r...

.r.... we

{~3‘\..l
1.... I
.

.3.

”5
Ln”.
.... a: ..

:.. ..\ «
hml xn.
'.'I‘.&.,I , I,

. .4.
:.-amr.fi.;
3 ad?
. ..

I 11‘ n
v: 2va:n::vlu a
LPhflrnfinflu Iv .:!I(hu§tl.
3‘3; §$A:O.~1.ifuyl’
r: . risk. tfluo!
U313.
. ..i
'
le‘isaalttx
. I .
'9'.4upflwnu~ :nlflxfi.

aWh,§...Mnu .
v!..fi...r1¥.ie:£l
. vmfis—E

.\ .A. ~ to $9»! to
9:83.“: ”(it‘s—.9 t
to .

anus
1..

x
341...! Illll‘ttb.‘
3|}..."fll! 1P!-
I...“{I-..:
«an.
it

.v 3:19.65.
Izai- .1 n»

 

3% fig.

..:l
.5. z

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the
thesis entitled

GENDER EFFECTS OF COMMUNICATOR AND
RECIPIENT
IN RELATIONSHIP VIOLENCE PREVENTION
MESSAGES

presented by

KAOUTAR TBATOU

has been accepted towards fulfillment
of the requirements for the

MA degree in Advertising

 

 

 

.p—S’,

I

 

\
War Professor’s Signature‘
57/? / M

Date

MSU is an Affinnative Action/Equal Opportunily Employer

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KzlProleccalPres/CIRCIDateDue.indd

 

GENDER EFFECTS OF COMMUNICATOR AND RECIPIENT
IN RELATIONSHIP VIOLENCE PREVENTION MESSAGES

By

Kaoutar Tbatou

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF ART
Advertising

2010

 

ABSTRACT

GENDER EFFECTS OF COMMUNICATOR AND RECIPIENT
IN RELATIONSHIP VIOLENCE PREVENTION MESSAGES
By
Kaoutar Tbatou
This study examined main and interaction effects of communicator (voice-over)

and recipient’s genders on the response to relationship violence prevention public

I

service announcements (PSA). Message response refers to source credibility,
emotional response, cognitive response, attitude toward the issue, attitude toward the
communicator and attitude toward the PSA, and behavioral intention. Participants
(N=223) took an online survey measuring their response to two versions of a
relationship violence prevention PSA. Each version included either a male or female
voice-over. The results suggest that the gender of the communicator significantly
affects perceived source credibility, such that the female communicator was perceived
as more credible than the male communicator. The gender of the recipient has a
significant impact on only one aspect of emotional response, fear, and on attitude

toward the issue of relationship violence. In both cases, female recipients had higher

degree of fear and a more positive attitude issue than males did. No communicator’s or

recipient’s gender impacts were found on behavioral intention. Also, no gender
interaction effects were found on any response variable.
The topic requires further examination, but the findings have a set of academic and

professional implications. Future research recommendations are also discussed.

DEDICATION

Al hamdu Iillah.
To my loving family and friends.

I never would have reached my goals without your unwavering trust and support.

 

ACKNOWLEDGEMENTS

I would like to express my gratitude to Dr. Hye-J in Pack, my thesis advisor, for
her continuous encouragement and guidance throughout the progress of this thesis.
Her insight, enthusiasm and commitment have been a great source of inspiration to me
as I made my first steps in research.

I also extend my deep thanks to my two committee members Dr. Keith Adler for
his generous psychological support and academic guidance, and Dr. Thomas Hove for
always providing time and thought in order to better the style and substance of my
thesis.

My thanks also go to Dr. Mira Lee, Dr. Karen Mishra and Dr. Thomas Hove for
allowing the participation of their students in this study.

Finally, I wish to thank my volunteers and friends — Angela Gongqi Yu, Kana
Aoki, Maria Pintar and Onur Agirseven — for the time and effort they have dedicated

to make my work easier.

TABLE OF CONTENTS

LIST OF TABLES ......................................................................... vii
LIST OF APPENDICES ................................................................... viii
CHAPTER 1
INTRODUCTION 1
l .1 . Background ...................................................................... 1
1.1. 1. Relationship violence as a serious social issue ...................... 1
1.1.2. Definition of relationship violence ...................................... 3

1.1.3. The need for communication research on relationship violence. .. 4

 

1.2. An overview of the thesis ..................................................... 5
1.2.1. Purpose of the study and general research question ................ 5
1.2.2. Theoretical framework .................................................... 5
1.2.3. Overview of the study design .......................................... 6
1.2.4. Significance of the study ...............................................

1.2.5. Organization of the thesis .............................................. 16
CHAPTER II
BACKGROUND 17

2.1. Relationship violence as a serious social issue ............................ 17
2.1.1. Forms, causes and effects ............................................. 17
2.1.2. Action about relationship violence .................................... 22

2.2. The need for communication research on relationship violence ........... 25

CHAPTER III
THEORETICAL FRAMEWORK 31
3.1. Communicator and recipient genders: 32
Main effects on message response ..........................................
3.1.1. Effects of the communicator’s gender on message response ........ 32
3.1.2. Effects of the recipient’s gender on message response ............ 37

3.2. The match-up hypothesis ........................................................ 40
3.2.1. Congruence: a main match-up mechanism ............................. 41
3.2.2. Can gender be a match-up factor? ................................................... 42
3.2.3. The match-up hypothesis within the present study .................... 43

