps required to execute the brand strat-

egy successfully.

Maritz Travel Company has made critical strategic business

and communications decisions based on the research.

www.marcusthomasllc.com;www.maritzglobalevents.com

IDIs Help Restructure Maritz Travel

>snapshot

©Ryan McVay/Getty Images

schi18939_ch06_122-147.indd 136 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

137

Managing the IDI

Participants for individual depth interviews are usually chosen not because their opinions are representa- tive of the dominant opinion, but because their experiences and attitudes will reflect the full scope of the issue under study. Participants for individual depth interviews also need to be verbally articulate in order to provide the interviewer with the richness of desired detail. Primary Insights Inc. developed its CUE methodology to help marketers understand the performance cues that consumers use to judge a product. It uses purposive sampling to recruit individuals “with a specific interest in and aptitude for analytical thinking and discovering how things work.” CUE combines in-home product use with a diary pretask, followed by individual depth interviews that extract what the participant saw, felt, heard, smelled, and sensed when interacting with the product. What evolves is a hierarchy of sensory cues that clients may use when modifying products to improve customer satisfaction.19

Individual depth interviews are usually recorded (audio and/or video) and transcribed to provide the researcher with the rich detail for which the methodology is used. Interviewers are also themselves de- briefed to get their personal reactions to participant attitudes, insights, and the quality of the interview. Individual depth interviews use extensive amounts of interviewer time, in both conducting interviews and evaluating them, as well as facility time when premises are occupied for interviews. And while some respondents feel more comfortable discussing sensitive topics or sharing their own observations, behav- iors, and attitudes with a single person, others are more forthcoming in group situations.

Group Interviews A group interview is a data collection method using a single interviewer with more than one research participant. Driven by the belief that the data extracted will be richer because of the interaction, group interviews are one of the few research techniques in which the participants are encouraged to interact. Group interviews can be conducted in-person, on the phone or other mobile device, or via the Internet; the moderator works harder to keep participants engaged with each other in the latter two types of groups.

Group interviews can be described by the group’s size or its composition. Group interviews vary widely in size: dyads (two people), triads (three people), minigroups (two to six people), small groups (focus groups—6 to 10 people—unarguably the most well known of group interview techniques), or supergroups (up to 20 people). Smaller groups are usually used when the target population is small, when the topic or concept list is extensive or technical, or when the research calls for greater intimacy. Dyads also are used when the special nature of a friendship or other relationship (e.g., spouses, superior–subordinate, siblings) is needed to stimulate frank discussion on a sensitive topic. Dyads and triads are also used frequently with young children who have lower levels of articulation or more limited attention spans and are thus more diffi- cult to control in large groups. A supergroup is used when a wide range of ideas is needed in a short period of time and when the researcher is willing to sacrifice a significant amount of participant interaction for speed.

In terms of composition, groups can be heterogeneous (consisting of different individuals; variety of opinions, backgrounds, actions) or homogeneous (consisting of similar individuals; commonality of opinions, backgrounds, actions). Groups also can comprise experts (individuals exceptionally knowl- edgeable about the issues to be discussed) or nonexperts (those who have at least some desired informa- tion but at an unknown level).

A group interview’s structure and process include moderator interaction with the group and probing of the group to clarify responses. As a result, the moderator may create bias in the results by sending verbal and nonverbal signals that some responses are more favorable than others. The moderator might also direct discussion down paths that are least likely to help the client. Only training, and subsequent experience, can overcome these potential weaknesses of group interviews. The skilled researcher helps the manager determine an appropriate number of group interviews to conduct. The number of groups is determined by

• The scope of the issue(s) being studied: The broader the issue(s), the more groups needed.

• The number of a target population’s distinct segments of interest: The larger the number and the greater the distinctions, the more groups needed.

• The number of new ideas or insights desired: The larger the number, the more groups needed.

• The level of detail of information: The greater the level of detail, the more groups needed.

schi18939_ch06_122-147.indd 137 2/14/18 3:12 PM

>part II The Design of Business Research

138

>Exhibit 6-9 Factors Influencing Participant Contributions in Group Interviews

Positive/Facilitators

Recognition/ego enhancement

Moderator’s expressed appreciation for participant contributions that contribute to issue understanding; participant’s open agreement with other participant comments.

Personal contribution Participant’s desire to be helpful and perception that his or her contributions are helpful.

Validation Participant’s need to have his or her feelings, attitudes, or ideas validated.

Catharsis/load-sharing Participant’s need to share something negative or bothersome with others.

Personal growth Participant’s desire to increase knowledge or understanding through new per- spectives; participant’s desire for new experiences.

Socialization Participant’s desire to meet new people and make new friends in a “safe” environment.

Expectations Participant’s accurate understanding of the purpose of the group discussion.

Extrinsic rewards Participant’s value perception of fee for participation.

Negative/Inhibitors

Use of abstract terminology Moderator or participant’s use of terminology or unfamiliar jargon.

Ego threats Participant’s challenging another participant’s knowledge of the subject.

Political correctness Participant’s withholding comments for fear that his or her contributions might be perceived as disrespectful of another’s knowledge or opinions.

Ego defense Participant’s withholding a comment for fear that it will make him or her appear unintelligent or that the opinion will be unpopular with the group.

Memory decay Participant’s failure to remember incidents or details of incidents.

Embellishment Participant’s creative additions to memories of behaviors in order to participate fully or inflate status.

Inarticulation/rambling accounts

Participant’s inability to express ideas quickly or concisely.

Confusion Participant’s lack of understanding of the issue under discussion.

Reticence Participant’s need to be invited to participate (rather than actively volunteering comments).

Time Participant’s concern about other obligations.

Dominating/monopolizing Participant’s attempting to take leadership or the spotlight, thus blocking contri- butions of others.

• The level of geographic or ethnic distinctions in attitudes or behavior: The greater these influ- ences, the more groups needed.

• The homogeneity of the groups: The less homogeneity, the more groups needed.

The general rule is: Keep conducting group interviews until no new insights are gained. Often a limited number of groups will suffice, or sometimes the number might grow to 8 or even 12. It is often preferable, depending on the topic, to run separate group interviews for different subsets of the target population. For example, a study on nutritional advice may begin with separate consumer and physician groups to determine the best ways to provide the advice. This type of homogeneous grouping tends to promote more intense discussion and freer interaction.20

Researchers caution against forming groups solely on demographic descriptors; they favor “natural” groups (like families, co-workers, church members, etc.) where the participants share an affinity base.21 For customer groups, however, consideration should be given to such factors as gender, ethnic- ity, employment status, and education because culture is a primary determinant of perception. In a

schi18939_ch06_122-147.indd 138 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

139

recent exploratory study of discount shoppers, the attitudes about the economy and personal finances expressed by East Coast respondents and West Coast respondents diverged widely. The research spon- sor was able to use information from group interviews to build a strategy tailored to each geographic area.22

Regardless of group composition, it is the moderator who sets the tone of the group. Homogenous groups often discover their similarities early and get along well. But with heterogeneous groups, the moderator must provide the ice-breaker activities that get the participants interacting with each other. As with individual depth interviews, the moderator is responsible for developing the recruitment screener and the group discussion guide. Exhibit 6-9 summarizes the facilitators and inhibitors of individual participation in group interviews. A closer look at one of the best known of group interviews, the focus group, may clarify these distinctions.

Focus Groups The term focus group was first coined by R. K. Merton in his 1956 book, The Focused Interview. The focus group is a panel of people (typically made up of 6 to 10 participants), led by a trained moderator, who meet for 90 minutes to two hours. The facilitator or moderator uses group dynamics principles to focus or guide the group in an exchange of ideas, feelings, and experiences on a specific topic. You’ll find a sample focus group discussion guide for an in-person focus group in Appendix B.

Focus groups are often unique in research due to the research sponsor’s involvement in the process. Most facilities permit the sponsor to observe the group and its dynamics in real time, drawing his or her own insights from the conversations and nonverbal signals he or she observes. Many facilities also allow the client to supply the moderator with new topics or questions in real time that are generated by those observing. This option is generally not available in an IDI, other group interviews, or survey research.

Focus groups typically last about two hours but may run from one to three hours. Facilities usu- ally provide for the group to be isolated from distractions. Thus, the famous, or infamous, mirrored window allows those who are interested to observe the group while they avoid interfering with the group dynamics. Some facilities allow for product preparation and testing, as well as other creative exercises.

Fewer and lengthier focus groups are becoming common, especially as we use online panels and communities to draw participants. As sessions become longer, activities are needed to bring out deeper feelings, knowledge, and motivations. Many of the projective techniques used in IDIs are also used in focus groups:23

Focus groups are often used as an exploratory technique but may be a primary methodology. In two such cases, a small college used focus groups to develop a plan to attract more freshmen applications, and a blood center used a focus group to improve blood donations.24 Focus groups are especially valuable in the following scenarios:25

• Obtaining general background about a topic or issue.

• Generating research questions to be explored via quantitative methodologies.

• Interpreting previously obtained quantitative results.

• Stimulating new ideas for products and programs.

• Highlighting areas of opportunity for specific managers to pursue.

• Diagnosing problems that managers need to address.

• Generating impressions and perceptions of brands and product ideas.

• Generating a level of understanding about influences in the participant’s world.

Groups best enable the exploration of surprise information and new ideas. Discussion guides can be modified as the research team moves on to the next focus group. Even within an existing focus group, an adept facilitator can build on the ideas and insights of previous groups, getting to a greater depth of understanding.

schi18939_ch06_122-147.indd 139 2/14/18 3:12 PM

>part II The Design of Business Research

140

Qualitative research covers a wide range of techniques, but

they share important qualities. Every technique is an attempt

to elicit meaning. Every technique digs deep but uses a smaller

sample size to deliver that depth. Every technique depends

on communication, albeit in a variety of formats. Qualitative re-

searchers have used the Internet for decades, but they have

embraced the smartphone as penetration has improved. More

than 77 percent of U.S. adults now own a smartphone, up from

35 percent just six years earlier; making penetration near 100

percent among younger adults. Internet broadband is now

used by 88 percent of U.S. adults. Dan Weber, CEO of Canada-

based itracks, has been taking notice. “We’ve seen increased

adoption of online qualitative and specifically mobile qualitative

methods with increased penetration of mobile devices," shared

Weber. “Participants are communicating a strong preference for

mobile-friendly research experiences.”

itracks GO is a complete suite of online and mobile tools

for the qualitative researcher, covering IDIs and group discus-

sions. The newest version of this software has solved some

problems that have plagued qualitative researchers. itracks

has designed the itracks Board platform to engage with peo-

ple who may have intermittent or limited access to broad-

band or WiFi Internet by coming up with a mobile solution

that allows participants to respond to qualitative questions

and post video responses without being connected to the

Internet. When their device returns to an environment with

broadband or WiFi Internet access, the responses, including

videos, will be automatically posted. “Stability is no longer an

issue regarding mobile research,” shared Weber. “If you’re

conducting research using an Internet-based platform and

the participant’s daughter downloads a movie, you could lose

them. itracks Board will accommodate intermittent discon-

nections to the Internet by uploading of responses and vid-

eos once Internet access is restored. Research conducted on

smartphones now allows participation from any environment

including backyards, stores, or in remote locations regardless

of Internet capability.”

In an era of incredible time pressure, itracks GO, as part

of the software’s back end, builds in participant calendar no-

tices and reminders, as well as reminder notices to sponsor-

ing managers. Those managers can observe from their office

or home what the qualitative moderator sees and hears. In an

era of global research, itracks GO provides its suite in almost

any language (or can have a new language ready in fewer

than seven days).

Qualitative research-

ers generally want to re-

cord participants whether

in the midst of a group dis-

cussion, IDI, testing a new

product in their home or in

a store making a purchas-

ing decision. Any projec-

tive techniques generate

video. Video clips provide

an extensive amount of

additional information be-

yond word choice, includ-

ing body language and

tone of voice. These clips

also provide additional

context, which can be

observed in the video’s

background. But upload-

ing these high-quality,

full-length videos was

problematic until itracks

designed a way to com-

press and fragment a single video into thousands of smaller

clips that could upload quickly, then be reconnected just as

quickly. It also incorporates marking software for tagging clip

sentiments such as positive, neutral, or negative and offers

“VideoVault” for storing and editing videos into video highlight

reels. These highlight reels provide important validation sup-

port for research findings and insights in research reports.

Some qualitative subjects are highly sensitive. Weber noticed

an uptick in researcher concern that subjects who had agreed to

participate were dropping out of groups or IDIs; these participants

were uncomfortable broadcasting their own image. itracks GO

now includes an audio-based IDI tool that allows respondents to

participate without being filmed and includes media shares and

markup tools to enhance participant engagement.

Gen Y and Gen Z cohorts (born 1960–2000) are great tex-

ters, so itracks GO incorporates a text-based chat platform. Tran-

scripts of these sessions can be instantly downloaded and later

analyzed by text-analytic software, such as OdinText.

“Mobile is playing a critical role in Qual,” claimed Weber. He

expects that role will be even more important as our depen-

dence on the smartphone continues to grow.

www.itracks.com

Qualitative Research in the Era of Smartphones

>snapshot

Courtesy of itracks

schi18939_ch06_122-147.indd 140 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

141

Telephone or online focus groups have proven effective in the following scenarios:

• When it is difficult to recruit desired participants—members of elite groups and hard-to-access individuals such as experts, professionals, physician specialists, high-level executives, and store owners.

• When target group members are rare, “low-incidence,” or widely dispersed geographically— directors of a medical clinic, celebrities, early adopters, and rural practitioners.

• When issues are so sensitive that anonymity is needed but respondents must be from a wide geo- graphic area—people suffering from a contagious disease, people using nonmainstream products, high-income individuals, competitors.

• When you want to conduct only a couple of focus groups but want nationwide representation.

Additionally, online focus groups work well when:

• Teens or young adults are the desired participants.

• Desired participants are technically employed segments of the market, those essentially comfort- able with computer use.

• A computer-based application, such as software or a game, is the topic of group discussion.

• Visual stimuli are necessary to discuss the topics of interest.

Like telephone focus groups, videoconferencing and online focus groups offer significant savings over in-person focus groups. By reducing the travel time for the moderator and the manager, coordinating such groups can be accomplished in a shorter time and at a lower cost.

Group Interview Drawbacks and Their Solutions  However, given time constraints, group interviews permit spending only limited time extracting detail from each participant.26 This problem is magnified when a group interview is structured to cover numer- ous questions or topics.

Another drawback of the group interview is the increased difficulty recruiting, arranging, and co- ordinating group discussions that require travel. But this aggravation—which can be subcontracted to a specialist research supplier—is deemed a small price to pay for the insights that often are revealed by group interaction.

Interviewers are tested by the challenge of managing the group’s conversation while avoiding inter- jecting themselves into the group’s process. It is also the moderator’s job to control the extrovert or dom- inant personality and ensure meaningful contributions from all others, including the most introverted or private thinkers. When control is not maintained, some members’ opinions may be suppressed and valuable insights lost. Sometimes an individual will be more honest with a neutral interviewer than with a group of peers. One example is a group of small-business owners being unwilling to divulge competitive strengths and weaknesses. A skilled researcher can anticipate which topics are more likely to obtain good results with an individual or a group interview.

Recording, Analyzing, and Reporting Group Interviews In face-to-face settings, some moderators use large sheets of paper on the wall of the group room to record trends; others use a personal notepad. Facility managers produce both video- and audiotapes to enable a full analysis of the interview. The verbal portion of the group interview and moderator debriefing sessions are transcribed and added to moderator notes. These are analyzed across several focus group sessions using content analysis. This analytical process provides the research sponsor with a qualitative picture of the respondents’ concerns, ideas, attitudes, and feelings. The preliminary profile of the content of a group interview is often done with computer software in content analysis. Such software searches for common phrasing and words, context, and patterns of expression on digitized transcripts.

schi18939_ch06_122-147.indd 141 2/14/18 3:12 PM

>part II The Design of Business Research

142

In a researcher’s search for meaning, he often combines the two primary qualitative methodologies of in- terview and observation. Two of these combinations, the case study and action research, deserve special mention as they are used frequently in business research.

Case Study27

The case study, also referred to as the case history, is a powerful research methodology that combines individual and (sometimes) group interviews with record analysis and observation. Researchers ex- tract information from company brochures, annual reports, sales receipts, and newspaper and maga- zine articles, along with direct observation (usually done in the participant’s “natural” setting), and combine it with interview data from participants. The objective is to obtain multiple perspectives of a single organization, situation, event, or process at a point in time or over a period of time. Case study methodology—or the written report from such a research project, often called a case analysis or case write-up—can be used to understand particular processes. For example, one study might evaluate new product development processes for similarities, especially the use of outside consultants, ideational techniques, and computer simulation. Another study might examine in detail the purchaser’s response to a stimulus like a display. The results of the research could be used to experiment with modifications of the new product development process or with display selection and placement processes to gener- ate higher-value transactions. The research problem is usually a how and why problem, resulting in a descriptive or explanatory study.

