Communication Channels The Effects Of Frequency, Duration, And Function On Gratification Obtained

Communication Channels: The Effects Of Frequency,

Duration, And Function On Gratification Obtained

Chapter 1: Introduction to the Study

Communication involves the exchange of thoughts, ideas, or emotions

between two or more people. It can be verbal or nonverbal, written, or oralh.

Effective communication occurs when information and mutual

understanding pass between the sender and receiver, thus conveying

meaning and possibly producing an appropriate or desired reaction (Wrench,

Thomas-Maddox, Richmond, & McCroskey, 2008). In an appropriate

reaction, the communication exchange is effective and produces gratification

for both the sender and receiver. Gratification from effective communication

does not come from merely sending messages; it comes only after the

message is understood and appropriately interpreted by the receiver. Thus,

gratification from any communication derives from the receiver’s

interpretation. The closer the interpretation is to the intended interpretation,

the more gratification the sender obtains (Wrench et al., 2008) and the more

likely it is that a successful communication exchange has taken place.

Communication channels are the methods and techniques used to send

messages, like the telephone, letters, reports, meetings, or the Internet.

People choose between communication channels for numerous reasons, such

as heuristics, ease of use, experience, or simple preference, and the

communication channel may contribute to the success of the overall message

(Wrench et al., 2008). A successful message in this context could be judged

by the appropriateness of the receiver’s reaction as desired by the sender.

For example, if a new member of staff must be introduced to other

employees in the company, a manager may choose to call a meeting, send an

e-mail, write a memo, or introduce the new employee through a series of

face-to-face encounters. Any one of these communication channels can be

effective, depending on the overall goal and available resources. While oral

channels, such as formal meetings or informal face-to-face encounters may

be preferable, they are time-consuming. An e-mail, by contrast, is quick to

write, and all employees with e-mail accounts have the chance to see it. A

memo, also easy to write, takes more time and may not be seen if posted in a

place with limited access. Thus, for the manager, sending an e-mail may be

the most effective and gratifying communication channel choice, saving time

and ensuring that everyone has been informed. Gratification comes when

other employees greet the newcomer by name and are prepared to start

working with him/her. In this example, the choice of e-mail was the most

gratifying.

Communication channel choice can have a significant effect on

current and future relationships between sender and receiver and can be

costly if not done correctly. Each communication exchange must be weighed

and evaluated before choosing the best channel for the message. Dobos

(1992) referred to this evaluation of communication channels as strategic

communication, and for Leonard, van Scotter, and Pakdil (2009), such

communication is important for establishing relationships and working

effectively.

Using the right channel with a clear message may change receivers’ attitudes

and encourage the desired reaction. Time and money may be wasted if

senders do not consider their previous experience with the channel, the

communication partner, and the topic, resulting in misinterpreted messages

(Klyueva, 2010).

To improve communication, companies spend money on new

communication channels, but these are only as effective as the people using

them. Misinterpretation of messages may come from inadequate

communication skills (Hargittai, 2010). If employees do not have adequate

oral or written skills, the channel is irrelevant. Communication channels can

go unused if employees are not properly trained or do not receive adequate

gratification when using them (Dobos, 1992)

In this study, I examined how employees selected communication

channels in the workplace. My initial purpose was to identify how

employees chose between communication channels to send their messages

and what motivated their choice. I asked whether employees considered a

specific communication function (production, maintenance, or innovation)

for the content and audience of their message. I questioned how

communication has changed with the arrival of new, digital alternatives such

as Skype, the Internet, and video conferencing. I aimed to identify the most

frequently used communication channels in the workplace and focused on

the gratification senders obtain from using specific channels and at what

point they reject one communication channel and replace it with another.

The overall premise was that frequency of use, duration, and function predict

gratification and how employees choose communication channels in the

workplace. It was proposed that they choose the communication channel

which gives them the most gratification.

Chapter 1 begins with a general discussion of the study of

communication channels, research purpose and problem statements, and the

nature of the study. An outline of the research design, including the research

questions and hypotheses, is presented. The theoretical framework, uses and

gratifications theory of Blumler and Katz, is discussed with its relevance to

the present study. Assumptions, limitations, and delimitations are addressed,

leading to a final section on the significance of the present study.

Background

Many companies have addressed ineffective communication by

training employees after recruitment. However, another solution might be to

improve their communication skills before they enter the workplace (e.g., in

higher education). Various authors have called for higher education courses

focusing on communication technology. Hargittai (2010) found that

experience, access, and differences in technology competency directly

influenced students’ choice of communication channels. Neuman and

Brownell (2009) noted the growing importance of communication

technology competencies in the hospitality workplace and their need to be

included in university curricula. These two studies are pertinent as the

population examined in the present study consisted of alumni from an

international hospitality school in Switzerland who have a range of

experience, access, and competency in respect of communication

technology. Junco and Cotton (2011) noted the negative effects of

multitasking between communication channels on students’ academic

performance. Multitasking is equally relevant for employees who may have

to juggle several communication tasks at the same time. Moran, Seaman, and

Tinti-Kane (2011) employed uses and gratifications theory to study the

impact of social media networks on higher education. Smith and Wolverton

(2010) examined the communication competencies needed to make higher

education graduates more effective communicators in the workplace.

Communication technology, multitasking, and communication competencies

are reviewed in Chapter 2 in relation to uses and gratifications theory and

communication channel choice.

By identifying communication channels, which provide the most

gratification obtained (GO), future business communication curricula could

be designed to prepare undergraduates to be efficient communicators in the

workplace. Too many communication courses focus only on traditional

communication channels and do not reflect the reality of the workplace

(Neuman & Brownell, 2009). It is necessary to evaluate traditional and

modern communication channels in general business communication to

reflect more accurately the communication competencies these students will

need. In this study, I aimed to provide an understanding of the

communication skills young graduates need that could be used to prepare

new business communication curricula for higher education.

Other authors have addressed the need to improve employees’

communication competencies in the workplace. Kasavana, Nusair, and

Teodosic (2010), Napoli (2010), and Peng and Zhu (2011) examined the

influence of traditional and modern communication channels on the

relationships between senders and receivers. Im, Kim, and Han (2008)

addressed perceived risks of communication technology in the workplace,

finding that when new communication technology is introduced, employees

must be trained in its use, or they may refuse to use it. Ahmad et al. (2010),

Nordin, Halib, and Ghazali, (2011) and White, Vane, and Stafford (2010)

examined the link between employee satisfaction and the amount and quality

of information employees receive. Gratification may be affected by the

volume and quality of messages being sent.

The link between time spent and gratification was examined in the present

study when assessing frequency of use and duration (time spent) on each

communication channel.

All of the studies reviewed have examined U.S. populations, for the

most part, U.S. students. The population chosen for the present study

consisted of alumni of an international hospitality school in Switzerland who

worked in different positions, organizations, and countries around the world.

This population included 84 different nationalities who studied in one of two

languages, French or English. With such a diverse population, the present

study gave a broad view of GO from communication channel choice from an

international population.

Uses and gratifications theory has evolved over time. Include a topic

sentence.

Katz, Blumler, and Gurevitch (1973-1974), McQuail (1984), Rubin (1993),

and Siraj (2007) provided a historical background of uses and gratifications

theory in mass media communication research and suggested its application

to other communication channels. The theory has since been applied to

modern communication channels, such as the

Internet and social media networks. Bagdasarov et al. (2010), Urista, Dong,

and Day (2008), and Kink and Hess (2008) applied uses and gratifications

theory to various communication channels, including the Internet, television,

and social media networks to examine how users choose among channels.

Although the Internet was tested in the present study, many other modern

channels such as teleconferencing, Skype, instant messaging, and mobile

phone were also included as they occur in workplace communication.

Previous scholars have targeted a few communication channels or grouped

channels into categories such as written, oral, or electronic. However, uses

and gratifications theory has not been applied to the range of traditional and

modern communication channels commonly used in the workplace, and for

this reason, the present study filled a gap in the current knowledge on this

subject.

Although uses and gratifications theory has been applied

predominantly to mass or electronic media, the concept of GO is applicable

to any communication channel. Each user seeks some gratification in

sending a message and chooses a communication channel accordingly. In

mass media research the user is the receiver, but, in general communication,

the user is generally considered to be the sender (Bagdasarov et al., 2010;

Dobos, 1992; Kink & Hess, 2008; Urista et al., 2008). Once a channel no

longer provides adequate gratification, the sender seeks out an alternative

(Dobos, 1992). In the literature review in Chapter 2, I discuss research into

sender’s motivations for choosing one communication channel amongst

many alternatives in the workplace. In the present study, I aimed to provide

a more detailed examination with an extensive list of 15 communication

channels currently used in the workplace.

Problem Statement

In this study, I examined choices among traditional and modern

communication channels in the workplace. Traditional communication

channels include telephone, letters, faxes, business reports, presentations,

and face-to-face meetings. Modern communication channels include the

Internet, e-mail, instant messaging, teleconferencing, and video conferencing

(e.g., Skype). While the modern channels are all dependent on technology,

they differ in their purpose. For instance, the Internet may be used as a

research engine to find and share information, while Skype represents a

bundle (voice, text message, video) of individual communication

possibilities. Many studies have been directed at specific communication

channels, such as social media networks (Kasavana et al., 2010), instant

messaging (Junco & Cotton, 2011), the Internet, and e-mail (Dobos, 1992;

Neuman & Brownell, 2009). However, no research has been found on the

link between duration, frequency of use, function, and GO when choosing

amongst individual communication channels.

Previous researchers have concentrated on individual communication

channels or groups of channels. For instance, Newman and Brownell (2009)

examined positive and negative sides of two communication technologies, e-

mail and instant messaging, and Kasavana et al. (2010) defined online social

networking and its implications on the hospitality industry. Hargittai (2010)

examined university students’ skills and experiences in using the Internet,

while Dobos (1988) studied three communication functions (production,

maintenance, and innovation) as applied to face-to-face meetings, written

memos, and electronic media. These and other studies (D’Urso & Rains,

2008; Dobos, 1992; Timmerman, 2010) included only a limited range of

communication channels. The present study widened the range of channels

examined to encompass all those used in the workplace, which were

assessed in terms of the GO each channel provides. Thus, it served as an

initial step in understanding how one group (i.e. hospitality professionals)

chose communication channels based on frequency of use, duration,

function, and GO.

Purpose of the Study

The purpose of the present study was to measure GO derived from the

different functions (production, maintenance, and innovation) of traditional

and modern communication channels used in the workplace. GO was also

measured through frequency of use and duration of each communication

channel. In uses and gratifications theory in mass media, researchers focuses

on the gratification obtained by the receivers of messages (i.e., the

audience). In contrast, I focused upon GO received by the senders of

messages, an approach which has also been used by other authors in general

communication research (Dobos, 1992; Kink & Hess, 2008; Urista et al.,

2008). The independent variables in this study were frequency of use,

duration, function (production, maintenance, or innovation). The dependent

variable was the gratification obtained when choosing these channels.

Nature of the Study

The present study was a quantitative, cross-sectional survey, which

aimed to provide an overview of the current state of employees’

communication channel choice at one time and place (Frey, Botan, & Kreps,

2000; Keyton, 2011). The population to whom this survey was administered

was drawn from the alumni of an international hospitality school in

Switzerland who were employed in different positions and companies

around the world. According to G*Power, a minimum of 77 cases were

necessary to run multiple linear regression using three predictor variables

(Faul, Erdfelder, Buchner, & Lang, 2009). Population and sample size will

be discussed in more detail in Chapter 3. It was foreseen that this one-off

survey of the way these employees chose communication channels would

provide sufficient information to respond to the research questions.

Independent variables included duration, frequency of use, and function

(production, maintenance, or innovation) of communication channels in the

workplace. The dependent variable was GO. Linear regression was

employed to establish how well each independent variable predicts GO for

communication channels.

Research Questions and Hypotheses

The overarching research question (RQ) was: using multiple linear

regression, can Y (gratification obtained) be predicted in terms of three

independent variables (frequency of use, duration, and function)? The

corresponding H0 can be stated as:

H01: R = 0; linear regression is a good fit.

H02: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y.

Furthermore, the RQ and hypotheses were analyzed in terms of

various demographic characteristics, namely gender and work experience.

Therefore, lower level RQs and corresponding hypotheses were specified.

An example is provided for Gender: RQ1: Are there gender differences

when determining whether Y (gratification obtained) can be predicted in

terms of three independent variables (frequency of use, duration, and

function)?

H01M: R = 0; using only Male data, linear regression is a good fit.

H01F: R = 0; using only Female data, linear regression is a good fit.

H02M: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using Male data only.

H02F: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using Female data only.

Other examples include

RQ2: Does the number of years of work experience affect whether Y

(gratification obtained) can be predicted in terms of three independent

variables

(frequency of use, duration, and function)?

H01W (work experience): R=0; using ranges of work experience

(expressed in years), linear regression is a good fit.

H02W: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using work experience (expressed in years).

RQ3: Does the communication channel chosen affect whether Y

(gratification obtained) can be predicted in terms of three independent

variables (frequency of use, duration, and function)?

H01C (communication channel): R=0; using the communication

channel chosen, linear regression is a good fit.

H02C: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using the communication channel.

Variables

The independent variables were frequency of use, duration (time spent

in hours) for each communication channel, and function (production,

maintenance, or innovation). The dependent variable was the GO when

choosing individual communication channels in the workplace (rated on

Likert scale 1-7). Multiple linear regression was run to test the hypotheses

and how well the independent variables predict the dependent variable based

on gender, work experience, and communication channel.

Theoretical Base

Uses and Gratifications Theory

Uses and gratifications theory assumes that communication users

make active, rational choices between alternatives to maximise the

gratification obtained (GO) (Katz, Blumler, & Gurevitch, 1973). The theory

has been much applied to mass media, where the users, i.e. the audience,

have the ability to choose one mass medium alternative over another. In the

present study, the same logic has been applied to channels of general

communication channels, where the users are the senders, rather than the

recipients of messages. It is assumed that they choose which channel to use

on the basis of GO, avoiding those which are less gratifying.

Mass media to which uses and gratifications theory has been applied

include television, radio, and, more recently, the Internet, where gratification

has been measured through process, content, and social use. According to

Katz et al. (1973), mass media audiences seek gratification through finding

entertainment (diversion), maintaining close personal relationships, creating

an online personal identity, and controling others’ behavior (surveillance).

Uses and gratifications theory assumes that an audience continues to use a

communication channel until they fail to derive sufficient gratification from

it. This same theory was applied in the present study to senders of messages

in the workplace, who are assumed to choose communication channels as

long as they provide sufficient gratification.

Dobos (1988) expanded upon previous research by applying uses and

gratifications theory to communication channels other than mass media. She

examined what she called three functions of communication channels:

production or the process of giving and receiving information, maintenance

or maintaining social relationships with others, and innovation or

brainstorming innovative and creative new ideas. She found that different

communication channels were perceived to offer different levels of

gratification depending on their predominant function (Dobos, 1988). For

both mass media and general communication, channel choice is made after

considering the function and subsequent GO (Katz et al., 1973; Dobos,

1992).

According to uses and gratifications theory, people make rational

choices by identifying their media needs and taking action to gratify them

(Katz et al., 1973). For instance, a person who wants to know the latest news

may choose between newspapers, radio, or television, basing their choice on

availability and accessibility. On a train, the newspaper might be the best

option; in the kitchen, the radio; in the family room, the television. Early

uses and gratifications theorists examined why an audience chooses one

medium, or channel, over the alternatives. Katz et al. (1973) defined uses

and gratifications theory as the audience’s relationship between gratification

sought (GS) and GO when choosing between media channels to acquire

information or to be entertained. An active audience will consciously choose

the channel which provides the most gratification and will continue to use

this as long as the gratification exists. When GS and GO are mismatched, the

audience will seek out new, alternative media channels to satisfy them.

Uses and gratifications theorists assume an active audience who

expect to be gratified with their choice. In this way, each medium competes

with other sources of need satisfaction and is goal-directed (Blumler, 1979;

Palmgreen, Wenner, & Rosengren, 1985). Media may be chosen for the

potential GO that can be derived from the content, exposure, and/or social

situation. Each medium has the potential to fulfill a wide range of

gratifications (Blumler, 1979; Palmgreen et al., 1985). The active audience

evaluates whether there is sufficient GO and, if not, they seek out another

medium to fulfill their media needs.

Individuals fulfill some need when choosing exposure to a mass media

channel. The need may be straightforward (e.g., it is raining, and there is

nothing to do, so a user turns on the television to pass some time).

Individuals may also choose a channel because of the information it provides

(e.g., a documentary or a news program). An individual who obtains

gratification from this choice will continue to use this mass media channel.

In this study, I tested whether this is also true of general communication

channels used in the workplace.

Early uses and gratifications researchers strove to distinguish between

GS and GO (Palmgreen et al., 1985) in the use of mass media channels. An

active audience must first be aware of the gratifications available from using

the medium (their GS) before evaluating their GO. Assuming these variables

can be measured accurately, the extent to which GO matches GS will

reinforce or reduce continued use of this medium (Stone, Singletary, &

Richmond, 1999). However, if GO does not match GS, individuals will seek

out other channels for gratification. Individual needs will prompt different

media choices, and even those who choose the same medium may derive

different GO from it

(Stone et al., 1999).

Early researchers examined the relationship between motivation and

behavior in mass media users. However, because uses and gratifications

theory relates channel choice to motives and needs, it is equally relevant to

workplace communicators. The theory is not limited to traditional

communication channels or to traditional needs, such as social interaction,

passing time, information, habit, and entertainment (Siraj, 2007). New media

needs have been added, including time-shifting, on-line meetings,

networking, and virtual workplaces (Siraj, 2007). Thus, uses and

gratifications theory has been shown to be relevant to all communication

channels and all communication needs which fit the purpose of the present

study.

There are many motivations for choosing between alternatives. Bagdasarov

et al.

(2010) used uses and gratifications theory to identify viewers’ motives for

making particular television channel choices and how these choices satisfied

viewers’ needs. Preference, previous knowledge, and timing contribute to

this choice process. For instance, viewers may decide to watch a new

television program because it stars a wellknown actor, is a police drama, and

is shown at 9:00 p.m. for 1 hour. While watching, viewers may be drawn

into the action, script, and storyline. After the program, they relive the

experience by discussing the plot and trying to imagine what will happen as

the series unfolds. They conclude that the program is worth watching and

plan to watch it again next week. The gratification they receive from

watching the program leads to an intention to watch it again the following

week. Thus, one gratifying communication experience may lead to future

choices of the same TV channel. Senders of communications in the

workplace are similarly more likely to continue choosing a channel for

future communication exchanges if they have experienced effective

communications with it

(Bagdasarov et al., 2010; Dobos, 1992).

If the television program in the example above produced a negative

experience for the audience (e.g., the police drama was unexciting, acting

was substandard, and the plot was transparent), the audience would be

expected to choose an alternative program the following week. As long as a

channel produces gratification, it will continue to be used, but once it has

disappointed the user, it will be replaced. In the workplace, the sender who

uses e-mail to send a message but does not receive a timely response may

replace that channel with a telephone call for the next exchange.

Gratification can come from two sources: process gratification and

content gratification (Rubin, 1985). Process gratification relates to the

experience of using media for entertainment, distraction, or companionship

(Rubin, 1985) and is not linked to the content of the message, but to the act

of using the medium itself. Content gratification, on the other hand, is based

on the message and its intrinsic value for the reader (Rubin, 1985) and the

process (i.e., the channel) may be irrelevant, as long as the message brings

ultimate value to the recipient. For example, a birthday message might be

sent via Facebook, e-mail, or traditional mail, the receiver being satisfied

just to receive the message itself regardless of the medium. Content

gratification is important in terms of transmitted meaning and future

communication exchanges. However, n this study, I focused on the sender

and was concerned with process gratification only.

Functions.

Mass media researchers who have employed uses and gratifications

theory propose four functions of media: diversion, personal relationships,

personal identity, and surveillance (Katz et al., 1973). In a similar way,

Dobos (1992) identified three functions of workplace communication:

production, maintenance and innovation. Production includes specific tasks

that provide information using communication channels. Maintenance

relates to the social aspect of communication and entails starting new

relationships or maintaining existing ones in both positive and negative

communication exchanges. Innovation denotes creative processes and the

ability to generate new ideas. Each of these functions has been evaluated in

the present study of senders’ GO when choosing common communication

channels in the workplace.

Models.

Uses and gratifications theorists have created many models to support

the overarching theory. McQuail (1985) proposed two models for uses and

gratifications theory; a cultural and a cognitive model. The cultural, or

affective, model considers that the primary motivation is the expectation of

involvement between two parties and that satisfaction comes from emotional

experiences such as arousal, empathy, or wonder. The cognitive model

includes interest or curiosity as the main source of motivation and

satisfaction as coming from guidance, surveillance, application, or social

exchange (McQuail, 1985). Blumler (1979) suggested that cognitive

motivation facilitates information gain as the person who seeks information

from a medium will be more likely to attain it, while affective motivation

will include emotional pleasure or arousal. It is the involvement with the

medium which is likely to produce and promote reinforcement effects

(Blumler, 1979).

According to uses and gratifications theory, the motivation to use

mass media is based on social norms and culture as people seek

reinforcement of what they appreciate, stand for, or value (Blumler, 1979).

The audience expects to see a reflection of their perception of “real life”. For

example, early television series, like Little House on the Prairie or The

Waltons, showed wholesome examples of U.S. life, and each episode dealt

with moral issues which could be discussed with children afterward. In these

shows, the television families had mothers and fathers, sisters and brothers,

and a pet or two. There was little violence and minimal, if any, vulgar

language. As the years went by, however, the status quo changed. Programs

began to involve new types of family; for instance, The Cosby Show showed

a wealthy, African American family with two working parents and more

risqué themes such as living together out of wedlock, teenage sex, and drugs.

Today’s programs reflect still newer images of the U.S. status quo, showing

divorced couples, recomposed families, scenes of violence, and vulgar

language. As social culture and norms evolve, so do the television programs

which mirror them.

Changes in the cultural status quo make themselves apparent in

general communication channels as well and are further influenced by new

technology and communication culture, as well as social norms. Meetings no

longer need to be conducted in person, employees no longer need to go to

the office, and documents no longer need to be sent during work hours only.

Hand-written letters were sequentially augmented and replaced by the

telegraph and telephone, and the latter is currently being challenged by

the Internet and e-mail. Changes in the perceived social status quo have

contributed to the evolution of communication channels and the way

audiences use them. Variables linked to status quo including curiosity,

diversity, personal identity, and surveillance will be discussed further in

Chapter 2.

Uses and gratifications theorists have often compared models to define

the best one. Palmgreen and Rayburn (1985) compared six possible uses and

gratifications models by defining the terms beliefs, gratifications, and

satisfaction. Beliefs are ‘the subjective probability that a media object

possesses a particular attribute in the general sense’ (p. 339). Gratification is

“some cognitive, affective, or behavioral outcome of media behavior”

(Palmgreen & Rayburn, 1985, p. 339). While both beliefs and gratifications

are operationalized as expectancies, they have different referents.

Satisfaction refers to a “general feeling of fulfillment as the result of

repeated exposure to a particular content genre” (Palmgreen & Rayburn,

1985, p. 339). This comparison shows the strengths and weaknesses of the

six models when applying current research design methodology. For

example, the discrepancy models had low predictive power and were prone

to reliability problems. The absolute value model was rejected for its

inability to differentiate between overobtaining and underobtaining

gratifications (Palmgreen & Rayburn, 1985, p. 343). According to their

discussion, the best models to use are the gratification obtained models and

especially the expectancy value model. Their findings were considered when

elaborating on the research design in Chapter 3.

Users

The role of the user, or audience, has been essential when applying

uses and gratifications theory in mass media research. According to

Palmgreen et al. (1985), a key dimension for the audience is the ability to

give meaning to a message, so that different audiences do not necessarily

perceive messages in the same manner. Experiences, heuristics, emotions,

and position may affect how the audience interprets the meaning of the

message. This is problematic, and McLeod and Becker (1973) warned that if

receivers can interpret messages in any way they like, the content itself

becomes irrelevant. To further complicate the issue, Katz et al. (1973)

suggested that any one content may serve many purposes. This may explain

why communication continues to be problematic in the workplace.

The audience poses a greater challenge in applying uses and

gratifications theory as people may not always be clear about the extent to

which they have been gratified (McGuire, 1974). For instance, a program

which gratifies a need for entertainment may not gratify the need for

information. The audience is active and individuals’ participation may

influence, positively or negatively, the GO they receive (Levy & Windahl,

1985). McQuail and Gurevitch (1973) posited three categories which may

affect audience gratification: (a) personality characteristics, (b) social roles

and social experience, and (c) variations in environmental and situational

circumstances (p. 289). These categories could affect how gratifying a given

medium is in a specific time and place. The GO from watching a football

match could be enhanced by watching it with a group of fellow supporters,

but it could be limited if the crowd cheers for the other team and the other

team wins. The outcome of the match and the environment in which it is

watched both affect how much gratification the audience obtains. The

audience, as users, may be individually affected by personality, social roles,

and environment which can cloud their judgment of the medium, message,

or GO. These same elements were considered in the present study where

employees rated the GO from 15 communication channels used in the

workplace.

