ppt

Account Planning & Research Nikki McClaran

Week 4: Quantitative Research

Recap:

2McClaran, ADV342, SM19

•Last week = qualitative research •Important notes: • Questions meant to be open-ended • Looking for answers, not confirming them • Differences between focus groups and individual interviews • Ethnographies

Refresh: What is Quantitative Research?

3McClaran, ADV342, SM19

•Research conducted to examine relationships between variables using statistics •Relationships can be • Correlational or descriptive • Causal or cause-and-effect •No relationship

4McClaran, ADV342, SM19

The type of relationship you seek to test influences the type of method you need • Correlation: surveys, observational • Can be positive (move in same direction) or negative (move in

opposite direction) • Causation: experiments • Some also say longitudinal surveys • This is when an independent variable and dependent variable become

relevant

Variable Relationships

5McClaran, ADV342, SM19

Example positive relationship:

Example negative relationship

Variable Relationships

6McClaran, ADV342, SM19

Two important concepts when thinking about quantitative research include: •Sampling •Validity and reliability

Important Concepts

7McClaran, ADV342, SM19

Sampling

8McClaran, ADV342, SM19

Sampling • Population: a group of people or objects that

are the main focus of scientific query • You, as the researcher, define the population • Can be as big or small as you define (e.g.,

humans, MSU students, females, etc.)

Sample: a subset of the population that is intended to represent the population as a whole

POPULATION

SAMPLE

9McClaran, ADV342, SM19

Sampling Error

SAMPLE

• Anytime you conduct a sample, you have to be aware of how accurately your sample actually reflects your population.

m m

m m

m

m

m

m

m

m Population

m m m m Sample

Is my sample accurate? Why or why not?

10McClaran, ADV342, SM19

Types of Sampling •Two main types of sampling •Probability or “random” sampling: every person in

the population has an equal chance of being selected •Only kind of sampling that is “generalizable”

•Non-probability or “non-random” sampling: everyone does not have an equal chance

11McClaran, ADV342, SM19

Non-probability sampling Four main methods of non-probability sampling: • Convenience: sample from those easiest to sample • Purposive (judgement): sample from a target population of interest • Quota: sample of a particular subset of people • An extension of purposive sampling, in which you are trying to

make sure a certain demographic or characteristic is measured • Snowball: you recruit people, who then recruit others • Typically used for hard-to-reach, hard-to-find, groups of people

12McClaran, ADV342, SM19

Can you easily reach your population?

Yes No

Snowball sampling

Trying to reach a certain target audience?

Yes No

Convenience sampling Trying to fulfill certain frequency of characteristics? (50% female;

50% male)

Yes No

Purposive samplingQuota sampling

13McClaran, ADV342, SM19

Why does sampling matter?

14McClaran, ADV342, SM19

Random Sampling Learn more here: https://www.pewresearch.org/fact- tank/2017/05/12/methods-101-random- sampling/

15McClaran, ADV342, SM19

Validity: did you measure what you intended to measure? •Internal: is your data factually sound? •External: is your data representative/generalizable?

Validity and Reliability

16McClaran, ADV342, SM19

Reliability: is your data stable and consistent?

Validity and Reliability

17McClaran, ADV342, SM19

1. Develop research questions/hypotheses 2. Construct measurement tool 3. Sample population 4. Conduct research 5. Analyze results 6. Present results

Conducting Quantitative Research

18McClaran, ADV342, SM19

At the most broad level, there are open and closed questions

At a more specific level, there are different levels of measurement

Constructing questionnaires

19McClaran, ADV342, SM19

Levels of Measurement

Nominal:

Ordinal:

Interval:

Ratio:

20McClaran, ADV342, SM19

Beyond the typical multiple choice and T/F questions, there are also •Likert scales •Semantic differential scales

•These are interval/ratio level questions

Types of questions

21McClaran, ADV342, SM19

Types of questions Likert Scale:

Semantic differential

22McClaran, ADV342, SM19

Dictates the type of statistics you can conduct: •Nominal and ordinal: descriptive statistics •Mean, mode, frequencies, etc.

•Interval and ratio: inferential statistics •Important when making claims about

relationships between variables

Why does this matter?

23McClaran, ADV342, SM19

Answer the following three questions using the following scale:

1. Wouldn’t you agree that because Starbucks has higher quality coffee, it is a good coffee chain?

2. I go to MSU and am a big Spartan fan. 3. I’ve used my phone for at least one hour every

day, for the past month.

In-Class Activity

24McClaran, ADV342, SM19

•Provide explicit instructions •Use simple, clear, and active language •Don’t use “jargon” •DON’T: Of the following, what are things that you may care

for while you go on a shopping experience that is typical?” •DO: Please select what matters to you during a typical

shopping experience:

Constructing “Good” Questions

25McClaran, ADV342, SM19

•Avoid Bias: • Leading questions: suggesting a “correct” answer • DON’T: How dumb is it that Pitaya doesn’t have a Facebook? • DO: Indicate your views on Pitaya’s social media presence?

• Loaded questions: forcing a certain answer • DON’T: Where do you enjoy drinking beer? • DO: Do you drink beer? Where do you enjoy drinking beer?

Constructing “Good” Questions

26McClaran, ADV342, SM19

•Avoid Bias: •Double-barreled questions: asking two questions in

one •DON’T: How supportive are you of the government and

U.S. foreign policy? •DO: How supportive are you of the government? How

supportive are you of the U.S. foreign policy?

•Use reasonable time frames

Constructing “Good” Questions

27McClaran, ADV342, SM19

•Make them mutually exclusive, meaning someone shouldn’t be able to choose more than one of your answers • DON’T: What is the weather currently like: sun, rain, snow, or fog. • DO: What is the weather currently like (select all that apply): sun,

rain, show, or fog

Constructing “Good” Answers

28McClaran, ADV342, SM19

•Be exhaustive, meaning all available options are provided •DON’T: How many times do you eat out per month? 1-5

times; 6-10 times; more than 10 times •DO: How many times do you eat out per month? 1-5 times; 6-

10 times; more than 10 times, I do not eat out

Constructing “Good” Answers

29McClaran, ADV342, SM19

•http://msu.qualtrics.com

Let’s practice how to make questions and analyze the results

30McClaran, ADV342, SM19

•Considerations: •Should complement or enhance your written/verbal

report •Don’t do it just because you can •Keep them simple •Make sure it is integrated

A note on tables/figures

31McClaran, ADV342, SM19

• Ideal for presenting exact numbers • Includes: • Descriptive title • Columns are clearly labeled • Number or respondents in each group is shown • Numbers are reasonably rounded • Columns add to appropriate total • Fonts are consistent and not distracting

Numeric Tables

32McClaran, ADV342, SM19

• Ideal for comparing groups Bar Charts

33McClaran, ADV342, SM19

• Ideal only when looking at one point in time, but you want to show the share or a segment compared to the whole

Pie Charts

34McClaran, ADV342, SM19

• Ideal for displaying trends, especially over time •Remember, no causation

Pie Charts