Researchers select the specific organizations or situations to profile because these examples offer critical, extreme, or unusual cases. Researchers most often choose multiple cases, rather than a single

Founder and principal researcher Robert W. Kahle of Kahle

Research Solutions Inc., in his book Dominators, Cynics, and

Wallflowers, dissects typical focus group participants to illu-

minate ways to modify their problem behaviors. Dominators

are all-knowing, quick to answer, and choose a seat location

in order to challenge the moderator for control. Cynics display

negative behaviors and deride the ideas of others. Hostiles

have an agenda of their own and seek corrective action; they

are often angry and combative. Intoxicateds are under the

influence of something, fidgety and incoherent. Proselytizers

cannot accept that others hold opposing opinions and try to

persuade others to their opinion. Blatherers offer long, off-

topic answers and ignore moderator cues. Jokers find every

comment source material for a new joke, story, or comical fa-

cial expression. Followers tend to repeat others’ opinions.

Wallflowers withdraw both physically and verbally. Finally,

Co-moderators often engage participants before a discus-

sion starts, ask questions of their own, and seek to befriend or

support other participants.

Why is each of these behaviors a problem and how would

you handle each of these problem participants?

www.kahleresearch.com, www.paramountbooks.com

Problems within Focus Groups

>snapshot

©Spencer Grant/PhotoEdit/Getty Images

>Combining Qualitative Methodologies

schi18939_ch06_122-147.indd 142 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

143

case, to study because of the opportunity for cross-case analysis. In studying multiple cases, a deeper understanding emerges. When multiple cases are chosen, it is because they offer similar results for pre- dictable reasons (literal replication) or contrary results for predictable reasons (theoretical replication). While theoretical sampling seems to be common, a minimum of 4 cases with a maximum of 15 seems to be favored.

In the case study, interview participants are invited to tell the story of their experience, with those chosen representing different levels within the same organization or different perspectives of the same situation or process to permit depth of perspective. The flexibility of the case study approach and the emphasis on understanding the context of the case being studied allow for a richness of understanding sometimes labeled thick description.

During analysis, a single case analysis is always performed before any cross-case analysis is con- ducted. The emphasis is on what differences occur, why, and with what effect. Prescriptive inferences about best practices are concluded after completing case studies on several organizations or situations and are speculative in nature.

Business students are quite familiar with studying cases as a means of learning business principles. In Search of Excellence, a book by Tom Peters and Robert Waterman, was developed using case study meth- odology.28 Other similar studies profiled in books written on Procter & Gamble and Disney have also used this methodology. In the business arena, such case studies have examined changes in new product development, sales processes, hiring practices, and training programs.

Action Research Managers conduct research in order to gain insights to make decisions in specific scenarios. Action research is designed to address complex, practical problems about which little is known—thus no known heuristics exist. So the scenario is studied; a corrective action is determined, planned, and implemented; the results of the action are observed and recorded; and the action is assessed as ef- fective or not. The process is repeated until a desired outcome is reached, but along the way much is learned about the processes and about the prescriptive actions being studied. Action researchers investigate the effects of applied solutions. Whatever theories are developed are validated through practical application.29

Suppose a restaurant that had never received a customer complaint earns its first challenge by a disgruntled diner. If no general rule existed about how to treat unhappy patrons, the organization could study the situation and come up with alternative actions. It might:

• Ignore the problem. (Its lack of experience would prevent it from knowing that negative word of mouth—negative buzz—would be the likely result.)

• Do whatever is necessary to replace the unsatisfactory meal within the shortest period of time.

• Accept the current circumstance as uncorrectable, apologize to the customer, and remedy the situation by picking up the table’s full dining tab and offering the customer a free meal to get him or her back in the restaurant another day.

In action research, one of those alternatives would be chosen and implemented, and then the results recorded. Was the customer happy when he or she left? Did the customer return to dine another evening or never return again? Over the next three months, what was the customer’s full revenue value? If the customer didn’t return, the next time a disgruntled customer voiced dissatisfaction, a different action would be chosen, implemented, and then assessed in comparison to the first option’s results.

> Merging Qualitative and Quantitative Methodologies

Triangulation is the term used to describe the combining of several qualitative methods or com- bining qualitative with quantitative methods. Because of the controversy described earlier, quali- tative studies may be combined with quantitative ones to increase the perceived quality of the

schi18939_ch06_122-147.indd 143 2/14/18 3:12 PM

>part II The Design of Business Research

144

research, especially when a quantitative study follows a qualitative one and provides validation for the qualitative findings. Four strategies for combining methodologies are common in busi- ness research:30

1. Conducting qualitative and quantitative studies simultaneously.

2. Conducting multiple waves of quantitative studies, measuring changes in behavior and attitudes over time, while a qualitative study is ongoing.

3. Conducting a qualitative study prior to conducting a quantitative study. A second qualitative study then might follow the quantitative study, seeking more clarification.

4. Conducting a quantitative study before conducting a qualitative study.

An example of the first strategy would be the combination of a public opinion poll at the time focus groups are being held to discover ways to sway a particular public’s opinion on a union election process. For the second strategy, we might collect life histories while multiple waves of questionnaires are measuring the response to differing training tactics. For the third, we could perform a qualitative study to identify peoples’ behaviors and perceptions with respect to furniture shopping processes and interior decorating; then we could use that information to develop a quantitative study to measure the actual frequency of behaviors and attitudes. And, fourth, we might survey people’s behavior and attitudes toward a change in company mission and find we need some IDIs to explain findings that are unclear.

Many researchers recognize that qualitative research compensates for the weaknesses of quantitative research and vice versa. These forward thinkers believe that the methodologies complement rather than rival each other.

>Ethical Issues and Their Solutions As with any design, the participants, the research sponsors, and the researchers have responsibilities in the planning of qualitative research. These responsibilities are focused on privacy, confidentiality, qual- ity, and safety. Qualitative studies often cover numerous activities over a span of days or weeks; partici- pants have the responsibility of completion. They are also asked to share life experiences and anecdotes; they have a responsibility of truthfulness.

Managers often don’t have the skill set to conduct qualitative research because many of the techniques require specialized educational background and training. Managers who want qual- ity research have the responsibility of full disclosure and lack of coercion when hiring that expertise.

Because qualitative research often tackles sensitive subjects and probes for meaning, researchers have the responsibility to protect confidentiality of both participant and the manager sponsor, choose ap- propriate methods and activities within those methods, and train moderators/interviewers within those methods to ensure quality.

To address these concerns, researchers often go to extremes to make participants feel comfortable and keep them engaged. Participants are sharing their innermost thoughts and personal experiences with the researcher. One of the strengths of qualitative research is the ability to use a participant’s voice, video, photos, collages, etc., as evidence to support researcher insights. During the process, the par- ticipants need to feel that their identity is protected. Thus researchers must deliver on the participants’ right of privacy and the right of consent before using their image, words, and audio and video files. The researcher must also protect the insights gathered, sharing these only with the manager-sponsor. It is especially important in qualitative research to respect the data’s limitations: not to project results of a small sample to a target population.

schi18939_ch06_122-147.indd 144 2/14/18 3:12 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

145

action research 143

authority figure 133

ambiguities and paradoxes 133

cartoons or empty balloons 133

case study (case history) 142

collage 133

completion/fill in the blank 133

component sorts 133

computer-assisted personal interview (CAPI) 134

content analysis 141

convergent interviewing 135

creative innovation role play 133

critical incident technique 135

cultural interviews 135

data saturation 127

ethnography 135

expert group 137

focus group 139

grounded theory 135

group interview 137

heterogeneous group 137

homogeneous group 137

imaginary universe 133

imagination exercises 133

individual depth interview (IDI) 134

interview 129

interview guide (discussion guide) 131

laddering or benefit chain 133

life histories 135

>keyterms

LO6-1 Qualitative research includes an array of interpretive tech- niques that seek to describe, decode, translate, and other- wise come to terms with the meaning, not the frequency, of certain more or less naturally occurring phenomena in the social world. Qualitative research methodologies differ from quantitative methodologies based on the focus of the research; its purpose; researcher involvement; sampling design; sample size; research design, including participant pretasking; data source, type, and preparation; methods of data analysis; level of insights and meaning extracted; research sponsor involvement; speed of the research; and data security. A qualitative methodology may be used alone to address organizational problems or in combina- tion with other qualitative or quantitative methodologies.

LO6-2 Qualitative research is designed to tell the researcher how (process) and why (meaning) things happen as they do. It is appropriate for all business disciplines and is used extensively in marketing, HR, finance, and general management. In business planning and deci- sion making, qualitative methodologies are used in job analysis, advertising concept development, productiv- ity enhancement, new product development, retail design, process understanding, market segmentation, sales analysis, and union representation.

LO6-3 Qualitative methodologies used in decision making evolved from techniques used in anthropology, sociol- ogy, psychology, linguistics, communication, economics, and semiotics. Common among these methods are the individual depth interview (IDI) and the group interview, as well as observation combination methodologies (e.g., eth- nography, action research, case analysis, and grounded theory). Within group interviews, the focus group is the most widely used methodology. Qualitative research often uses creative exercises to probe deeply into motivations,

attitudes, opinions, feelings and experiences. These exer- cises can include projective techniques, which encourage the participant to share their own thoughts while project- ing them on another. Among these techniques are word or picture association, completion, cartoons or empty balloons, the Thematic Apperception Test, imagination exercises, and sorting and writing exercises.

LO6-4 The interviewer is central to great qualitative research and should be chosen for his or her skills. The interviewer is not just a data collector, but a collaborative consultant on the project, with numerous contributions, including recommending method and level of structure; developing the criteria for sample participants, developing the recruit- ment screener, and recruiting participants; controlling the interview and managing special facilities, if necessary; developing pretasking exercises; preparing the interview/ discussion guide; preparing instructions, research stimuli and tools; debriefing observers; supervising the transcrip- tion process; contributing to data analysis and insight for- mation; and preparing or directing the preparation of the research report, including extracting video or audio clips.

LO6-5 The ethical issues of qualitative research include con- fidentiality (both sponsor and participant), quality, and safety. The interviewer/data collector shares these responsibilities with the research designer. Due to the nature of the sensitive subjects sometimes discussed in qualitative research, protecting the participant is para- mount. Obtaining informed consent is the best strategy for working with this issue, as well as adhering to strict privacy guidelines. Participants often agree to engage in projects that cover a period of days and sometimes weeks; they have the responsibility to complete research once it is started, as well as to share truthfully their life experiences, thought, attitudes, and opinions.

>summary

schi18939_ch06_122-147.indd 145 3/3/18 2:37 PM

>part II The Design of Business Research

146

moderator 129

nonexpert group 137

nonprobability sampling 127

oral history (narrative) 135

personification 133

pretasking 127

projective drawing 133

projective techniques 133

qualitative research 124

quantitative research 124

recruitment screener 131

role playing 133

semistructured interview 130

sensory sorts 133

sequential interviewing 135

storytelling 133

structured interview 130

Thematic Apperception Test 133

third-party projection 133

triangulation 143

unstructured interview 130

visitor from another planet 133

word or picture association 133

write a letter 133

Terms in Review

1 How does qualitative research differ from quantitative research?

2 What is data saturation, and how does it influence qualita- tive research?

3 What is the nature of data collected in qualitative research?

4 Why do senior executives feel more comfortable relying on quantitative data than qualitative data? How might a qualita- tive research company lessen the senior-level executive’s skepticism?

5 Distinguish among structured, semistructured, and unstruc- tured interviews.

Making Research Decisions

6 Assume you are a manufacturer of small kitchen electrics, like Hamilton Beach/Proctor Silex, and you want to deter- mine if some innovative designs with unusual shapes and colors developed for the European market could be suc- cessfully marketed in the U.S. market. What qualitative re- search would you recommend, and why?

7 NCR Corporation, known as a world leader in ATMs, point- of-sale (POS) retail checkout scanners, and check-in kiosks at airports, announced in June 2009 that it would move its world headquarters from Dayton (OH) to Duluth (GA), a suburb of Atlanta, after more than 125 years. An employer of 1,200 mostly high-salaried, professional workers in Dayton, NCR was enticed to move by Georgia’s offer of more than $56.9 million in tax credits; its fast-growing, educated 25- to 34-year-old population cohort; international offices for 10 European state governments; and the busiest international airport (Atlanta) in the world.

a What qualitative research might NCR have done to reach this decision?

b NCR will use its move to Georgia to downsize its world headquarters workforce. What qualitative research could help NCR determine which of its 1,200 employees will be offered positions in Duluth?

From Concept to Practice

8 Use Exhibit 6-6 to develop the recruitment screener for the research you described in your answer to question 6.

9 Conduct a focus group among students in your class on one of the following topics:

a The department’s problems offering requirements and electives essential for meeting your graduation expectations.

b Entertainment sponsored by your university to bring the community on campus.

From the Headlines

10 Lately, airlines have been having a rough time, in terms of legal actions and PR issues, with consumers openly expressing outrage at being bumped from—or forcibly re- moved from—flights. Design a qualitative study to reveal the suppressed (as opposed to surface) issues that are contrib- uting to this rage.

a What are some of the surface issues?

b Who will you want to participate and how will you recruit them?

c What qualitative method(s) will you choose. Be specific about any exercises you will incorporate.

>discussionquestions

schi18939_ch06_122-147.indd 146 3/3/18 2:37 PM

>chapter 6 Stage 2: Data Collection Design: Qualitative Research

147

>cases*

Akron Children’s Hospital

Covering Kids with Health Care

Lexus SC 430

NCRCC: Teeing Up a New Strategic Direction

Ohio Lottery: Innovative Research Design Drives Winning

Open Doors: Extending Hospitality to Travelers with Disabilities

Ramada Demonstrates Its Personal BestTM

Starbucks, Bank One, and Visa Launch Starbucks Card Du- etto Visa

USTA: Come Out Swinging

*You will find a description of each case in the Case Abstracts section of this textbook. Check the Case Index to determine whether a case provides data, the research instrument, video, or other supplementary material. Cases and case supplements are available in Connect.

schi18939_ch06_122-147.indd 147 2/14/18 3:12 PM

LO7-1 When observation studies are most useful.

LO7-2 The distinctions between monitoring nonbehavioral and behavioral activities.

LO7-3 The strengths of the observation approach in research design.

LO7-4 The weaknesses of the observation approach in research design.

LO7-5 The three perspectives from which the observer–participant relationship may be viewed in observation studies.

LO7-6 The various designs of observation studies.

>learningobjectives After reading this chapter, you should understand . . .

Stage 2: Data Collection Design: Observation Research

>chapter 7 “No longer do

companies study consumers’ psyches only

by asking people what they think about technology and how

they use it. Now they conduct observational research, dispatching

anthropologists to employ their ethnographic skills by interviewing, watching and videotaping

consumers in their natural habitats.” Katie Hafner,

author

schi18939_ch07_148-169.indd 148 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

149

Research designs are classified by the approach used to gather primary data: Researchers can monitor (observe) or communicate. Behavioral scientists define observation research in terms of animal or human behavior, but this is too limiting for observation as a methodology in business research. We prefer that you define observation as a method that collects data through all our primary senses; it involves listening, reading, smelling, tasting, and touching.

It’s important to remember that businesses are using research to identify paths to innovation. Struc- tured communication studies sometimes stifle the very participants businesses hope to use to open doors to such innovation; they interrupt participants’ natural thought processes and behaviors with questions that might seem irrelevant to the participant. During observations, participants engage in actual behav- iors with processes, objects, environments, and people. Participants offer nonverbal cues about their feel- ings through body language and verbal cues through spontaneous comments. Trained observers often identify participant innovations and the motivations for those innovations (e.g., participants combine products or products and services to tackle a problem; they change processes to work more efficiently, save time, or reduce waste). Participants in communication studies have proven to be highly unreliable reporters of their own behavior, but observers work with actual actions, not memories. Observers can be trained not only in what is but in what might be feasible. Exhibit 7-1 describes the conditions under which observation is an appropriate method for data collection, contrasting those conditions with ones from the communication modes (e.g., interviews, surveys, etc.).