The needs of audience members may be classified into three types: (a)

selectivity, (b) involvement, and (c) utility (Levy & Windahl, 1985, p. 113).

Selectivity means the ability to choose amongst various alternatives,

involvement includes perceived connection between audience and content,

and utility is how individuals use or anticipate using communication

channels for social and psychological purposes (Levy & Windahl, 1985).

This typology is temporal, depending on whether audience activity occurs

before, during, or after exposure (Levy & Windahl, 1985). The following

example applies the same typology to the sender. A manager needs a quick

response to an urgent message.

The manager selects e-mail as the most appropriate channel for the message

(selectivity). The manager knows that the reader also has and uses e-mail as

they have exchanged many e-mails in the past (involvement). The format,

content, context, language, and register are based on previous exchanges and

should provoke a positive response (utility). However, how the message is

interpreted will depend on the personal characteristics and experiences of the

receiver, as discussed above. Levy and Windahl (1985) cautioned that only

certain messages or parts of messages may be considered as pertinent by the

audience. Thus, merely following the typology and selecting the

communication channel does not guarantee more comprehension or GO than

any other alternative. Other factors may still impede communication.

One factor which may impede communication is the audience’s

interpretation of the message. Blumler (1979) defined audience activity as

utility, intentionality, selectivity, and imperviousness to influence.

According to Katz (1974), individuals choose a medium after evaluating

alternatives which could fulfill their media needs. For Rubin (1993), people

vary in terms of Blumler’s four activity factors, and this may influence

“whether messages even have the opportunity to affect a person’s

cognitions, attitudes, or behavior” (p. 100). Thus, although communication

includes reflection, assessment, and final choice of communication channel,

users may differ in the gratification they seek from a given medium.

Users may choose between althernatives based on certain stages of

activity when using a specific medium. According to Lin (1993), there is a

significant relationship between the activity of the viewers and the level of

GO they experience. Rubin (1993) suggested that an involved audience

actively acquire and process information from the environment through a

preinvolvement stage which includes preexisting attitudes, beliefs, and

motivations and an involvement stage that occurs during message reception

and consists of participation, attention, and emotion. Lin posited three such

phases: selectivity prior to exposure, involvement during exposure, and use

of media content for cognitive processing after exposure. Siraj (2007) also

identified three activity phases: preactivity, duractivity, and postactivity. In

fact, preactivity involves selection of the content, duractivity includes

psychological attentiveness and personal involvement, and postactivity

consists of behavior after the experience, such as discussion or reflection

(Siraj, 2007). An active audience takes a decision to participate in a specific

communication experience and is capable of evaluating this experience in

terms of a specific communication channel.

In mass media research, uses and gratifications theorists have focused

on the audience as users who chose the medium which gave the most

gratification. When applying uses and gratifications theory to

communication in the workplace, however, the focus shifts to a different

user, the sender, who chooses the communication channel which provides

the most gratification for the communication exchange. The latter

gratification can be considered effective communication and could

encourage or discourage future use of the communication channel. Receivers

are not targeted, as they do not choose a communication channel for the

exchange, but wait passively for the choice to be made by the sender. For

this reason, in the present study, the sender of workplace messages was the

focus of the research.

Alternatives

According to uses and gratifications theory, the audience makes

conscious decisions when choosing a mass media channel. However,

audience effectiveness at judging a channel could be skewed by personal

motivations, but they may also be overwhelmed or constrained by the

number of possible choices (Palmgreen et al., 1985). Therefore, both the

available channels and those expected by receivers must be considered

(McQuail & Gurevitch, 1974). For instance, a company may post all daily

news messages on the company Intranet. Although such messages might be

communicated in the company newsletter or shown on in-house television

screens, the employees do not look for them there because the expectation is

that they will be posted on the Intranet. In this example, the Intranet is the

appropriate channel for this message and provides the most gratification to

the sender who uses it.

In principle, the needs satisfied by one mass media channel could be

satisfied in other ways or through other channels (Carey & Kreiling, 1974).

Rosengren (1973) suggested that motives for seeking functional alternatives

include change, compensation, escape, or vicarious satisfaction (Rosengren,

1973). Thus, if one mass media channel does not offer sufficient GO,

audience members seek out alternatives. The same concept can be applied to

communication channels. Although the alternatives are numerous and ever

expanding, employees must choose which channel is most effective for the

message that needs to be sent. A channel that does not offer sufficient GO

will be replaced by another one.

Alternatives may also be chosen if the original purpose of the message

changes. For instance, a communication channel that is customarily used to

convey information may prove inadequate for brainstorming creative ideas,

and the sender would be obliged to choose another channel, such as a face-

to-face meeting for this new function (Rubin, 1985). According to Rubin

(1985), motives may need to be realigned to meet situational constraints, or

messages may not be successfully communicated and may not produce

acceptable GO for the sender.

Social Change

Social change can come from effective communication. For

Rosengren (1974), the way messages are perceived is influenced by

individuals’ personal agendas and by societal variables. Individuals have

needs, motivations, heuristics, and experiences which affect the

interpretation of messages received, but they are also subject to societal

norms and expectations. When communication is done well, GO is high and

the chances that this communication channel will be further used in the

future increases. This provides an opportunity for social change. When

audiences are gratified and recognized for their good deeds, like corporate

social responsibility (CSR), more good deeds are likely to follow. The

challenge is finding the correct channel to announce the good deeds and

encourage others to follow the example. This is currently often done through

the Internet, as will be discussed in greater detail in Chapter 2.

There are many ways in which mass media could be used to

encourage social change. According to Katz et al. (1973), mass media

channels could be used to ease conflicts and tension, bring awareness of

existing problems, offer real-life opportunities to satisfy certain needs,

embody and promote positive values, and provide a field of expectations for

the audience. Palmgreen et al. (1985) suggested that opportunities for social

change can come from an active audience who use media systems to

encourage further social change. For Jensen (2002), communication media

can also be used to promote public events, unify institutional communication

processes, and publicize certain cultural practices and worldviews. Audience

members with similar values seek out a mass media channel for gratification,

which springs from a need to match personality characteristics, social roles,

and environmental circumstances. For example, after Hurricane Katrina, the

Internet was used to mobilize thousands of people to rebuild houses and

entire neighborhoods. This example of positive social change has since

inspired others to use the Internet for other positive actions.

Evolution of Uses and Gratifications Theory

Uses and gratifications theory has evolved over several decades.

According to

Ruggiero (2000), “until the 1970’s, uses and gratifications theory

concentrated on gratifications sought, excluding outcomes, or gratifications

obtained” (p. 6). Palmgreen and Rayburn (1985) reviewed the way uses and

gratifications theory has been combined with other theories, such as

expectancy value theory, to create several models of GS and GO. Amongst

these, models based solely on GO have been reported to be the most reliable

(Ruggiero, 2000). Dobos (1988, 1992) repeated this exercise of testing

various combinations of GS and GO models and confirmed that GS alone

does not account for significant variance in channel choice and makes a

trivial contribution to predicting future communication channel choice.

Thus, for the present study on choosing communication channels, only GO

was examined.

GS has been defined as the needs or expectations of users, while GO

is the outcome from using specific media or communication channels

(Dobos, 1988; Dobos, 1992; Lin, 1999; Palmgreen & Rayburn, 1985;

Ruggiero, 2000). Dobos (1992) stated that GS is based on mutually shared

expectations, company culture, and social environment, while GO represents

actual fulfillment and is, therefore, more appropriate for predicting

communication channel choice. In the present study, respondents were asked

to self-report their GO with specific functions of communication channels.

They were not asked to report on their expectations or those of the company.

For this reason, it was pertinent to examine only GO in the present study, as

this represents actual fulfillment from using specific communication

channels.

In the present study, respondents were asked to quantify actual time

spent (i.e., duration) and frequency of use of individual communication

channels in the workplace. These factors are also based on the outcomes of

communication choice (GO), rather than expectations (GS). In fact, it could

prove difficult, if not impossible, to predict how long or how often a channel

may be used in the future. For this reason, it was prudent to ask respondents

to base their responses on outcomes (i.e., how much time and how

frequently they actually use specific communication channels in the

workplace).

In the workplace, communication is often strategic, and choosing

between alternatives may be crucial to the success of the communication

exchange. According to Dobos (1992), risks can be minimized by basing

decisions on GO from channels used in previous exchanges, and for similar

functions, to choose the ideal alternative for the next exchange. This was

tested in the present study.

Uses and Gratifications Theory and the Workplace

Uses and gratifications theory has been applied not only to mass

media such as television, but also to modern communication channels like

the Internet. The paradigm of uses and gratifications theory is based on one

question: “Who uses which media, under what circumstances, for what

reasons and with what effects?” (Rosengren, 1974, p. 269).

Uses and gratifications theory is based on the assumption that people know

what they need and can satisfy and verbalize these needs (Elliott, 1973;

Wenner, 1985). In mass communications, the content should provide the

user (audience) with sufficient GO (McGuire, 1974) as a result of which the

audience may choose a mass media channel for diversion, personal

relationships, personal identity, or surveillance. As long as the GO is

adequate, the medium will continue to be used. In this study, however, I

sought to understand the GO a sender receives when choosing a specific

communication channel. I assumed that senders choose appropriate channels

by function, use them frequently, obtain gratification from them, and

continue to use the same channels in the future.

Workplace communication may affect many people at the same time.

It includes one-on-one, departmental, company-wide and corporate

communication, and at each level, employees may have different roles and

different interests. For example, a department manager may speak one-on-

one to another department manager who is on equal terms; this is referred to

as horizontal communication (Keyton, 2011). The same manager may then

call a departmental meeting with employees who have lower status and

speak to them as their boss (e.g., downward communication) (Keyton,

2011). When needing time off, the manager may go to the boss’s office to

make a request (i.e., upward communication; Keyton, 2011). In one day, the

same manager may employ numerous communication strategies in the

workplace. The individual seeks gratification through communication.

Alternatives are a key feature of uses and gratifications theory. In the

modern workplace, many alternatives are available for communicating

messages, and managers may be inundated with new technology which

promises quick and efficient communication. They must choose between

traditional communication channels like letters or faxes, which could be

slower than their newer counterparts, and new technology, which could be

difficult to use or expensive to implement, or have unforeseen effects. For

instance, a new Microsoft Office version may be introduced in the

workplace. While it may promise more applications, simpler formats, and

ease of use, managers must adapt to this new technology and, initially, may

spend more time learning to use it than doing their work. The more time

perceived as wasted on learning to use a new program, the less GO.

Managers may continue using an older version with which they are fluent,

even if a new, more efficient alternative might ultimately offer more GO. As

time spent on a given medium is the most frequently used variable to

measure GO in uses and gratifications research (McLeod & Becker, 1973),

time spent, or duration, using specific communication channels was also

adopted as a variable in the present study.

One of the most significant factors in choosing a communication

channel is accessibility for both the sender and receiver. For an international

project, all participants need to have access to the same communication

technology. It would be futile to have video-conferencing in one building,

but no access in another. Further, the sender and the receiver must have

similar preferences and behavior toward specific technology. Thus,

frequency of use is often associated with accessibility, since accessible

technology tends to be more frequently employed than less convenient

channels and provides greater GO (Weibull, 1985). For this reason,

frequency of use and GO were tested in the present study.

Accessibility requires that users perceive a communication technology

as relevant and also that they possess the technical skills necessary to use it

(Jensen, 2002). This is, perhaps, most salient when introducing new

communication technology in the workplace. If new technology is not

perceived as more useful or user-friendly, staff may refuse to try it. Jensen

(2002) suggested that knowledge gaps must be addressed so that the new

technology does not increase, rather than diminish, social inequalities or

skills differentials. A hotel in Chicago provides an example. When e-mail

accounts were introduced, all messages from management to staff were sent

by e-mail, rather than (as before) by a printed memo posted in the staff

room. However, not all staff (for instance housekeepers) could conveniently

refer to their e-mail account throughout the day. There was a computer in the

staff room, but these workers preferred to eat or relax during breaks, not

look at e-mails, and they missed some pertinent messages. Once the problem

was identified, messages were duplicated in memos on the staff room

bulletin board. The potential GO of e-mail as a quick and effective

communication channel was not realized; instead, more GO was obtained

from the printed memos.

Company norms and policy may facilitate communication (e.g., all

pertinent messages sent by e-mail) or hinder it (e.g., employees overloaded

with e-mails so they stop reading them; Elliott, 1973). The perceived

meaning of communication policy rules may vary in terms of interpretation

and therefore outcome (Lull, 1985). Company norms or policies may further

complicate an already complex communication issue.

Professional people’s working behavior is related to the specific demands of

their day-today work environment, and this is likely to extend to

communication behavior and communication choices (Elliot, 1973).

In the workplace, upward communication is used when employees are

making requests or seeking information. Information-giving, however, tends

to be directed downward (e.g., management announces new changes or

policies to the workforce). Communication between colleagues is horizontal

as each worker is seen as an equal. Katz and Lazarsfeld (2006) suggested,

however, that employees attempt to communicate their needs through the

person with the highest status. Many employees engage in personal

communication with managers who, in turn, address a larger group of

employees, rather than responding to individual concerns. Person-to-person

messages flow horizontally between colleagues and may be sent upward

individually to management, but management often responds by

communicating downward to the entire group in a less personal manner

(Katz & Lazarsfeld, 2006). Katz and Lazarsfeld explained that when there is

a shared concern, a message may be more effective if given to a larger group

at one time. This may, in fact, be one reason why managers communicate

top-down messages through staff meetings or mass e-mails. By targeting the

larger group, managers do not need to address the same issue numerous

times and can limit unsolicited questions.

The manager’s power also influences the nature and outcome of

communications. Katz and Lazarsfeld (2006) compared two groups of

workers, one of which was arranged in a circle so that no one person had

more perceived power than another, while the second had one person placed

in the middle of a wheel shape. The wheel facilitated effective

communication but limited discussion as group members simply waited for

the person in the middle to make final decisions. For a manager, using a

wheel may be more gratifying, as fewer errors are made and more work is

accomplished. Employees, on the other hand, received more gratification in

the circle where no one was the leader, and all participants had equal say.

Thus, from a uses and gratifications’ perspective, there may be a disparity

between what gratifies management and what gratifies employees. This will

be addressed further in Chapter 2.

Uses and Gratifications and Future Research in Communications

Uses and gratifications theory can be applied to any form of

communication. There have been calls upon researchers to build the theory

to cope with the predicted expansion in communication technology

(Williams, Phillips, & Lum, 1985, p. 241). For these authors writing in 1985,

new technologies included communication satellites, videotape, computers

and storage media, mobile telephones, teleconferencing, fiber optics, and

video, some of which have already become near obsolete. However, the

spirit of their statement stands. Uses and gratifications theory should be

constantly adapted to gain further understanding of the ever-changing

communication channel options available to users and the gratification they

receive from using them (Williams et al.,

1985). The present study contributed to that need.

Definition of Terms

Communication: A process which involves the exchange of messages

and meanings between sender and receiver (Berger, 2011)

Communication channels: Oral and written forms of communication

which target individuals or groups (Berger, 2011).

Communication technologies: Technologies such as Internet, e-mail,

instant messaging, telephone, Skype, or video-conferencing which are

dependent on the tools used to communicate (Berger, 2011).

Gratifications: “Potential rewards offered, whether by the media

content or by exposure per se, or by the social settings in which exposure to

the media typically occurs” (McGuire, 1974, p. 167).

Mass media: Means of sending messages to a large number of people

who are not necessarily chosen and may interpret the messages differently

based on their own needs (Berger, 2011).

Modern communication channels: These include Skype, the Internet,

e-mail, instant messaging, and video-conferencing. These channels are

defined by and depend upon specific technologies which must be used in the

communication process.

Traditional communication channels: These are taken to include

meetings, faceto-face discussions, and written correspondence such as

memos, letters, reports, agendas, and minutes. Each traditional channel has

clear structure and format, and has been used for organizational

communication for decades. Traditional channels are taught in business

communication courses and have been the basis for workplace

communication for many years.

Assumptions, Limitations and Scope and Delimitations

Assumptions

The present study was based on the assumption that employees make

conscious decisions when choosing one communication channel over

another. I assumed that the user of a channel, the sender of a communcation,

reflects on the function, content, purpose, audience, previous relationship,

tone, and register of a message when drafting it. However, senders may

choose the simplest communication channel at their disposal, based on

preferences, heuristics, experiences, accessibility, company policy, or

perceived GO. The present study on communication channels through uses

and gratifications theory ensured that the theoretical foundation accurately

reflected the phenomenon being studied

(Dusick, 2011) as it was based on comparable previous studies.

Methodological assumptions were that the population was appropriate

and the sample was valid and sufficient for the statistical tests which were

conducted. The population and sample were carefully selected to provide the

data required and match the regression procedure that was used in order to

test the hypotheses. This meant a randomly distributed sample of a

calculated minimum size. Although the population of alumni from an

international hospitality school in Switzerland is specific and was

purposefully chosen, as representing a typical business community, the

sample taken from this group was randomly distributed. A further

assumption was that the population chosen was representative of the

population at large. It is argued that although a sample drawn from the

10,000 active alumni of the one school may not represent hospitality

employees in their entirety, it does offer a meaningful glimpse of the kind of

problems faced by the larger business population.

The variables for the present quantitative study were defined and

measurable. The instrument used consisted of 19 questions from Downs and

Hazen’s (1977)

Communication Satisfaction Questionnaire (CSQ), which were used to

measure satisfaction with communication functions, and all items in Dobos’s

(1988) study on choice of new media and traditional channels, slightly

modified to test individual communication channels (rather than grouping

them into three categories as was done in her studies). These questionnaires

have been previously tested and have proven reliable and valid instruments

for measuring communication channel use and gratifications obtained.

Limitations

Limitations of a study are the elements a researcher cannot control.

The present study was limited by whether the theoretical foundation

accurately reflected the phenomena/variables being studied. The variables

duration, frequency of use, and function supported the theoretical foundation

and were defined and measurable. Of the variables defined for the present

quantitative study, GO, proved to be the most difficult to measure.

Respondents were asked to self-report their GO with both traditional and

modern communication channels which is something they may have never

done or even considered in the past.

The process of testing itself may change the variables being measured

or the hypotheses being tested (Frankfort-Nachmias & Nachmias, 2008). A

survey on choosing communication channels may take respondents by

surprise. They may even wonder about the relevance of rating

communication on the basis of gratification obtained.

Although the survey was based on existing questionnaires, which have

been proven to be valid and reliable, the two instruments use different scales.

Downs and Hazen’s (1977) CSQ used a 7-point scale; Dobos’s (1988) study

employed a 4-point scale. In the survey questionnaire for the present study,

GO was rated for all variables on a scale of 1-7 to ensure consistency across

the instrument and comparability of the results between variables.

According to Dusick (2011), the results of a study are limited by the

ability of statistical procedures to find statistical significance. It was

considered that regression procedures were the most appropriate for testing

the research hypotheses and that these would find adequate statistical

significance in the data. As mentioned above, the population of 10,000

active alumni was sampled randomly in order to satisfy the requirements for

regression analysis.

Another potential limitation may be a lack of respondents to test the

hypotheses. According to Dusick (2011), the results of a study are limited by

the ability of statistical testing to detect significant differences/relationships

if they exist in the population; if there are no differences in the population to

begin with, there will be no differences in the analysis. There is no way to

ensure that all respondents or a true representation of the overall population

will respond, especially to a survey sent by e-mail. Lack of time or little

interest in the topic could lead to low response rates which might influence

or bias final results. A clear introduction to the purpose and importance of

the topic could motivate employees to respond to this survey questionnaire.

Respondents need to see the value in responding (e.g., that they might

become better communicators) and the potential social change that could

come out of this research. They must understand how this present study is

relevant to them.

Delimitations

Participation in this study was limited to active alumni from an

international hospitality school in Switzerland who graduated from one of

three academic programs over the period of the past 40 years and who were

actively employed. Unemployed or retired individuals were excluded from

this study. Alumni come from more than 84 different nationalities and lived

and worked all over the world and, therefore, potentially represent a wide

range of outlooks and experiences, although presumably related mostly to

the hotel and restaurant industries. Although they speak many native

languages, they all followed an academic program in English or French and

could be addressed in one of the two languages. I am fluent in both

languages. The questions were designed and pretested to accommodate this

wide spread of nationalities, and professional translation and back-

translation was used to ensure that the questions had the same significance in

both English and French.

The present study was limited to the function, frequency of use of,

duration, and GO from choosing specific communication channels in the

workplace. Communication channels used for personal purposes were not

considered. Only those who chose the communication channels, assumed to

be the senders of messages, were examined.

GO was measured using Likert scales which were piloted before

administration. The results of the study are applicable to alumni from one

international hospitality school in Switzerland, regardless of the year, who

are actively employed. The results are not applicable for alumni who are

currently unemployed or retired.

Significance of the Study

Previous scholars have addressed factors which contribute to the

success of the communication exchange: namely, communication channel

choice, accessibility, ease of use, context, motivations, heuristics,

experience, and GO derived from them. While all of these factors may be

relevant, it was beyond the scope of this study to analyze so many. For this

reason, duration, frequency of use, function, and GO were chosen as the

variables to be tested. Many articles have been written evaluating specific

communication channels, but none of them have addressed as many

channels as the present study.

Some researchers have evaluated employee satisfaction with

communication based on clarity, transparency, pertinence, and timeliness of

the content of the messages sent or the frequency with which employees are

informed. These scholars did not, however, examine the specific GO when

using individual communication channels. Uses and gratifications models

have been applied to one or several communication channels (Dobos, 1988;

Dobos, 1992; McQuail, 1984; Palmgreen & Rayburn, 1985). While each of

these models contributes to communication research by examining how

communication in general takes place, they do not examine so many

communication channels used in the workplace or the GO from them.

In this study, I attempted to define GO from communication channels

used in the workplace on the basis of duration, frequency of use, and

function. The initial purpose of this research was to understand how

employees chose the communication channels they used. When the most

frequently used and most gratifying communication channels are

established, future researchers might aim to create a communication model

reflecting the channels being used by employees in the workplace.

Summary and Transition

Chapter 1 began with an introduction to the present study on

communication channels. Limitations, assumptions, and the potential for

social change were also addressed. Uses and gratifications theory was

examined in terms of communication channels, audience, alternatives,

workplace, and future implications.

Chapter 2 will address communication research and the uses and

gratifications paradigm, in terms of duration, frequency of use, function, and

GO from specific technologies and communication channels.

Chapter 2: Literature Review

Introduction

A study of the GO that senders obtain from using specific channels

may shed light upon channel choice, which is the topic of this dissertation.

Factors identified by researchers as influencing GO included audience,

alternative channels, ease and frequency of use, and functions (i.e.,

production, maintenance, and innovation). This approach was applied to

mass media channels in early uses and gratifications research, and it is

considered also applicable to other communication channels. Thus, for the

purposes of the present study, uses and gratifications theory offers an

effective tool for understanding why employees choose communication

channels by examining duration, frequency, function, and GO from channel

choices. This chapter presents an extensive review of communication

research literature, in which each section is subdivided to reflect the

variables discussed above (i.e., frequency and duration, function, and

gratification obtained). I also discuss the (predominantly quantitative)

research designs and variables used in previous research and those authors’

comments about future research.

Reviewed literature derived primarily from recent, peer-reviewed

articles, although relevant early articles on uses and gratifications theory

(e.g.,Katz, Blumler, & Gurevitch, 1973,1974; McQuail, 1985; Rubin, 1985,

1993 ) were included, as well as communication research from prominent

researchers such as Dobos (1988, 1992) and Lin (1993, 1999). Although

some articles were written decades ago, they are relevant to the present of

the dissertation through their application of uses and gratifications theory to

communication channels and their suggestions for future research. Key

words for searching the literature included uses and gratifications,

communication, channels, satisfaction, higher education, traditional and

modern communication channels, and workplace. These terms were

identified as frequently used descriptors in the literature of the themes

discussed in Chapter 1.