While it might appear that the communication approach dominates, observational research has never been more important to business. We monitor everything—from the movement of goods from factory floor to retail store to the performance of every part in a fast-moving train; from the performance of workers in emergency rooms to the shelf life of food in the refrigerators in homes; from the behavior of patrons in restaurants to the readers’ eye movements on web pages. Where once organizations might

>Exhibit 7-1 Selecting the Data Collection Method

C

O M

M U N I C

A T

I

O N

O B S E R

V A

T

I O

N

Respondent has suppressed orburied information Sensitive material needed

Only behavior data needed

M em

ory decay

a m ajor factor

Limit ed ac

cess to

respo nden

t

Ex ten

siv e a

mo un

t

of inf

orm atio

n n ee

de d

Re sp

on de

nt n

ee ds

tim e

to a

ns w

er

Fu tu

re in

te nt

io ns

n ee

de d

Respond ents

widely d istributed

Info rma

tion ne

ede d q

uic kly

Pu rp

os e

m us

t b e

di sg

uis ed

Re sp

on de

nt u

nw ill

in g

to p

ar tic

ip at

e

N at

u ra

l s e

tt in

g im

p e

ra tiv

e

N ee

d to

a vo

id

m es

sa g

e fil

te rin

g

Anonymity needed or desired

Attitude/opinion inform ation

needed

Expectations needed M

o tivatio

n need ed

P ast b

e h

avio r n

e e

d e

d

Fo llo

w- up

s,

ca llb

ac ks

n ec

es sa

ry

D ev

ic es

e xi

st fo

r ac

cu ra

te re

co rd

in g

Investigative Questions

schi18939_ch07_148-169.indd 149 2/14/18 3:14 PM

>part II The Design of Business Research

150

have relied solely on human observation, then later enhanced that human observation with camera foot- age, today some observation increasingly relies on embedded sensors or software as collection methods. In the past few years, we’ve reached a tipping point where computerized sensors in devices—the Internet of Things (IoT)—are expanding our observation data at a rate that exceeds our ability to analyze it. And more changes in observational research are coming.

Much of what we know comes from observation. We notice co-workers’ reactions to political intrigue, the sounds in the assembly area, the smell of perfume, the taste of office coffee, the smoothness of the vice president’s marble desk, and a host of other stimuli. While such observation may be a basis for knowledge, the collection processes are often haphazard. Observation research, as scientific inquiry, is conducted specifically to answer a research question, must be systematically planned and executed, must use proper controls, and must provide a reliable and valid description of what happened. The versatility of observation makes it an indispensable primary method and a supplement for other meth- ods. Exhibit 7-2 depicts the use of observation in the research process.

Sometimes, while observation is the right approach, we need to change the observation protocol to achieve our research objectives. One example might be how we use observation research to under- stand planes and catastrophic airline accidents. Currently, airlines collect critical plane operation data and minute-by-minute cockpit conversations with sensors and digital recording devices that store this

>Exhibit 7-2 Observation and the Research Process

Design the Research

Prepare Error Reduction Plan

Case available

Case unavailable

Measurement Questions

Design Observer Training

Train Observer(s)

Investigative Questions

Collect and Prepare Data

Pretest Checklist & Protocols

Prepare Observation Checklist

Verify Observation Protocol

Select Observation

Approach

Reschedule Observation

Select Case and Schedule Observation

Select or Build the Sample Frame

ProbabilityNon-ProbabilitySelect Type of Observation (see Exhibit 8-3)

Stage 1

Stage 2

Stage 3

Data Collection Design Sampling Design

schi18939_ch07_148-169.indd 150 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

151

information in “black boxes” mounted in the plane. Once collected, we might analyze these data to schedule maintenance, understand part design or performance, or explain plane failure. But in one year, we lost hundreds of lives when three planes crashed and the black box—and its data—was not retrieved. Following one such accident, a CBS News story reported that current technology could capture such information in real time and store it on cloud servers. Such a method would allow the airline to know precisely what happened to a downed plane and when and where it happened. As with all research deci- sions, the cost of using this newer methodology (estimated at as much as $10,000 per year per plane) is a concern, but also to be considered is the coverage of ground-based radar (estimated at 30 percent of the world), the availability of satellite-based WiFi and GPS, and the high cost of black-box retrieval in water-based crashes. Are observation data captured via on-plane black boxes sufficient as an observation method to meet the research objectives?1

>Observation Research Design Numerous factors influence research design, as summarized in Exhibit 7-3, including:

• Choice of sampling design

• Choices within data collection design: • Content of observation • Environment • Observational condition • Data collection protocols • Observer presence • Observer concealment

>Exhibit 7-3 Factors Influencing Observation Research Design

Factors Options

Who/What Sampling Designs ● Event sampling

● Time sampling

What Content of Types ● Factual

● Inferential

● Both

What Types of Observation ● Nonbehavioral observation

● Behavioral observation

When/where Environment ● Field (natural setting)

● Laboratory (artificial setting)

How Data Collection Protocols ● Simple observation

● Systematic observation

How Observer presence ● Direct

● Indirect

How Observation concealment ● No Concealment

● Concealment

● Partial Concealment

How Observer involvement ● No Involvement

● Participant observation

schi18939_ch07_148-169.indd 151 2/14/18 3:14 PM

>part II The Design of Business Research

152

© M

at H

ay w

ar d/

S h

ut te

rs to

ck

>Sampling Design What qualifies a case—a participant activity, record, object, event—to be observed? The answer is in the investigative questions. If a case possesses that information, it qualifies.

Acts or behaviors, time frames or people can represent target populations in observation. While some forms of observation can use probability sampling (e.g., analysis of records, video records of behaviors), many forms of observation must rely on nonprobability sampling.

Within these two broad categories, there are two types of sampling designs that are specially designed for observation: event sampling and time sampling.2 In event sampling, the researcher records that a specific behavior or act occurs; the duration of the observation is determined by when the act begins and ends. In some instances, the researcher may also capture its antecedents—events and conditions im- mediately before the behavior of interest—and the consequences—what happens immediately after the event. For example, assume a machine malfunctions or stops; the event of interest is the stoppage itself. When did it occur, what was the machine producing, what or who stopped the machine (an internal device, a person operating the machine, etc.), who was operating the machine at the time it stopped, etc. While it is unlikely we could hire an observer to wait and watch for machines to malfunction, we can use records of such events. Either event or time sampling can be used when observing human behavior.

In time sampling, the choice of cases is based on some time interval. There are three types of time sampling: a time-point sample, continuous real-time samples, or a time-interval sample. For time-point sampling, recording occurs at fixed points for a specified length of time, the variable of interest is mea- sured if it falls within that time frame from all cases. For example, a manager might be interested in how often her employees check their email during the day. With the assistance of her IT department, she monitors email access during three times: 2:00, 3:00, and 4:00 for 10 minutes each time. If employees ac- cess their email, she measures not only access, but duration of access (were they on email for 2 minutes, 7 minutes, or all 10 minutes). From this study, an assessment about the disruptiveness of email might be determined. Over a prolonged period, if the samples are drawn randomly, time sampling can give a good estimate of the pattern of activities.

With continuous real-time sampling, any and all behavior during the specified time of the behavior is recorded; behavior of the sampled case as well as behavior by others. For example, imagine you plan professional conferences for a living. You are interested in how attendees use breaks between conference sessions. You might record the spaces outside the conference sessions during breaks and study the various behaviors (do attendees exchange contact information, discuss on conference topics, make plans to meet for dinner or drinks; contact their office or make personal calls, work on tablets or desktops, enter rest-

room facilities; return to sessions or remain in break spaces, etc.) Any and all behaviors are captured, and you likely have anticipated some of these behaviors, so you can check off how often each occurs, even recording what men and women attendees do. In this study, we might assess the placement of re- freshments, seating capacity in break spaces, and time frame allotted for breaks.

Time-interval sampling records every de- sired behavior in real time but counts the behavior only once during the interval for each case. In essence, you are interested in whether the behavior occurred, not how often or its many variations. So using our conference planner, she might record whether any man or woman used the rest- room facilities or checked his or her phone, etc., during the break time; if just one did, for each activity, then that act would be verified as occurring during break times.

schi18939_ch07_148-169.indd 152 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

153

If we use people as sample cases, then their selection is often because of some prior behavior or characteristic. For example, an observation study might involve product use. We observe a person purchase an item in a store (e.g., box of cake mix) and recruit them for a home-use study. We arrange a time to watch them bake their cake. Do they follow directions precisely or make modifications; what tools (pans, spatulas, bowls) do they use in the preparation of cake batter; what additional products do they use to complete the cake (e.g., fruit preserves or spreads, icing, candies, sprinkles or decorations, candles); do they have difficulties or is the cake a masterpiece? From such a study, we might recommend changes in packaging directions, packaging design, or new products.

Sampling design must also answer the following: Who are the contacts to gain access to the sample cases? These are often intermediaries who help with introductions and serve as contacts to reach if conditions change or trouble develops. These intermediaries also provide access to desired records, recording devices, etc.

>Data Collection Design The data collection design specifies the details of the observation. In essence it answer the questions of what, when, where, and how.

Content Type Specific conditions, events, or activities that we want to observe determine the observational re- porting system (and correspond to measurement questions). To specify the observation content, we should include both the major variables of interest and any other variables that may affect them. From this cataloging, we then select those items we plan to observe. For each variable cho- sen, we must provide an operational definition if there is any question of concept ambiguity or special meanings. Even if the concept is a common one, we must make certain that all observers agree on the measurement terms by which to record results. For example, we may agree that vari- able W will be reported by count, while variable Y will be counted and the effectiveness of its use judged qualitatively.

Observations may be either facts (overt actions, surface indicators, spontaneous utterances) or inferences (interpretations of facts). Let’s use a salesperson’s presentation as what is observed. Maybe the researcher is trying to determine the effectiveness of the current training program. What is mea- sured may be defined as an act. Any of the following could be defined as an act for an observation study:

• A single expressed thought

• A physical movement

• A facial expression

• A motor skill

• An outcome

Although acts may be well defined, they often present difficulties for the observer. A single statement from a sales presentation may include several thoughts about product advantages, a rebuttal to an objec- tion about a feature, or some remark about a competitor. The observer is hard-pressed to sort out each thought, decide whether it represents a separate unit of observation, and then record it quickly enough to follow continued statements.

Exhibit 7-4 shows how we could separate the factual and inferential components of an observation for the salesperson’s presentation. This table is suggestive only. It does not include many other variables that might be of interest, including factual data on customer purchase history; company, industry, and general economic conditions; the order in which sales arguments are presented; and specific words used to describe certain product characteristics. The particular content of observation will also be affected by the nature of the observation setting.

schi18939_ch07_148-169.indd 153 2/14/18 3:14 PM

>part II The Design of Business Research

154

Types of Observation Researchers using observation monitor behavioral activities and nonbehavioral conditions, which, as shown in Exhibit 7-5, can be classified roughly as follows:

Nonbehavioral Observation

• Record analysis

• Physical condition analysis

• Physical process analysis

Behavioral Observation

• Nonverbal analysis

• Linguistic analysis

• Extralinguistic analysis

• Spatial analysis

Nonbehavioral Observation

A prevalent form of observation research is record analysis. This may involve historical or current records and public or private records. These may be written, printed, sound-recorded, photographed, film, or videos. Historical statistical data may be the only sources used for a study. Analysis of current financial records and economic data also provide major data sources for business studies. Other examples of this type of observation are the content analysis of competitive advertising and the analy- sis of personnel records. The major problem associated with nonbehavioral observation is the lack of consistency of data recorded over time. If a business failed to use consistent operating definitions of concepts and constructs, it makes direct comparison difficult.

Physical condition analysis is typified by store audits of merchandise availability, studies of plant safety compliance, and analysis of inventory conditions. Process analysis (or activity analysis) includes time/motion studies of manufacturing processes and analysis of traffic flows in a distribution system, paperwork flows in an office, and money (digital and currency) flows in the banking system. Both of these types of observations can use observers to record overt actions and conditions or record analysis, or a combination.

>Exhibit 7-4 Content of Observation: Fact versus Inference

Fact Inference

Introduction/identification of salesperson and customer Credibility of salesperson; qualified status of customer

Time and day of week Convenience for the customer; welcoming attitude of the customer

Product presented Customer interest in product

Selling points presented per product Customer acceptance of selling points per product

Number of customer objections raised per product Customer concerns about features and benefits

Salesperson’s rebuttal of objection Effectiveness of salesperson’s rebuttal attempts

Salesperson’s attempt to restore controls Effectiveness of salesperson’s control attempt; conse- quences for customer who prefers interaction

Length of interview Customer’s/salesperson’s degree of enthusiasm for the interview

Environmental factors interfering with the interview Level of distraction for the customer

Customer purchase decision General evaluation of sales presentation skill

schi18939_ch07_148-169.indd 154 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

155

>Exhibit 7-5 Selecting an Observation Approach

Observation Approach

Attitudes, motivations,

intentions, expectations

Conditions, behaviors,

events, processes Communication

Approach

Behavioral

Nonbehavioral

Nonverbal Analysis

Linguistic Analysis

Extralinguistic Analysis

Spatial Analysis

Record Analysis

Physical Condition Analysis

Physical Process Analysis

Personal

Mechanical/Digital

Direct versus indirect

Open versus concealed

Participant versus nonparticipant

Design the Research

Investigative Questions

Behavioral Observation

The observational study of persons can be classified into four major categories.3 Nonverbal behav- ior, the most prevalent of these, includes body movement, motor expressions, and even exchanged glances. At the level of gross body movement, one might study how a salesperson travels a territory. At a fine level, one can study the body movements of a worker assembling a product or time-sample

schi18939_ch07_148-169.indd 155 7/2/18 2:03 PM

>part II The Design of Business Research

156

One emerging technology that is gathering interest among re-

searchers and their clients is eye-tracking; more than 58 percent

of researchers in the annual GRIT study indicate they are currently

or plan to use the technology. Sticky, a leading visual optimization

platform, provides a biometric solution that helps digital content

developers measure and optimize the visual web. Participants,

often members of client panels or Sticky’s own panel called “The

Crowd,” are introduced to visual stimuli via the Internet via their

laptop, desktop, or tablet. “Sticky measures facial expressions

and eye movements in 3D, allowing for the most accurate tracking

possible, even when people move, lighting changes, or the face is

partially obstructed,” shared Hans Lee, CEO of Sticky. “When mea-

suring the impact of video and other content, the intensity of emo-

tions defines message resonance—and therefore the likelihood that

the content goes viral.” Businesses curate a lot of content on the

web. Each is intensely interested in what the eye sees and tracks

and in having their content shared. “Earning attention leads to

market success. With Sticky, a firm can measure the number of

seconds a medium actually earns—an industry first,” claimed Lee.

For example, Sticky’s online eye-tracking methodology

tested one of its own planned ads—Guessing Game—and re-

vealed that while almost all of the ad’s audience quickly saw

the key visual and the headline message, the ad delivered only

average brand recognition. Without brand recognition, an ad is

unlikely to drive prospects to the company. (See the generated

heat map in action on Connect.)

How your eye tracks through visual stimuli translates not only

to engagement with an organization’s social network sites, but

also to message recall, website click-through, and ultimately, pur-

chase. Sticky can aggregate survey data with its eye-tracking and

facial coding data to understand a participant’s level of emotional

engagement with visual stimuli. The technology—once only

available in specially equipped labs but now available no mat-

ter where a participant is located—is used in website design,

package design, ad design, email communications, and emotion

understanding.

www.sticky.ad

>snapshot Visual Content Gets Sticky

the activity of a department’s workforce to discover the share of time each worker spends in various ways. More abstractly, one can study body movement as an indicator of interest or boredom, anger or pleasure in a certain environment. Motor expressions such as facial movements can be observed as a sign of emotional state. Eyeblink rates are studied as indicators of interest in advertising mes- sages. Exchanged glances are of interest in studies of interpersonal behavior.

Courtesy of Sticky by Tobii Pro

schi18939_ch07_148-169.indd 156 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

157

Linguistic behavior is a second frequently used form of behavior observation. One simple type familiar to most students is the tally of “ahs” or other annoying sounds or words a professor makes or uses during a class. More serious applications are the study of an executive’s presentation content when addressing financial analysts or the study of what, how, and how much information is conveyed in a training situ- ation. A third form of linguistic behavior involves interaction processes that occur between two people or in small groups. In some observations, spontaneous utterances are as important as the overt acts of behavior measured.4

Verbal behavior also may be analyzed as extralinguistic behaviors. One author suggests there are four dimensions of extralinguistic activity:5 (1) vocal, including pitch, loudness, and timbre; (2) temporal, including the rate of speaking, duration of utterance, and rhythm; (3) interaction, including the tenden- cies to interrupt, dominate, or inhibit; and (4) verbal stylistic, including vocabulary and pronunciation peculiarities, dialect, and characteristic expressions. These dimensions could add substantial insight to the linguistic content of the interactions between supervisors and subordinates or salespeople and customers.

A fourth type of behavior study involves spatial relation- ships, especially how a person relates physically to others. One form of this study, proxemics, concerns how people or- ganize the territory about them and how they maintain dis- crete distances between themselves and others. A study of how salespeople physically approach customers and a study of the effects of crowding in a workplace are examples of this type of observation.

Often in a study, the researcher will be interested in two or more of these types of information. Such studies may re- quire more than one observer. In these forms of behavior study, it is also important to consider the relationship be- tween observers and participants. If you observed this behavior, what would you record?

© im

tm ph

ot o/

S h

ut te

rs to

ck

Courtesy of Sticky by Tobii Pro

>snapshotcont’d

schi18939_ch07_148-169.indd 157 2/14/18 3:14 PM

>part II The Design of Business Research

158

Observation Environment Observation studies can be designed for the field or laboratory. In business research, field studies may take place at a customer’s home, the shopping environment, an employee work area (plant, office, distri- bution center), a supplier’s location, and more. Field studies, offering a natural setting, are most likely to obtain unaltered behavior, especially when the observer isn’t directly involved.

Laboratory studies are most likely to provide data protection. When specialized equipment is needed for observation (e.g., eye-tracking cameras, heart rate monitors, galvanic skin response machines, etc) laboratory settings are often the choice. We’ve had some success with employing eye-tracking via a subject’s laptop and tablet cameras. Laboratory settings are obviously more expensive, usually involve smaller sample sizes, and pose more difficulties in recruiting subjects. Laboratory observations can be part of an experimental design.

If you read or watch the news, you’ll know that urban areas have

had a sharp increase in questioned—if not questionable—police

actions with regard to excessive use of force, and bystander vid-

eos have played an increasing role in judging police actions. In

2015, the U.S. Supreme Court, in Graham v. Connor, held that an

officer’s actions, however, “must be judged from the perspective of

a reasonable officer, rather than with the 20/20 vision of hindsight.”

In an article in The Atlantic, Seth W. Stoughton, a law professor and

former police officer, and Geoffrey Alpert, a professor of criminol-

ogy, both at the University of South Carolina; along with Jeff Noble,

a police consultant based in Orange County, California, write, “The

aversion to what officers derisively refer to as “ ‘second-guessing’

. . . [makes] officers less receptive to a critique of their actions . .

. [and] makes them reluctant to provide their own complete and

honest critiques.” Yet nationwide, we’ve seen a demand for police

to change and for police decisions to be more transparent, result-

ing in a clamoring for use of police body and cruiser cameras.

Do you believe you get a true picture of an incident when you

see a body-mounted or dash-mounted video? Stoughton, who

also consults with law enforcement agencies, has choreographed

a series of videos to demonstrate the answer to this question. His

use parallels an observation study based on respondents watch-

ing video footage of mock police incidents. Using a series of

chest-mounted or dash-mounted cameras and bystander videos,

he shows just how difficult it is to arrive at an accurate conclusion

when using only a police-cruiser or body-cam video.

Chest-mounted cameras during interactions or pursuits often

create jerky movements and wildly distorted images. Stoughton

calls this “deceptive intensity”; it creates the impression that the

officer is under attack when he might not be. In an interception

incident video, using a dash-mounted camera involving a fleeing

suspect and Taser use by an officer, accuracy is related to van-

tage point. The body camera doesn’t reveal the use of a Taser

or the absence of a gun, while video shot by a bystander does.

“When video allows us to look through someone’s eyes, we

tend to adopt an interpretation that favors that person,” explains

Stoughton, a psychological phenomenon known as “camera

perspective bias.” Stoughton’s research also reveals that the

degree to which the viewer trusts or distrusts the police influ-

ences his or her video interpretation. So while the bystander

might not know any facts leading up to the incident, his or her

camera has its own bias. He concludes video evidence inter-

pretation depends, therefore, on perspective as well as bias.

So in observation research, should we consider the camera

as an accurate, unbiased observer? See for yourself. Check out

the videos on Connect.

>snapshot Observation and Police Cameras

©Aaron Roeth Photography

schi18939_ch07_148-169.indd 158 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

159

Data Collection Protocol Observation is found in almost all research studies, at least at the exploratory stage. Such data collection is known as simple observation. Its data collection protocols (procedures used to record observations) are not standardized, as one would expect, because of the discovery nature of exploratory research. The decision to use observation as the major data collection method may be made as early as the moment the researcher moves from research questions to investigative questions—the specific questions the researcher must answer with collected data. If the study is to be something other than exploratory, systematic observation employs standardized procedures, trained observers, schedules for recording, and other devices for the observer that mirror the scientific procedures of other primary data methods. Systematic studies vary in the emphasis placed on recording and encoding observational information:

At one end of the continuum are methods that are unstructured and open-ended. The observer tries to provide as complete and nonselective a description as possible. On the other end of the continuum are more structured and predefined methods that itemize, count, and categorize behavior. Here the investigator decides beforehand which be- havior will be recorded and how frequently observations will be made. The investigator using structured observation is much more discriminating in choosing which behavior will be recorded and precisely how [it is] to be coded.6

One author classifies observation studies by the degree of structure in the environmental setting and the amount of structure imposed on the environment by the researcher,7 as reflected in Exhibit 7-6. The researcher conducting a class 1, completely unstructured study would be in a natural or field setting endeavoring to adapt to the culture. A typical example would be an ethnographic study in which the researcher, as a participant-observer, becomes a part of the culture and describes in great detail every- thing surrounding the event or activity of interest. One researcher took a punch press job in a factory to describe the rituals that a small work group relied on to make their highly repetitive, monotonous work bearable.8 With other purposes in mind, business researchers may use this type of study for hypothesis generation.

Class 4 studies—completely structured research—are at the opposite end of the continuum from completely unstructured field investigations. The research purpose of class 4 studies is to test hypoth- eses; therefore, a definitive plan for observing specific, operationalized behavior is known in advance. This requires a measuring instrument, called an observation checklist; think of this as similar to the interview discussion guide or the questionnaire but for observation studies. These are developed in much the same way—working from the measurement questions to determine what should be ob- served. For human observations—those following behaviors and processes—the checklist may contain maps of interior spaces and diagrams to show where behaviors, or deviations from current processes, occurred. Exhibit 7-7 shows the parallels between survey design and checklist development.

Checklists should possess a high degree of precision in defining relevant behavior or acts and have mutually exclusive and exhaustive categories. The coding is frequently predetermined, thereby simplifying data recording and analysis. The participant groups being observed must be comparable and the labora- tory conditions identical. The classic example of a class 4 study was Bales’s investigation into group inter- action.9 Many team-building, decision-making, and assessment center studies follow this structural pattern.

The two middle classes of observation studies emphasize the best characteristics of either researcher- imposed controls or the natural setting. In class 2, the researcher uses the facilities of a laboratory— videotape recording, two-way mirrors, props, and stage sets—to introduce more control into the environment while simultaneously reducing the time needed for observation. In contrast, a class 3 study takes advantage of a structured observation checklist in a natural setting.

>Exhibit 7-6 Classification of Observation Studies

Class Purpose Structure Environment Research Tool

1 Generate hypothesis Completely unstructured Natural setting

2 Unstructured Laboratory

3 Structured Natural setting Observation checklist

4 Test hypothesis Completely structured Laboratory Observation checklist

schi18939_ch07_148-169.indd 159 2/14/18 3:14 PM

>part II The Design of Business Research

160

Data collection protocols must answer questions of when observations are to be made. Remember, observers collect with all their senses, so these decisions reflect not just what can be seen, but what can be heard, touched, smelled, etc. In terms of when, is the time of the observation important? Assume we are studying the out-of-stock conditions in a supermarket. The exact times of observation may be impor- tant. Inventory is shipped to the store on certain days only and at certain times, and buying peaks occur on various days and times. The likelihood of a given product being out of stock is a function of both time-related activities. Or assume we are observing worker break habits and patterns to determine how a space set aside for this purpose is redesigned. Would use patterns differ in the middle of shift verses closer to end of shift?

Protocols must also answer questions of how. If there are two or more observers, how will they divide the task? How will the results be recorded for later analysis? How will the observers deal with various situations that may occur—when expected actions do not take place or when someone challenges the observer in the setting?

Protocols must also answer questions of where. Within a spatial confine, where does the act take place? In a retail traffic pattern study, the proximity of a customer’s pause space just inside the door be- fore choosing a direction of movement to the entry door, a display, or directional sign might be recorded. Must the observation take place from a particular location within a larger venue? The location of the ob- server plus the location of the act observed (e.g., a sales approach observation within a store) can signifi- cantly influence the acts recorded. Observers face unlimited variations in conditions and must respond to these during the time frame of a study. When the plans are thorough and the observers well trained, observation research can be responsive. Fortunately, problems that do occur rarely occur simultaneously.

>Exhibit 7-7 Flowchart for Observation Checklist Creation

Measurement Questions

Participant ID

Observer ID

Data for Observation Classification

Administrative Notations

Checklist Design Process

Observation Group B

Observation Group A

Observation Group C

Observation location

Observation timing

Observation conditions

Revise

Target Observations

Pretest Observation

Protocol

Collect and Prepare Data

Pretest Checklist & Protocols

Prepare Observation Checklist

Verify Observation Protocol

schi18939_ch07_148-169.indd 160 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

161

Observer Presence Human observation studies present a clear opportunity for observer bias. Careful consideration must be made in determining the relationship between observer and participant. Consider these three choices:

• Whether the observation is direct or indirect.

• Whether the observer’s presence is known or unknown to the participant.

• What role the observer plays.

Direct observation occurs when the observer is physically present and personally monitors and re- cords what takes place. This approach is very flexible because it allows the observer to react to and report subtle aspects of events and behaviors as they occur. He or she is also free to shift places, change the focus of the observation, or concentrate on unexpected events if they occur. A weakness of this approach is that observers’ perception circuits overload as events move quickly, and observers must later try to reconstruct what they were not able to record.

Indirect observation occurs when a recording is made by mechanical, photographic, or elec- tronic means. For example, a special camera that takes one frame every second may be mounted in a department to study employee movement. The observer then extracts data from the images. Indirect observation is less f lexible than direct observation but is also much less biasing and may be less erratic in accuracy. Another advantage of indirect observation is that the permanent record can be reanalyzed to include many different aspects of an event. Electronic recording devices, which have improved in quality and declined in cost, are used frequently in observation research.

Observer Concealment Should the participant know of the observer’s presence? Observers use concealment to shield them- selves from the object of their observation. Often, technical means such as one-way mirrors, hidden cameras, or microphones are used. When the observer is known, there is a risk of atypical activity by the participant (reactivity response). The initial entry of an observer into a situation often upsets the activity pat- terns of the participants, but this influence usually dissipates quickly, especially when participants are engaged in some absorb- ing activity or the presence of observers of- fers no potential threat to the participants’ self-interest. Hidden observation is a form of spying, however, and the propriety of this ac- tion must be reviewed carefully. The potential bias from participant awareness of observers is always a matter of concern. It also raises ethical issues. Would participants have given their consent to be observed? After being de- ceived and having their privacy invaded, what further damage could come to the partici- pants if the results became public? This issue needs to be addressed when concealment and covert participation are used.

Unobtrusive Measures

To minimize reactivity response, some researchers suggest using innovative observational procedures that can be both nonreactive and inconspicuously applied. Called unobtrusive measures, these approaches encourage creative and imaginative forms of indirect observation, archival searches, and variations on

© D

re am

P ic

tu re

s/ S

h an

n on

F au

lk /G

et ty

I m

ag es

schi18939_ch07_148-169.indd 161 2/14/18 3:14 PM

>part II The Design of Business Research

162

simple and contrived observation.10 Of particular interest are measures involving indirect observation based on physical traces that include erosion (measures of wear) and accretion (measures of deposit).

Natural erosion measures are illustrated by using the frequency of replacement of vinyl floor tile in front of certain museum exhibits as an indicator of exhibit popularity. The study of wear and tear on book pages could be used is a measure of library book use. Counting the remaining brochures in a car dealer’s display rack after a favorable magazine review could be used to suggest consumer interest. Physi- cal traces also include natural accretion such as discovering the listenership of radio stations by observ- ing car radio settings as autos are brought in for service. Another type of unobtrusive study involves estimating liquor and magazine consumption by collecting and analyzing family trash. One interesting study compared beer consumption reports acquired through interviews with the findings of sampled trash. The validity of interview data was questioned when the beer can count from trash supported a 77 percent consumption rate, not the 15 percent reported.11

By excavating landfills, one researcher who gained insight into human behavior and cultural patterns noted “people will describe their behavior to satisfy cultural expectations, like the mothers in Tucson who unanimously claimed they made their baby food from scratch, but whose garbage told a very different tale.”12

Physical trace methods present a strong argument for use based on their ability to provide low-cost access to frequency, attendance, and incidence data without contamination from other methods or reac- tivity from participants. They are excellent “triangulation” devices for cross-validation. Thus, they work well as supplements to other methods. Designing an unobtrusive study can test a researcher’s creativity, and one must be especially careful about inferences made from the findings. Erosion results may have occurred because of wear factors not considered, and accretion material may be the result of selective deposit or survival.

Partial Concealment

A modified approach involves partial concealment. The presence of the observer is not concealed, but the objectives and participant of interest are concealed. For example, a study of selling methods may be conducted by sending an observer with a salesperson who is making calls on customers. However, the observer’s real purpose may be hidden from both the salesperson and the customer (e.g., the observer may pretend she is analyzing the display and layout characteristics of the company’s merchandise in the stores they are visiting).

Observer Involvement Should the observer enter the observation setting and act as both an observer and a participant— called participant observation? We have several situations where observers might be involved in the observation act: (1) The behavior we want to measure occurs infrequently, so we create a scenario to simulate the behavior within our research time frame; (2) the behavior we want to measure has many variations, and we need to measure each variation; and (3) the recruited observer is part of the process or behavior we need to measure, so excluding them is not practical. Sometimes he or she is known as an observer to some or all of the participants; at other times the observer’s true role is concealed. In-home- and in-work-environment ethnography studies (immersion by the observer into the culture of the participant) are used in business; the participant is aware of the observation and has fully consented to the protocols that are used. Some examples are product use studies and work process studies.

Participant observation has two major problems: The observer’s role may inf luence the way others act, and it places a dual role on the observer (recording can interfere with participation and participation can interfere with observation and recording). These problems are the major reason that business research avoids participant observation, when possible. Training can ad- dress both these problems, but also causes participant observation to be more costly or less reliable.

schi18939_ch07_148-169.indd 162 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

163

Observer Training and Supervision The data collector/observer is a critical aspect of observation research. We want the best observers pos- sible. What we look for in observers are:

• Experience: Ability to extract the most from an observation.

• Concentration: Ability to function in a setting full of distractions.

• Detail-oriented: Ability to remember details of an experience.

• Unobtrusive presence: Ability to blend with the setting and not be distinctive.

The observers should be thoroughly versed in the requirements of the specific study. Each observer should be informed of the outcomes sought and the precise content elements to be studied. Observer trials with the instrument and sample videotapes should be used until a high degree of reliability is appar- ent in their observations. If observation is at the surface level and involves a simple checklist or coding system, then experience is less important. Inexperience may even be an advantage if there is a risk that experienced observers may have preset convictions about the topic. Only intensive videotaped training relieves these problems.

When a study calls for multiple observers, there may be interpretative differences between observers. These must be reconciled. Training can address these differences. Another approach to dealing with problems related to variability between observers is to turn the problem into an opportunity—use teams. IDEO used a team comprised of a human factors expert, an engineer, and a designer to go into office buildings when it was working on an empathic design project for office-equipment supplier Steelcase. Each used their inherent differences in training and educa- tion to observe the same workers in their respective environments, but each brought back differ- ent data.13

Observers may be subject to halo effects (cognitive bias that influences the observer’s impression of the person, object, event or act he or she is observing),14 observer drift (gradual alterations over time in observation documentation),15 and fatigue. If a halo effect bias is at work, an obviously attractive subject might get more positive inferential observations. In such an instance, an observer would use the global evaluation of beauty to make positive judgments on other traits, like friendliness, socially adeptness, efficiency, ability to maintain strong relationships, etc. Halos are like any other heuristic; they shorten our time to decision, but they may also be false, leading us to the wrong conclusion. Ob- servers can have halos relating to gender, age, ethnicity, job performance, etc. Observer drift causes decay in reliability or validity of observations over time that affects the coding of categories; this is a serious problem if not caught early.16 Both halo effects and observer drift can be addressed in training and by having supervisors verify that observation protocols are being strictly followed. Only multiple observers, rest periods during observation, and shortening the observation study time frame can ad- dress fatigue.

>Evaluation of the Observation Method Strengths Observation is the only method available to gather certain types of information. The study of records, mechanical processes, and young children, as well as other inarticulate participants, falls into this cat- egory. Another value of observation is that we can collect the original data at the time they occur. We need not depend on reports by others. Researchers can avoid memory decay (loss of detail about an event or prior behavior caused by passage of time or interfering events, such as learning)17 and selective filtering or selective attention (due to sheer volume of information collected through our senses, we pay attention to only a small proportion of information we believe is critical).18 Observation overcomes many of these deficiencies of questioning.

schi18939_ch07_148-169.indd 163 2/14/18 3:14 PM

>part II The Design of Business Research

164

The design of a behavioral observation study follows the same

pattern as other research. Once the researcher has specified

the investigative questions, it is often apparent that the best way

to conduct the study is through observation. Guidance for con-

ducting a behavioral observation and translating the investiga-

tive question(s) into an observation checklist is the subject of this

Closeup. We first review the procedural steps and then explain

how to create a checklist.