Literature databases accessed through online libraries of Walden University

and a

Swiss-based international hospitality school included ABI/Inform Global,

Google Scholar, ProQuest Central, and SAGE. In addition, the following

journal titles were searched, which are known to include articles on

communication research:

Communication Research, Journal of Applied Communication Research,

Communication

Quarterly, Management Communication Quarterly, Computers and

Education,

Communication Studies, The Uses of Mass Communication, Journal of

Business Communications, and Media Gratifications Research. Two

researchers, Michael Hecht and Erik Timmerman, were contacted directly by

e-mail. Michael Hecht gave permission to use his instrument, Interpersonal

Communication Satisfaction Inventory (1978). Erik Timmerman sent a

research article (Timmerman, 2010) containing details of his research

instrument, which could not be accessed through databases without paying.

An attempt to contact Jean Dobos was unsuccessful as, unfortunately, she is

deceased. This review is divided into five parts, starting with a historical

overview of communication research and the ways in which uses and

gratifications theory have been applied to it. I examine communication in

educational and business settings, in each case addressing the duration,

frequency of use, and predominant functions of different communication

channels, as well as the GO obtained from them. I then discuss the potential

of communication channels to promote positive social change by keeping the

public informed, examining duration, frequency of use, function, and GO

associated with appropriate communication channels. The review concludes

with a discussion of research designs used by other communication

researchers and a justification of the design chosen for the present study.

Communication Research

Communication researchers examine the process by which meanings

are created and interpreted (Rubin, Rubin, Haridakis, & Piele, 2010). The

general discipline of communication is held to include 10 main content

areas: communication and technology, group communication, health

communication, instructional communication, intercultural and international

communication, interpersonal communication, language and symbolic

codes, mass communication, organizational communication, and public

communication (Rubin et al., 2010). In this study, I focused on interpersonal

and organizational communication in the workplace (i.e., how employees

communicate within and outside of the company).

Mass media has been defined in numerous ways. Rubin et al. (2010)

and Berger (2011) identified mass media as one facet of a wider set of

communication channels or technologies. Berger and Jensen (2002) also

identified mass media as a means of communication. However, for the

present study, it was necessary to differentiate between mass media and

other modes of communication.

Technology is one topic that researchers have used to differentiate

between mass media and other types of communication. Berger (2011) noted

that the technology upon which the mass media depend targets an

unspecified audience and that feedback is either delayed or nonexistent.

However, the message is often controlled and perhaps modified by

gatekeepers whose approval influences the way the public receive it. In

contrast, most other modes of communication target a specific audience and,

because they are essentially private in nature, there are no gatekeepers.

Technology may be required for some modes of communication, but not for

others. Unlike mass media, feedback in general communication such as that

in the workplace is not only possible, but quick (Berger, 2011). Because of

these differences, mass media and interpersonal communication are often

separated (Jensen, 2002), although some technologies can be used for both.

While I focused on communication processes in the workplace rather

than mass media, it should be noted that there may be an overlap between

the two when the Internet is considered. Researchers have studied the

Internet (including social media networks, company websites, and intranets)

in the contexts of both mass media and general communications. For

example, Dobos (1988, 1992) examined the gratification obtained by users

from oral, written, and electronic channels, Timmerman (2010) applied

media richness theory to electronic and traditional communication channels,

and D’Urso and Rains (2008) tested channel expansion theory with new and

traditional communication media. D’Urso and Rains combined both

concepts, demonstrating the overlap between mass media and general

communication research. For the present study, however, the term

communication channel was preferred to technology or media because not

all channels in the workplace are media-based or rely on the use of

technology. In the present study, the Internet was not examined through a

mass media perspective, but viewed as a communication channel available

to employees in the workplace.

Communication is often based on elements called modes or flows.

According to Jensen (2002), three modes of communication (one-to-one,

one-to-many, and many-tomany) occur in three types of communication

flow: information flow, user flow, and context flow. These modes and flows

have been applied to mass media, but they are equally relevant in general

communication. In that context, information flow is the content or message

that a sender wants to communicate to one or several receivers. User flow

involves accessibility (i.e., a sender can only use a specific channel if the

receiver (s) also has access). Context flow, which brings people together,

relates to the workplace culture or to relationships between sender (s) and

receiver (s). This is essential in global workplaces where clients or

employees may be stationed throughout the world. With email, for instance,

a pertinent message can be sent to any or all employees and will be

instantaneously received. Although the feedback, or response, may be

delayed, the sender’s role is fulfilled at the outset of the communication

process.

Communication researchers have based their work primarily on three

theories: uses and gratifications theory, channel expansion theory, and

expectancy value theory.

These theories differ somewhat in content, but Griffin (2009) regarded them

as sharing 10 common factors: motivation, self-image, credibility,

expectation, audience adaptation, social construction, shared meaning,

narrative, conflict, and dialogue. These factors are rationalized in the

following discussion. Motivation comes from a basic human need for

affirmation and control, as a result of which senders are motivated by a need

to communicate with one or many receivers. Communication affects and is

affected by senders’ and receivers’ sense of identity, (i.e. their self-image

within their cultural context;Griffin, 2009). In the workplace, this is

demonstrated by the roles each employee plays within and outside of the

company. In each role, the employee has a self-image to preserve as well as

that of the company.

Griffin (2009) also noted that all messages, whether verbal or non-

verbal, are validated or discounted by others. A manager who demands a

report the following day will be more likely to receive it than a peer asking

another peer for a favor on a similarly short deadline. Thus, communication

may be validated or discounted on the basis of a hierarchical relationship.

Expectation involves the anticipation of action. A client contacts a company

for a service. The expectation is that the company will provide the service in

an efficient and polite manner. If the exchange is less than the expectation,

the client may choose to go to another provider of that service. Expectation

also affects perceptions, interpretations, and response (Griffin, 2009).

Griffin’s (2009) next three characteristics, audience adaptation, social

construction, and shared meaning, reflected the mutuality of general

communication processes. To be effective, a given message must target a

specific audience and be based on a shared social reality, in order for its

actual interpretation to tally with the intention of the sender. All of these

depend upon sender and receiver having a shared sociocultural perception of

the world. The last three, narrative, conflict, and dialogue, to some extent,

underlie the functional content of all communication. Griffin (2009) noted

that conflict may arise between employees, management, and clients

whenever they have incompatible values and goals or are competing over

scarce resources. Communication channels often carry an element of social

contact, so the choice of channel may produce a positive or negative

emotional effect, which in turn may be especially helpful or damaging

during a time of conflict or crisis. Narrative and dialog contribute to the

social effectiveness of a communication channel, by enhancing the human

touch of the encounter, by presenting the content as a story, or through the

mutual responses of the sender and receiver of the communication (Griffin,

2009).

Communication Channels and Uses and Gratifications

Theory

Mass media channels such as television have provided a fruitful field

for research based on uses and gratifications theory for McQuail (1984),

Rubin (1993), and Siraj (2007), but the theory has also been applied to

Internet-based media. For instance, Kink and Hess (2008) applied uses and

gratifications theory to the relationship between search engines and

traditional information sources. They found that gratification sought (GS)

only develops if the user believes that a channel possesses particular

practical attributes (Kink & Hess, 2008). They examined such variables as

place and frequency of Internet use, experience, speed of the Internet

connection, frequency of using other media, age, gender, and education.

They found that online search engines outweighed older established ways of

retrieving information (specifically encyclopedias, yellow pages, and

telephone directory assistance) in regard to the gratification obtained. Kink

and Hess’s (2008) study suggested users might obtain more gratification

from using modern communication channels than from traditional ones.

Other researchers have applied uses and gratifications theory to

general Internet use (Roy, 2007) and MySpace and Facebook (Urista et al.,

2008). Roy (2007) identified different motivations for using the Internet,

which was based on previous uses and gratifications studies. These included

information seeking, economic incentives, selfimprovement, companionship,

diversion, escapism, self-expression, amusement, establishing status, and

peer pressure. Urista et al. (2008) demonstrated empirically that respondents

use social media networks like MySpace and Facebook for perceived

efficiency and convenience, curiosity about others, popularity, relationship

forming, and reinforcement. Roy (2007) and Urista et al. (2008) examined

one variable which was relevant to the present study, (i.e. function), in the

form of tasks completed through communication channels.

In a study of television viewing, Lin (1993) created a composite GS

variable based on five factors: informational guidance, interpersonal

communication, social interaction, entertainment, and diversion. Other

variables examined were demographics, parental mediation, number of

siblings, and the ability to make viewing decisions. Lin (1993) was able to

demonstrate that more motivated, captivated, and engaged viewers tended to

derive greater gratification from their experience, but suggested a need for

further studies on the relationship between different media technologies and

their content. The present study also examined the relationship between

communication channel choice and function, (i.e. how users choose

appropriate channels for the message which needs to be communicated).

Communication Channels and Education

Because today’s students are tomorrow’s workforce, it is relevant to

understand how students use communication channels. Increasingly since

the 1970s, students have been expected to use technology in and outside the

classroom to communicate, conduct research, and complete assignments.

Although practically all university campuses expect that students will be

able, or will learn during their studies, to use modern communication

technology, what Junco and Cotton (2012, p. 506) referred to as digital

inequalities still persist, as students’ technological skills tend to vary with

gender, racial origin, and socioeconomic status. Students are most

disadvantaged by poor access to technology or by gaps in their skills or

knowledge (Hargittai, 2010). Communication researchers have also

addressed students’ engagement and satisfaction with modern

communication channels, teachers’ and students’ use of communication

channels, and multitasking through communication channels.

Junco and Cotton (2012) reported that of 36,950 university students

they surveyed, 99% had either a laptop (84%) or a desktop (15%) computer,

90% made regular use of social networking websites, and 73% sent regular

instant messaging (IM) texts. However, as Hargittai (2010) pointed out, not

all students have the same skill set in using these technologies or, perhaps,

obtain the same satisfaction from them. Various researchers (Junco &

Cotton, 2011; Kasavana et al., 2010; Neuman & Brownell, 2009) have

shown that university students use modern communication technology,

especially social media networks, to keep in touch and reinforce existing

social connections, but there was a discrepancy in the literature regarding the

positive or negative effects of using these technologies in the classroom.

Junco and Chickering (2010) found that the use of modern technologies in

the classroom was positively related to academic and psychosocial

outcomes, but a later study showed that using technology during class time

had a negative impact on students’ overall performance (Junco & Cotton,

2012).

Frequency and Duration

To study the impact of technology in the classroom upon students’

academic performance, Junco and Cotton (2012) asked how much time

students spent online, how much they spent overall on studying, and how

often they felt they were multitasking. They found students spent over 2

hours per day on the Internet (Junco & Cotton, 2012) which agreed with

previous studies (Junco & Cotton, 2011; Turner & Reinsch, 2010). Hargittai

(2010) measured Internet skills using a 27-item scale and was able to

compare actual online abilities with students’ perceived skill level, high

school GPA (a recognized predictor of overall college GPA), and parental

education (as a proxy for socioeconomic status). Of these variables, time

spent and frequency of use were relevant to the present study as they may be

key indicators of how employees choose communication channels.

Although Junco and Cotton (2012) concluded that the Internet had a

detrimental effect on academic performance, they found no such negative

correlation with other communication technologies, including e-mail, cell

phone, or SMS messaging. These findings suggested that in the workplace,

the Internet might have an analogous detrimental effect on work

performance while other technologies, including traditional communications

like memos or the telephone, may not. This assumption was indirectly tested

when comparing function and GO from choosing traditional or modern

communication channels in the workplace.

A study by the same authors examined the relationship between

instant messaging (IM), multitasking, high speed Internet access, and

sociodemographic factors. On average, IM users spent 120 minutes per day

actively chatting, but considered that much of this time was spent

multitasking (Junco & Cotton, 2011). Unlike their 2012 study, these

researchers found in an earlier study that IM use had a detrimental effect on

academic performance, of which the students were unaware (Junco &

Cotton, 2011). This is interesting, in that students presumably obtained

gratification while they were chatting on IM, but felt less gratification when

their poor academic performance manifested itself. Junco and Cotton (2012)

also found that multitasking had a negative impact upon academic

performance. They defined multitasking in learning situations as divided

attention between the learning objective and other ill-defined tasks which

distract from the original learning objective Junco and Cotton (2012)

considered multitasking as causing a task overload of the human information

processing system. It was also relevant to the present study, since a range of

different channels may compete for employees’ attention and perhaps

distract them from attending to priority tasks.

Turner and Reinsch (2010) found that the most common multitasking

combination used by professionals was speaking on the phone while reading

and even writing e-mails at the same time. Thus, multitasking, often given as

a positive characteristic in résumés, may be disadvantageous to performance

as students or workers’ attention is divided among tasks, none of which

receives their full concentration (Turner & Reinsch, 2010). As Junco and

Cotton (2011) noted: “no matter how good individuals become at

multitasking, they might not ever be as effective and efficient as when they

do one thing at a time” (p. 372). Students and employees may believe they

are skilled multitaskers when, in fact, they are distracted or slacking, and

impeding their performance in the long run (Galluch & Thatcher, 2011;

Turner & Reinsch, 2010). The GO sought by multitaskers is linked to self-

expression, since many messages may be communicated to different people

at the same time. However, this actually hinders twoway communication

flow as the sender is primarily concerned with sending rather than with

responding. The sense of urgency to respond to multiple parties

simultaneously is not equivalent to either efficiency or productivity and

could produce a negative response from the receivers (Turner & Reinsch,

2010).

Junco (2011) studied the relationship between Facebook use and

students’ engagement in the educational process. Junco (2011) proposed that

the more time students spent on social media networks, the more involved

they were in campus organizations and real world issues. This scholar

assessed students’ use of technology, their frequency of using Facebook

overall, and their engagement in various types of Facebook activities (Junco,

2011). In addition, they surveyed the time students spent preparing for class

and on cocurricular activities. Gender, ethnicity, and socioeconomic status

were used as control variables. Junco (2011) reported that students spent

101.09 minutes on Facebook per day and checked the site 5.75 times per

day. Contrary to the two articles by Junco and Cotton (2011, 2012) showing

negative effects of modern communication technology on academic

performance, these researchers showed a positive relationship between

Facebook use and student engagement. This finding was relevant to the

present study since it opened the possibility that social media networks could

increase employees’ engagement in their workplace.

Junco and Chickering (2010) acknowledged the positive influence of

modern technology on student engagement and community ties. However,

they urged caution about negative effects such as shortened communication,

misinterpretation of tone, and missed communication cues. Other negative

social effects included a misguided belief in online privacy, net records,

which may remain online forever, and cyberbullying (Junco & Chickering,

2010). They suggested that educational institutions formulate institutional

policies on modern communication technology (Junco & Chickering, 2010).

Function

Although the Internet has revolutionized knowledge communication,

Altbach, Reisberg and Rumbley (2009) posited that it continues to

exacerbate differences in access and skills between nations and

socioeconomic classes. Much previous literature addressed how modern

communication technology can be used in the classroom to alleviate this

problem (DiVerniero & Hosek, 2011; Galluch & Thatcher, 2011; Mondi,

Woods, & Rafi, 2008; Moran et al., 2011). In a study of how faculty use

social media, Moran et al., (2011) showed that the level of awareness did not

vary with age or stage of career, although the level of use was greater for

younger instructors. DiVerniero and Hosek (2011) reported that students

they surveyed expected their instructors to be familiar with modern

technology both in and out of the classroom. This was problematic for older

faculty who, according to Moran et al. (2011), posted information

significantly less often than their younger colleagues. According to Junco

and Cole-Avent (2008), the digital divide also extends to faculty, especially

in terms of technological skills.

The volume and detail of information posted online has had

implications for both students and faculty (DiVerniero & Hosek, 2011).

Potential employers may refer to students’ online records, including those on

social networks, when making a hiring decision. Both students and

instructors may feel ill at ease if they encounter each other on social media

networks like Facebook. The balance between students wanting instructors

to use modern technology in the classroom and being privy to their private

online lives is precarious. Knowing some details about instructors’ lives

allows students to perceive them as more human and approachable, but

information about their personal problems or opinions may seem

uncomfortable and dissonant (DiVerniero & Hosek, 2011). Although

viewing instructors’ online profiles made them easier to understand as

people, students still preferred instructor disclosures that presented them in a

positive light (DiVerniero & Hosek, 2011).

Moran et al. (2011) noted that online video is the most common type

of modern communication technology used in higher education, and they

reported that students often watched YouTube videos relating to the

environment, politics, and world events. Although information that was once

accessible only through books can now be accessed at any time on the

Internet, faculty members reported reluctance using it due to the time it took

to find appropriate videos and the feeling that they lacked appropriate

training.

Privacy was also a potential issue when using online technology inside

and outside the classroom, and DiVerniero and Hosek (2011) suggested a

system of rules to determine when private information should be revealed or

concealed, which they called communication privacy management theory.

For instance, sharing class notes might be acceptable, but sending a

completed paper to another student could incite copying. Although new

technology offered more possibilities for engaging students academically,

this technology also offered more opportunities for plagiarism, and other

unethical behavior, including cyberbullying. Moran et al. (2011) defined the

challenge as finding ways in which modern communication channels can be

used in the classroom to the greatest positive effect.

Galluch and Thatcher (2011) noted that the Internet can be used

beneficially to update course materials, post course materials and grades,

communicate daily with students, and serve as a platform for online

examinations. On the other hand, the Internet has intensified grade culture

among both faculty and students and has made possible new forms of

cheating, as well as being a focus of distraction during class, cyberloafing,

and cyberslacking (Galluch & Thatcher, 2011). Students who were criticized

in the past for loafing or slacking in the classroom were now using the

Internet to loaf or slack because it was difficult for the instructor to tell

whether their laptop was being used to take notes or to view Facebook

accounts or less academic YouTube videos.

To understand the growth of such maladaptive IT use in the

classroom, Galluch and Thatcher (2011) examined effort expectancy,

performance expectancy, perceived opportunities, social norms, and

perceived threats. They found that the first three variables were significantly

related to appropriate uses of IT. Social norms showed a positive

relationship with students’ intention to cyberslack while perceived threats

showed a negative effect (Galluch & Thatcher, 2011). In terms of uses and

gratifications theory, the first three factors provided enough GO to

encourage repeated appropriate use of IT in the classroom. Ironically, social

norms also provided GO, but that GO motivated them negatively, and was

likely to lead to further cyberslacking.

Junco and Cole-Avent (2008) found that males and females used the

Internet in equal proportions but in different ways. Females were more likely

to chat on Facebook, while males were more likely to play games online.

However, within these parameters, age, socioeconomic status, education,

and broadband access also produced significant differences in Internet use.

For example, young adults were more likely to use the Internet than those

over the age of 40. Lower income students tended to use computers in the

academic setting predominantly for basic tasks such as testing or

assignments, while upper income students might learn how to program

computers. Specific Internet applications such as social media were

predominantly used by university students (Junco & Cole-Avent, 2008).

Although most professionals and university students had cell phones and

access to the Internet, university students preferred text messages to e-mail

(Junco & Cole-Avent, 2008). GO from using e-mail or text messaging,

among other modern communication channels, was tested in the three

hypotheses of the current study. Pirani and Sheehan (2009) examined

how e-mail, telephone, and messaging are accommodated in emergency

communication strategies in a university setting. They found that while e-

mail remains preeminent for official communication, institutions were

increasingly considering its replacement with text messaging in emergency

situations.

Junco and Cole-Avent (2008) noted university students’ preference for text

messaging, which is more likely to command their attention in an emergency

situation. It is possible that this will eventually spread to other situations,

and, since these students will become employees, a similar shift from e-mail

to text messaging may also occur in the workplace. Traditional phone use

was predicted to remain important, although telephone wiring and private

branch exchanges could be replaced by Voice over Internet Protocol (VoIP)

(Pirani & Sheehan, 2009).

Pirani and Sheehan (2009) reported that e-mail was the only

communication channel studied, out of 11, in which users said they would

have confidence in an emergency. Institutions were confident that

emergency notifications via e-mail would reach their intended recipients, but

they were less confident that the messages would be received in time for

recipients to take appropriate action (Pirani & Sheehan, 2009). Sending a

message seems like a certainty, but it is only effectively communicated if the

receiver checks their e-mail regularly. An emergency occurring at the

weekend may remain unknown to the receiver until Monday. This may be

one critical reason to consider using text messaging for emergency messages

in the future in both schools and the workplace.

Bembridge, Levett-Jones, and Jeong (2010) noted that young

graduates often suffer from reality shock due to a gap between

undergraduate programs and the realities of the workplace, especially in

terms of information and communication technology skills. Jones (2011)

questioned whether communication skills taught in colleges and universities

adequately reflect workplace requirements. Company managers assume that

new graduates have sufficient digital literacy to cope with information in

digital formats (Nelson, Courrier, & Joseph, 2011). Bembridge et al. (2010)

referred to this generation as digital natives, since they have grown up in a

world of technology and should be adept at using it. The problem is the

variation in computer proficiency and online skills between different

graduates. Jones (2011) reported that employers were only marginally

satisfied with communication competencies of new employees, including

three levels of digital literacy: (a) digital competence (skills, concepts,

approaches, attitudes); (b) digital usage (application of competences within

specific professional domains); and (c) digital transformation (digital usage

that is developed to enable innovation and creativity, and stimulate

significant change within the professional domain; Nelson et al., 2011).

Nelson et al. (2011) administered a questionnaire to faculty asking

what specific digital literacy aspects students needed to have for their

respective fields. Using factor analysis and ANOVA, these authors were able

to determine commonalities and differences between colleges regarding

digital literacy. They found that all colleges showed similar results for

information search and retrieval, information validation, and information

communication. The most significant difference came in MIS Skills required

between the College of Business and the College of Social Sciences,

Mathematics, and Education (Nelson et al., 2011). Digital literacy, while

pertinent to current workplace communication, was not the direct focus of

the present study. For this reason, specific digital skills were not addressed

in the current research project although they might be inferred from

employees’ choice of communication channel.

Gratification Obtained

In principle, students can obtain gratification from any electronic

media. Mondi et al. (2011) applied uses and gratifications expectancy theory

(UGE), an offshoot of uses and gratifications theory, to determine how and

why students use e-learning, or electronically supported learning, and their

perceptions of its utility. They found that affective, personal integrative, and

entertainment UGE positively influence the perceived e-learning experience

and commented “the more the medium has to offer, the more useful it will

become” (p. 255). In principle, communication channels that offer more

options to the user will be employed more frequently and create greater GO.

While the five types of UGE mentioned above could be tested on specific

communication channels used in the workplace, it was beyond the scope of

this research project.

Numerous articles have examined student skills and motivations in

respect of specific communication channels (Hulea, 2009; Junco & Cole-

Avent, 2008; Nelson et al., 2011; Pirani & Sheehan, 2009). The focus has

mainly been on the skills students have upon entering university, acquire at

university, and are expected to be able to use effectively upon graduation.

Jones (2011), however, noted that little is known about the technology-based

computer skills employers expect of new employees, or whether technology-

related skills are more important than traditional ones (p. 248). The present

study focused on hospitality employees, who were all graduates of the same

international hospitality school in Switzerland. Thus, the communication

skills they acquired in their academic program, both traditional and modern,

and the skills they have since acquired were relevant to this study. It will

certainly be relevant for future research on communication courses and

necessary curricula changes in this hospitality school. According to

Hulea (2009), expectations of communication (GS) are influenced by

external stimuli, such as the social environment or the organizational culture.

These elements affect content, language, tone, register, and channel chosen,

to communicate a specific message in a concrete situation. A sender’s GO

will be high if the receiver responds promptly and as expected to the initial

message (Hulea, 2009). The objective of the message, then, is to express the

motivation underlying its sending (Hulea, 2009). The sender’s GO in the

initial exchange will influence future choices of that same communication

channel. However, GS was not directly explored in the present study, which

concentrated upon GO, outcome, and future communication channel choice.

Future Research

Junco (2011), Junco and Cotton (2011) and Junco and Cotton (2012)

called for future research to clarify how the frequency and mode of use of

modern communication technology relate to academic outcome. Galluch and

Thatcher (2011) suggested further investigation into cyber-slacking and its

effects on academic performance. These studies could easily be adapted to

the workplace from the original educational settings.

Nelson et al. (2011) pointed out the need for further research on digital

competence within life situations, including the workplace. Bembridge et al.

(2010), Jones (2011), and Nelson et al. (2011), proposed a review of

existing curricula, to prepare students more effectively to communicate in

the workplace by digital and nondigital means. Bembridge et al. (2010)

identified a need to examine the transferability of communication

technology skills acquired at university to specific workplace environments.

In this context, the present study aimed inter alia to identify the

communication preferences and behavior of employees, which could be used

to inform existing hospitality curricula to prepare students better for

communication in the workplace.

Communication Channels and the Workplace

As seen in the previous section, higher education is supposed to

prepare graduates with appropriate communication and information

technology skills for the workplace. The reality, however, is that as many as

20% of managers in the USA are reported to be deficient in communication

skills, and some have never even learned to type (Hagler, Erthal, Walzer, &

Anderson, 2009). In the past, secretaries typed written documents, but since

the 1990s, managers have been expected to draft their own documents,

produce their own presentations, and write their own e-mails.