Most studies that use behavioral observation follow a gen-

eral sequence of steps that parallel the research process. (See

Exhibit 7-2.) Here we adapt those steps to the terminology of the

observation method:

● Define the content of the study. ● Develop a data collection plan that identifies the

observational cases, sampling strategy, and acts

(operationalized as a checklist or coding scheme). ● Secure and train observers. ● Collect the data. ● Analyze the data.

Assume you work for MarketViews, a specialty research

firm hired to conduct a study for HomeExtravaganza, which has

been experiencing declining repeat visits at its new superstore

units. Preliminary results from a periodic survey indicate that

customer confusion in the megastore is discouraging customers

from returning, so an observation study is planned to see how

employees and store design elements contribute to or solve the

problem. The research questions might be:

● What do employees do to reduce or eliminate customer

confusion? ● What do employees do that contributes to customer

confusion? ● Which design elements diffuse customer confusion? ● Which design elements contribute to customer

confusion?

Further assume that the survey indicates that customers

who feel confused and cited this confusion as their reason for

not returning had entered the store looking for a variety of mer-

chandise, stocked in various locations throughout the vast store.

They described their hopelessness as evolving from the experi-

ence of not knowing where to start the process of searching.

Such customers showed no particular similarities or differences

in terms of ethnic background, age, or education level. Some

had the assistance of greeters or floor assistance associates,

while others did not.

The observation cases will be twofold: shoppers entering

the store through the main entrance and employees serving

as greeters or floor assistance associates. Customers who re-

quest assistance from the main-entrance greeter or who consult

the directional location signs will be the primary target. If they

approach a floor assistance associate, the employee will also

become a case.

Survey research reveals some inconsistency by time of

day and type of merchandise sought, so MarketViews plans to

select cases during four primary day parts—early morning, midday,

afternoon, and early evening—as well as at all three “directional

locations,” where signage describing the store and the shop-

per’s current “You are here” location is noted.

A preliminary tour of the store helps identify the acts to record.

During a subsequent meeting, it is decided that MarketViews will

record the customer’s seeking of assistance, either personal or

signage; his or her consulting of the directional signage and its

location; the customer’s path to the desired merchandise; and

whether a purchase is completed. The assistance acts of the floor

assistance associate will also be recorded. These are determined

to be assistance versus no assistance, providing the customer

with copy of a directional map on which the associate marks the

location of the merchandise and the path to get there, inquiry

of other staff for location assistance, passing the customer to

another floor assistance associate, accompanying the customer to

the correct aisle, accompanying the customer to actual merchan-

dise shelf location, and providing verbal assistance to selection

of the appropriate product from the array of products provided.

It is determined that MarketViews will create and test an

observation checklist, making any necessary changes after

the test. The checklist developed is shown in Exhibit CU-1. The

foremost concern is that either the customer or the associate

will discover that he or she is being observed and will change

behavior. Human observers will be used to trace the path of

observational targets. By means of the store’s security cam-

eras, researchers will record customers flowing through the

main entrance and past the greeter location and stopping at

directional location signs. Counts of brochure store maps dis-

tributed from the directional signage locations and copysheet

directional maps used by floor assistance associates will

also be used as a measure of customers seeking directional

assistance.

Designing the Observation Study

>closeup

schi18939_ch07_148-169.indd 164 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

165

>closeupcont’d

Day: M T W Th F Sa SuTime

Target Shopper Characteristics: � Male � Female � Child � Child-teen+ � Adult � Senior

Target Customer Interception Location: � Main Entry � Directional Location Sign: � #1 � #2 � #3

Shopping Cart Used: � No � Yes

Greeter verbal interaction with target: � No � Yes Greeter No. Action � Point to directional sign � Verbal directions

Floor Assistance Associate Interaction: � No � Yes Interception location: Aisle # Crossway # Associate # Assistance given: � No � Yes Action: � Verbal direction only � Verbal direction plus pointing � Store directional copy-map with marked mdse location � Store directional copy-map with mdse location & path � Inquire of other sta� � Pass customer to another FAA � Accompany customer to aisle location � Accompany customer to mdse shelf location � Product selection assistance o�ered

Directional Sign Interaction: � No � Yes Sign location: � #1 � #2 � #3

Purchase: � No � Yes: Item Sought Assistance for: � No � Yes

Customer Path:

Date Store No. Observer #

Shopper Companion(s): � Alone � With others: � Other adult No. M No. F � Child/children No.

1

2

3 G

>Exhibit CU-1 Sample Checklist for HomeExtravaganza Study

schi18939_ch07_148-169.indd 165 2/14/18 3:14 PM

>part II The Design of Business Research

166

A third strength is that we can secure information that most participants would ignore either because it is so common and expected or because it is not seen as relevant. For example, if you are researching buying activity in a store, there may be conditions important to the research study that the shopper does not notice or consider important, such as: What is the weather? What is the day of the week or the time of the day? How heavy is customer traffic? What is the level of promotional activity in competing stores? We can expect to learn only a few of the answers to these questions from most participants in a communication study.

The fourth advantage of observation is that it alone can capture the whole event as it occurs in its natural environment. Whereas the environment of an experiment may seem contrived to partici- pants, and the number and types of questions limit the range of responses gathered from respon- dents in a survey or interview, observation is less restrictive than most primary collection methods. Also, the limitations on the length of data collection activities imposed by surveys or experiments are relaxed for observation. You may be interested in all the conditions surrounding a confrontation at a bargaining session between union and management representatives. These sessions may extend over time, and any effort to study the unfolding of the negotiation is facilitated by observation. Communication studies could seldom provide the insight of observation for many aspects of the negotiation process.

Finally, participants seem to accept an observational intrusion better than they respond to commu- nication study. Observation is less demanding of participants and normally has a less biasing effect on their behavior than does communication. In addition, it is also possible to conduct disguised and unob- trusive observation studies much more easily than disguised communication studies.

Limitations The observation method has some research limitations, yet any consideration of the merits of observa- tion confirms its value when used with care and understanding.

Observation is a slow and expensive process that requires either human observers or costly surveil- lance equipment. The data collector may be needed at the scene of the event when it takes place, yet it is often impossible to predict where and when the event will occur. One way to guard against missing an event is to observe for prolonged periods until it does occur; in such situations, observer fatigue, bore- dom, and distracting events can reduce the accuracy and completeness of observation.

Reliable observations are restricted to information that can be learned by overt action or surface indicators. To go below the surface, the observer must make inferences. Two observers will probably agree on the nature of various surface events, but the inferences they draw from such data are much more variable.

Observation is more suited to subjective assessment and recording of data than to controls and quantification of events. When control is exercised through active intervention by the researchers, their participation may threaten the validity of what is being assessed. Even when the sampled number of cases is small, the observation records can be disproportionately large and difficult to analyze.

Observation is limited as a way to learn about the past. It is similarly limited as a method by which to learn what is going on in the present at some distant place.

Finally, it is difficult to gather information on intentions, attitudes, opinions, or preferences with observation.

>Ethical Issues and Their Solutions When we conduct a communication-based study, we obtain consent from the participant. But when we conduct an observation using human cases, we conceal our presence and their participation. Yes, we can let them know they were a subject after the fact, but they might still feel violated. While it is not illegal to observe people in public places or in their work environment, their presence in such places does not remove their right for some privacy.

schi18939_ch07_148-169.indd 166 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

167

The right of quality research is not inherent in observation research, but it must be protected by develop- ing an appropriate observation checklist and extensive observer training. The observer has a much more involved role in not only data collection, but also in data interpretation and developing insights. Pretesting the checklist and observation protocols, as well as supervising observers in their execution of observation protocols and debriefing observers following data collection, are valuable steps in assuring quality research.

Observation research frequently uses nonprobability sampling. In such instances, assuring quality also means not projecting small-sample, nonrespresentative results to a target population and drawing conclusions that can’t be supported by the data collected. Using such research to form hypotheses rather than test hypoth- eses is appropriate. When a census or probability sample is used, observation can be used to test hypotheses.

LO7-1 Observation is one of the few options available for studying records, mechanical processes, lower animals, small children, and complex interactive pro- cesses. We can gather data as the event occurs and can come closer to capturing the whole event than with communication. On the other hand, we have to be present to catch the event or have some recording device on the scene to do the job.

LO7-2 Sampling design in observation includes choosing the target population (people, acts or events, behaviors, etc.) and the appropriate approach (probability vs. nonproba- bility). Two sampling methods are unique to observation: event and time sampling. Event sampling measures all factors involved with a single specified event, while time sampling chooses a time interval and records desired variables within that time frame. There are three types of time sampling (time point, continuous time, and time in- terval). In data collection design, we must address seven different factors: content type (facts or acts, inferences of those acts, or both), type of observation (nonbehavioral or behavioral), the environment (field or laboratory), data collection protocol (simple observation or systematic ob- servation), observer presence (direct or indirect observa- tion), observation concealment (full, partial, or none), and observer involvement (participant observation or none). The measurement instrument for observation research is the observation checklist.

LO7-3 Observation includes a variety of monitoring situations that cover nonbehavioral and behavioral activities. Both types may be conducted with human or digital/ mechanical observers, but all require human coding and interpretation. Nonbehavioral observation includes record analysis, physical condition analysis, and physical process analysis. The most common nonbehavioral ob- servation is record analysis. Lack of consistency in creat- ing records can limit the effectiveness of record analysis. Behavioral observation includes nonverbal analysis, lin- guistic and extralinguistic analysis, and spatial analysis.

LO7-4 The strengths of observation as a data collection method include securing information about people or events that cannot be derived from communication

approaches, avoiding memory decay and selective filtering by securing information as it occurs, securing information participants might ignore or not think to report, securing environmental context information, securing data in a less intrusive way, and being more flexible about data collection time periods.

LO7-5 Observation may be limited by any of the following: the difficulty of predicting when specific desired events or acts might occur, causing long periods of observa- tion; slow nature of data collection; possible reactivity response if observation is known by participants; small sample size; high costs, especially when specialized equipment or facilities are used; recording demands of high-speed events; questionable reliability of infer- ence for capturing attitudes, motivations, and cognitive processes; coding for disproportionately large records; reconciling variability between observers on facts and inferences from the same act or record; observer halo effects, drift, and fatigue; excessive responsibilities caused by the dual role required during participant observations; and lack of express consent and result- ing violation of participant’s right to privacy. Unobtrusive measures offer ways to reduce reactivity response in observation research, either to confirm the findings from other methods or operate as singular data sources. Most limitations may be addressed by digitally recording acts of interest, providing observer training, pretesting obser- vation checklists and observation protocols, and verify- ing observation protocols during actual data collection.

LO7-6 The ethical issues are focused on the participant’s right to privacy and the manager’s right to quality research. The first might be addressed by debriefing participants after the research if observation was concealed. The right of quality is not inherent in observation research unless the researcher takes special care with the development of the observation checklist, develops checklist-specific training, pretests both the checklist and the observation protocols, and verifies those protocols during the obser- vation research. Observation can be used to generate and test hypotheses, so knowing which is appropriate for a given study also protects research quality.

>summary

schi18939_ch07_148-169.indd 167 2/14/18 3:14 PM

>part II The Design of Business Research

168

concealment 161

direct observation 161

event sampling 152

extralinguistic behavior 157

halo effect 163

indirect observation 161

linguistic behavior 157

memory decay 163

nonverbal behavior 155

observation 149

observation checklist 159

observer drift 163

participant observation 162

physical condition analysis 155

physical traces 161

process (activity) analysis 155

reactivity response 161

record analysis 154

selective filtering 163

simple observation 159

spatial relationships 157

systematic observation 159

time sampling 152

unobtrusive measures 161

>keyterms

Terms in Review 1 Compare the advantages and disadvantages of the obser-

vation to the communication approach. Under what circum- stances could you make a case for using observation?

2 What ethical risks are involved in observation? In the use of unobtrusive measures?

3 Based on present or past work experience, suggest man- agement problems that could be resolved by using obser- vation-based data.

4 Distinguish between the following:

a Nonverbal, linguistic, and extralinguistic analysis.

b Factual and inferential observation.

c Memory decay and selective filtering.

Making Research Decisions 5 The observer–participant relationship is an important con-

sideration in the design of observation studies. What kind of relationship would you recommend in each of the following cases?

a Observations of professional conduct in the classroom by the student author of a course evaluation guide.

b Observation of retail shoppers by a researcher who is in- terested in determining customer purchase time by type of goods purchased.

c Observation of a focus group interview by a client.

d Effectiveness of individual farmworker organizers in their efforts to organize employees of grape growers.

6 Assume you are the manufacturer of modular office systems and furniture as well as office organization elements (desk- top and wall organizers, filing systems, etc.). Your company

has been asked to propose an observation study to exam- ine the use of office space by white-collar and managerial workers for a large insurance company. This study will be part of a project to improve office efficiency and paperwork flow. It is expected to involve the redesign of office space and the purchase of new office furniture and organization elements.

a What are the varieties of information that might be observed?

b Select a limited number of content areas for study, and operationally define the observation acts that should be measured.

c Develop a checklist to be used by observers in the previ- ous study. Determine how many observers you need, and assign two or three to a specific observation task.

d Compare the results of your group members’ checklists for stability of recorded perceptions.

7 Amazon’s Kindle (its electronic book, magazine, and news- paper reader) has been far more successful than its SONY competitor. This could be at least somewhat influenced by the “Oprah factor”; the Kindle was first introduced on the Oprah show and strongly endorsed by the mega-mogul. But the original Kindle had some physical problems—button placement and page-turn speed among them. Correcting these problems resulted in the Kindle2, released in 2009 to glowing reviews, and ultimately to the Kindle Fire. What observation research might have been used in the redesign of the original Kindle?

8 You wish to analyze the pedestrian traffic that passes a given store in a major shopping center. You are interested in determining how many shoppers pass by this store, and you

>discussionquestions

schi18939_ch07_148-169.indd 168 2/14/18 3:14 PM

>chapter 7 Stage 2: Data Collection Design: Observation Research

169

would like to classify these shoppers on various relevant di- mensions. Any information you secure should be obtainable from observation alone.

a What other information might you find useful to observe?

b How would you decide what information to collect?

c Devise the operational definitions you would need.

d What would you say in your instructions to the observers you plan to use?

e How might you sample this shopper traffic?

9 Walmart got a very mixed reaction when it announced that after 30 years it was eliminating greeters from its overnight shift in more than 3,800 stores, redefining the role of day- shift greeters, and moving them further into the store and away from the door. What observation research would you have conducted to reach these decisions?

From Concept to Practice 10 Using Exhibit 7-5, identify the type of study described in

each of the Snapshots featured in this chapter.

From the Headlines 11 In May 2017, LAX, the world’s fourth-largest airport, relo-

cated 15 airlines in a three-night (five hours per night) pe- riod in order to accommodate Delta’s expansion. During the period, Delta operated more than 200 flights a day out of four terminals. In Terminal 2 and 3, everything moved. It relocated ticket counters, kiosks, computer sys- tems, and more than 1,000 signs. The process was in the planning stages for 15 months. Delta picked up the tab for the move for all airlines involved. If you were Delta, what observation research would you have conducted during this massive endeavor to judge the effectiveness of the move planning?

Akron Children’s Hospital

Envirosell

Net Conversions Influence Kelley Blue Book

State Farm: Dangerous Intersections

>cases*

* You will find a description of each case in the Case Abstracts section of this textbook. Check the Case Index to determine whether a case provides data, the research instrument, video, or other supplementary material. Cases and case supplements are available in Connect.

schi18939_ch07_148-169.indd 169 2/14/18 3:14 PM

LO8-1      How an experiment is different than other primary data collection methods.

LO8-2   The seven steps of a well-planned experiment.

LO8-3   Internal and external validity with experimental research designs.

LO8-4   The three types of experimental designs and the variations of each.

LO8-5   The advantages and disadvantages of the experimental method.

LO8-6  Ethical issues in experiments and their solutions.

>learningobjectives After reading this chapter, you should understand . . .

Stage 2: Data Collection Design: Experiments

>chapter 8

“It’s not an experiment

if you know it’s going to work.”

Jeff Bezos, CEO,

Amazon

schi18939_ch08_170-193.indd 170 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

171

> Experiments and Causation Why do events occur under some conditions and not under others? Experiments are research studies designed to explain why something happens; in these studies the researcher proposes a causal hypoth- esis (a speculation that one variable causes a change in another variable). The researcher then ma- nipulates the independent variable (IV) and observes whether the dependent variable (DV) is affected as hypothesized.

Causation Evidence The essential element of causation is that A “produces” B or A “forces” B to occur. Thus, there is at least one independent variable (IV) and one dependent variable (DV) in a causal relationship. Empirically, we can never demonstrate with any certainty an A-B causality. Causation needs evidence to support an inductive conclusion; we cannot demonstrate a causal relationship deductively. As such, in a causal study we estimate the probability that A “produces” B based on what we observe and measure.