Frequency and Duration

Hagler et al. (2009) examined factors contributing to productivity in

the creation of business e-mail messages. They chose e-mail because its use

in business, estimated at twice that of the telephone and e-mailing, was

reported to take up approximately 25% of an administrators’ working day

(Hagler et al., 2009). Their study found that a significant correlation between

text keyboard speed and the quality of the e-mail message that is produced.

Their work was relevant to the present study as e-mail is widely used in all

workplaces and has been demonstrated to have an impact on productivity

(Hagler et al,

2009).

Previous empirical studies showed no difference in the use of

telephone by different age groups (Junco & Cotton, 2012; Lee, Leung, Lo,

Xiong, & Wu, 2011). However, Clark and Roberts (2010) reported that older

individuals tended to prefer faceto-face communication while younger ones

preferred texting. Quan-Haase and Wellman (2005) introduced the term

hyperconnected to describe people who use seven or more communication

devices in the workplace; Godfrey, Seiders, and Voss (2011) referred to this

as multichannel communication. Presumably, the more communication

channels available, the more complex is the choice that employees must

make to match channel to message.

Godfrey et al. (2011) studied the effect of telephone, e-mail, and

postal mail contacts with customers of an automobile service company. They

found that the value of repurchase per customer peaked at approximately

three telephone contacts, between three and four e-mail contacts, and

between nine and ten postal contacts. Increasing contact volume beyond

these points reduced rates of repurchase. Godfrey et al. (2011) also

examined the interaction of different communication channels, finding that if

one telephone contact was made, between five and six additional e-mail

contacts would be tolerated up to the ideal point. However, if more (up to

five) telephone contacts were made, the acceptable number of additional e-

mail contacts fell to between two and three (Godfrey et al., 2011). The

notions that communication channels obey a law of diminishing returns and

that use of different channels may affect the acceptance of a message both

positively and negatively are crucial considerations that may influence

interpretation of data in the present study.

Function

Lee et al. (2011) noted that the absence of nonverbal cues and lack of

warmth and interaction that typically characterize Internet communication

tended to result in impersonality, shallow interactions, and difficulty in

building social support. This absence of nonverbal cues was found to

contribute negatively to the perceived quality of users’ lives because they

had a much weaker role in social interaction than offline (i.e., person-to-

person) communications.

Lee et al. (2011) identified time and displacement as two possible

causes of weakened social ties. People substituted poorer quality (online)

social relationships for richer (face-to-face) ones and tended to spend more

time on the poorer relationships. The Internet allowed them more apparent

contact with other people for a fraction of the time that would be spent in

face-to-face contact and periods when they could guarantee to be free. Age,

marital status, and education also affected the frequency of Internet use and

thus the quality of life. For instance, young, single, highly educated users

tended to spend more time on Internet. Although the quality of life itself was

outside the scope of the present study, a relationship may exist between

gratification obtained by using communication technology and perceived

quality of life.

Gratification Obtained

Relevant articles in this regard were that of Cho, Ramgolam, Schaefer, and

Sandlin (2011), who examined communication overload and channel

synchronicity and Fonner and Roloff’s (2010) study on teleworkers and job

satisfaction. Hyperconnected employees tended to be overloaded with the

volume of messages they received and might lack the time to process them

(Cho et al., 2011; Fonner & Roloff, 2010). Cho et al. (2011) examined the

relationship between channel use and communication overload among

traditional forms of communication (including face-to-face meetings,

telephone, and memos) and newer forms such as e-mail, cell phones, instant

messaging, SMS, and blogs. They found that perceived communication

overload applied to channels possessing both high and low synchronicity but

that increased organizational identification could create a positive

relationship between communication overload and job satisfaction. This

research was compelling for the present study as it linked both traditional

and modern communication channels to GO.

Nordin et al. (2011) reported that new employees tended to seek task-

oriented information to complete their jobs, while established workers were

more interested in organizational information such as updates on

organizational events, rules, and goals. Their study found that employees

had been adequately informed about their job requirements and

responsibilities (Nordin et al., 2011). Job information represents another

source of GO that employees may be expected to seek and was potentially

relevant to the present study.

Lee et al. (2011) applied a new term, hyperpersonal, to people who

created multiple impressions and managed multiple relationships online.

These self-edited and constructed personae may be heavily amended

representations of their real lives, and it has been noted elsewhere that

Internet users may portray themselves in any way they like, regardless of

any prevailing socially constructed reality (Beddows, 2008).

According to Lee et al. (2011), this may be a release, (e.g., allowing an

extremely shy person to communicate without face-to-face interaction), or a

risk, preventing users from knowing with whom they are communicating.

Josgrilberg (2011) noted the growing symbiosis between humans and

technology, an apparently unavoidable relationship which affects even the

most remote areas of the Earth. Humans must be prepared to acknowledge

both cyberspace and physical space to make sense of their existence in the

21st Century (Josgrilberg, 2011). The idea of symbiosis, which by definition

provides gratification to its participants, was thus relevant to the present

study.

Employee Satisfaction

Job satisfaction is an emotional response towards one’s employment

situation, including pay, promotion, coworkers, and customers (Ahmad et

al., 2010). Among other things, employees’ perceptions of management

communication are directly linked to their satisfaction and retention.

Philippe, Helpling, and Koehler (2009) investigated the content of

managers’ messages and the gratification obtained from them by employees.

Their study included six areas: providing feedback, explaining the

organization vision, giving reasons for change, communicating reward

systems, differences between words and actions, and communication as a

guide for employees’ actions. Their survey showed that employees were

most concerned with consistency, organizational vision, plausible

explanations (e.g., of change), reward structures, and feedback (Philippe et

al., 2009). They found that job satisfaction was significantly correlated with

communication and that consistency and (perceived) sharing of information,

together with appropriate, well-timed feedback were indispensable aspects

of communication.

A number of authors have reported that employees identify more

strongly with organizations that they feel communicate openly and honestly

(Ahmad et al., 2010; Al Eslami Kandlousi, Ali, & Absollahi, 2010;

Bakanauskiene, Bendaraviciene, & Krikstolaitis, 2010; Lowry, Romano,

Jenkins, & Guthrie, 2009; White et al., 2010). For White et al. (2010), the

CEO ultimately sets the tone for internal communication and thus has the

most influence over mutual trust between management and staff.

Communication may succeed in making employees identify with the goals

and values of an organization, but communicating too little creates a vacuum

and too much may result in information overload (White et al., 2010), an

inverted U relationship that recalled Jones’s (2011) study discussed above.

White et al. (2010) suggested that this relationship encompasses both the

amount of information and the appropriateness of the content communicated.

For Al Eslami Kandlousi et al. (2010), the effectiveness of workplace

communication was a matter of fulfilling employees’ informational needs. If

information is not provided officially, employees will seek information from

alternative sources, most notably the hearsay and speculations of colleagues

(the grapevine) which provide ample, though often inaccurate, information

(Al Eslami Kandlousi et al., 2010; Nordin et al., 2011).

Informants in a qualitative study by White et al. (2010) said they

found e-mail highly convenient, but still preferred face-to-face

communication because of the immediate feedback it provided. This

concurred with the finding of Lee et al. (2011) that face-to-face

communication strengthens social ties more than the Internet. According to

White et al. (2010), administrators were often satisfied with information

flow, but unsure of what information needed to be conveyed to employees at

a lower level in the institution and, although they placed ample information

on the company website, it often went unseen because employees did not

seek information from that channel.

Employees at the highest and lowest organizational levels said they

were satisfied with the information they received while middle managers

wanted more information, even if it did not directly affect their daily tasks

(White et al., 2010). When they received information from upper

management, middle managers felt respected by being kept in the loop and

they were more inclined to engage with the company. White et al. (2010)

concluded that e-mail was appropriate for quick notices and updates, printed

paper signified importance, and websites were used as information archives.

These conclusions were tested in the present study when examining the

functions for which employees use specific communication channels.

Rogelberg, Allen, Shanock, Scott, and Shuffler (2010) reported that

more than 10 million meetings occurred daily in the U.S., frequently through

conference calls and video conferencing. Meetings were reported to be a key

element in employee satisfaction offering an opportunity for employees to

come together to share ideas, exchange information, and brainstorm new

concepts, and also a forum for complaints, where frustrated employees could

find empathy or sympathy in a socially acceptable venue (Ahmad et al.,

2010; Rogelberg et al., 2010). Rogelberg et al. (2010) found that satisfaction

with meetings significantly predicted overall job satisfaction, but was

unrelated to gender, job tenure, organizational type, or the number of hours

worked. However, satisfaction from meetings was correlated with job level,

recalling the findings of White et al. (2010) that middle managers welcomed

more communication than employees at the top or bottom of the

organization. Accordingly, job level and organizational type and size were

added as demographic questions in the present study. Fonner and Roloff

(2010) found that employees who worked from home (teleworkers) reported

higher job satisfaction than office workers because they were not required to

participate in so many meetings, which they characterized as unsuccessful,

unproductive, and time-consuming. Working at home allowed people to

limit contact with others and escape from office politics, gossip, and

unplanned meetings, but reduced organizational commitment (Fay & Kline,

2011). Fonner and Roloff’s (2010) study showed that telework significantly

affected stress and the frequency of information exchange, but neither of

these was significantly related to job satisfaction. Thus, worklife balance

may be a better predictor of teleworkers’ satisfaction than stress from

meetings or information exchange. Fay and Kline (2011) suggested that

teleworkers’ satisfaction with informal communication was positively

related to family talk, socializing talk, the degree to which they liked their

co-workers, and organizational commitment, but found that teleworkers’

overall job satisfaction was not increased by the

perceived quality of either their relationships or their informal

communication.

Employees spent much of their time communicating with colleagues

through meetings, formal correspondence, or informal conversations. Those

who had strong, positive working relationships were found to be more

inclined to engage with the organization’s goals and less inclined to leave

(Madlock & Booth-Butterfield, 2012). At times, they even turned to

colleagues for interpersonal needs normally filled by family members, or

engaged in relational maintenance strategies to keep the peace. Madlock and

Booth-Butterfield (2012) reported that these strategies, which included task

sharing, positive expressions of attitude, and conflict management, predicted

employees’ organizational commitment and job satisfaction.

Future Studies

Much of the research on communication and job satisfaction has been

quantitative in approach, but Al Eslami Kandlousi et al. (2010) suggested

that qualitative studies on communication and employee satisfaction would

be useful. Perceptions of communication and communication skills are

subjective, and studies based on interviews or focus groups might illuminate

current knowledge about how people communicate in the workplace. They

also felt it might be profitable to compare the effectiveness of formal and

informal communication channels and the satisfaction obtained from using

them.

Communication Channels and Social Change

Frequency and Duration

Modern technology tends to extend and democratize communication

and in principle should therefore promote positive social change. One

problem with this is the digital divide, the uneven distribution among

societal groups and nations of access to knowledge and technology (Jeffres,

Neuendorf, & Atkin, 2012). However, Jung (2008) tested the relationship

between Internet connectedness and various social and technological factors.

Of the three, access, especially ownership of a home computer, years of

Internet experience, and the number of Internet access places at home or

work had the most significant effect on Internet connectedness (Jung, 2008).

Jung (2008) suggested there is a need for future research into the

connectedness of other media.

Function

Communication channels differ in their effectiveness, depending upon

the circumstances of communication and the social environment (Servaes,

Polk, Shi, Reilly, & Yakupitijage, 2012). This is true when companies

attempt to communicate their corporate social responsibility (CSR) deeds

and social change projects. For Clark and Roberts (2010), communicating

CSR involved numerous stakeholders such as external investors who wanted

to see profits, and internal employees who had to be encouraged to

participate in such actions. Clark and Roberts (2010) noted the conflict

between social and financial performance, which are essentially short-term

concerns and corporate social responsibility which operates on a much

longer timescale. For Jeffres et al. (2012), CSR communication was much

wider than the workplace, as actions which affect social change can

potentially affect everyone in a specific community.

Although companies believe that socially sustainable development is

essential, they are often unsure how to communicate effectively that they are

doing it (Du, Bhattacharya & Sen, 2010). According to Servaes et al. (2012)

such communication must be open, inclusive, and participatory and CSR

actions are easily under-publicized or overexaggerated, both of which result

in unsuccessful communication. Einwiller, Carroll, and Korn (2010) found

that the volume of reportage on CSR actions influenced public perception

and the reputation of the company. In fact, Einwiller et al. (2010) have

shown that CSR activities are one of five cognitive dimensions that people

use to evaluate a company as customers or employees. They found that the

reputation dimension of social and environmental responsibility exerted a

significant influence on emotional appeal, which was in turn related to

people’s intentions to purchase goods or services, or to apply for a position

in the company. Thus, the public are favorably influenced by reports of a

company’s CSR actions as long as an ideal volume of such communication

is not exceeded, and CSR actions appear to be in line with the company’s

prime source of

revenue.

Public skepticism was reported to interfere with effective

communication of CSR issues since customers and others tended to distrust

a company if they suspected ulterior or self-serving motives (Du et al.,

2010). For instance, McDonald’s was treated with suspicion when it

publicized healthy menus, but communications about Ronald McDonald

Houses for sick children (children being perceived as the prime consumer of

fast food) have been positively received. Receivers scanning such

communications ignored anything they perceived as irrelevant (Kiyatkin,

Reger, & Roger-Baum, 2011) and perceptions of relevance tended to reflect

individuals’ roles (e.g., as employee, consumer, investor). In addition,

messages were misleading as the same word often had different meanings

for different target audiences (Kiyatkin et al., 2011). This might be relevant

where companies use one term to denote both actions which are not CSR

and those that are. Receivers may still misunderstand or misinterpret a

message even if the amount of information provided and the communication

channel used are appropriate. Senders of CSR messages must be aware of

these issues when choosing information and communication channels to

communicate CSR.

Servaes et al. (2012) used a mixed methods approach to examine

communication of CSR messages in health, education, environment, and

local government. These scholars demonstrated the complexity of such

communication processes and especially that CSR cannot be communicated

unless the intended audience accesses and uses channels in specific ways.

Latzer (2009) noted how legality became blurred as new technology was

introduced. For instance, programs that were broadcast over the Internet

were not legally considered television and Skype was not considered a

telephone service. Other nascent legal issues included the protection of

intellectual property, taxation of Internet trading, and the regulation of

domain-name systems (Latzer, 2009), to which Clark and Roberts (2010)

added the privacy of personal information, for instance, where employers

consulted social networking sites as part of their assessment of job

candidates. This was a difficult area since HR departments had to

demonstrate due diligence in their assessment of applicants’ suitability, but

were at risk of overstepping privacy boundaries (Clark & Roberts, 2010).

The evolution of telecommunications and media policy and the law varied

from country to country have made it nearly impossible to apply one policy

to all of the countries and communities involved (Latzer, 2009).

On one hand, the Internet can affect, promote, and encourage social

change; on the other hand, new technologies rarely fulfill their potential as

they are dependent on complex social and political issues which may exceed

the utility of the technology itself (Jeffres et al., 2012). In principle, the

Internet offers an ideal platform to communicate with millions of people

simultaneously, in real time, with a common voice, but as Clark and Roberts

(2010) have noted, information posted online is already in the public

domain. In fact, posting honest personal observations or opinions could be

detrimental if an individual is regarded as a representative of a company

both inside and outside the workplace (Clark & Roberts, 2010).

Gratification Obtained

Jeffres et al. (2012) showed that the gratification audiences obtained

from different mass media channels depended upon differences in volume

and content and in the way individuals processed the messages they

received. Unlike other mass media channels, the Internet has the potential to

reduce knowledge gaps through both instrumental learning (seeking specific

information) and incidental learning (gaining background information at the

same time as pursuing other tasks) (Jeffres et al., 2012).

Hummert (2009) reported that effective communication was directly related

to people’s psychological wellbeing and social adjustment and suggested

that changing individuals’ communication behavior might improve their

social health. By clarifying the nature of workplace communication, I

attempted to improve current knowledge about workplace communication

and perhaps communicators’ social health.

Future Research

Jeffres et al. (2012) called for further research into the relationship

between social status and the use of emerging technologies. Kiyatkin et al.

(2011) identified a need to distinguish between what is communicated and

what is understood regarding CSR and understand why organizations focus

more or less on particular social issues. Hummert (2009) suggested that

communication research should investigate ways of encouraging effective

and beneficial communication practices in the community. According to

Frey (2009), there is a need to research how communication can contribute

to public understanding of issues such as alternative energy, climate change,

or positive social change. Clark and Roberts (2010) recommended drafting

company policy and guidelines on how to use information obtained online,

and Latzer (2009) suggested that future research may contribute to bridging

currently identified gaps in government policy and legislation on

communication technology. This legislation could be a step towards

encouraging and promoting positive social change.

Research Design and Justification

Theoretical Models

Uses and gratifications theory was chosen as the principal basis for

this dissertation, because it has been tested through numerous studies

employing uses and gratifications theory in the study of mass media and

general communication (Bagdasarov et al., 2010; Blumler, 1979; Katz et al.,

1973; Kink & Hess, 2008; Lin, 1999; Siraj, 2007). Although various

versions of the theory have emerged, for instance that of McQuail (1984),

who incorporated cultural and cognitive dimensions, the basic theory

outlined in Chapter 1 and discussed below, has sufficient features to make it

suitable for the present study.

Over the decades, uses and gratifications theory has evolved and been

embellished with other theories. Communication privacy management

theory (CPM) was used by DiVieniero and Hosek (2011) to study students’

perceptions of instructors’ online self-disclosure. Uses and gratifications

theory was combined with expectancy value theory by Mondi et al. (2008),

producing what they called uses and gratifications expectancy theory (UGE),

which they used to evaluate students’ perceived e-learning experiences. Uses

and gratifications theory and the technology acceptance model (TAM) were

used together by Galluch and Thatcher (2011) to research maladaptive IT

use in the classroom. Each of these approaches investigated communication

channels in terms of users’ GO.

The section: Communication Channels and the Workplace, revealed

six situations where other theories have been combined with uses and

gratifications theory to examine employee choice of communication

channels. Einwiller et al. (2010) and Jung (2008) applied media system

dependency theory (MSD) to research projects into the effect of media

influence on corporate reputation and the relationship of interconnectedness

with the social environment, respectively. Madlock and Booth-Butterfield

(2012) used

Shultz’s (1958) theory of interpersonal needs to investigate how employee

relationships with colleagues influenced job satisfaction. Lowry et al. (2009)

used a different approach based on the Computer Mediated Communication

Interactivity Model (CMCIM) to examine how process satisfaction was

linked to system adoption and continuance. Timmerman (2010) applied

media richness theory in a study of communication channels while D’Urso

and Rains (2008) used channel expansion theory to examine how employees

use new and traditional communication media (e-mail, telephone, IM, and

face-to-face) in the workplace. Original exponents of uses and gratifications

theory, including Katz, Blumler, Gurevitch, and Rubin suggested that it was

applicable to general communication channels, and the present study

intended to use it to assess the selection of channels of general

communication in the workplace. The research questions and hypotheses did

not require the uses and gratifications model to be augmented with other

theories as some other researchers have done. It should be borne in mind that

those studies dealt with extremely specific issues among a limited range of

communication channels. Adapting or augmenting uses and gratification

theory was not appropriate in the present study, which intends to apply a

broader perspective across a greater range of communication channels.

Surveys and Measurement Tools

Jones (2011) used Warner’s (1995) skills inventory to examine the

written communication skills of new graduates. The Communication

Satisfaction Questionnaire (CSQ) of Downs and Hazen (1977) has been

employed by several groups to assess satisfaction with communication in the

workplace (Al Eslami Kandlousi et al., 2010;

Bakanauskiene et al., 2010; Nordin et al., 2011; Rogelberg et al., 2010).

Rogelberg et al. (2010) used the Job Description Index (JDI) to assess job

satisfaction in the context of meetings, and Madlock and Booth-Butterfield

(2012) measured organizational commitment using the Organizational

Commitment Questionnaire (OCQ) of Mowday et al. (1979); job satisfaction

through Abridged Job in General Scale (AJIG); and communication

satisfaction with Hecht’s (1978) Interpersonal Communication Satisfaction

Inventory (ICSI).

Measurement tools such as CSQ, JDI, or AJIG, could have been

applied or adapted to the present study. They have been tried and tested in

numerous research projects on communication and are relevant for the

present study for two reasons: (a) they are established measures that have

been successfully used in the past; and (b) their use would allow the findings

of the present study to be compared with those from previous studies. The

final version of the survey questionnaire for the present study, used 19

questions from the CSQ.

Variables

Many variables were examined in the literature including frequency of

use, tasks performed, function, satisfaction, age, previous knowledge, place,

gender, education, motivation, level of awareness, perceptions, and Internet

skills. Variables selected as relevant to the present study were as follows:

independent variables were duration, frequency of use, and function, and the

dependent variable was respondents’ perception of gratification obtained.

Other variables from previous research which might have been added

to the present research design included job position, company size, and

gender. According to the literature, different communication channels tend

to be chosen by employees holding different roles. Although job position or

company size could affect communication channel choice, they were not

tested as independent variables for the present study. Instead the study

sought to determine whether there were differences in communication

channel choice according to gender and number of years of work experience.

Education level, previous knowledge, and Internet skills were not

considered, although they might be tested in a subsequent study.

This review has examined literature dealing with the frequency of use,

duration, and function of communication channels and the GO that users

derive from them. Different researchers have applied these broad principles

to education, the workplace, and to positive social change. Frequency of use

and duration have frequently been tested by gender, level of awareness,

previous knowledge, education, or Internet skills (Junco, 2011; Junco &

Cotton, 2010; Junco & Cotton, 2012; Moran et al., 2011). Function was

examined through tasks performed (DiVerneiro & Hosek, 2011; Dobos,

1988; Dobos, 1992; Godfrey et al., 2011; Junco & Cole-Avent, 2008; Kink

& Hess, 2008; Lee et al., 2011; Moran et al., 2011; Philippe et al., 2001;

Pirani & Sheehan, 2009; Turner & Reinsch, 2010). Satisfaction or GO with

one or several communication channels was rated by respondents using

Likert scales (Ahmad et al., 2010; Bagdasarov et al., 2010; Fay & Kline,

2011; Kink & Hess, 2008; Madlock & Booth-Butterfield, 2012; Nordin et

al., 2011; Rogelberg et al., 2010; Siraj, 2007).

Research Designs in the Literature

The majority of the studies reviewed in this chapter used a

quantitative, crosssectional survey design. Most of the authors

acknowledged that a limitation of this design is the difficulty of determining

causality (Al Eslami Kandlousi et al., 2010; Cho et al.,

2011; Einwiller et al., 2010; Junco, 2011; Junco & Cotton, 2011; Junco &

Cotton, 2012; Mondi et al., 2011). Junco and Cotton (2012) suggested the

need for future longitudinal and controlled studies, from which the causality

of observed relationships might be determined.

Another limitation of the survey approach was self-reporting (Fonner

& Roloff, 2010: Junco, 2011; Junco & Cotton, 2011; Junco & Cotton, 2012).

People were prone to over or underestimate their skills, to please the

researcher, or give a correct response. They might also misremember or

misreport the time they spent and the frequency with which they used

channels, as well as sometimes being unclear about tasks, and competences.

However, it would be difficult to record these variables more accurately

unless users could be tracked electronically while they used communication

channels. This might be practicable for telephone, e-mail, or the Internet,

since accounts can be automatically controlled. With written correspondence

it is more complicated, for instance it might require timing subjects with a

stopwatch and such unavoidably overt observation might change the way

people work or the time they spend on tasks. Thus, even measuring time

under strict conditions may still not accurately determine time spent on tasks

and channels.

Research Design for this Study

Like many of the studies reviewed above, I used a cross-sectional

survey design, which is acknowledged as the method of choice for showing

the status quo in a real-life situation (Frey et al., 2000). The instrument used

for this survey contained a demographic section, together with elements

from existing instruments developed by other research groups. Nineteen

questions from Downs and Hazen’s (1977) CSQ were used unchanged. The

questionnaire items used by Dobos (1988) to study the selection of

communication channels were modified to identify individual channels,

rather than groups of channels. These instruments have been used

extensively elsewhere and were considered reliable and valid for the present

study.

Summary and Transition

In this chapter, I examined the relevant literature in depth through the

categories of duration, frequency of use, function, and GO in

communication research studies. I also identified various authors’

suggestions for extending their research, some of which this study proposed

to undertake. In the next chapter, I examine the intended research design,

including discussions of cross-sectional survey methodology and of the

measurement tools that were employed. Reliability and validity issues are

also addressed, as well as research ethics and IRB requirements. The chapter

concludes with an explanation on how the survey questionnaire was drafted.

Chapter 3: Research Method

Introduction

This chapter begins with a description of the chosen research design as

it relates to the problem statement. I then discuss the research setting and

sample and justify the sampling strategy. I address data collection and

analysis in relation to the hypotheses and variables that formed the basis of

the study and consider the statistical procedures used to test the research

hypotheses. The chapter concludes with sections on the protection of

participants and dissemination of findings.

Research Design and Approach

Research Design and Research Questions

The overarching RQ was: using multiple linear regression, can Y

(gratification obtained) be predicted in terms of three independent variables

(frequency of use, duration, and function)? Lower level RQs are as follows:

RQ1: Are there gender differences when determining whether Y (GO)

can be predicted in terms of three independent variables (frequency of use,

duration, and function)?

RQ2: Does the number of years of work experience affect whether Y

(GO) can be predicted in terms of three independent variables (frequency of

use, duration, and function)?

RQ3: Does the communication channel chosen affect whether Y (GO)

can be predicted in terms of three independent variables (frequency of use,

duration, and function)?

I used a cross-sectional survey design to address the research

questions. A crosssectional design involves a one-off survey where a random

sample of individuals responds to a set of questions about their backgrounds,

experiences, and attitudes (Frankfort-Nachmias & Nachmias, 2008). The

research presented here was based on a sample of employees who had both

present and past experience using communication channels and were,

therefore, likely to have opinions (e.g., attitudes) about different channels.

While these employees graduated from hospitality programs in an

international hospitality school in Switzerland, their jobs, positions,

responsibilities, and environments differed. Like those of hundreds of

hospitality management programs throughout the world, this population was

eclectic; graduate positions extended beyond the hospitality milieu including

positions in finance, banking, hotels, restaurants, events management,

marketing, and so on. The population in this study provided an overview of

employees in various industries and positions and their perceptions of GO

when choosing communication channels, thus offering an accurate account

of the current communication channel choice in the workplace. I assumed

that all participants were currently employed, used communication channels

at work, and had opinions about them.

There are positive and negative reasons to choose a cross-sectional

design. According to Frey et al. (2000) and Neuman (2007), a cross-

sectional design is the simplest and least costly alternative for conducting

social research studies and is, therefore, the most frequently used. Cross-

sectional surveys typically describe the status quo, but do not permit

inferences about social processes or change because they refer only to one

point in time, and in principle the same population may respond differently

if the survey is repeated. Nevertheless, it is the approach used most

frequently in communication research based on uses and gratifications

theory, but two limitations have been repeatedly cited: (a) that it is

impossible to determine causality, and (b) that participants’ self-reporting

may produce inaccurate results. For instance, respondents might over or

underestimate time spent using communication channels, especially as their

state of mind at the time of taking the survey can affect the response

(Ruggiero, 2000). In addition, cross-sectional surveys measure intended or

reported behavior, rather than actual behavior (Jensen, 2012).

A cross-sectional survey was appropriate for two reasons: (a) I aimed

to provide a general overview of the status quo of communication channel

choice in the workplace, and (b) a study over time might have complicated

the situation because new communication channels are continually

introduced into the workplace. To summarize, the present study on

communication in the workplace was in the domain of social science

research and comprised a one-off survey of a random sample of individuals

from a larger population of graduates of an international hospitality school in

Switzerland. I intended to record their attitudes toward the specific topic at

one particular moment in time.

Advantages of a cross-sectional design include the practicality of

using a real-life situation such as communication in the workplace, and not

requiring the random assignment of individual cases because all employees

are capable of responding to the survey (Frankfort-Nachmias & Nachmias,

2008). However, to conduct linear regression, a random sample was

necessary. As no specific position, tenure, title, or company was targeted,

any graduate who was in employment could have been chosen at random to

respond to the survey questionnaire. For example, the electronic mailing list

of over 10,000 active alumni was used, from which the sample necessary for

conducting the linear regression was selected at random. The population and

sample will be discussed further in a dedicated section below. Disadvantages

of this approach include poor control of rival explanations (for instance,

gratification may not be the most important factor in employees’ choice of

communication channels) and the fact that the direction of causality must be

inferred (Frankfort-Nachmias & Nachmias, 2008). Inferred causality will be

addressed in Chapter 5.

To test the questions for the final survey questionnaire, a focus group

was undertaken with employees who worked at the hospitality school where

the study was conducted. Participants in the focus group, some of whom

were alumni, were currently employed (at the school) and had used

communication channels extensively in their workplace, thus making them

an appropriate population for the focus group. Their feedback, which led to

changes in the initial survey questionnaire, will be discussed in greater detail

later in Chapter 3 and Chapter 4.

The questionnaire used in the present study asked participants to rate

the GO they obtained from specific communication channels on a series of

Likert scales. There was, therefore, a risk that they might under or

overestimate their gratification with each channel or were not aware of their

gratification. Duration and frequency of use might also have been under or

overestimated. These risks were considered when the results were analyzed,

but the practical solution adopted was to offer ranges of time, such as

increments of 30 minutes, or ranges of frequency, such as every day, >once

a day, once a week, or once a month. The time increments made it easier for

participants to estimate what they had done, although this did not directly

solve the problem of under or overestimation, and it reduced variance in the

data.

Another concern was that employees might use the survey for ulterior

reasons, for instance, to complain about communication channels,

colleagues, or bosses. Employees may have felt misled that the survey did

not allow them to complain about existing problems, such as poor workplace

communication. They may have also felt obliged by management to respond,

thus doing so with a negative attitude. They might also have given socially

correct responses to appear more efficient, for instance, by underestimating

the duration of communication tasks or overestimating their communication

skills. For all of these reasons, the purpose of the survey was made clear on

the consent form and was stated in the introduction preceding the

questionnaire (See Appendix E). This consent form served to introduce the

researcher and the purpose of this research project. It explained how the

findings would later be used. Random participants who received the email

directly from the alumni association who agreed to participate in my survey

questionnaire were immediately directed to the consent form. It was only by

clicking on the link to the survey at the end of the consent form that

participants had full access to the survey questionnaire.

As well as estimating duration and frequency of use for each

communication channel, participants were asked to rate GO according to the

three functions of communication channels discussed under Research Design

in Chapter 2: production, maintenance, and innovation. This required that

participants understood each of these functions. Based on feedback from the

focus group, the original three functions were defined on the survey

questionnaire as five potential responses: for production (giving information,

receiving information), for maintenance (establishing new relationships,

maintaining relationships), and for innovation (brainstorming new ideas).

Participants were asked to tick as many of these as necessary for each

communication channel.

It was necessary to reassure employees of their anonymity, explain the

survey’s purpose, and state how it might affect them. Employees were

informed in the consent form how the survey results would be used and were

offered an opportunity to read the results at the end of the study (See

Appendix E). The focus group was also able to add further comments and

opinions about the survey questionnaire before the final version was sent out

by e-mail.

Variables

Previous communication researchers, discussed in Chapter 2,

examined a number of variables including frequency of use, tasks

performed, function, satisfaction, age, previous knowledge, place, gender,

education, motivation, level of awareness, perceptions, and Internet skills. In

the present study, I measured three independent variables: frequency of use,

duration, and function, which were tested for their predictive value of GO.

Previous scholars have limited the number of communication channels

investigated, or have aggregated channels into groups. For my study, I

included as many as possible of the individual communication channels

actually used in the workplace, which were tested individually.

Independent variables.

The overarching research question (RQ) was: using multiple linear

regression, can Y (gratification obtained) be predicted in terms of three

independent variables (frequency of use, duration, and function)? The

corresponding H0 can be stated as:

H01: R = 0; linear regression is a good fit.

H02: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y.

Furthermore, the overarching RQ and hypotheses were broken down

into lower level RQs and hypotheses which were analyzed in terms of

various demographic characteristics, namely gender and work experience.

Independent variables for the present study were frequency of use, duration,

and function. These are specified below:

Frequency of use is an independent variable on a ratio scale that is

discrete (ranging from never to >once a day). I measured frequency of use

with 15 individual communication channels and tested frequency for all

hypotheses using multiple linear regression. Illustrates this process, showing

how the independent variable, frequency of use (discrete), was tested, where

it appeared on the survey questionnaire (Part 4 A, 15 communication

channels), how it was assessed (1-6), and which statistical test was used

(multiple linear regression). Figure 1 also shows how the dependent

variable, GO, (discrete) was tested, where it appeared on the survey

questionnaire (Part 4 B, 15 communication channels), how it was assessed

(1-7 Likert scale), and which statistical test was used (multiple linear

regression). The values for this variable were coded from 1 for never to 6 for

>once per day. Dummy coding was used with 1 for never up to 6 for >once

per day. I conducted multiple linear regression using the sum of the scores of

frequency of use of all channels, duration and function, to determine whether

frequency of use predicted GO. After analyzing the overall results of the

three independent variables and GO for the overarching hypotheses, the

results for gender, work experience, and individual communication channels

were analyzed as well.

Duration is an independent variable, also measured on a ratio scale

and discrete (ranging from never to >4 hours per week). I measured duration

with 15 communication channels and tested it in all hypotheses through

multiple linear regression. Figure 2 shows how the independent variable,

duration (discrete), was tested, where it appeared on the survey questionnaire

(Part 3A, 15communication channels), how it was assessed (1-6), and which

statistical test was used (multiple linear regression). Figure 2 also shows

how the dependent variable, GO, (discrete) was tested, where it appeared on

the survey questionnaire (Part 3B, 15 communication channels), how it was

assessed (1-7 Likert scale), and which statistical test was used (multiple

linear regression). The values for this variable were coded from 1 for never

to 6 for >4 hours per week. Dummy coding was used by giving the value of

1 for never and 6 for >4 hours per week. Multiple linear regression was

initially run with the sum of the scores for duration of 15 communication

channels, frequency of use, function, and GO to determine if duration

predicted GO. After analyzing the overall results of the three independent

variables and GO for the overarching hypotheses, the results for gender,

work experience, and individual communication channels were analyzed as

well.

Functions (production, maintenance, and innovation) is an

independent variable on a nominal scale. Figure 3 shows how the

independent variable, functions (discrete), was tested, where it appeared on

the survey questionnaire (Part 5A, 15 communication channels), how it was

assessed (1-5), and which statistical test was used (multiple linear

regression). Figure 3 also shows how the dependent variable, GO, (discrete)

was tested, where it appeared on the survey questionnaire (Part 5B, 15

communication channels), how it was assessed (1-7 Likert scale), and which

statistical test was used (multiple linear regression). I measured function

with 15 communication channels and tested it in all hypotheses through

linear multiple regression. In order to run the linear multiple regression,

functions were defined as five tasks: giving information, receiving

information, establishing new relationships, maintaining relationships, and

brainstorming new ideas. The sum of the number of functions for each

communication channel was calculated and used as a variable to indicate

function, the assumption being that the greater the number of functions used

for a given communication channel, the better function predicts GO for that

channel. Multiple linear regression was initially run with the sum of function

scores for all 15 communication channels, duration, frequency of use, and

GO to determine if function predicts GO. After analyzing the overall results

of the three independent variables and GO for the overarching hypotheses,

the results for gender, work experience, and individual communication

channels were analyzed as well.

Dependent variable.

The dependent variable was GO from communication channels in the

workplace which could be measured on an interval scale. I measured GO for

each communication channel through 19 questions in Part 2 of the survey

questionnaire (Q 1-19). As discussed above, this was derived from the CSQ

of Downs and Hazen (1977), which measured GO on a scale of 1-7 from

very dissatisfied to very satisfied. GO for each of the 15 communication

channels was measured using the scale of 1-7 on the survey questionnaire in

part 3B for duration, 4B for frequency, and 5B for function, the values being

coded from 1 for not gratifying at all to 7 for extremely gratifying. Multiple

linear regression was used to test all hypotheses.

Time and Resource Constraints

The survey questionnaire was sent out by e-mail in March 2014,

participants being given two weeks to respond to the initial e-mail request.

After the two week time frame, 332 responses were received. Response rates

are discussed in more detail in the section on data collection in Chapter 4.

Survey instruments appropriate for this study were identified from the

literature review in Chapter 2 and will be discussed in more detail below.

Since they were required in both English and French, the translation

department at the international hospitality school in Switzerland ensured

accurate translation and back-translation of the questionnaire.

Methodology: Setting and Sample

Population

The population for the present study consisted of individuals who

graduated from a single international hospitality school in Switzerland and

were currently in employment. As an employee, I contacted the school’s

Alumni Coordinator who has access to the school’s database of alumni and

the facilities to send an e-mail survey directly to their most recently recorded

address. The population for the present study was therefore chosen on a

convenience basis, but was purposeful to the extent that all participants were

employed in jobs likely to require regular and extensive business

communication. Their exact job titles, geographic locations, and ages were

irrelevant to the present study. The only two factors taken into account were

alumni membership and active employment.

The number of registered alumni was over 25,000 at the time in

question, but this study addressed a subgroup who had chosen to stay in

contact with the school, called the active alumni, which numbered about

10,000. Members of this group participate in on and off campus alumni

events throughout the world and they tend to take an interest in changes and

developments in the school and its academic programs. They often employ

students for internship positions in their companies and are generally

motivated to support the school and its faculty’s activities. For these reasons,

active alumni were the best group to target for this survey research. To

ensure the sample was random, each of the individuals on the database was

given an equal probability of being selected

(Frankfort-Nachmias & Nachmias, 2008).

The alumni department can provide a full list of alumni e-mail

addresses or create various sublists of alumni for specific research projects

on request. For example, they are able to limit mailings to only general

managers or only F&B staff. In the case of the present study, there was no

need to specify such criteria. Any of the alumni could be selected at random.

Furthermore, the alumni department sent the survey on my behalf using a

random sample, so I had no role in sending out the e-mail questionnaire.

They also have strict rules forbidding researchers access to their alumni

database. The alumni database is frequently solicited for research projects,

so the alumni department limits the number of surveys sent to one per month

and they do not send out reminder e-mails as they feel this is invasive and

annoying. The final response rate for my survey questionnaire sent by e-mail

with no reminder was just under 4%.

Sampling and Sampling Procedures

Because the population of 10,000 active alumni was sampled on a

random basis, there was no predisposition toward type of position, sector,

company, or country; any position in any industry in any country was

acceptable. Because all respondents had an equal chance of being chosen,

sampling was considered random. This was a necessary prerequisite for

conducting the regression analyses necessary to test the research hypotheses.

Sample Size

The G*Power tool was used to define an optimal sample size of 77 for

multiple linear regression (Faul, Erdfelder, Buchner, & Lang, 2009). G*

Power is a free tool which computes statistical power analysis for many

different tests including multiple linear regression

(http://www.gpower.hhu.de/). To calculate sample size, I followed the three

steps in the G*Power manual:

In Step 1 for choosing the statistical test, I chose Linear multiple

regression: Fixed model, R² deviation from zero.

In Step 2, from the Type of power analysis menu, I chose the first

item which displayed input and output parameters appropriate for an a

priori power analysis. In A priori, sample size is computed as a function of

three things: (a) the required power lever (1 – β); (b) the pre-specified

significance level α; and (c) the population effect size to be detected with

probability 1 – β (Faul, et al., 2009, p.3). In this case, I chose a medium

effect size of .15.

In Step 3, I provided the input parameters required for the multiple

linear regression analysis (See Figure 4.). The main output parameter of the

type of analysis selected in the main window is by default selected as the

dependent variable y. In an a priori analysis, for instance, this is the sample

size as seen below (Faul, et al., 2009).

Thus, to calculate this number, the effect size was .15, Alpha =.05,

Power = .80 and predictors (or independent variables) = 3 (See Figure 4.).

The actual power is .8018 which is an acceptable number for multiple linear

regression analysis (Faul et al., 2009).

As seen above in the G*Power analysis, 77 respondents were

necessary for each variable which was being measured (Faul et al., 2009).

To assess whether there were gender differences in communication channel

choice and GO, the ideal sample of 77 was multiplied by 2 (for the 2

genders) giving a requirement of 154 respondents. To assess the effect of

work experience on communication channel choice and GO, the ideal

sample of 77 was multiplied by 6 (for 6 work experience ranges) giving a

requirement of 462 respondents. To reach the number of respondents

necessary to test work experience (462), a far greater number than 462

survey questionnaires would have needed to be sent out to ensure the ideal

sample of 77 per independent variable. With the actual response rate of 4%,

11,550 participants would have been necessary to reach the number of 462

respondents. This exceeds the total number of active alumni and was not

possible. With a random sample, there was no way to ensure that work

experience and gender were equally distributed.

Colleagues who have conducted research among the same active

alumni sample have reported response rates in the range 5-16%. For a

response rate of only 5%, it would have been necessary to send

questionnaires to 9,240 participants to get the 462 respondents necessary. At

the average response rate (10.5 %), 4,400 questionnaires

would have been required. In fact, the survey questionnaire was sent out to

8,467 active alumni and a 3.92% response rate was obtained.

Eligibility of Participants

All participants were equally eligible to respond as long as a working

e-mail address was available. The questionnaire was sent in French and

English and respondents were allowed to respond in the language with

which they were most comfortable. It was assumed that all were adults, over

the age of 21. Alumni who were currently unemployed were not considered,

but any active alumnus who was currently employed, regardless of job title

position, or company was eligible to participate in this study. In the

introduction to the project, it was made clear that the study was intended for

active alumni who were currently employed. Individuals who did not fit this

description did not continue completing the questionnaire. Question 3 on

whether the participant was employed was an eliminatory question.

Respondents who answered no to Question 3 immediately received a

message which thanked them for their time and participation. They were not

permitted to continue completing the survey questionnaire.

Characteristics of the Selected Sample

As mentioned above, the survey questionnaire was available in both

English and French. Although the selected sample of active alumni

graduated from the same international hospitality school in Switzerland, they

might have followed different academic programs or studied in English or

French. Until 1996, there was no English program, so the preponderance of

participants in this research project may have been graduates of the French

programs. After conducting the survey, this was found to not be the case as

more French participants clicked on the initial survey link (6.37% for

French;

5.77% for English), but more English participants completed the final survey

(181 English respondents; 151 French respondents).

The school was the first hotel school in the world, founded in 1893,

and had a predominantly male student population for its first 70 years. It was

only after the introduction of housekeeping and receptionist programs that

women began to attend the school. Since the late 1990s, gender numbers

have leveled and the school administration now ensures an equal proportion

of men and women in its academic programs. Thus, more recent graduates

might be men or women while older graduates would be predominantly

male. There was a possible risk of bias. There was also a risk that gender

categories would not have sufficient numbers of participants, 77 for each

gender.

Instrumentation and Materials

Data Collection Tools/Instruments

Nineteen questions from the CSQ (Downs & Hazen, 1977) were chosen for

inclusion in the questionnaire used in the present study. This instrument is in

the public domain, and no supplementary permission was needed to use it.

Although it focuses on communication and the satisfaction (or GO) which

comes from it, it has been used to test individual communication channels.

The CSQ has demonstrated consistently strong reliability in previous studies

where communication channels were grouped (Al Eslami

Kandlousi et al., 2010; Bakanauskiene et al., 2010; Nordin et al., 2011;

Rogelberg et al., 2010) and it has been shown to be appropriate for

measuring satisfaction with communication functions. The 19 questions

from the CSQ were selected as Part 1 of the present research questionnaire

(See Appendix B) as they provided information about the information

function of communication channels and the satisfaction derived from them.

Following the 19 CSQ questions in Part 1, the questionnaire then

targeted 15 specific channels through duration, frequency of use, functions,

and GO (See Appendix

C). For function, items from Dobos’s (1988) questionnaire were used to

measure satisfaction with specific communication channels (See Appendix

A). This originally employed a 4-point scale to measure satisfaction based

on three organizational functions: production, maintenance, and innovation.

Dobos’s study did not examine the GO obtained from each channel. Dobos

analyzed the responses in three aggregated channel categories: oral, written,

and electronic. The present study examined the GO derived from individual

communication channels. Permission to use Dobos’s questionnaire in a

doctoral thesis was granted on August 7, 2012 by the Routledge Taylor &

Francis publishing house.

The CSQ measures satisfaction with communication functions using a

7-point scale. I retained the same scale for the present study in order to

enable comparisons with previous research. To measure frequency of use,

participants were asked to choose from a list of responses from never to

>once a day in a typical work week. Duration was measured in hours from

never to >4 hours per day. As seen earlier, the three functions were defined

as five tasks: giving information, receiving information, establishing new

relationships, maintaining relationships, and brainstorming new ideas in.

Participants could choose as many functions for each channel as applicable.

GO was rated on a 7point Likert scale. The 4-point scale used in Dobos’s

instrument to measure GO was replaced with a 7-point scale in the present

questionnaire in order to preserve comparability between the different

sections.

Reliability

Uses and gratifications theorists employed Downs and Hazen’s

(1977) CSQ scale items and their reliability and validity in their research

projects were clearly established.

Cronbach’s Alpha is the most common measure of scale reliability, which

calculates the variance within the item and covariance between a particular

item and other items in the scale and evaluates how strongly the individual

items on the scale are connected (Frankfort-Nachmias & Nachmias, 2008).

Alpha values of .7 to .8 are generally regarded as acceptable, and a value

substantially lower than this indicates an unreliable scale

(Field, 2009). Coefficient alpha values for Downs and Hazen’s eight

dimensions of satisfaction with communication have been consistently high,

ranging from .72 to .96 (Rubin et al., 2004).

Focus group.

To adapt existing survey designs to the purpose of this research

project, I made various changes. First of all, participants were required to

self-report the estimated time spent on each of the communication channels

(duration) and the frequency of use. A Likert scale response was not

appropriate for this type of inquiry. To test these new elements and ensure

that no communication channels were omitted, a focus group was conducted

prior to drafting the final survey. In an informal 90-minute discussion with

employees at the international hospitality school in Switzerland, the

following questions were asked:

1. How would you define a communication channel?

2. How many do you use each day?

3. Name the top three you use (by frequency) in the workplace.

4. Name the top three you use (by frequency) at home.

5. How would you use communication channels for production?

6. How would you use communication channels for maintenance?

7. How would you use communication channels for innovation?

8. How would you define the three functions of communication

channels:

production, maintenance, and innovation?

9. What does the word gratification mean to you?

10. Which communication channel offers you the most gratification in

the workplace? How do you rate this?

11. With all things equal, if you could only choose one communication

channel for the workplace, which one would you choose?

12. Is there a communication channel that you would never use? If so,

why not?

Based on the focus group feedback in February 2014, I readjusted the

questionnaire items as necessary. The changes included defining the

functions into five subcategories: giving information, receiving information,

establishing new relationships, maintaining relationships, and brainstorming

new ideas. A second change involved removing 24 questions from the CSQ

which were not relevant to this study, as they asked respondents to evaluate

supervisors or personal relationships at the workplace. Only the 19 questions

from the CSQ regarding communication function were selected for the

present study.

The actual questions were piloted with a convenience sample of

colleagues who worked at and were also alumni of the international

hospitality school in Switzerland. This exercise confirmed that the survey

questionnaire took approximately 20 minutes to complete. Once their

feedback was addressed, participants in the initial focus group received the

edited version for validation. They confirmed that the changes reflected their

comments.

Survey design.

It was proposed to retain the 19 questions of the original Downs and

Hazen’s (1977) CSQ survey instrument which were linked to the function of

communication channels. From Dobos’s (1988) study on new media and

traditional channel choice, the original functions of production, maintenance,

and innovation were retained but defined as the following specific tasks: for

production (giving information, receiving information), for maintenance

(establishing new relationships, maintaining relationships),and for

innovation (brainstorming new ideas). Respondents were asked to rate the

GO of each communication channel based on the importance of each of the

functions when choosing a communication channel. The latter was important

for comparing specific results in production, maintenance, and innovation

with overall GO.

Part 1 of the questionnaire elicited demographic information,

including year of graduation, nationality, size of company, years in the

company, and language (French or English) in which the respondent studied

when in school. Question 3 in Part 1 asked if the respondent was currently

employed. If the respondent was not currently employed and responded no

to Question 3, the questionnaire stopped there as it was an eliminatory

question. This was a setting chosen in SurveyMonkey to stop participants

who were ineligible to answer the entire survey questionnaire. The 19

questions on function from

Downs and Hazen’s (1977) CSQ questionnaire were included in Part 2. Part

3 included the hours per week spent on each communication channel from

never to >4 hours and a 7-point scale to rate GO for duration. Part 4 offered

a choice of communication channels and the frequency of use of each

expressed in ranges from monthly to >once a day and a 7-point scale to rate

GO from frequency of use of communication channels. In Part 5,

participants were asked to choose among 5 functions for each channel as

well as rate GO from function on a scale of 1-7. The final version of the

questionnaire for this research project can be found in Appendix C.