In testing a causal hypothesis, we seek three types of evidence:

1. Covariation between A and B • Do we find A and B occur together in the way hypothesized? • When A does not occur, is there also an absence of B? • When there is more or less of A, does one also find more or less of B?

2. Time order of events moving in the hypothesized direction • Does A occur before B?

3. No other possible causes of B • Can one determine that C, D, and E do not covary with B in a way that suggests possible causal

connections?

In addition to these three conditions, successful inference-making in causal designs must meet two other requirements. All factors, with the exception of the independent variable, must be held constant and not confounded with another variable that is not part of the study (control). Each case studied must have an equal chance for exposure to each level of the independent variable in the study (random assignment of cases to groups).

Causal Conclusion The Mill’s method of agreement,1 illustrated in Exhibit 8-1, helps rule out some variables as irrelevant. If we can find Z and only Z in every case where we find C, and no others (A, B, D, or E) are consistently found with Z, then we can conclude that C and Z are causally related. In Exhibit 8-1, A, B, D, and E are unlikely to be causes of Z. However, there is an implicit assumption that there are no variables to consider other than A, B, C, D, and E. One can never accept this supposition with certainty because the

>Exhibit 8-1 Mill’s Method of Agreement

No. 1

Descriptive Factors Outcome

No. 2

Therefore

A B C Z

C D E Z

C Z

schi18939_ch08_170-193.indd 171 2/14/18 3:24 PM

>part II The Design of Business Research

172

number of potential variables is infinite. In addition, while C may be the cause, it may instead function only in the presence of some other variable we haven’t studied.

The negative canon of agreement states that where the absence of C is associated with the absence of Z there is evidence of a causal relationship between C and Z. Together with the method of agreement, this forms the basis for the method of difference (see Exhibit 8-2): “If there are two or more cases, and in one of them observation Z can be made, while in the other it cannot; and if variable C occurs when observa- tion Z is made, and does not occur when observation Z is not made; then it can be asserted that there is a causal relationship between C and Z.”2

There are three major types of relationships that can occur between two variables:3

• Symmetrical

• Reciprocal

• Asymmetrical

A symmetrical relationship is one in which two variables fluctuate together, but we assume the changes in neither variable are due to changes in the other. Symmetrical conditions are most often found when two variables are alternate indicators of another cause or independent variable. Assume a business spon- sors a camping club for its workers and occasionally this club sponsors weekend camping events. The business experiences high absenteeism on Mondays following a camping club event among those who are members. If the relationship is symmetrical, a correlation between low work attendance and active participation in a company camping club would be dependent on another factor. Many research studies reveal symmetrical relationships.

A reciprocal relationship exists when two variables mutually influence or reinforce each other. This could occur if the reading of an advertisement leads to the use of a brand of product. The usage, in turn, sensitizes the person to notice and read more of the advertising of that particular brand. Many research studies reveal reciprocal relationships.

Research analysts often look for asymmetrical relationships. With these we hypothesize that changes in one variable (the independent variable—IV) are responsible for changes in another variable (the de- pendent variable—DV). For example, we hypothesize that participation in the company-sponsored camp- ing club event (IV) is responsible for high absenteeism at work on the following Monday (DV). The identification of the IV and DV is often obvious (as in our absenteeism scenario), but sometimes the choice is not clear. In these latter cases, we evaluate independence and dependence on the basis of:

1. The degree to which each variable may be altered. The relatively unalterable variable is the inde- pendent variable (IV) (e.g., age, social status, camping club membership, present manufacturing technology).

2. The time order between the variables. The independent variable (IV) precedes the dependent vari- able (DV).

Exhibit 8-3 describes four types of asymmetrical relationships: stimulus-response, property-disposition, disposition-behavior, and property-behavior. Experiments usually involve stimulus-response relation- ships. Property-disposition relationships are often studied in business and social science research.

>Exhibit 8-2 Mill’s Method of Difference

Descriptive Factors Outcome

Therefore

A B C Z

A B No C No Z

C Z

schi18939_ch08_170-193.indd 172 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

173

Definitions: A stimulus is an event or force (e.g., drop in temperature, crash of stock market, product recall, or explosion in factory). A response is a decision or reaction. A

property is an enduring characteristic of a participant that does not depend on circumstances for its activation (e.g., age, gender, family status, religious affiliation, ethnic

group, or physical condition). A disposition is a tendency to respond in a certain way under certain circumstances (e.g., attitudes, opinions, habits, values, and drives). A

behavior is an action (e.g., consumption practices, work performance, interpersonal acts, and other kinds of performance).

Although experimental methods neither ensure discovery of all relevant variables nor provide certain proof of causation, they help advance our understanding of causality by eliminating inadequate causal arguments.4

Causal inferences are going to be made. Although they may be neither permanent nor universal, these inferences allow us to build knowledge of presumed causes over time. Such empirical conclusions provide us with successive approximations to the truth.

One classic example of an experiment is the study of bystanders and thieves.5 In this experi- ment, participants were asked to come to an office where they had an opportunity to see a person steal some money from a receptionist’s desk. A confederate of the experimenter, of course, did the stealing. The major hypothesis concerned whether people observing a theft will be more likely to report it (1) if they are alone when they observe the crime or (2) if they are in the company of someone else.

The independent variable in our example was the state of either being alone when observing the theft or being in the company of another person. The dependent variable was whether the participants reported observing the crime. The “theft” had to occur before the “report of the crime,” so the time order of events is upheld. The results suggested that bystanders were more likely to report the theft if they observed it alone rather than in another person’s company. Researchers concluded it was based on this factor as there was no other variable to affect the reporting of the crime. All other variables were controlled.

Ex post facto research designs, in which a researcher interviews respondents or observes what is or what has been, also have the potential for discovering causality. The distinction between these methods and an experiment is that the researcher is required to accept the world as it is found, whereas an experiment allows the researcher to alter systematically the variables of interest and observe what changes follow.

>Exhibit 8-3 Four Types of Asymmetrical Causal Relationships

Relationship Type Nature of Relationship Examples

Stimulus-Response An event or change results in a response from some object.

● A change in work rules leads to a higher level of worker output.

● A change in government economic policy restricts corporate financial decisions.

● A price increase results in fewer unit sales.

Property-Disposition An existing property causes a disposition.

● Age and attitudes about saving.

● Gender and attitudes toward social issues.

● Social class and opinions about taxation.

Disposition-Behavior A disposition causes a specific behavior.

● Opinions about a brand and its purchase.

● Job satisfaction and work output.

● Moral values and tax cheating.

Property-Behavior An existing property causes a spe- cific behavior.

● Stage of the family life cycle and pur- chases of furniture.

● Social class and family savings patterns.

● Age and sports participation.

schi18939_ch08_170-193.indd 173 2/14/18 3:24 PM

>part II The Design of Business Research

174

In a well-executed experiment, researchers must complete a series of activities to carry out their craft suc- cessfully. Although the experiment is the premier scientific methodology for establishing causation, the resourcefulness and creativeness of the researcher are needed to make the experiment live up to its poten- tial. The researcher must accomplish following to make the endeavor successful, as shown in Exhibit 8-4:

1. Select relevant variables.

2. Specify the treatment levels.

3. Control the experimental environment.

4. Choose the experimental design.

5. Select and assign the participants.

6. Pilot test, revise, and test again.

7. Analyze the data.

> Conduct an Experiment6

If we could get overweight employees to lose weight and smok-

ers to quit, employers would have lower health care costs. That’s

the motivation behind health-based incentives among employ-

ers. A recent experiment sheds light on which incentive ap-

proaches are more likely to work—and which aren’t.

Mitesh Patel, a professor of medicine at the University of

Pennsylvania’s Perelman School of Medicine and a professor

of health care management at Wharton, claims, “About 80% of

large employers in the United States use some form of finan-

cial incentive for health promotion,” but little research has been

done to demonstrate effectiveness of these programs. Patel

conducted a year-long experiment to test whether or not finan-

cial incentives delivered through premium adjustments (offering

to lower an employee’s contribution to their health insurance if

health goals are met) work to help people lose weight.

In the experiment, “we offered people $550 if they could lose

5% of their weight over the course of the year. They would get

that as a premium adjustment, dispersed across 26 payments”

(each paycheck would increase by $21 each pay period). One

group got the incentive starting the next year. Another group

got the incentive immediately. A third group got cash incentives

rather than the premium adjustment. And the control group got

nothing. The employees in the control group essentially lost no

weight, and neither did any of the employees assigned to the

experimental groups.

Patel concluded from the research that the standard ap-

proach of using premium adjustments didn’t work. He hypoth-

esized that the reason is that the incentives are often hidden.

“People tend to prefer immediate gratification; they want re-

wards now. It’s really hard to get someone to lose weight and do

something that’s hard to do now for a reward that’s next year.”

Behavioral economics studies recommend that incentives be

immediate and noticeable to be effective.

How would you diagram this experiment? What are the

experimental variables? What are the dependent variables?

What were the weaknesses and strengths of this experimental

design?

Source: “Why Some Workplace Health Incentive Programs Don’t Work-and Some Do,”

Knowledge@Wharton, April 4, 2016, downloaded April 6, 2016 (http://knowledge

.wharton.upenn.edu/article/workplace-health-incentive-programs-dont-work/?utm

_source=kw_newsletter&utm_medium=email&utm_campaign=2016-04-06).

>snapshot Experiments in Improving Employee Health

©imtmphoto/Alamy

schi18939_ch08_170-193.indd 174 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

175

Select Relevant Variables The researcher’s task is to translate an amorphous management problem into the question or hypoth- esis that best states the objectives of the research. Consider the following research question as we work through the seven points listed above:

Does a sales presentation that describes product benefits in the introduction of the message lead to improved reten- tion of product knowledge?

Because a hypothesis is a speculation about the outcome of the study, it might take this form:

Sales presentations in which the benefits module is placed in the introduction of a 12-minute message produce better retention of product knowledge than those where the benefits module is placed in the conclusion.

The researchers would need to select variables that best operationalize the concept’s sales presenta- tion, product benefits, retention, and product knowledge. The product’s classification and the nature of the intended audience should also be defined. In addition, the term could be better operationalized statistically by means of a significance test. The number of variables in an experiment is constrained by the project budget, the time allocated, the availability of appropriate controls, and the number of partici- pants being tested. For statistical reasons, there must be more participants than variables.7

Design the Research

Conduct the Experiment

Pilot Testing & Revision

Sampling Strategy

Select participants

Assign participants to treatment levels

Data Collection Design

Control the experimental environment

Choose the experimental design

Specify levels of treatment

Select relevant variables

ExplorationExploration

Clarify the Research Question Stage 1

Stage 2

Stage 3

>Exhibit 8-4 Experiments in the Research Process

schi18939_ch08_170-193.indd 175 2/14/18 3:24 PM

>part II The Design of Business Research

176

For decades, we’ve seen robots replacing humans on the man-

ufacturing floor. Where repetitive tasks require highly precise

maneuvers, robots seem to have worked well. The push toward

a higher minimum wage—$15 is being implemented in some

states and counties—is also increasing interest in robotics. Has

artificial intelligence advanced enough that we might use ro-

bots in retail situations?

In robot-enthusiastic Japan, CBS correspondent Seth Doane

spent a night in Henn na Hotel, a low-cost, futuristic facility near

the city of Nagasaki, where everything from baggage handling

to checking in to room service is handled by robots. The move

toward human replacement in jobs was driven by economics.

“For five-star hotels that are selling high-end service, human staff

are essential. But for three- or four-star hotels, you need com-

fortable lodging, and a basic level of communication at a rea-

sonable price,” claimed Hideo Sawada, the manager at Henn’na

Hotel. The robots have reduced labor costs by 70 percent com-

pared to a hotel staffed by humans.

The founding partner of Peppers and Rogers Group,

Don Peppers, warns that the trade-off of machines and hu-

mans should consider more than economics. “Even the most

routine and repetitive tasks can still be improved, altered,

upgraded, or modified into better, more efficient tasks, and

machines will never suggest those kinds of changes. Human

beings will.”

Computers get exponentially more powerful at a lower cost

each year. This makes cost-effective robotics more likely for

many repetitive tasks. And the experiments continue. In one San

Francisco store, Target is testing the use of a robot to spot in-

ventory vacancies or out-of-place merchandise on shelves. In

a Whole Foods store, if you drink tea, you can participate in an

experiment where tea is ordered from and brewed by a robot.

Wendy’s and McDonald’s are testing self-serve kiosks in some

restaurants.

But inside Toyota Motor Corp.’s oldest plant, there’s a corner

where humans are forging crankshafts. “We cannot simply de-

pend on the machines that only repeat the same task over and

over again,” Mitsuru Kawai, a senior managing officer at Toyota

revealed in a Bloomberg article. “To be the master of the ma-

chine, you have to have the knowledge and the skills to teach

the machine.”

Sources: “A night in Japan’s robot hotel,” CBS News, July 22, 2015, downloaded May 23,

2016 (http://www.cbsnews.com/news/inside-japan-robot-hotel-hennna-where-staff

-are-robots/); Craig Trudell, Yuki Hagiwara and Ma Jie, “Herald Toyota’s Vision of

Future,” Bloomberg, April 7, 2014, downloaded May 23, 2016 (http://www.bloomberg.

com/news/articles/2014-04-06/humans-replacing-robots-herald-toyota-s-vision-of-

future); Judith Aquino, “Automate or Humanize? The Great Customer Service De-

bate,” 1to1 Media, May 23, 2016, downloaded May 23, 2016 (http://www.1to1media.

com/view.aspx?docid=35796&utm_content=title&utm_source=1to1weekly&utm_

medium=email&utm_campaign=05232016); Larry Downes, “Happy birthday to

Moore’s Law,” The Washington Post, April 16, 2015, downloaded May 29, 2016

(https://www.washingtonpost.com/news/innovations/wp/2015/04/16/happy-birthday-

to-moores-law/); Matt McFarland, “Ex-McDonald’s CEO says raising the minimum

wage will help robots take jobs,” Washington Post, May 25, 2016, downloaded May

29, 2016 (https://www.washingtonpost.com/news/innovations/wp/2016/05/25/ex-mc-

donalds-ceo-says-raising-the-minimum-wage-will-help-robots-take-jobs/); Julie Vogt-

man and Agata Pelka, “Minimum Wage Update: State and Local Highlights,” National

Women’s Law Center Blog, June 5, 2015, downloaded May 23, 2017 (https://nwlc.

org/blog/minimum-wage-update-state-local-highlights/); Julie Vogtman, “The Fight for

$15 is Winning for Women and Families,” National Women’s Law Center, April 2017,

downloaded May 23, 2017 (https://nwlc.org/wp-content/uploads/2017/04/The-Fight-

for-15-is-Winning.pdf ).

>snapshot Robotic Experiments

©Colin Anderson/Blend Images LLC

The selection of measures for testing requires a thorough review of the available literature and instru- ments. In addition, measures must be adapted to the unique needs of the research situation without compromising their intended purpose or original meaning.

schi18939_ch08_170-193.indd 176 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

177

Specify Treatment Levels In an experiment, the researcher manipulates the independent variable, called the experimental treat- ment. The treatment levels of the independent variable are the arbitrary or natural groups the re- searcher chooses within the independent variable of an experiment. For example, if salary (IV) is hypothesized to have an effect on employees’ exercising of stock purchase options (DV), salary might be divided into high, middle, and low ranges to represent three treatment levels of the independent variable.

The levels assigned to an independent variable should be based on simplicity and common sense. In the sales presentation example, the experimenter should not select 8 minutes and 10 minutes as the start- ing points to represent the two treatment levels if the average message about the product is 12 minutes long. Similarly, if the benefits module is placed in the first and second minutes of the presentation, ob- servable differences may not occur because the levels are too close together. Thus, in the first trial, the researcher is likely to position the midpoint of the benefits module the same interval from the end of the introduction as from the end of the conclusion (see Exhibit 8-5).

Under an entirely different hypothesis, several levels of the independent variable may be needed to test order-of-presentation effects. Here we use only two. Alternatively, a control group could provide a base level for comparison. The control group is composed of participants who are not exposed to the independent variable(s), in contrast to those who receive the experimental treatment.

Control the Experimental Environment At this stage, we are principally concerned with environmental control, holding constant the physical en- vironment of the experiment. The introduction of the experiment to the participants and the instructions would likely be videotaped for consistency. The arrangement of the room, the time of administration, the experimenter’s contact with the participants and so forth, must all be consistent across each administration of the experiment.

Other forms of control involve subjects and experimenters. When participants do not know if they are receiving the experimental treatment, the experiment is characterized as blind. When the experiment- ers do not know if they are giving the treatment to the experimental group or to the control group, the experiment is characterized as double blind. Both approaches control unwanted complications such as participants’ reactions to expected conditions or experimenter influence.

Body of Sales Presentation ConclusionIntroduction

Benefits module

IV1 =

Benefits module

Product Knowledge

Product Knowledge

IV2 =

DV1 =

DV2 =

Hypothesis: Sales presentations in which the benefits module is placed in the introduction of a 12-minute message produce better retention of product knowledge by the customer than those in which the benefits module is placed in the conclusion.

E�ect = DV1 – DV2

>Exhibit 8-5 Experiment of Benefits Module Placement of within a Sales Presentation

schi18939_ch08_170-193.indd 177 2/14/18 3:24 PM

>part II The Design of Business Research

178

Data privacy and security regulations are very different

throughout the world, so finding a model that allows a tele-

com to monetize its extensive customer data is an impor-

tant task. Zeotap, a German startup that interfaces multiple

telecom companies’ data with advertisers wishing to tar-

get specific prospects, just completed an experiment with

Mercedes-Benz.

When you purchase a mobile device, you provide your car-

rier with your name and address. They match this with your plan

details, including the number of phones, your phone models,

your plan details (voice, data), and your monthly plan cost. Over

time, they add your actual usage, including how and where you

use your various mobile devices. What Zeotap does is take ano-

nymized, non–personally identifiable telecom data and targets

mobile ads to segments of those consumers through its own

platform and those of ad tech partners (like Zenith).

Because the E-class is an expensive sedan, Mercedes

wanted to reach “individuals between the ages of 40 and 60

who spent 100 euros or more on their mobile bill each month, or

those who spent 100 euros or more and paid for more than one

mobile line” and was able to do so with Zeotap’s technology.

How would you design and evaluate the experiment?

www.mercedes-benz.com; www.zeotap.comwww.zeotap.com

>snapshot Zeotap Experiments with Mercedes Benz

©Lars A. Niki

In our sales presentation experiment, extraneous variables have the potential for distorting the effect of the treatment on the dependent variable and must be controlled or eliminated. These might include differences in age, gender, race, dress, communications competence, and many other characteristics of the presenter, the message, or the situation.

Choose the Experimental Design Experiments have their own unique designs. In conducting an experiment, the researcher uses his or her experience and knowledge to select one design that is best suited to the goals of the research. Judicious selection of the design improves the probability that the observed change in the dependent variable was caused by the manipulation of the independent variable and not by another factor. It simultaneously strengthens the generalizability of results beyond the experimental setting.

Select and Assign Cases The participants or participants selected for the experiment should be representative of the target popu- lation to which the researcher wishes to generalize the study’s results. In the sales presentation example, corporate buyers, purchasing managers, or others in a decision-making capacity would provide better generalizing power than, say, undergraduate college students if the product in question was targeted for industrial use rather than to the consumer.

The procedure for random sampling of experimental cases is similar in principle to the selection of participants for a survey or an observation study. The researcher first prepares a sampling frame and then assigns the participants for the experiment to groups using a randomization technique. However, if randomization is used, those assigned to the experimental group are likely to be similar to those assigned to the control group. Random assignment to the groups is required to make the groups as comparable

schi18939_ch08_170-193.indd 178 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

179

as possible with respect to the dependent variable. Randomization does not guarantee that if a pretest of the groups was conducted before the treatment condition, the groups would be pronounced identi- cal; but it is an assurance that those differences remaining are randomly distributed. In our example, we would need three randomly assigned groups—one for each of the two treatments (benefits module early, benefits module late) and one for the control group.

When it is not possible to randomly assign cases to groups, matching may be used. Matching employs a nonprobability quota sampling approach. The object of matching is to have each experimental and control case matched on every characteristic used in the research. This becomes more cumbersome as the number of variables and groups in the study increases. Because the characteristics of concern are only those that are correlated with the treatment condition or the dependent variable, they are easier to identify, control, and match.8 In the sales presentation experiment, if a large part of the sample of participants was composed of businesswomen who had recently completed communications training, we would not want the characteristics of gender, business experience, and communication training to be disproportionately assigned to one group.

Some authorities suggest a quota matrix as the most efficient means of visualizing the matching process.9 In Exhibit 8-6, one-third of the cases from each cell of the matrix would be assigned to each

Business Experience

Training in communication skills

Women

Category Frequencies before Matching

Men

No Business Experience

Business Experience

No Business Experience

No training in communication skills

45

39

84

18 9

69 9

9

15

9

3

3

2

6

3

3

3

5

3

3

2

6

3

3

3

5

3

3

2

6

3

3

3

5

Experimental Groups

Group Composition after Matching

Control GroupX1 X2

842828 28

>Exhibit 8-6 Quota Matrix Example

schi18939_ch08_170-193.indd 179 2/14/18 3:24 PM

>part II The Design of Business Research

180

of the three groups. If matching does not alleviate the assignment problem, a combination of matching, randomization, and increasing the sample size would be used.

Pilot Test, Revise, and Pretest The procedures for this stage are similar to those for other forms of primary data collection. Pilot testing is intended to reveal errors in the experiment’s design and improper control of extraneous or environmental conditions. Pretesting the measurement instruments permits refinement before the final test. This is the researcher’s best opportunity to revise scripts, look for control problems with laboratory conditions, and scan the environment for factors that might confound the results. In field experiments, researchers are sometimes caught off guard by events that have a dramatic effect on cases: the test marketing of a competitor’s product announced before an experiment, or a reduction in force, reorganization, or merger before a crucial organizational change. The experi- ment should be timed so that subjects are not sensitized to the independent variable by factors in the environment.

How might future work spaces be different than those of today

or 10 years ago? According to Carlo Ratti, director of MIT’s SEN-

SEable City Lab, work, living, and recreational spaces of today

will all be the work spaces of tomorrow. “Charles-Édouard

LeCorbusier, [the father of modern architecture and instrumental

in the present design of urban spaces], claimed cities had four

functions: dwelling, work, recreation, and transportation. Some

current researchers predict that cities will effectively disappear.

But cities are not obsolete; they are thriving with the merger of

the digital and physical worlds. And we are collecting data to

understand that combination.”

“In 1979 there were no cell phones. It’s estimated that more

than 7 billion connections on mobile devices will be made in

2016, with more than 50 billion connections by 2020. With

sensors being embedded in all facets of our lives through

the IoT, that number will only climb.” SENSEable has done ex-

periments on work, living, and leisure spaces in such diverse

places as Dubai shopping centers, the Louvre museum in

Paris, and the campus of MIT. MIT was one of the first cam-

puses to be totally WiFi enabled (pre-2000), so it should be

no surprise that MIT is monitoring where and when that WiFi

network is accessed. It is also experimenting using its WiFi

network to monitor when spaces are occupied and adjusting

the heating/cooling environment on demand—what they call

“local warming.”

SENSEable’s research reveals that work isn’t stationary, it

shifts to various spaces throughout the day. Thus, a space no

longer has to have a single, dedicated use—a space that is

empty when that use is no longer immediate. One European

furniture manufacturer is experimenting with modular furniture

made up of a single-size pod that can adjust its height. When

several pods are massed together, they can become desks,

chairs, sofas, beds, and even a pub table with stools. Some

cities are embracing this concept of flexible spaces by offer-

ing “co-working spaces” for individuals who work remotely

but don’t want to work from home. Starbucks may be attrac-

tive for some, but soon, such workers will have many different

options.

senseable.mit.edu

>snapshot MIT SENSEable City Lab

©Blend Images LLC

schi18939_ch08_170-193.indd 180 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

181

Analyze the Data If adequate planning and pretesting have occurred, the experimental data will take an order and struc- ture uncommon to surveys and unstructured observational studies. It is not that data from experiments are easy to analyze; they are simply more conveniently arranged because of the levels of the treatment condition, pretests and posttests, and the group structure. The choice of statistical techniques is com- mensurately simplified. Researchers have several measurement and instrument options with experi- ments. Among these are:

• Observational checklists and coding schemes.

• Self-report instruments with open-ended or closed questions.

• Scaling techniques (e.g., Likert scales, semantic differentials, Q-sort).

• Physiological measures (e.g., galvanic skin response, EKG, voice pitch analysis, eye dilation).

> Validity in Experimentation Even when an experiment is the ideal research design, it is not without problems. There is always a ques- tion about whether the results are true. Although there are several different types of validity (whether a measure accomplishes what it claims), here only the two major varieties are considered: internal validity (do the conclusions we draw about a demonstrated experimental relationship truly imply cause?) and external validity (does an observed causal relationship generalize across persons, settings, and times?).10 Each type of validity has specific threats we need to guard against.

Internal Validity Among the many threats to internal validity, we consider the following seven:

• History

• Maturation

• Testing

• Instrumentation

• Selection

• Regression toward the mean

• Experimental mortality

We’ll use a design called a preexperiment to discuss several of these. Assume a company’s manage- ment wishes to find the best way to educate its workers about the financial condition of the company be- fore this year’s labor negotiations. To assess the value of such an effort, managers design an experiment of employee knowledge of the company’s finances (O

1 ). Then they present the educational campaign (X)

to a sample of employees, after which they again measure their knowledge level (O 2 ). This design can be

diagrammed as follows:

O

1

X

O 2

Pretest

Manipulation

Posttest

In this design, we take a control measurement (O 1 ) of the dependent variable before introducing the manipu-

lation to the independent variable (X). After the manipulation, we take an after-measurement (O 2 ) of the

dependent variable. Then the difference between O 1 and O

2 is the change that the manipulation has caused.

Between O 1 and O

2 , however, many events could occur to confound the effects of the education effort.

History

During the time that an experiment is taking place, some events outside the company’s control may occur that confuse the relationship being studied. A major corporation may experience a strike. A newspaper

schi18939_ch08_170-193.indd 181 2/14/18 3:24 PM

>part II The Design of Business Research

182

article about companies with financial problems might appear. A union meeting at which this topic is discussed might be held, or another occurrence could distort the effects of the company’s education test.

Maturation

Changes also may occur within the participant that are a function of the passage of time and are not specific to any particular event. These are of special concern when the experiment covers a long time, but they may also be factors in tests that are as short as an hour or two. A participant can become hungry, bored, or tired in a short time, and this condition can affect response results. In our experiment, a participant could begin a course in finance or accounting at the local university, between O

1 and O

2 .

Testing

The process of taking a test can affect the scores of a second test. The mere experience of taking the first test can have a learning effect that influences the results of the second test. So in our experiment, the act of testing financial knowledge can actually increase employee sensitivity to such information, especially to those answers they did not know.

Instrumentation

This threat to internal validity results from changes between observations in either the measuring instru- ment or the observer. Using different questions at each measurement is an obvious source of potential trouble, but using different observers or interviewers also threatens validity. There can even be an instru- mentation problem if the same observer is used for all measurements. Observer experience, boredom, fatigue, and anticipation of results can all distort the results of separate observations. As we would likely measure financial knowledge with a survey or interview, if we change either the method or the instru- ment, we threaten the experiment’s internal validity.

Selection

An important threat to internal validity is the assignment of participants to experimental and control groups. Validity considerations require that the groups be equivalent in every respect. If participants are randomly assigned to experimental and control groups, this selection problem can be largely overcome. Ad- ditionally, matching the members of the groups on key factors can enhance the equivalence of the groups. If we chose only participants from the financial or accounting departments for one group and only workers from the manufacturing floor for another, we would create internal validity problems for our experiment.

Regression toward the Mean

This factor operates especially when groups have been selected by their extreme scores. Suppose we measure the output of all workers in a department for a few days before an experiment and then conduct the experiment with only those workers whose productivity scores are in the top 25 percent and bottom 25 percent. No matter what is done between O

1 and O

2 , there is a strong tendency for the average of the

high scores at O 1 to decline at O

2 and for the low scores at O

1 to increase. This tendency results from

imperfect measurement that, in effect, records some persons abnormally high and abnormally low at O 1 .

In the second measurement, members of both groups score more closely to their long-run mean scores.

Experimental Mortality

This occurs when the composition of the experimental groups changes during the test. Attrition is es- pecially likely in the experimental group, and with each dropout, the group changes. Because members of the control group are not affected by the testing situation, they are less likely to withdraw. In a com- pensation incentive study, some employees might not like the change in compensation method and may withdraw from the test group; this action could distort the comparison with the control group that has continued working under the established system, perhaps without knowing a test is underway.

schi18939_ch08_170-193.indd 182 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

183

All the threats mentioned to this point are generally, but not always, dealt with adequately in experi- ments by random assignment. However, five additional threats to internal validity are independent of whether or not one randomizes.11 The first three have the effect of equalizing experimental and control groups, something no researcher wants.

1. Diffusion or imitation of treatment. If people in the experimental and control groups talk, then those in the control group may learn of the treatment, eliminating the difference between the groups.

2. Compensatory equalization. Where the experimental treatment is much more desirable, there may be an administrative reluctance to deprive the control group members. Compensatory actions for the control groups may confound the experiment.

3. Compensatory rivalry. This may occur when members of the control group know they are in the con- trol group. This may generate competitive pressures, causing the control group members to try harder.

4. Resentful demoralization of the disadvantaged. When the treatment is desirable and the experi- ment is obtrusive, control group members may become resentful of their deprivation and lower their cooperation and output.

5. Local history. The regular history effect already mentioned impacts both experimental and con- trol groups alike. However, when one assigns all experimental persons to one group session and all control people to another, there is a chance for some idiosyncratic event to confound results. This problem can be handled by administering treatments to individuals or small groups that are randomly assigned to experimental or control sessions.

External Validity Internal validity factors cause confusion about whether the experimental treatment (X) or extraneous factors are the source of observation differences. In contrast, external validity is concerned with the interaction of the experimental treatment with other factors and the resulting impact on the ability to generalize to (and across) times, settings, or persons. Among the major threats to external validity are the following interactive possibilities:

• Reactivity of testing on X

• Interaction of selection and X

• Other reactive factors

The Reactivity of Testing on X

The reactive effect refers to sensitizing participants via a pretest so that they respond to the experimental stimulus (X) in a different way. A before-measurement of a participant’s knowledge about the ecology programs of a company will often sensitize the participant to various experimental communication ef- forts that might be made about the company. This before-measurement effect can be particularly signifi- cant in experiments where the IV is a change in attitude.

Interaction of Selection and X

The target population from which one selects cases may not be the same as the population to which one wishes to generalize results. Suppose you use a selected group of workers in one plant for a test of the piecework incentive system. The question may remain as to whether you can extrapolate those results to all production workers within the firm. Or consider a study in which you ask a cross section of a population to participate in an experiment but a substantial number refuse. If you conduct the experiment only with those who agree to participate (self-selection), can the results be generalized to the total population?

Other Reactive Factors

The experimental settings themselves may have a biasing effect on a participant’s response to X. An artificial setting can obviously produce results that are not representative of larger populations. Suppose

schi18939_ch08_170-193.indd 183 2/14/18 3:24 PM

>part II The Design of Business Research

184

the workers who are given incentive pay are moved to a different work area to separate them from the control group. These new conditions alone could create a strong reactive condition.

If participants know they are participating in an experiment, there may be a tendency to role-play in a way that distorts the effects of X. Another reactive effect is the possible interaction between X and participant characteristics. An incentive pay proposal may be more effective with persons in one type of job, with a certain skill level, or with a certain personality trait (competitiveness).

Problems of external validity may not be as easily alleviated with careful design selections. External validity is largely a matter of generalization, which, in a logical sense, is an inductive process of extrapo- lating beyond the data collected. In generalizing, we estimate the factors that can be ignored and that will interact with the experimental variable. Assume that the closer two events are in time, space, and measurement, the more likely they are to follow the same laws.

As a rule of thumb in experimental design choices, we first seek internal validity. Researchers then try to secure as much external validity as is compatible with the internal validity requirements by making experimental conditions as similar as possible to conditions under which the results will apply.

> Experimental Research Designs The many experimental designs vary widely in their power to control contamination of the relation- ship between independent and dependent variables. The most widely accepted designs are based on this characteristic of control: (1) preexperiments, (2) true experiments, and (3) field experiments (see Exhibit 8-7).

Preexperimental Designs All three preexperimental designs are weak in their scientific measurement power—that is, they fail to control adequately the various threats to internal validity. This is especially true of the after-only study.

After-Only Design

This may be diagrammed as follows:

X

O

Treatment or manipulation

of independent variable

Observation or measurement

of dependent variable

(1)

An example is an employee education campaign about the company’s financial condition without a prior measurement of employee knowledge. Results would reveal only how much the employees know after the education campaign, but there is no way to judge the effectiveness of the campaign. How well do you think this design would meet the various threats to internal validity? The lack of a pretest and control group makes this design inadequate for establishing causality.

One-Group Pretest–Posttest Design

This is the design used earlier in the educational example. It meets the various threats to internal validity better than the after-only study, but it is still a weak design. How well does it control for history? Matura- tion? Testing effect? The others?