Testing and Validity

I conducted multiple linear regression on all hypotheses to ascertain

how well frequency of use, duration, and function predict GO. The premise

of the present study was that the greater the frequency, the greater the

duration, and the more the functions used for a given channel, the greater

would be the GO for that channel.

Regression makes four assumptions which must be addressed, namely

variance, linearity, independence, and normality. Residuals, i.e. differences

between observed and predicted values, can be used to check for violations

in the assumptions. To determine how much variance of the dependent

variable can be accounted for by each independent variable, an F test was

run to calculate multiple R2 and see how well the model fitted in the

population (Norusis, 2008). To check normality, linearity, and

independence, histograms were created by plotting the residuals to see if the

distribution was normal with a horizontal band of residuals. If there were

patterns, or changes over the range of the independent variable, the

relationship might not have been linear (Norusis, 2008) and the assumptions

might have been violated. In the case of violated assumptions, the data

would have needed to be transformed. A Durbin-Watson test was also run to

test correlation of adjacent residuals (Norusis, 2008).

Threats to Validity

For quantitative research design to be valid, an instrument must

measure what it claims to measure. Common internal threats include history,

maturation, instrumentation, and selection, while external threats often

include population generalization, environment and time (Frankfort-

Nachmias & Nachmias, 2008). I will address various kinds of validity,

including external, internal, construct, and statistical conclusion validity and

potential solutions for reducing these risks in the following sections.

External validity.

When conducting research based on a survey questionnaire, some of

the threats to external validity include population generalization,

environment, and temporal concerns. Race and cultural bias were not issues

in the present study because the topic of choosing communication channels

in the workplace was not felt to be dependent on these factors. Group power

or the influence of one participant over another was likewise not considered

a threat, because of the isolation and heterogeneity of respondent group

members.

There was a risk that participants who were chosen randomly from the

database of alumni might have worked in the same department or company

or have graduated within a close timeframe. Potentially, this might have

been the greatest threat to external validity, and although in principle it could

have skewed the results, it was a necessary corollary of random sampling. It

seemed relatively unlikely to occur as questionnaires were randomly

distributed among active alumni throughout the world. The topic of

communication channels is pertinent, but not controversial. Participants were

not asked to judge other people, as for the most part they were self-reporting

the gratification they receive when choosing communication channels. Thus,

there was little risk that this aspect would cause validity issues.

Internal validity.

Threats to internal validity in survey research include maturation and

experimental mortality, history, instrumentation and human error, statistical

regression, selection, diffusion and imitation of treatment. In this study, the

most important consideration was what might go wrong during the research

process. In previous research on uses and gratifications theory, causality

could not be determined, the direction of causation had to be logically or

theoretically inferred (Frankfort-Nachmias & Nachmias, 2008). This was a

weakness or limitation noted in many of the previous studies. Of the

internal validity threats, maturation and experimental mortality were not

issues as the survey was sent out at one point in time, and, once it had been

completed, the data were analyzed and finalized. History, however, could

have been an issue if a participant had had a negative communication

experience immediately prior to completing the survey. For instance, a new

communication channel, which was time consuming and difficult to use,

may have been introduced into the workplace. If employees who were faced

with that were asked to respond about gratification received from

communication channels, the response might have been negatively skewed

by the experience of using a new channel. However, such individual effects

were minimized by statistical treatment. Instrumentation was a minimal risk

as the survey was based on tested questionnaire items from prior research,

and the questionnaire was discussed and piloted with typical respondents

before use. Selection was also a risk even though a random sample was used.

From the original population, all participants had an equal chance of being

selected to respond to the survey. It was possible that the population

breakdown resulted in a disproportionate number of surveys being sent out

to the same country or company which may have skewed the results. This

could not be controlled as the survey was sent randomly.

Issues associated with cross-sectional design have already been

discussed. Another concern was that participants, who inevitably self-select

to some extent, might not have accurately reflected the greater population of

this hospitality school’s alumni. Further, they might not have accurately

reflected the views of hospitality school graduates as a whole.

Construct and statistical conclusion validity.

Construct validity denotes the fit between what an instrument intends

to measure and what it actually measures. The logical process of construct

validity is composed of four steps: (a) suggesting an instrument measures a

certain property; (b) inserting the proposition into a theory; (c) predicting the

properties which should be related to the instrument, and (d) collecting data

to confirm or reject predicted relations (FrankfortNachmias & Nachmias,

2008). If the relationships cannot be demonstrated, the instrument may be

considered invalid. These problems were, to a large extent, avoided by

basing the instrument for this research on past survey questionnaires, the

validity and reliability of which have been confirmed by many studies.

Statistical conclusion validity implies that conclusions drawn from the

statistical analysis accurately reflect reality. I attempted to do this by

ensuring random sampling of the general population of hospitality program

graduates, which satisfied the requirements for the statistical tests employed

as well as ensuring as far as possible that responses of the sample accurately

reflected those of the population. As much as possible, the population used

in the present study reflected the reality of alumni from one hospitality

school in Switzerland.

Data Collection and Analysis

Data Analyses

Data were analyzed using SPSS for Windows 20. Analysis involved

the calculation of descriptive statistics such as means, standard deviations,

and modes, and cross-tabulation with demographic information, including

age, gender, nationality, language (French or English), graduation year,

number of years working in the present company, and job position or title.

Hypotheses

Regression was used to test the research hypotheses, listed below:

The overarching research question (RQ) was: using multiple linear

regression, can Y (gratification obtained) be predicted in terms of three

independent variables (frequency of use, duration, and function)? The

corresponding H0 can be stated as:

H01: R = 0; linear regression is a good fit.

H02: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y.

Furthermore, the RQ and hypotheses were analyzed in terms of

various demographic characteristics, namely gender and work experience.

Therefore, lower level RQs and corresponding hypotheses were specified.

An example is provided for Gender: RQ1: Are there gender differences

when determining whether Y (gratification obtained) can be predicted in

terms of three independent variables (frequency of use, duration, and

function)?

H01M: R = 0; using only Male data, linear regression is a good fit.

H01F: R = 0; using only Female data, linear regression is a good fit.

H02M: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using Male data only.

H02F: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using Female data only.

Other examples include

RQ2: Does the number of years of work experience affect whether Y

(gratification obtained) can be predicted in terms of three independent

variables

(frequency of use, duration, and function)?

H01W (work experience): R=0; using ranges of work experience

(expressed in years), linear regression is a good fit.

H02W: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using work experience (expressed in years).

RQ3: Does the communication channel chosen affect whether Y

(gratification obtained) can be predicted in terms of three independent

variables (frequency of use, duration, and function)?

H01C (communication channel): R=0; using the communication

channel chosen, linear regression is a good fit.

H02C: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using the communication channel.

Table 1

Hypotheses and variables

Hypotheses Gender Work

Experience

(in years)

Channel Frequency

of use

Duration Function GO

H01M X X X X X X

H01F X X X X X X

H02M X X X X X X

H02F X X X X X X

H01W X X X X X X

H02W X X X X X X

H01C X X X X X

H02C X X X X X

I used regression to assess significant relationships between

independent and dependent variables in all the hypotheses as seen in Table

1. To test the independent variables as to which has the most significant

relationship and is the best predictor of GO, multiple linear regression was

used and β values compared. To test all hypotheses, the same multiple linear

regression model analysis was run on frequency of use, duration, function,

and GO for 15 communication channels used in the workplace. To test the

lower level hypotheses, multiple linear regression was run on the same

variables, but the sample was split into two genders and six work experience

ranges respectively.

Data Collection and Analysis

The survey questionnaire was sent by e-mail in March 2014 after

which the data were entered and stored in SPSS. The first tests run were

descriptive analyses to get reports on data status. No major problems with

the accuracy or quality of measurement appeared. Initial data screening

ensured that responses were legible and complete, all relevant questions

were answered, and relevant contextual information was included.

Data were coded to show where and how it can be accessed. Once the

data were entered, statistical tests were conducted to confirm or disprove the

null hypotheses. As the present study was based on rating gratification and

assessing the duration and frequency of use of communication channels, all

responses were numeric. Variables were named in a clear, coherent manner

so that the tests could be run. For example, frequency of use was named as

Frequency; duration was Duration. The three functions

(production, maintenance, and innovation) represented by five specific tasks

were aggregated and their total was named as Functions. Gratification

obtained was named as

GO. Under the label column, a longer explanation of each variable was

entered. Variables such as frequency of use and duration, work

experience ranges, communication channels, and gender were coded as seen

in Table 2. On the survey, duration was worded as never to > 4 hours. In

SPSS, never was defined as 1, < 1 hour as 2, 1-2 hours as 3, 2-3 hours as 4,

3-4 hours as 5, and > 4 hours as 6. The same procedure was used with

frequency of use which ranges from never to > once per day. Work

experience was aggregated into 6 groups ranging from <5 years (1) to >30

years (6).

Gender was coded as 1 for male and 2 for female.

Table 2

Example of entering variables in SPSS

Name Type Label Measure

1 Y Numeric GO Scale

2 X1 Numeric Frequency of Use Scale

3 X2 Numeric Duration Scale

4 X3 Numeric Function Scale

5 Survey ID Numeric Participant ID Scale

6 Gender Numeric Gender Nominal

7 YrsWorkExperience Numeric Ranges work experience Scale

8 Communication

channels

Numeric 15 channels Nominal

Once the data were entered, multiple linear regression was run in the

following steps:

Step 1: I clicked on Analyze, Regression, and Linear. Gender, work

experience, communication channel, frequency of use were entered into

independent box. GO was placed in the dependent variable box.

Step 2: I clicked on Statistics. In this box, Confidence Intervals (95%),

Descriptives, Estimates, and Model Fit were selected. Once this was done, I

clicked Continue and OK. This provided the output for the multiple linear

regression.

Step 3: To analyze, the slope was assessed to see if the population slope was

equal to 0. I looked at the ANOVA table. If p value > alpha, we cannot reject

the null hypothesis. Thus, linear regression is not a good fit.

Step 4: I created a scatterplot by clicking Graphs, Legacy Dialogs, Simple

Scatter, and Define. For H1, GO was moved in the Y-axis box; frequency of

use was moved into the X-axis box before clicking OK. A regression line

was created by clicking on Elements, Fit Line at Total, and Close. The

scatterplot allowed me to assess how accurately the regression equation

predicted the dependent variable scores.

Step 5: I created a plot of predicted and residual values. Once the data were

entered, and the specified statistical tests were run, graphic representations,

including histograms and scatter-plots were produced. Histograms were

useful for recognizing violations of linear regression assumptions.

Scatterplots were used to examine relationships between variables. I clicked

on Analyze, Regression, Linear (with the same settings as Step 1). Then I

clicked on Plots in the linear regression dialogue box. I moved GO into Y-

axis and Frequency of use into X-axis before clicking Continue and OK. As

no apparent pattern appeared, I deduced that no assumptions have been

violated (Green & Salkind, 2011).

Protection of Human Participants

IRB International

The active alumni of the international hospitality school in

Switzerland live and work throughout the world, so IRB regulations for

international research had to be respected in this study, the definition being

any study intentionally designed to target individuals outside the U.S.

Although the language of informed consent is mentioned in international

regulations, this was not an issue for the present study since alumni studied

at the school in either French or English and were able to respond to the

survey in these languages. The school’s official, certified translator was

contracted to translate the survey. According to IRB, translated documents

must respect three criteria: (a) backtranslation confirms the accuracy of the

translation; (b) qualifications of translator and back-translator are

documented; and (c) translation and back-translation procedures are

documented. All of these steps were respected.

Other elements, such as permission to use human subjects, or consent

from parents were not issues in the present study, as there was no testing and

these surveys were sent out to participants who graduated from a Western

culture establishment, who could choose to respond to the survey if they

wanted to. While cultural differences must be respected, the topic of

gratification from communication channels is not dependent on participants’

social or economic backgrounds. Also, it is not offensive and does not

promote stereotypes in any way. Since all potential participants followed

academic programs at the same higher education institution, their abilities to

respond to the survey were equal.

Since survey questionnaires were sent by e-mail, no local resources

were required and no physical presence was needed. Participants were made

aware of the purpose of the research project and where and how the data will

be used. Specific ethical concerns are dealt with in the next section.

Ethical Concerns

Total anonymity could not be guaranteed as the identity of alumni

from the international hospitality school in Switzerland could be traced

through their e-mail addresses, although personal information such as names

was not requested. No attempt was made to identify individuals and

identification was not necessary for the study. Further, as the survey

questionnaire was sent out directly from the alumni department, I had no

knowledge of which alumni received the survey questionnaire or which ones

responded. To help ensure honest responses, a signed confidentiality

agreement was sent with the questionnaires (See Appendix D). In this way,

participants were assured that no third party will be privy to or use their

results in any way and that their data would only be used in the present

study. Confidentiality was not expected to be a serious problem when

participants are asked to assess their gratification with communication

channels. Although there were no correct answers to these questions,

participants may have been concerned that their responses will be made

known to company management. They were assured that this was not the

case.

Dissemination of Findings

An outline of this proposal was presented in September 2012 at the BiTS

Communication and Media Conference, at the Business and Information

Technology School in Iserlohn, Germany. In due course, it is intended to

disseminate results of the research at other European conferences. One such

conference will be hosted by ECREA, European Communication Research

and Education Association. As a member of this group, I am invited to

participate in conferences and submit papers. Another opportunity to present

aspects of the work may be the next annual EuroCHRIE (a professional

association that links hospitality and education research) conference. The

present study is relevant as it relates to the development of communication

in the hospitality management curriculum.

When this dissertation is complete, its findings will be made available

on the Walden database and there may be opportunities to publish the

findings in academic journals such as Communication Theory or the

Communication Quarterly. All findings will be made available to those who

participated in the study.

Summary and Transition

Chapter 3 began with a detailed description of the research design for

the present study. The population was comprised of alumni of a single

international hospitality school in Switzerland. I analyzed the data using

SPSS and evaluated the threats to validity which were minimal in this

research design. The measurement tools chosen were based on accepted and

valid tools used in the past by Downs and Hazen (1977) and Dobos (1988).

A section on protection of human participants addressed IRB requirements

for international research, ethical concerns, and confidentiality. The

appendices include the survey questions for the present study as well as the

sample confidentiality agreement.

Chapter 4: Results

Introduction

The overall purpose of this study was to measure how well the

independent variables frequency of use, duration, and function predict the

dependent variable, GO, when choosing communication channels. The

sample was derived from alumni from an international hospitality school in

Switzerland who responded to a survey on the communication channels they

used in the workplace. Chapter 4 begins with a discussion of the pilot

study/focus group and changes made to the survey questionnaire based on

their feedback. I then describe the data collection process. In the results

section, descriptive statistics are reported, followed by the findings of

multiple linear regression studies conducted using SPSS. Chapter 4

concludes with a summary of the responses to the research questions.

Research Questions and Hypotheses

The overarching research question (RQ) was: using multiple linear

regression, can Y (gratification obtained) be predicted in terms of three

independent variables (frequency of use, duration, and function)? The

corresponding H0 can be stated as:

H01: R = 0; linear regression is a good fit.

H02: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y.

Furthermore, the RQ and hypotheses were analyzed in terms of

various demographic characteristics, namely gender and work experience.

Therefore, lower level RQs and corresponding hypotheses were specified.

An example is provided for Gender:

RQ1: Are there gender differences when determining whether Y

(gratification obtained) can be predicted in terms of three independent

variables (frequency of use, duration, and function)?

H01M: R = 0; using only Male data, linear regression is a good fit.

H01F: R = 0; using only Female data, linear regression is a good fit.

H02M: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using Male data only.

H02F: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using Female data only.

Other examples include

RQ2: Does the number of years of work experience affect whether Y

(gratification obtained) can be predicted in terms of three independent

variables

(frequency of use, duration, and function)?

H01W (work experience): R=0; using ranges of work experience

(expressed in years), linear regression is a good fit.

H02W: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using work experience (expressed in years).

RQ3: Does the communication channel chosen affect whether Y

(gratification obtained) can be predicted in terms of three independent

variables (frequency of use, duration, and function)?

H01C (communication channel): R=0; using the communication

channel chosen, linear regression is a good fit.

H02C: b1=0, b2=0, b3=0; are the 3 independent variables needed in

explaining the variation in Y using the communication channel.

Pilot Study

After receiving IRB approval on January 28, 2014, I conducted a 90-

minute focus group on February 6, 2014 as a pilot study with colleagues who

worked at the international hospitality school in Switzerland whose

graduates were to be participants in this study. Following the protocol,

which was submitted to IRB, after the participants had read and signed the

consent form, I began with the 10 focus group questions already discussed in

Chapter 3, on communication channels and gratifications. I found that the

participants understood the terms communication channels and gratifications

as defined on the consent form. They were able to define communication

channels and gratifications and their definitions were considered sufficiently

similar to those listed on the consent form not to require changes to the

consent documents (except for adding the IRB approval number and

expiration date, according to IRB regulations).

Focus group participants were also able to list the communication

channels they used in the workplace. They understood the differences

between the three main functions of communication channels (i.e.,

production, maintenance, and innovation). However, when completing the

survey, they had difficulty remembering what the functions were or how

they should be assessed. A solution to this problem is addressed later in this

chapter.

After completing the survey, participants were invited to assess each

section of the survey and suggest how it might be improved. Two areas for

development were discussed, and are addressed in the next two paragraphs.

The first area concerned questions from the original survey

instruments from Downs and Hazen and Hecht, which I had included in their

entirety in the initial version of my survey. The participants did not

understand how all of the questions linked to communication channels or to

this research project. After recording their comments and speaking to my

mentor, I reduced the number of questions from the Downs and Hazen’s

CSQ survey questionnaire from 43 to 19 questions, choosing only those

which involved specific communication functions in the workplace

(production, maintenance, or innovation) and discarding those seeking

opinions about communication with coworkers or supervisors. I also chose

to omit the questions from Hecht’s survey asking for opinions about

communication with direct supervisors. Communications with a supervisor

were not tested in my hypotheses, so these questions were not considered

relevant to this survey questionnaire.

The second improvement area involved Part 6 of my original survey

questionnaire, which asked participants to assess the relevance of Dobos’s

three functions (production, maintenance, and innovation) for each

communication channel. Participants found this section long and confusing

and had problems remembering what the functions were and how they

should be rated. Two participants said they would have stopped completing

the survey at that point if they had received it at home because it was too

long. Two suggestions were made to improve this section. The first was to

add a definition of each function above the scale to remind them of their

meaning. The second suggestion was to change the format completely.

As discussed briefly in Chapter 3, I chose the second option. I defined

the three functions in terms of five separate activities and asked respondents

to tick the box (es) if they used the communication channel in these

specified ways. This was the only question where more than one response

was required. The functions were listed above the response format as Giving

information, Receiving information, Establishing new relationships,

Maintaining relationships, and Brainstorming. These are the definitions of

each of the function, but in a more comprehensible and user-friendly format

for the respondents. Respondents could choose as many functions as

appropriate for each communication channel. The sum of the choices was

entered as the independent variable, Functions Sum, in the multiple linear

regression. Thus, the functions to be tested in the hypotheses were present,

but in a clearer format for the participants.

A request for approval of these changes was ratified by IRB on

February 19, 2014, and the full revised survey is presented in Appendix C.

The survey questionnaire was also translated and back-translated by a

certified linguist, as discussed in Chapter 3.

Data Collection

Demographic Variables

The survey questionnaire was sent by e-mail via a SurveyMonkey link

(https://de.surveymonkey.com/) on March 14, 2014 and closed on April 4,

2014. The survey was sent out by the alumni office to 8,467 random e-mail

addresses of active alumni. Of the total, 8,460 reached the recipients

successfully and 3,484 e-mail invitations were opened. Five hundred and

twelve participants clicked on the link to SurveyMonkey (6.37% for French

survey; 5.77% for English survey). The final number of participants was 332

(181 from English survey; 151 from French survey) giving a 3.92% response

rate in total.

Those who responded to the English survey were not necessarily

American or British. As stated in Chapter 3, the international hospitality

school alumni population was derived from over 84 different nationalities.

Thus, those who responded to the English survey may have followed the

academic program at the school in English or just preferred to complete the

survey in English, and it was impossible to verify their nationality or where

they were located. As geographic location was not a variable in my study, it

was not necessary to analyze this further.

Table 3 shows the demographic characteristics of the study sample.

Among the 332 participants who opened the survey invitation and clicked on

the link, 207 were male, 117 were female, and eight did not give their

gender. Since gender was one of the variables to be tested in H01M and

H01F, it was necessary to consider the difference in gender numbers when

analyzing the results. Question 3 asked if the respondents were currently

employed. This was an eliminatory question to which 23 respondents gave a

negative response and were automatically excluded from the rest of the

survey. The number of respondents who did not finish the survey was 161.

Adding the number of respondents who were unemployed (23) gives 184

and this, subtracted from the original 332, left 148 complete questionnaires

suitable for use in the regression analysis. A breakdown of these is shown in

Table 3.

For the number of years of work experience which were tested in

H01W, 47 had worked <5 years; 28 had worked 6-11 years; 28 had worked

12-19 years; 13 had worked 20-24 years; 15 had worked 25-29 years; and 17

had worked >30 years. This was important when analyzing the results of

each range, as the majority of the respondents (n= 104; 70.3%) fell into the

first three categories (i.e., less than 20 years of work) while only 45 (30.4%)

occupied the other three categories. As ranges of work experience were

tested in the regression as a general predictor of GO, these numbers could

have caused reliability problems for this study.

Table 3

Demographic information of the 148 respondents

Demographic Number Percentage

Gender Male

Female

Not given

65%

35%

Years of work experience

<5 years

6-11 years

12-19 years

20-24 years

25-29 years

>30 years

32%

19%

8.8%

10.1%

11.5%

Type of organization

you work for

Hospitality

High Tech

Manufacturing

Education

Civil Service

Government

Other

58.8%

3.4%

2.4%

7.4%

1.3%

26.4%

Total 148 100.0%

Some conclusions can be made from the demographic information in

Table 3, relating to the hypotheses and overall purpose of this study. For the

hypotheses concerning gender, one issue was that the number of male

respondents was almost double that of females (95 male (65%) to 53 female

(35%)). This difference in number will be addressed later in Chapter 4.

However, the biggest concern came from numbers in the work experience

categories. Out of 148 respondents,103 (70%) were in the first three ranges

of work experience, suggesting that respondents were young professionals

with up to 20 years of work experience. For a dissertation on communication

channels including both traditional and modern channels, this high

percentage of young professionals might have distorted the results if they

were particularly keen on modern communication channels such as IM, e-

mail, and teleconferencing and used them significantly more frequently than

would older respondents. This was addressed when analyzing the statistics

relating to years of that work experience later in Chapter 5.

The most useful statistic from the demographic information was the

proportion of respondents who work in the hospitality industry. Out of 148

respondents who gave their workplace industry, 87 (58.8%) identified the

hospitality industry as their place of employment. While job title, position,

and company were not relevant for the hypotheses tested in the present

study, the overwhelming majority of respondents who have stayed in the

hospitality industry since graduation was significant. With this majority of

87 (58.8%) who have remained in the hospitality industry, the study could be

taken as a true representation of hospitality school graduates who often

choose to work in other industries as discussed in Chapter 3 under the

characteristics of sample population. A breakdown of the industries named

by respondents is shown in Table 3. The second highest choice, Other, n=39

(26.4%), could not be analyzed because it was not further specified.

Research Tools

For each of the hypotheses, I conducted multiple linear regression

with one set of predictors using the Enter method (Field, 2009) as the

variables did not need to be tested in a specific order. For the overarching

hypotheses, H01and H02, frequency of use, duration, and functions were

entered as independent variables while GO was entered as the dependent

variable. I conducted multiple linear regression and reported R2 change,

descriptives, part and partial correlations, collinearity diagnostics, Durbin-

Watson, and Casewise diagnostics with outliers outside 2 standard

deviations. I entered ZPred in X axis and ZResid in the Y-axis to create

histogram and part and partial plots.

For H01M and H02M and H01F and H02F, I split the cases into male

and female data before conducting the multiple linear regression. To test

H01W and H02W regarding ranges of work experience, the original data

were split to be able to compare the six different ranges. Finally, for H01C

and H02C, the sum of duration, frequency of use, and function for each

communication channel was entered separately into the independent variable

box for multiple linear regression.

Outliers

When initially conducting multiple linear regression with the model of

three independent variables (duration, frequency of use, and function) and

sample of 148 respondents, outliers did not appear to be problematic until I

saw the results. According to the original responses, the R2 showed a

variance or predictive power of .35 or 35% for the original model with all

variables and all cases included. For this reason, I chose to omit the outliers

defined in Figure 5 which depicts Cook’s Distance and run the multiple

linear regression again. This resulted in R2 of .52 or a 52% prediction rate for

GO. It also resulted in a better significance for duration (from .998 in the

original model to .167) and function (from .662 to .527). While duration may

still be above .1, a better result for the overall model is shown.

After defining the outliers above, I returned to the data set and,

starting from the bottom, removed each of the respondents from highest

number (311) to lowest (24). By working my way up the list, I was able to

guard the pertinent responses while omitting the outliers. My final data set

resulted in 132 responses. The sample size, n, decreased from 148 to 132,

but the latter still fell within the appropriate range for running multiple linear

regression (n = 104 + 3= 107; n=132 > n=107).

The initial results also showed that frequency of use was the only

significant predictor of GO (significance = .000). Linear regression was also

conducted using frequency of use only. This resulted in only a slight

improvement on the R2 (from .35 to .38). For this reason, I decided to report

both the results of the multiple linear regression for the entire model with the

outliers removed and the results from frequency of use only.

Reliability Estimates

Cronbach’s alpha was used to assess the reliability of dependent and

independent variables for the overarching hypotheses. Cronbach’s alpha

showed acceptable reliability at .765. The number of valid cases is 132

(n=132), the final number of respondents used in the regression analysis.

The number of final participants is greater than the calculated acceptable

sample for multiple linear regression, n= 104 + number of variables. In this

case, there are three predictor variables, so n= 107. Thus there are sufficient

respondents to run the multiple linear regression. Table 4 shows the

reliability of the overarching and lower level hypotheses.

Table 4

Reliability and Cronbach’s Alpha for Hypotheses

Hypotheses Cronbach’s alpha for duration, Cronbach’s alpha

for

frequency of use, and function frequency of use only

H01and H02 .765 .760

H01M and H02M .772 .789

H01F and H02F .750 .626*

H01W and H02W .703

<5 years

6-11 years

12-19 years

20-24 years

25-29 years

>30 years

.753

.835

.630*

.794

.819

.811

.730

.644*

.900

.728

.896

H01C and H 02C .872 . 679*

*<.700; Denotes a potential reliability problem

As seen in Table 4, Cronbach’s alpha for the overarching and lower

level hypotheses were greater than .700 showing good reliability except for

the work experience group of 12-19 years in the model including frequency

of use, function, and duration. Cronbach’s alpha for 6-11 years, 25-29 years,

>30 years, and for individual channels showed even better reliability.

Cronbach’s alpha if Item Deleted was above .3 for all variable except

frequency of use for male data only (.294) and female data only

(.259). In the work experience ranges, Cronbach’s Alpha if Item Deleted was

above .3 for all variables except frequency of use for >5 years (.274). For

work range experience 6-11 years, 20-24 years, 25-29 years, and >30 years,

Cronbach’s Alpha if Item Deleted was greater than .3 for all variables. For

work range experience 12-19 years, both frequency of use (.145) and

duration (.261) were well below .3.

As seen on Table 4, Cronbach’s alpha for the overarching and lower

level hypotheses were greater than .700 except for female data only (.626),

12-19 years (.644), and individual channels (.679) using frequency of use

only. This showed weak reliability in these groups with frequency of use

only. The strongest reliability was noted in work experience ranges 20-24

years (.900) and >30 years (.896) using frequency of use only.

Assumptions

Table 5 shows how the assumptions of this regression model were

tested for the overarching and lower level hypotheses.

Table 5

Assumptions for all hypotheses

Hypotheses Durbin- VIF

Watson

All All

variables

Tolerance

All

variables

DurbinWatso

Frequency of

use only

VIF

Frequency

of use only

Tolerance

Frequency

of use

only

H01 and H02 1.758* Averag

e 1.59

All

variables

>.1

1.828* 1.000 1.000

H01M and

H02M

1.716* Averag

e 1.55

All

variables

>.1

1.606* 1.000 1.000

H01F and H02F

2.195**

Averag

e 1.72

All

variables

>.1

2.148** 1.000 1.000

H01W and

H02W

<5 years

6-11 years

12-19 years

20-24 years

25-29 years

>30 years

2.080**

1.980*

1.584*

1.176*

2.677**

2.512**

2.02

2.63

1.34

1.38

1.90

1.58

All

variables

>.1

1.953*

2.569**

1.815*

1.135*

1.859*

2.360**

1.000 for

all ranges

1.000 for

all ranges

H01C and

H02C

1.922* 1.87 All

variables

>.1

1.836* All

variables

>.1

1.41

*<2.00 = positive correlation **>2.00= negative correlation

According to Field (2009), an average VIF of greater than one may

suggest multicollinearity problems; however, the average VIF sums in Table

5 were relatively low and were not a cause for major concern. Also, the sums

for tolerance and all variables were consistently greater than .1 which

showed there were no serious problems of multicollinearity.

For the Durbin-Watson for all variables and the hypotheses, positive

relationships were found in the overarching hypotheses, H01M and H02M,

and H01C and H02C. In the work experience ranges, positive relationships

were established for 6-11 years, 12-19 years, and 20-24 years, while

negative relationships were found for <5 years, 25-29 years, and >30 years.

The Durbin-Watson also showed a negative relationship for H01F and H02F.

Part and partial results were positive for both frequency of use and function

showing that the predictor and criterion variables are directly related.

For the Durbin-Watson for frequency of use only and the hypotheses,

positive relationships were found in the overarching hypotheses, H01M and

H02M, and H01C and H02C. In the work experience ranges, positive

relationships were established for <5 years, 12-19 years, 20-24 years, and

25-29 years, while negative relationships were found for 6-11 years and >30

years. The Durbin-Watson also showed a negative relationship for H01F and

H02F. Part and partial results were positive for the hypotheses except H01C

and H02C showing that the predictor and criterion variables are directly

related. In the case of the H01C and H02C with the individual communication

channels and frequency of use, meetings, agendas, and e-mail produced

negative part and partial results.

The histograms and p-plots in figures 8, 9, and 16 showed normally

distributed residuals. The histogram is a bell-shaped curve which shows the

shape of the distribution (Field, 2009). The straight line in these figures

represents normal distribution where the points lie on the same line (Field,

2009). The points were randomly and evenly dispersed for the overarching

hypotheses and male and female data which indicates that the assumptions

of linearity and homoscedasticity have been met (Field, 2009). Based on the

histograms and plots shown in these figures, normality may be assumed in

the data and hence the relevant assumptions for multiple linear regression

were met. However, in the figures for work range experience, the points did

not lie on the line; rather, they swerved in an s-shape around the line which

could suggest a slight deviation from normality. The histograms in the

ranges of work experience could also suggest problems with distribution.

Descriptive statistics

Descriptive Statistics for Independent and Dependent Variables

This section begins with the descriptive statistics and reliability results

for the overarching hypotheses, H01 and H02. For these hypotheses, the cases

for all 15 channels were added together to give a sum for duration,

frequency, function, and GO to produce the variables Duration SUM,

Frequency SUM, Functions SUM, and GO SUM. Multiple regression was

run using these sums which explains the high numbers for mean and std.

deviation.

Table 6

Descriptive statistics for independent and dependent variables

Mean Std.Deviation N

GO SUM 243.52 58.62 132

Functions SUM of all

channels

30.85 9.37 132

Duration SUM for all

channels

44.19 9.43 132

Frequency SUM for

all channels

54.03 9.34 132

As seen in Table 6 of descriptive statistics for the overarching

hypotheses, the final sample size was 132 from the 332 participants who

opened the survey. For multiple linear regression, a minimum sample size of

n=104 + number of variables tested must be met (Field, 2009). With three

variables, the sample size should be n=107. With n=132, the minimum

number was exceeded. From Chapter 3, the sample size calculated using

G*Power of 77 cases per variable (for gender and work experience ranges)

was not reached. However, with a confidence level of 95%, sample size =

132, population = 8460, the confidence interval was 8.46. This showed that

the margin of error was relatively small. Thus, the actual sample represented

the overall population well.

Multiple Regression Model Statistics

Using SPSS, I ran Multiple regression with forced entry of the three

predictor variables (frequency of use, duration, and function) and the

outcome variable (GO). As seen in Table 7, the multiple correlation

coefficient for the overarching hypotheses was .72. The R2 which showed

the variance in the outcome was .52; otherwise stated, 52% of the variance

was accounted for when using the model of the independent variables of

frequency of use, duration, and function predict the dependent variable, GO.

Table 7

Multiple Regression Summary of all Variables and all Hypotheses

H01 and H02 .719 .517 .506 41.21 GO 1.00

Frequency of

use .713

Duration .391

Function .315

H01M and H02M .743 .552 .536 41.02 GO 1.00

Frequency of

use .736

Duration .390

Function .330

H01F and H02F .670 .449 .407 43.10 GO 1.00

Frequency of

use .660

Duration .392

Function .288

H01W and H02W

<5 years

6-11 years

12-19 years

20-24 years

25-29 years

>30 years

.692

.737

.703

.725

.788

.907

.479

.543

.494

.526

.620

.823

.437

.478

.422

.289

.507

.779

40.31

45.48

47.81

53.42

40.08

29.86

All

positive

All

positive

Negative for

functions

Hypotheses R R 2

Adjusted R 2 Standard Error Pearson Correlation

All

positive

All

positive

All

positive

H01C and H 02C .693 .481 .436 50.79 All positive

As seen in Table 7, the Pearson correlations were greater than 0

showing that changes are made in the same direction. The only exception is

the work experience range of 12-19 years (-.001). From the R2 in Table 8,

the variance or predictive power ranged from .449 or 45% for the model

using female data only to .62 or 62% for the work experience range of 25-29

years. While most of the R2 and adjusted R2 were close in number showing

they will reasonably predict the outcome when a different sample is used,

the work experience ranges showed the greatest differences in the two

numbers.

The greatest difference came in the work experience range of 20-24 years

(.237).

In Table 8, the results of conducting linear regression with only

frequency of use and GO were recorded.

Table 8

Multiple Regression Summary of Frequency of Use Only and All Hypotheses

H01and H02 .612 .375 .372 49.32 GO 1.00

Frequency of

use .612

H01M and H02M .652 .425 .420 46.10 GO 1.00

Frequency of

use .652

H01F and H02F .456 .208 .194 55.86 GO 1.00

Frequency of

use .456

H01W and H02W

<5 years

6-11 years

12-19 years

20-24 years

25-29 years

>30 years

.543

.575

.475

.818

.572

.812

.294

.331

.226

.670

.327

.659

.281

.331

.202

.645

.283

.643

48.00

48.59

61.90

42.54

48.76

40.33

GO 1.00

.543

.575

.475

.818

.572

.812

H01C and H 02C .647 .418 .361 46.79 All positive

Comparable to Table 7 where all variables were addressed, the

Pearson correlations in Table 8 using frequency of use only were greater

than 0 showing that changes are made in the same direction. From the R2 in

Table 9, the variance or predictive power ranged from .208 or 21% for the

model using female data only to .67 or 67% for the work experience range of

20-24 years. While most of the R2 and adjusted R2 were close in number

showing they will reasonably predict the outcome when a different sample is

Hypotheses R R 2

Adjusted R 2 Standard Error Pearson Correlation

used, the individual channels showed the greatest differences in the two

numbers (.055).

The results from the ANOVA were recorded in Table 9 to evaluate

the significance of the model using all variables and all hypotheses.

Table 9

Anova Results for All Variables and All Hypotheses

Hypotheses Regression df Residual df F Sig.

H01 and H02

3 131 45.68 .000**

H01M and H02M 3 84 34.47 .000**

H01F and H02F 3 40 10.85 .000**

H01W and H02W

<5 years

6-11 years

12-19 years

20-24 years

25-29 years

>30 years

11.63

8.32

6.83

2.22

5.45

18.60

.000**

.001**

.002**

.187

.018*

.000**

H01C and H 02C

*p< .05

** p< .01

3 128 45.68 .000**

Based on the ANOVA in Table 9, F(3,131) = 45.68, p<.01, this model

for the overarching hypotheses showed a significant relationship between the

variables. The between groups df is 3 and the within groups df was 131. The

independent variables predicted the dependent variable, thus, the null

hypotheses must be rejected. The model improved the variance or ability to

predict the dependent variable, GO, from 35% to 52%.

Based on the model summary in Table 9 using male data only, F=

(3,84)= 34.47, p<.01. The significance of .000 suggests this is a good model.

This model of all variables and male data only predicted the dependent

variable, GO. Using female data only, F=

(3,40)= 10.85, p<.01, the results of the F test of the multiple linear

regression showed a significance of .000, which suggests this model of all

variables and female data only predicted the dependent variable, GO. For

work experience ranges, all ranges showed significance of p<.01 except for

20-24 years (sig=.187) and 25-29 years (sig=.018). These results for gender

and work experience ranges must be considered with caution as the number

of respondents failed to reach the optimal number of respondents for

multiple linear regression (n= 107). Finally, for communication channels

entered separately into the model, F= (3,128) = 45.68, p<.01, the model with

all variables is significant in predicting GO.

When conducting multiple linear regression using only frequency of use and

GO,

I found similar results in the predictive power as seen in Table 10.

Table 10

Anova Results for Frequency of Use Only and All Hypotheses

Hypotheses Regression df Residual df F Sig.

H01and H02

1 178 106.87 .000**

H01M and H02M 1 114 84.38 .000**

H01F and H02F 1 60 15.73 .000**

H01W and H02W

<5 years

6-11 years

12-19 years

20-24 years

25-29 years

>30 years

22.10

16.29

9.33

26.38

7.30

42.43

.000**

.005*

.000**

.016*

.000**

H01C and H02C 15 152 7.29 .000**

*p< .05

** p< .01

Based on the ANOVA in Table 10, F(1,178) = 106.87, p<.01, this

model for the overarching hypotheses showed a significant relationship

between frequency of use and GO. The independent variable (frequency of

use) can be used to predict the dependent variable (GO). In fact, for each of

the split groups (gender, work experience range, communication channels),

there were significant results for predicting GO from frequency of use.

Nonetheless, these results, too, must be considered with caution as

respondents of female data only and all work experience ranges fell short of

the n=105 needed for multiple linear regression.

Regression Analysis Conclusions

H01 and H02.

For the overarching hypotheses, I conducted a multiple linear

regression analysis to evaluate how well duration, frequency of use, and

function predict gratifications obtained (GO) when using communication

channels in the workplace. The predictors were duration, frequency of use,

and functions while the criterion variable was GO. The overall linear

combination of the predictor variables was significantly related to the

criterion variable, F(3,131) = 45.68, p<.01. The sample multiple correlation

coefficient was .62 indicating that approximately 52% of the variance of GO

in the sample can be accounted for by the linear combination of duration,

frequency of use, and functions. The regression model overall predicts GO

significantly well, thus I must reject the null hypotheses although the

variance is only 52%.

H01M and H02M.

To test the hypotheses H01M and H02M, I conducted a multiple linear

regression analysis using only male data to evaluate how well duration,

frequency of use, and functions predict gratifications obtained (GO) when

using communication channels in the workplace. The predictors were

duration, frequency of use, and functions while the criterion variable was

GO. The overall linear combination of the predictor variables of only male

data were significantly related to the criterion variable, F(3, 84) = 34.47, p<

01. The sample multiple correlation coefficient was .74 indicating that

approximately 55% of the variance of GO in the sample can be accounted

for by the linear combination of duration, frequency of use, and functions for

male data only. The Durbin-Watson test showed 1.72 and a positive

correlation between the variables based on male data only. The regression

model using male data only had a slightly higher variance (55%) than that of

the original model (52%). Thus, the regression model using male data only

predicted

GO well, so I must reject the null hypotheses.

H01F and H02F.

To test the hypotheses H01F and H02F, I conducted a multiple linear

regression analysis using only female data to evaluate how well duration,

frequency of use, and functions predict gratifications obtained (GO) when

using communication channels in the workplace. The predictors were

duration, frequency of use, and functions while the criterion variable was

GO. The overall linear combination of the predictor variables of only female

data was significantly related to the criterion variable, F(3, 40) = 10.85, p<

01. The sample multiple correlation coefficient was .67 indicating that

approximately 45% of the variance of GO in the sample can be accounted

for by the linear combination of duration, frequency of use, and functions for

female data only. This was lower than the original regression model (52%).

The Durbin-Watson test showed 2.20 suggesting a negative relationship

between the variables based on female data only. The regression model

using female data only predicts GO significantly well, thus I must reject the

null hypotheses.

H01W and H02W.

To test the hypotheses H01W and H02W, I conducted a multiple linear

regression analysis using years of work experience ranges to evaluate how

well duration, frequency of use, and functions predict gratifications obtained

(GO) when using communication channels in the workplace. The predictors

were duration, frequency of use, and functions while the criterion variable

was GO. The overall linear combination of the predictor variables of the six

work experience ranges in Table 11 showed the following:

Table 11

Model Summary of Independent and Dependent Variables in Work

Experience Ranges

Work Experience Significance Durbin-Watson

<5= F(3, 38) = 11.63,

p<.01

Significant

6-11 years= F(3, 21) =

8.32, p<.01

Significant

12-19 years=F(3, 21) =

6.83, p< .01

Significant

20-24 years= F(3, 6) =

2.22, p> .01

Not

significant

25-29 years= F(3, 10) =

5.45, p<.01

Significant

>30 years-= F(3, 12) =

18.60, p<.01

Significant

According to Table 11 , the data for <5 years, 6-11 years, 12-19 years,

25-29 years, and >30 years were significantly related to the criterion

variable. The data for 2024 years were not significantly related to the

criterion. However, when all ranges of work experience were analyzed

together using multiple linear regression, F (4, 127) = 34.20, p< .01, they

showed significance as a whole group. Durbin-Watson for all work

experience equaled 1.732 showing a positive relationship between the

variables. DurbinWatson for the individual work experience categories

ranged from 1.18 to 2.68 which showed that different work experience

ranges affect the directionality of the relationship.

Cronbach’s alpha for the range of 12-19 years was extremely weak (.630),

while Cronbach’s alpha for the range of 6-11 years was .835 showing good

reliability. Thus the regression model using work experience range was

significant in predicting GO and the null hypotheses must be rejected. The

range with the lowest sample multiple correlation coefficient of .69 and

lowest variance of 48% was found in the < 5 years category.

H01C and H02C.

I conducted a multiple linear regression analysis using data from each

of the 15 communication channels individually to evaluate how well

duration, frequency of use, and functions predict gratifications obtained

(GO) when using communication channels in the workplace. The predictors

were duration, frequency of use, and functions while the criterion variable

was GO. The overall linear combination of the predictor variables of only

data from the 15 communication channels was significantly related to the

criterion variable, F (15, 128) = 45.68, p< .01. The regression model using

communication channels significantly predicted GO, thus the null

hypotheses must be rejected. The sample multiple correlation coefficient was

.69 indicating that approximately 48 % of the variance of GO in the sample

can be accounted for by the linear combination of duration, frequency of use,

and functions for data from the 15 communication channels which is lower

than the original regression model (52%).

Summary

The purpose of this multiple linear regression study was to examine

the relationship between duration, frequency of use, and function and GO,

while controlling for gender, years of work experience, and individual

communication channels. The results of the statistical analyses produced the

following general findings:

1. Multiple linear regression confirmed that there is a relationship

between the independent variables (duration, frequency of use, and

function) and the dependent variable (GO) for the overarching

hypotheses. The regression model overall predicted GO significantly

well with a variance of 52%. For this reason, we must reject the null

hypotheses.

2. Multiple linear regression was conducted with a split data file of male

and female to evaluate how well the independent variables predict

GO. There was evidence from both the male and female data that the

three predictors were significantly related to the criterion variable of

GO. However, the male data showed a positive correlation between

the variables and a variance of 55%, while the female data showed a

negative correlation and a lower variance of 45%. In both models, we

must reject the null hypotheses.

3. Multiple linear regression was conducted with a split file of ranges of

years of work experience to evaluate how well the independent

variables predict GO.

There was evidence that all ranges except 20-24 years were

significantly related

to the criterion variable. There were positive correlations for 6-11

years, 12-19 years, and 20-24 years and negative correlations for < 5

years, 25-29 years and >30 years. For this model, I must reject the

null hypotheses for all work experience ranges except the range of 20-

24 years. However, the low number of responses for each of these

categories (from 6 to 38 respondents) was problematic and may have

skewed the results. For example, the >30 years range showed a strong

variance of 82%, but there were only 12 respondents.

4. Multiple linear regression was conducted with each of the 15

communication channels individually to evaluate how well the

independent variables predict GO. There was evidence that overall

linear combination of the 15 channels were significantly related to the

criterion variable but had a variance of only 48%. I must reject the

null hypotheses.

5. The overarching research question (RQ) was: using multiple linear

regression, can

Y (gratification obtained) be predicted in terms of three independent

variables (frequency of use, duration, and function)? Based on the

findings and analysis of the present study, the response to the

overarching RQ was yes, although the variance is only 52% for the

regression model of the three independent variables of frequency of

use, duration, and function and the dependent variable, GO. Thus, the

multiple linear regression model overall predicted GO well.

In addressing the research questions:

The overarching research question (RQ) was: using multiple linear

regression, can Y (gratification obtained) be predicted in terms of three

independent variables (frequency of use, duration, and function)? When

tested together as a model, frequency of use, duration, and function, were

significant predictors of GO (p<.05), although frequency of use was the only

significant variable (p<.01). There was a positive correlation

(DurbinWatson= 1.76) between the predictor and criterion variables. I

rejected the null hypothesis and confirmed that the multiple linear regression

model overall predicts GO with a variance of 52%.

RQ1: Are there gender differences when determining whether Y (GO) can

be predicted in terms of three independent variables (frequency of use,

duration, and function)?

While both male and female data showed significant relationships

between the independent and dependent variables, male data showed a

positive correlation, while female data showed a negative correlation. This

may have been misleading, however, as there was double the number of

male respondents to female respondents. In both cases, the sample size was

below the acceptable size of n=107 (male = 84; female = 40). RQ2: Does

the number of years of work experience affect whether Y (GO) can be

predicted in terms of three independent variables (frequency of use,

duration, and function)?

The data were split into six work experience ranges and multiple

linear regression was run on each group. When analyzed together, work

experience ranges showed significant relationships between the independent

and dependent variables. When broken down into each range, however, all

ranges were significant except for 20-24 years. Like the gender above, none

of the sample sizes met the ideal sample of n=107. They ranged from 6 (20-

24 years) to 38 (<5 years). There were only 6 respondents in the

nonsignificant range which may have affected the results and led to an

overall interpretation which could be challenged.

RQ3: Does the communication channel chosen affect whether Y (GO) can

be predicted in terms of three independent variables (frequency of use,

duration, and function).

To respond to the third RQ and H01C and H02C, I ran multiple linear

regression by entering each of the 15 communication channels individually

into SPSS. The overall linear combination of the independent variables was

significantly related to the dependent variable. The sample size for this

regression was 189 which exceeded the ideal number of n= 119 (for the 15

channels). There was an overall positive correlation with these variables.

The final chapter, Chapter 5, includes an interpretation of the findings

as linked to the research questions. It continues with limitations of this study

and reflections on the importance of the topic. Both recommendations for

action and recommendations for further study are addressed. Chapter 5

concludes with implications for social change.

Chapter 5: Discussion, Conclusions, and Recommendations

Introduction

In this quantitative study, I examined how professionals chose

between traditional and modern communication channels used in business,

the premise being that users choose communication channels due to the GO.

The research design chosen was a crosssectional survey design. A

questionnaire was initially constructed using existing instruments from

Downs and Hazen, Hecht, and Dobos. After conducting the survey in a pilot

study, all questions from Hecht’s survey and certain questions from Downs

and

Hazen’s survey were omitted as they asked participants to rate

communication satisfaction with colleagues or supervisors. This was not the

objective of this dissertation. The final questionnaire was based on two

existing instruments: 19 questions from the CSQ of Downs and Hazen

(1977) and an instrument developed by Dobos (1988) to study functions of

communication channels. The final survey consisted of six parts: Part 1

included demographic questions; Part 2 included the 19 questions from

Downs and

Hazen’s CSQ; Part 3a asked participants to estimate how much time (in

hours) spent on 15 communication channels, and Part 3b asked them to rate

the GO of each channel based on this duration; Part 4a asked participants to

estimate the frequency with which they use each of the 15 channels (from

never to > once a day) and Part 4b asked them to rate the GO of each

channel based on frequency; Part 5a asked participants to choose the

functions (giving information, receiving information, establishing new

relationships, maintaining existing relationships, brainstorming) they do with

each of the 15 channels (from 1 to 5 functions possible); and Part 5b asked

them to rate the GO of each channel based on the functions they do with it.

The population consisted of currently employed alumni from an

international hospitality school in Switzerland, from whom 332 responses

were obtained. The survey data were downloaded from the SurveyMonkey

site as an Excel file and subsequently entered into SPSS 21. After

controlling for missing data, a final, usable sample of 148 was obtained.