O

X

O

Pretest Manipulation Posttest (2)

schi18939_ch08_170-193.indd 184 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

185

Static Group Comparison Design

This design provides for two groups, one of which receives the experimental stimulus while the other serves as a control. In a field setting, imagine this scenario. A forest fire or other natural disaster is the experimental treatment, and psychological trauma (or property loss) suffered by the residents is the measured outcome. A pretest before the forest fire would be possible, but not on a large scale (as in the California fires). Moreover, timing of the pretest would be problematic. The control group, receiv- ing the posttest, would consist of residents whose property was spared.

X

O

1

O 2 (3)

The addition of a comparison group creates a substantial improvement over the other two designs. Its chief weakness is that there is no way to be certain that the two groups are equivalent.

When multiple Xs and Os appear vertical to each other, this indicates that the stimuli and/or the observations take place simultaneously.

The Xs and Os in the diagram are read from left to right in temporal order.

O

R

X

O X O O

X O

O

X O

O

O X O

O _ _ _ _ _

Time

Time

Parallel rows that are not separated by dashed lines indicate that comparison groups have been equalized by the randomization process.

Those separated with a dashed line have not been so equalized.

q

An X represents an experimental manipulation of the independent variable presented to a group.

An R indicates that the group members have been randomly assigned to a group.

An O is a measurement of the dependent variable.

E An E represents the e�ect of the experiment and is presented as an equation.

>Exhibit 8-7 Key to Design Symbols

schi18939_ch08_170-193.indd 185 2/14/18 3:24 PM

>part II The Design of Business Research

186

True Experimental Designs The major deficiency of the preexperimental designs is that they fail to provide comparison groups that are truly equivalent. The way to achieve equivalence is through matching and random assignment of cases. With randomly assigned groups, we can employ tests of statistical significance of the observed differences.

It is common to show an X for the test stimulus and a blank for the existence of a control situation. This is an oversimplification of what really occurs. More precisely, there is an X

1 and an X

2 , and some-

times more. The X 1 identifies one specific independent variable, while X

2 is another independent variable

that has been chosen, often arbitrarily, as the control case. Different levels of the same independent vari- able may also be used, with one level serving as the control.

Pretest–Posttest Control Group Design

This design consists of adding a control group to the one-group pretest–posttest design and assigning the participants to either of the groups by a random procedure (R). The diagram is:

R

O

1

X

O 2

R

O 3

O 4 (4)

The effect of the experimental variable is:

E = ( O 2 − O

1 ) − ( O

4 − O

3 )

In this design, the seven major internal validity problems are dealt with fairly well, although there are still some difficulties. Local history may occur in one group and not the other. Also, if communication ex- ists between people in test and control groups, there can be rivalry and other internal validity problems.

Maturation, testing, and regression are handled well because one would expect them to be felt equally in experimental and control groups. Mortality, however, can be a problem if there are different dropout rates in the study groups. Selection is adequately dealt with by random assignment.

The record of this design is not as good on external validity, however. There is a chance for a reactive effect from testing. This might be a substantial influence in attitude change studies where pretests intro- duce unusual topics and content. Nor does this design ensure against reaction between selection and the experimental variable. Even random selection may be defeated by a high decline rate by participants. This would result in using a disproportionate share of people who are essentially volunteers and who may not be typical of the population. If this occurs, we will need to replicate the experiment several times with other groups under other conditions before we can be confident of external validity.

Posttest-Only Control Group Design

In this design, the pretest measurements are omitted. Pretests are well established in classical research design but are not really necessary when it is possible to randomize. The design is:

R

X

O

1

R

O 2 (5)

The experimental effect is measured by the difference between O 1 and O

2 :

E = ( O 2 − O

1 )

The simplicity of this design makes it more attractive than the pretest–posttest control group design. Internal validity threats from history, maturation, selection, and statistical regression are adequately controlled by random assignment. Because the participants are measured only once, the threats of test- ing and instrumentation are reduced, but different mortality rates between experimental and control groups continue to be a potential problem. The design reduces the external validity problem of testing interaction effect.

schi18939_ch08_170-193.indd 186 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

187

Field Experiments: Quasi- or Semi-Experiments12

Under field conditions, we often cannot control enough of the extraneous variables or the experimental treatment to use a true experimental design. Because the stimulus condition occurs in a natural environ- ment, a field experiment is required.

A modern version of the bystander and thief field experiment, mentioned at the beginning of the chapter, involves the use of electronic article surveillance to prevent shrinkage due to shoplifting. In a proprietary study, a shopper came to the optical counter of an upscale mall store and asked to be shown special designer frames. The salesperson, a confederate of the experimenter, replied that she would get them from a case in the adjoining department and disappeared. The “thief ” selected two pairs of sunglasses from an open display, deactivated the security tags at the counter, and walked out of the store.

Thirty-five percent of the participants (store customers) reported the theft upon the return of the salesperson. Sixty-three percent reported it when the salesperson asked about the shopper. Unlike previ- ous studies, the presence of a second customer did not reduce the willingness to report a theft.

This study was not possible with a control group, a pretest, or randomization of customers; but the information gained was essential and justified a compromise of true experimental designs. We use the preexperimental designs previously discussed or quasi-experiments to deal with such conditions. In a quasi-experiment, we often cannot know when or to whom to expose the experimental treatment. Usually, however, we can decide when and whom to measure. A quasi-experiment is inferior to a true experimental design but is usually superior to preexperimental designs. In this section, we consider a few common quasi-experiments.

Nonequivalent Control Group Design

This is a strong and widely used quasi-experimental design. It differs from the pretest–posttest control group design because the test and control groups are not randomly assigned. The design is diagrammed as follows:

O

1

X

O 2

O 3

O 4 (6)

There are two varieties. One is the intact equivalent design, in which the membership of the experi- mental and control groups is naturally assembled. For example, we may use different classes in a school, membership in similar clubs, or customers from similar stores. Ideally, the two groups are as alike as possible. This design is especially useful when any type of individual selection process would be reactive.

The second variation, the self-selected experimental group design, is weaker because volunteers are recruited to form the experimental group, while nonvolunteer participants are used for control. Such a design is likely when participants believe it would be in their interest to be a participant in an experi- ment—say, an experimental training program.

Comparison of pretest results (O 1 − O

3 ) is one indicator of the degree of equivalence between test and

control groups. If the pretest results are significantly different, there is a real question about the groups’ comparability. On the other hand, if pretest observations are similar between groups, there is more rea- son to believe internal validity of the experiment is good.

Separate Sample Pretest–Posttest Design

This design is most applicable when we cannot know when and to whom to introduce the treatment but we can decide when and whom to measure. The basic design is:

R

O

1

(X )

R

X

O 2 (7)

schi18939_ch08_170-193.indd 187 2/14/18 3:24 PM

>part II The Design of Business Research

188

One theory of job attitudes holds that “hygiene” factors, which

include working conditions, pay, security, status, interpersonal

relationships, and company policy, can be a major source of

dissatisfaction among workers but have little positive motiva-

tional power. This theory says that the positive motivator factors

are intrinsic to the job; they include achievement, recognition

for achievement, the work itself, responsibility, and growth or

advancement.a

A study of the value of job enrichment as a builder of job

satisfaction was carried out with laboratory technicians, or “ex-

perimental officers” (EOs), at British Chemical. The project was a

multiple group time series quasi-experiment. The project is dia-

grammed at the end of this Closeup.

Two sections of the department acted as experimental

groups, and two sections acted as control groups. It is not clear

how these groups were chosen, but there was no mention of

random assignment. One of the experimental groups and one

of the control groups worked closely together, while the other

two groups were separated geographically and were engaged

in different research. Hygiene factors were held constant during

the research, and the studies were kept confidential to avoid the

tendency of participants to act in artificial ways.

A before-measurement was made using a job reaction sur-

vey instrument. This indicated the EOs typically had low morale,

and many wrote of their frustrations. All EOs were asked to write

monthly progress reports, and these were used to assess the

quality of their work. The assessment was made against eight

specifically defined criteria by a panel of three managers who

were not members of the department. These assessors were

never told which laboratory technicians were in the experimen-

tal group and which were in the control group.

The study extended over a year, with the treatments intro-

duced in the experimental groups at the start of the 12-month

study period. Changes were made to give experimental group

EOs important chances for achievement; these changes also

made the work more challenging. Recognition of achievement

was given, authority over certain aspects was increased, new

managerial responsibilities were assigned to the senior EOs,

added advancements were given to others, and the opportu-

nity for self-initiated work was provided. After about six months,

these same changes were instituted with one of the control

groups, while the remaining group continued for the entire pe-

riod as a control. Several months of EO progress reports were

available as a prior baseline for evaluation. The results of this

project are shown in the exhibit below.

O O O X O O O O O O O O O O O

-------------------------------------------------

O O O O O O O O X O O O O O O

-------------------------------------------------

O O O O O O O O O O O O O O O

A Job Enrichment Quasi-Experiment13

>closeup

2 3 4 5 6 7 8 9 10 11 12 13 14

Months

M e

an S

co re

Expe rimen

tal gr oups

Cont rol g

roup s

Control group A

Control group B

Change introduced in experimental groups

Change introduced in control group A

9

8

7

6

5

4

schi18939_ch08_170-193.indd 188 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

189

The bracketed treatment (X) is irrelevant to the purpose of the study but is shown to suggest that the experimenter cannot control the treatment.

This is not a strong design because several threats to internal validity are not handled adequately. His- tory can confound the results but can be overcome by repeating the study at other times in other settings. In contrast, it is considered superior to true experiments in external validity. Its strength results from its being a field experiment in which the samples are usually drawn from the population to which we wish to generalize our findings.

We would find this design more appropriate if the population were large, if a before-measurement were reactive, or if there were no way to restrict the application of the treatment. Assume a company is planning an intense campaign to change its employees’ attitudes toward energy conservation. It might draw two random samples of employees, one of which is interviewed about energy-use attitudes before the information campaign. After the campaign the other group is interviewed.

Group Time Series Design

A time series design introduces repeated observations before and after the treatment and allows partici- pants to act as their own controls. The single treatment group design has before-after measurements as the only controls. There is also a multiple design with two or more comparison groups as well as the repeated measurements in each treatment group.

The time series format is especially useful where regularly kept records are a natural part of the envi- ronment and are unlikely to be reactive. The time series approach is also a good way to study unplanned events in an ex post facto manner. If the federal government were to suddenly begin price controls, we could still study the effects of this action later if we had regularly collected records for the period before and after the advent of price control.

The internal validity problem for this design is history. To reduce this risk, we keep a record of possible extraneous factors during the experiment and attempt to adjust the results to reflect their influence.

> An Evaluation of Experiments Advantages Causality cannot be proved with certainty, but the probability of one variable being linked to another can be established convincingly. The experiment comes closer than any primary data collection method to accomplishing this goal.

The foremost advantage is the researcher’s ability to manipulate the independent variable. Conse- quently, the probability that changes in the dependent variable are a function of that manipulation increases. Also, a control group serves as a comparison to assess the existence and potency of the manipulation.

The second advantage of the experiment is that contamination from extraneous variables can be con- trolled more effectively than in other designs. This helps the researcher isolate experimental variables and evaluate their impact over time.

Third, the convenience and cost of experimentation are superior to other methods. These benefits allow the experimenter opportunistic scheduling of data collection and the flexibility to adjust variables and conditions that evoke extremes not observed under routine circumstances. In addition, the experi- menter can assemble combinations of variables for testing rather than having to search for their fortu- itous appearance in the study environment.

Fourth, replication—repeating an experiment with different participant groups and conditions—leads to the discovery of an average effect of the independent variable across people, situations, and times. Fifth, researchers can use naturally occurring events and, to some extent, field experiments (a study of the dependent variable in actual environmental conditions) to reduce participants’ perceptions of the researcher as a source of intervention or deviation in their everyday lives.

schi18939_ch08_170-193.indd 189 2/14/18 3:24 PM

>part II The Design of Business Research

190

Disadvantages The artificiality of the laboratory is arguably the primary disadvantage of the experimental method. However, many participants’ perceptions of a contrived environment can be improved by investment in the facility. Second, generalization from nonprobability samples can pose problems despite random assignment. The extent to which a study can be generalized from college students to managers or execu- tives is open to question. And when an experiment is unsuccessfully disguised, volunteer participants are often those with the most interest in the topic. Third, despite the low costs of experimentation, many applications of experimentation far outrun the budgets for other primary data collection methods. Fourth, experimentation is most effectively targeted at problems of the present or immediate future. Experimental studies of the past are not feasible, and studies about intentions or predictions are difficult. Finally, management research is often concerned with the study of people. There are limits to the types of manipulation and controls that are ethical.

LO8-1      Experiments are studies involving intervention by the researcher beyond that required for measurement. The usual intervention is to manipulate a variable (IV)) and observe how it affects the variable being studied (DV).

LO8-2 Consideration of the following activities is essential for the execution of a well-planned experiment:

a Select relevant variables for testing.

b Specify the treatment levels.

c Control the environmental and extraneous factors.

d Choose an experimental design suited to the hypothesis.

e Select and assign participants to groups.

f Pilot test, revise, and conduct the final test.

g Analyze the data.

LO8-3 We judge various types of experimental research de- signs by how well they meet the tests of internal and ex- ternal validity. An experiment has high internal validity if one has confidence that the experimental treatment has been the source of change in the dependent variable. More specifically, a design’s internal validity is judged by how well it meets seven threats. These are history, maturation, testing, instrumentation, selection, regression toward the mean, and experiment mortality.

External validity is high when the results of an ex- periment are judged to apply to some larger population. Such an experiment is said to have high external validity regarding that population. Three potential threats to external validity are testing reactivity, selection interac- tion, and other reactive factors.

>summary

> Ethical Issues and Their Solutions The major ethical issues involved with experiments are participant welfare, participant privacy, research quality, and sponsor and result nondisclosure. In laboratory experiments, participants are often pre- recruited, so privacy would not be an issue as informed consent is a normal protocol. However, in field settings where participants don’t know they are part of an experiment, researchers do not have prior consent. This means that a participant’s privacy is violated for the sake of protecting the research objec- tive. Researchers deal with this by debriefing participants—sharing what details they can about the experi- ment—without damaging the continuing experiment.

In any experiment, when researchers manipulate one or more variables, participants are affected. In the classic experiment about bystanders reporting theft, the researcher can’t know what emo- tional trauma the participant experienced just by being an observer to a crime. The researcher should always judge the value of the information and insights against any possible harm to a participant. If the researcher has significant concerns, he or she can arrange for trained personnel (e.g., psychologists, counselors, etc.) to deal with expected or likely repercussions.

Research quality is always of paramount importance, especially when determining causation is the ob- jective. A researcher guarantees quality by carefully planning the design and execution of the experiment; this includes thorough training of any data collectors or parties executing the manipulation of variables.

Depending on the environment of the experiment, it may not be possible to keep the identity of the spon- sor concealed. But a researcher can always protect the results. Even if debriefing a participant is required in the protocol, it is inappropriate for the researcher or data collector to discuss results or research objective.

schi18939_ch08_170-193.indd 190 2/14/18 3:24 PM

>chapter 8 Stage 2: Data Collection Design: Experiments

191

LO8-4 Experimental research designs include (1) preexperi- ments, (2) true experiments, and (3) quasi-experiments. The main distinction among these types is the degree of control that the researcher can exercise over valid- ity problems. Three preexperimental designs were presented in the chapter. These designs represent the crudest form of experimentation and are undertaken only when nothing stronger is possible. Their weak- ness is the lack of an equivalent comparison group; as a result, they fail to meet many internal validity criteria. They are the (1) after-only study, (2) one-group pretest– posttest design, and (3) static group comparison.

Two forms of the true experiment were also pre- sented. Their central characteristic is that they provide a means by which we can ensure equivalence between experimental and control groups through random assign- ment to the groups. These designs are (1) pretest–post- test control group and (2) posttest-only control group.

The classical two-group experiment can be ex- tended to multigroup designs in which different levels of the test variable are used as controls rather than the classical nontest control.

Between the extremes of preexperiments, with little or no control, and true experiments, with random assignment, there is a gray area in which we find quasi-experiments. These are useful designs when some variables can be controlled, but equivalent experimental and control groups usually cannot be established by random assignment. There are many

quasi-experimental designs, but only three were cov- ered in this chapter: (1) nonequivalent control group design, (2) separate sample pretest–posttest design, and (3) group time series design.

LO8-5 An evaluation of the experimental method reveals several advantages: (1) the ability to uncover causal relationships, (2) provisions for controlling extraneous and environmen- tal variables, (3) convenience and low cost of creating test situations rather than searching for their app earance in business situations, (4) the ability to replicate findings and thus rule out idiosyncratic or isolated results, and (5) the ability to exploit naturally occurring events.

Some advantages of other methods that are liabili- ties for the experiment include (1) the artificial setting of the laboratory, (2) generalizability from nonprobability samples, (3) disproportionate costs in select business situations, (4) a focus restricted to the present and im- mediate future, and (5) ethical issues related to the ma- nipulation and control of human participants.