Multiple linear regression was used to assess statistically significant

relationships between the independent variables: frequency of use, duration,

and function, and the dependent variable GO. I confirmed that frequency of

use, duration, and function respectively predict GO.

This chapter begins with an interpretation of the findings as linked to

the research questions. It continues with an analysis which compares my

results to those from previous, similar research from the literature review,

and defines the gap which this research attempted to fill. In the third section,

I discuss limitations of the study and, in the fourth section, I summarize how

this study could be revised or improved for future studies. Implications for

social change are discussed, as well as how these findings can contribute to

positive social change. The final section of Chapter 5 is the conclusion for

both this chapter and the overall dissertation itself. It includes reflections on

the topic and the results from the multiple linear regression.

Interpretation of Findings

The overarching research question (RQ) was: using multiple linear

regression, can

Y (gratification obtained) be predicted in terms of three independent

variables (frequency of use, duration, and function)? Based on the findings

and analysis seen in Chapter 4, the null hypotheses were rejected. The three

independent variables were significantly related to the dependent variable.

Thus, the multiple linear regression model overall predicted GO, although

the variance was only 52% which suggests a weak model.

Previous scholars reviewed in Chapter 2 employed uses and

gratifications theory to test communication channels in the workplace. The

theory has been applied to mass media and general communication studies in

education and the workplace, often being adapted or augmented with other

theories to deal with specific issues or a limited range channels. I also

applied uses and gratifications theory, but in a broader perspective across a

range of communication channels. The basic premise was that employees

choose communication channels based on their GO.

Variables examined in previous studies included frequency of use,

tasks performed, functions, satisfaction, age, previous knowledge, gender,

education, motivation, perceptions, and Internet skills. These researchers

either focused on one to several channels or grouped the channels into

traditional and modern. No previous scholar tested as many channels with

predictor variables of GO as the present study. In my study, three predictor

variables, frequency of use, duration, and functions were used to show how

well they predicted GO for 15 communication channels. In general, I found

that the more frequently (frequency), the more time spent (duration), and the

more functions completed with the channel, the higher the GO. However,

while the model using these three predictor variables do predict GO with

these 15 communication channels, frequency of use was the only significant

predictor (p<.01).

Many of the researchers examined in Chapter 2 looked at uses and

gratifications theory from the perspective of the final user or made a

comparison between GS and GO. This was especially noted in studies on TV

or mass media channels. In this study, I focused on the GO of the sender

who chooses the communication channel, presumably for a specific purpose.

In my study, like previous studies, I was looking for gratification obtained,

not gratification sought.

RQ1: Are there gender differences when determining whether Y (GO)

can be predicted in terms of three independent variables (frequency of use,

duration, and functions)?

Previous scholars often found differences in the use of communication

channels between genders. In this study, both male data (F(3,84)= 34.47,

p<.01) and female data (F(3,40)= 10.85, p<.01) showed significant

relationships between the independent and dependent variables. To respond

to the research question, there were gender differences when determining

whether Y can be predicted in terms of frequency of use, duration, and

functions. In both cases, GO could be predicted from the independent

variables, but the directionality differed. Male data showed a positive

correlation between the variables, while female data showed a negative

correlation. For both male and female data, frequency was the highest

standard deviation, while duration and functions differed. For male data

only, duration was negative, while duration was positive for female data.

Regarding functions, male data only showed a positive value, while female

data showed a negative value. Thus, for the three predictor variables, there

were similar and positive standard deviations for frequency, yet different and

negative standard deviation values for duration and functions respectively.

Frequency was a positive predictor of GO, while there were gender

differences regarding the predictive power of duration and functions.

RQ2: Does the number of years of work experience affect whether Y

(GO) can be predicted in terms of three independent variables (frequency of

use, duration, and functions)?

Age was a variable tested frequently in previous research. While I did

not test age directly, the number of years of work experience could be used

to estimate age range. According to official school statistics, the average age

of a student starting their undergraduate studies was 19.2. Students following

the 4-year undergraduate program would be 23.2 years upon graduation.

With less than 5 years of work experience, this respondent must be less than

30-years-old. The average age for students entering the short 2-year program

was 23.4. Again, after 2 years and with less than 5 years of work experience,

they would be under 30-years-old. Conversely, someone with more than 30

years of work experience must be in their 50s or 60s, depending on their age

at graduation.

Age is relevant to this type of research when testing whether older

workers make different communication channel choices than younger

workers. Based on my findings, the all ranges showed significant

relationships among the variables except 20-24 years (p=.187; p>.05). This

finding was useful as many previous scholars showed that there are no or

few differences between age groups when it comes to choosing

communication channels, and I confirmed the results of these other studies.

There were no clear differences between younger and older generations

when it comes to choosing communication channels.

Six ranges of work experience were analyzed in Chapter 4. The results

were mixed. For three of the ranges, a positive and significant relationship

between the variables was found (6-11 years, 12-19 years, and 20-24 years).

The work experience ranges of <5 years, 25-29 years and >30 years showed

negative correlations between the criterion and outcome variables.

The response to RQ2 is complicated, but overall the answer seems to

be yes. All work experience ranges except 20-24 years were significant and

predicted GO in terms of frequency of use, duration, and functions. When

answering whether work experience ranges affect the prediction, the

response is affirmative. When all work experience ranges were run together

as work experience with the variables of frequency of use, duration, and

functions, the overall significance was F(4,127)=34.20, p<.01 with a Durbin

Watson of 1.732. Work experience and the independent variables predicted

GO significantly well. However, the low response rates in each of these

categories must be considered. The responses ranged from six in the 20-24

year category to 38 in the >5 year category. These numbers were far below

the acceptable number for multiple linear regression; thus, these results

could be questioned.

RQ3: Does the communication channel affect whether Y (GO) can be

predicted in terms of three independent variables (frequency of use, duration,

and functions)?

Previous studies discussed in Chapter 2 tended to concentrate on one

or several mass media communication channels used for entertainment such

as television, the Internet, or social media networks. However, the Internet

was only included in this study as a workplace communication channel,

since entertainment use was beyond the scope of the present study. Previous

researchers also focused on education and communication channels used by

students, which were likewise irrelevant to the present study. However, all

respondents derived from the same population of graduates from one

international hospitality school in Switzerland, so their education level,

which is a common variable in communication research, could be inferred.

At minimum, they have all earned a Swiss undergraduate hospitality degree

and have followed at least one course in communications in the higher

education system where they learned about various communication channels

as part of their curriculum. One future study could be to conduct a similar

project with current undergraduate students of this international hospitality

school in Switzerland to compare those findings with the ones stated in this

dissertation. This could be useful in bridging the gaps mentioned in Chapter

2 between communication competencies students actually have upon

graduation versus the communication competencies the professional world is

expecting of them.

Based on the results in Chapter 4, the model of independent and

dependent variables and communication channels was significant, thus

showing a relationship between the variables. There was a positive

correlation between the variables (Durbin-

Watson= 1.922). All channels had values above .3 for Cronbach’s Alpha if

Item Deleted which is good and the overall Cronbach’s alpha was .872

showing more reliability than the model for the overarching hypotheses

(Cronbach’s alpha = .765).

Limitations of the Study

In the present study, limitations derived from the survey questionnaire

itself, the sample size, self-reporting, and the variables chosen for this

model. In some cases, the limitations could be minimalized; in others, they

were uncontrollable.

Survey Questionnaire

A few of the respondents who randomly received the survey were

teaching colleagues of mine who were also former students of the school.

Feedback I received from them included issues with the length of the survey

and the perceived repetitiveness of trying to rate GO for each independent

variable (frequency of use, duration, and functions) and for each of the 15

communication channels. If I conducted the survey again, I would change

the format to make it easier for respondents to fill in.

Another issue involved gender and language. In Chapter 3, I discussed

both gender and language in the overall alumni population. There was no

program in English until 1996, so any graduates with over 18 years of work

experience had to follow the program in French. I was concerned that that

could lead to a majority of French-speaking respondents among the survey

results. French-speaking respondents came predominantly from France,

Switzerland, or Belgium, while English-speaking respondents represented 84

different nationalities. A majority of French language responses could have

been problematic for a survey which tried to examine hospitality

professionals on a global scale. In the final results, however, this was not the

case as 181 surveys were completed by alumni who studied in the English

program, while only 151 studied in the French program. Thus a balanced

mix of nationalities responded to the survey, meeting the initial goal of

giving a global overview of hospitality professionals.

For gender, however, the results were more problematic. In Chapter 3,

I discussed the introduction of women into this international hospitality

school in Switzerland in 1963. This fact alone does not preclude their

response rate. Women have attended the school long enough to fill all the

possible work experience ranges, including >30 years, although their

presence was minimal until the 1990s when gender numbers began to

equalize in this school. Over the past 24 years, there were roughly equal

numbers of female and male students, suggesting that there was an equal

possibility for respondents of both genders to fill all work experience ranges.

In Chapter 3, I addressed the possibility that there might be a larger

proportion of male respondents among the older graduates. In fact, the

survey data showed almost double the number of male (95) to female

respondents (53). The average age of male respondents was 41.9, while the

average age for female respondents was 33.3. This reinforced the

supposition in Chapter 3 that there could be a greater number of older men

who responded than women as they were the majority of students until the

1990s. Although the results for testing the model were significant for both

genders, the sample size was small for multiple linear regression. Also, as

seen in Chapter 4, there were minor collinearity problems with the data,

suggesting that the results from this small number could be questioned.

The final potential issue with the survey derived from the

demographic information about job title or industry. The present study was

based on how hospitality professionals choose communication channels in

the workplace. There would have been a problem had there not been a

majority of respondents with a hospitality background. In fact, more than

half of the respondents cited hospitality industry as their workplace. This

concurred with the reality of this international hospitality school’s official

statistics which reports approximately 40% of the school’s graduates do not

stay in the hospitality industry. Many choose other careers which allow them

to apply the competencies learned in the hospitality program. My results

reflected the current reality of these graduates.

Sample Size for Each Variable

Upon receiving the results from the survey, a few issues arose. First of

all, the number of respondents was much fewer than anticipated. As seen in

Chapter 4, the response rate was approximately 4%. As discussed in Chapter

3, colleagues who have conducted surveys with this same population of

alumni from this international hospitality school in Switzerland have

reported response rates in the range 5-16%. The low response rate received

for my study may be explained by lack of interest in the topic, lack of time,

an overload of surveys being sent out in the same month, or general

avoidance of responding to any survey which was several pages in length. I

spoke to the Alumni Coordinator who warned me that February was not the

best time to conduct a survey. It would have been better to send it in

September when relatively little information is sent out to this population. In

addition, a satisfaction survey was sent to the same alumni population the

week before my survey was available online and this may have affected the

response rate.

My survey questionnaires were sent out on a Friday, which may have

contributed to an increase in non-response rates. Further, the response rate

dropped dramatically after the first 48 hours. Without a reminder, which I

was not permitted to send, the survey may have been forgotten.

Initially, an incentive was considered to encourage participation, but

this was inadmissible on ethical grounds. While the researcher could have

offered a token prize, there was no feasible token prize which could be

distributed to all participants around the world, and therefore no prize was

offered. I relied solely on the interest and good will of the alumni, whose

assistance contributed to a research project at their alma mater.

Another issue which arose was the number of useable questionnaires

compared to the number of respondents. While the overall number of

respondents who clicked on the survey was 332, many of these surveys were

incomplete and contained gaps and only 148 were valid for multiple linear

regression. Some respondents did not complete the entire survey or did not

complete certain sections. The initial total of 332 respondents and the

number of valid cases of 148 exceeded the G*Power estimate of a total

sample size of 77 when the main hypotheses were tested with the whole data

set. The calculated sample size for conducting multiple linear regression is

n=104 + number of variables (Field, 2009). In my study, the sample size is

appropriate, n=148> n=107. However, this was only the case for the

overarching hypotheses. When the data were split into gender and ranges of

work experience, all samples were below n=107, varying between 13 and 95.

These low numbers could have skewed the results and reduced the

significance, but the collinearity statistics were sound for the overarching

hypotheses and no multicollinearity issues were found for the original

model.

Self-reporting on Frequency and Duration

One criticism of survey questionnaires examined in Chapter 3 was the

necessity to self-report. In my survey questionnaire, respondents were asked

to self-report duration in hours, frequency (days to months), functions (out

of five), and GO for each (Likert scale of 1-7). There may have been a link

between the length of the survey and self-reporting. For example,

respondents may have ticked boxes with less reflection later, as they started

to lose interest in responding to the questions. This was impossible to

control. The limitations of self-reporting are exemplified by the data

concerning fax, discussed below.

When looking at all 15 communication channels, the responses

regarding duration, frequency, and functions for modern channels such as e-

mail or mobile phone were as expected. Scores were high for all independent

variables and for GO. However, the responses of frequency and duration for

fax were higher than expected. This suggests two possibilities: either

respondents use faxes much more than other channels in their professional

capacity, or respondents ticked a box without reflecting on the channel. Fax

was the 9th channel on the list, and this may have led to respondent

complacency. Nonetheless, the fax data respected collinearity diagnostics

and had positive part and partial correlations. I must report what was found

regardless of my preconceptions about frequency of use or duration for

faxes.

Six of the 15 communication channels were significantly inter-

correlated: face-toface, telephone, letters, agenda, e-mail, and

teleconferencing. After considering the collinearity diagnostics, the most

reliable communication channels seemed to be letters, e-mail, and

teleconferencing. As discussed in Chapter 1, I defined face-to-face, letters,

telephone, and agendas as traditional channels, while e-mail and

teleconferencing were defined as modern channels. Previous research on

modern and traditional communication channels found little preference

between modern and traditional channels, and this has been confirmed by

my research. There were no major differences or preferences between

modern and traditional communication channels in the workplace.

As mentioned in Chapter 1, the significance of this study was to

evaluate specific communication channels used in the workplace and the GO

associated with their use. This study filled a gap in the communication

literature in three ways: (a) using an international population of hospitality

school alumni; (b) testing 15 communication channels used in the

workplace; and (c) establishing the relationship between frequency of use,

duration, and function to predict GO. This study confirmed previous

communication research results regarding traditional and modern

communication channels and offers new opportunities for further research

with other channels or other populations.

Variables for This Model

The variables chosen for this study, frequency of use, duration, and

function, did not provide an adequate model for predicting GO. In fact, only

frequency of use was significant (.000). After removing the outliers, the

significance of duration decreased to

.167, but that still exceeded the significance of .05 necessary to make it a

viable variable. The variables chosen for this study could have been tested

differently. For example, duration could have been an open question where

participants record how much time they actually spend on each channel.

Function could have been grouped differently such as the previous studies

by Dobos. For a future study, I would create a survey focusing on frequency

of use only as a predictor of GO. Although the variables tested here did not

produce a strong model, this may have been due to the survey design or

audience chosen. I would not dismiss using this combination of variables

again, as they have been used so often in existing communication research. I

would rework the survey design and methodology.

Other variables could have been chosen as well. As seen in the

literature review, other variables such as previous knowledge, job position,

education, motivation, perceptions, or existing skills could have been

examined. Gratifications sought could have been compared to GO for each

of the channels. These are potential variables which could be examined in a

future research project.

Reflections

In 2008, I spent a week at the Four Seasons Hotel, Chicago,

shadowing employees in all departments to establish which communication

channels they used. The purpose was to compare what is being taught in

business communication courses at one international hospitality school in

Switzerland to the reality of the hospitality industry. What I found was that

all types of communication channels, from traditional channels like faxes,

beepers, white boards, and bulletin boards to modern channels like e-mail,

Intranet, the Internet, and teleconferencing were being used on a daily basis

in the hotel. I interviewed department managers to discuss these channels

and found that each one had its specific place in the hotel for internal or

external communication. The visit not only affirmed that what I was

teaching was still relevant, but it also led to my interest in studying this topic

further in a doctoral program and, later, in future research. By conducting

my research with alumni from this hospitality school and analyzing their

results, I can adapt my curriculum for future business communication

courses.

Recommendations for Action

The results of this study are potentially applicable to graduates of this

international hospitality school in Switzerland. Although some have chosen

to follow careers outside the hospitality industry, the results are pertinent to

their current workplace as communication channels are relevant in all work

environments. In Chapter 1, the problem of poor communication skills was

addressed. This type of study can serve as a forum to discuss how

information is communicated and how communication could be improved.

For this reason, my results are also applicable to the hospitality school

where I am teaching and should be of interest to the management and

directors of the hospitality programs we offer. I can use these results to

develop new courses which will reflect both traditional and modern

communication channels used in the workplace. Future research could be

conducted with the current students of this international hospitality school

and potential employers to establish the communication gaps between what

students are learning in the program and what employers expect from young

graduates.

In order to disseminate the results of this study, the first step will be to

share this dissertation via Walden’s database and make it available to

management, staff, and students in this international hospitality school in

Switzerland. The next step will be to start new research projects within the

school on how students choose communication channels. I would like to

focus on a small range of modern communication channels, such as

telephone, IM, e-mail, and social media. Another study could be to assess

students’ social media use through the same independent variables of

frequency of use, duration, and function. Obtaining the PhD qualification

will enable me to obtain allocated time in my work schedule to do research,

so I will be able to conduct other studies on my own or in collaboration with

my colleagues. Communication research could easily be linked to other

topics such as marketing, strategy, finance, etc. Many of my colleagues have

already expressed an interest in collaborating on research projects of this

nature. I would eventually like to publish future research project results in

scholarly journals and present my results at one of the communication

conferences that I attend.

Recommendations for Further Study

According to the results in Chapter 4, all research questions for my

study showed positive and significant results, but with a predictive power of

only 52%. The variables I chose and the manner in which I tested them

resulted in a weak model. While multiple linear regression allowed me to

determine for the whole data set whether the independent variable (GO)

could be predicted in terms of three independent variables (frequency of use,

duration, and functions), the only variable which was significant was

frequency of use. For the corresponding lower level research questions

which specified gender, years of work experience, and individual

communication channels respectively multiple linear regression with the

relevant restricted data sets confirmed, albeit weakly, that the independent

variable (GO) could also be predicted in terms of three independent

variables (frequency of use, duration, and functions).

Upon finishing this study, there are three recommendations I can make

for further research: (a) conduct a qualitative study based on a series of

interviews to better understand why these channels are chosen; (b) replicate

Dobos’s earlier studies (1988,1992) using fewer channels or grouping them

as she did to compare with her results; (c) conduct future quantitative studies

on other populations such as students from the same international hospitality

school where this research data were gathered to compare results between

the alumni examined in this study and future graduates of the same program.

These options point to future studies which could continue to help fill gaps

in our knowledge of how communication channels are chosen and used. By

conducting these studies in Europe with a broader international population,

comparisons could be made with studies discussed in Chapter 2 that were

conducted entirely in the U.S. (Hargittai, 2010; Junco & Cotton, 2011; Junco

& Cotton, 2012; Kasavana et al., 2010; Neuman & Brownell, 2009).

Qualitative Research

In this doctoral study, I have been able to confirm that relationships

exist between independent variables (frequency of use, duration, and

functions) and a dependent variable (GO) when choosing communication

channels in the workplace. However, I could not gauge why these channels

are chosen for a specific message or task. The section from my survey

questionnaire on functions came the closest to indicating why channels are

chosen, as respondents were asked to choose the functions they do with each

channel. Analysis of the results showed that these responses left it unclear

what motivated employees to choose one channel over another. While I was

able to prove that there is a relationship between how often, how much time,

and how many functions communication channels are used for, the results do

not suggest why one channel is chosen over another or at which point one

channel is replaced by another one. For instance, the choice may have been

based on heuristics, facility of use, or company norms. In a future study, the

questions would be focused on how and not what. One way could be to ask

participants why they chose one channel over another for a specific

communication exchange. Their comments might offer valuable data which

would be useful for understanding how communication channels are chosen.

Dobos’s Study

Dobos (1988,1992) grouped, communication channels into face-to-

face, written, and electronic. Not surprising, my results showed similar,

significant statistics for electronic channels such as e-mail and the Internet.

But my findings also showed that the most significant communication

channels fall into the three groups: that is, face-to-face (p<.01), letter and

agendas (p<.05), and e-mail (p<.05). One recommendation could be to

replicate Dobos’s study using more channels, while respecting the same

groups. I was unable to find previous studies which tested so many

individual channels, which is why I attempted to fill this gap. To test the

independent variables (frequency of use, duration, and functions) and GO in

the overarching hypotheses, I entered the sum of all 15 channels into the

multiple linear regression equation. However, to test H01C and H02C, each

individual channel was entered into the multiple linear regression equation.

Perhaps conducting a test other than multiple linear regression, like the

discriminant factor analysis used by Dobos (1988,1992), could be suggested

for a future study.

Future Quantitative Studies

I imagine conducting future studies on how communication channels

are chosen with a completely different population, for example, lower level

employees versus managers. This type of study could help to answer the

questions about why there are so many communication breakdowns in the

workplace. One future study could be to reduce the number of channels to be

able to do a more in depth analysis of each one. I could choose the channels

most used in a more specific industry, like a hotel or restaurant, although

these would clearly have to be related to employee needs. Many employees

may not have used fax or teleconferencing in their daily job. A new study

might help to bridge gaps between the ways messages are communicated

between different levels of organizational hierarchy.

A study could also be conducted with current students enrolled in

hospitality schools in and outside of Switzerland to compare how

populations from different cultures choose communication channels. While

my initial intention is to conduct research with students at my international

hospitality school, a future research project could be imagined with other

international schools as well. Conducting research with university students

outside the U.S. would fill a gap in communication research which has not

yet been saturated in Europe.

Another study could be to take a channel, like the Internet, and break

it down into smaller subchannels, (e.g., types of sites used or purposes for

using them) or to take a channel which was not addressed directly in this

study like social media networks, and analyze its utility and subsequent

evolution in education or the workplace. Junco and Cotton (2011, 2012)

conducted many studies focusing uniquely on social media networks and

their studies could be replicated with an international student population as

well. Researchers have shown how much time and for what reasons U.S.

participants, both students and professionals use the Internet. A future study

might confirm or disprove these findings with an international population.

A further study could be conducted on innovative communication

technologies used in hospitality marketing, like Instagram, PinInterest,

Google+, or Twitter. These technologies are changing the ways hotels and

restaurants advertise offers and entice new clientele. As they are relatively

new, the research is not yet saturated and may offer another gap in

communication research. Gratifications sought could be examined with new

technologies before they are introduced into the workplace. Employers could

be asked what gratification they are seeking from a new communication

technology before implementation into their respective establishments.

Implications for Social Change

Previous scholars have shown gaps between the communication skills

possessed by young graduates and the skills that employers expect new

employees to have. My research on communication channels can contribute

to the literature and perhaps offer suggestions on how to adapt

communication courses in undergraduate programs to the reality of the

workplace. Since beginning my doctoral program at Walden, I have already

made changes in my own courses in regards to social change. I address the

potential for positive social change via communication channels including

the Internet, social media networks, and crowd-funding projects.

Through this dissertation, I have shown that the combination of

frequency of use, duration, and functions can be used to predict GO when

choosing communication channels in the workplace. Better understanding of

communication channels and how we communicate in the workplace could

lead to less wasted time, fewer misunderstandings, less aggravation, and,

eventually, greater GO. If curricula could be written to prepare

undergraduates how to choose communication channels for specific

communication situations encountered in the workplace, I believe there

would be more successful communication exchanges. Students would

become more efficient communicators, alleviating many of the

communication problems discussed in Chapter 2. The potential to make a

positive contribution to social change is inherent in communication.

Research such as mine could be the beginning of a new field linking

hospitality, communication, and marketing. Company executives often

introduce new communication technology into the workplace without

sufficient training. As a result, employees refuse to use new technology.

Studies on how communication channels could be best utilized could

contribute to the effective use of these communication channels and the

ultimate success of a company or brand.

A further social change element could be researching communication

channels and their link to a company’s CSR as seen in Chapter 2. With so

many channels to choose from, companies are often confused as to which is

the best one to communicate (and permit participation in) their social change

activities. I could test recent technologies in a future study to gauge how

other companies are using communication channels to report their CSR and

which ones are the most effective.

Conclusion

Through the present study, I have confirmed that duration, frequency

of use, and functions were significant in predicting GO in communication

channels used in the workplace, although the regression model proved to be

weak (52% variance). My findings confirmed previous research studies

which showed that age is not a factor when choosing communication

channels. Although there were some inevitable limitations (i.e., sample size),

the potential for future research is great. Various new sample populations

could be used for future studies. The focus could shift to minimizing the

number of channels studied to be able to analyze them in greater depth. I

have contributed to the literature on communication channels and could

conduct future studies to inspire users to promote positive social change

when using communication channels in education and the workplace.

Students also viewed