3.3. Gender schema theory: the role of gender schema in message response. 44

3.4. Hypotheses and research questions ............................................. 46
3.4.1. Main effects: hypotheses and research question ...................... 47
3.4.2. Interaction effects: hypotheses ........................................... 48
CHAPTER IV
METHOD 50
4.1. Study design ....................................................................... 50
4.2. Sample ............................................................................ 50
4.3. Procedure ........................................................................... 5 1
4.4. Stimuli .............................................................................. 52
4.5. Pretest .............................................................................. 53
4.6. Variables and measures .......................................................... 55
4.6.1. Independent variables ...................................................... 55
4.6.2. Dependent variables ...................................................... 55
4.6.3. Control variables ........................................................... 62
4.7. Analysis strategy ............................................................... 69
CHAPTER V
RESULTS 70
5.1. Descriptive statistics .............................................................. 70
5.2. Hypothesis and research questions testing .................................... 72
5.2.1 . Testing the main effects ................................................ 72
5.2.2. Testing the interaction effects ............................................ 78
CHAPTER VI
DISCUSSION 82
6.1. Overview of the findings ......................................................... 82
6.2. Discussion ......................................................................... 83
6.2.1. Discussing the significant findings ...................................... 83
6.2.2. Discussing the null findings .............................................. 87
6.3. Limitations of the study .......................................................... 92
6.4. Suggestions for future research ................................................. 94
6.5. Implications ........................................................................ 95
6.5.1. Practical implications ..................................................... 95
6.5.2. Theoretical implications ................................................. 97
CONCLUSION ............................................................................ 99
TABLES ..................................................................................... 101
APPENDICES .............................................................................. 123
BIBLIOGRAPHY .......................................................................... 134

vi

i

LIST OF TABLES

Crosstabulation of Communicator and Recipients’ Genders. . . . . 101
EFA of Scale Dependent variables .......................................... 102
List of Covariates and Affected Dependent Variables .................. 105
One-way Analysis of Variance for Effects of Likeability on Fear, 106
Anger and Attitude issue .......................................................

 

One-way Analysis of Variance for Effects of Perceived Similarity on 107

Fear ............................................................................

One-way Analysis of Variance for Effects of Prior Involvement in 108
Relationship Violence on Guilt and Behavioral Intention .............

Correlations of Major Variables (Pooled Sample) ........................ 109
Correlations of Major Variables (Male Participants) .................... 110
Correlations of Major Variables (Female Participants) ................. 11]
Means and standard deviations of dependent variables by gender 112
Analysis of Variance for Dependent Variables ............................ 115
Analysis of Covariance for Dependent Variables ......................... 117

vii

LIST OF APPENDICES

Appendix A: Wording of the survey questions by variables used... 122
Appendix B: Storyboard of the PSA ..................................... 128
Appendix C: IRB-approved consent form 131

viii

CHAPTER I - INTRODUCTION

1.1 — Background
1.1.1. Relationship violence as a serious social issue

Intimate relationship violence is one of the oldest social phenomena. For
centuries, both social and legal systems considered it a social taboo and a private
matter. Until the 19th century for example, husbands in Britain were allowed by law to
batter their wives — an act that was labeled as “lawful correction” (Hunt & Martin,
2001). However, interest from the media, in addition to governmental and non-
governmental organizations has grown considerably since the middle of the 20th
century — an advance in which women’s’ movements played a considerable role
(Huang & Gunn, 2001).

Since then, thousands of relationship violence cases have been reported every year,
in many countries around the world. In the United States alone, the US Department of
State estimates a yearly average of 4.8 million women experiencing physical assaults
and rapes from intimate partners (Tj aden & Thoennes, 2000). However, although
relationship violence cases are increasingly reported, and despite the growing public
awareness of it, the issue of relationship violence still persists, and even seems to get
more complicated and increasingly difficult to eradicate (Crowell, 1996).

Relationship violence has considerable consequences on individuals as well as
societies. On the individual level, the harm could be physical (e.g., injuries and
bruises) or psychological (e. g., depression, low self-esteem), and can affect both the

social and professional lives of victims as well as perpetrators.

The emerging importance of relationship violence has translated globally into
governmental and non-governmental advocacy, policy-making and communication
efforts. Among many objectives, these efforts aimed at raising awareness about
relationship violence and changing the behaviors of both perpetrators and victims.
Other efforts aimed to increase the involvement of the community by trying to change
the passive behavior of bystanders — people who are aware of a violent relationship
but do not try to stop it or help those involve in it.

There are many examples of large-reaching campaigns about the issue of
relationship violence. In the United States, the National Coalition against Domestic
Violence has maintained the tradition of organizing a yearly awareness month since
1987. Activities organized during those awareness months have aimed to sensitize
the public about the issue, support and guide victims, but also bring together advocates
to share experiences and ideas about their efforts to eradicate relationship violence.

On the global scale, Amnesty International launched in 2004 an international
campaign labeled “Stop Violence against Women”. Its objectives were to break the
silence around different forms of violence faced by women both inside and outside
households. As part of this campaign, Amnesty International supported the advance
of anti-violence legislation in many countries including the US.

Efforts like the ones described above fall within the category of professional
endeavors that thankfully have been able to make a difference. But besides
professional interest, a considerable amount of academic eflorts have been devoted to

the issue of relationship violence. This interest takes the shape of academic research

and experimental studies (See reviews by e.g., Gellcs & Conte, 1990; Pirog-Good &
Strets, 1989; and Johnson & Ferraro, 2000).

But as will be discussed in section 1.1.3 of this introduction, the issue of
relationship violence did not receive equal degrees of attention from all academic
disciplines. In fact, an important core of psychology and sociology research
scrutinized the issue of relationship violence. However, the academic field of
communication research still needs to devote more attention to the issue because such
research can help design effective relationship violence prevention campaigns.

Before discussing the scarcity of communication academic interest in relationship
violence, let us first briefly define the concept of relationship violence as it is used in
the present research.

1.1.2. Definition of relationship violence

The present study adopts the term “intimate relationship violence”, shortened as
“relationship violence” or “violence”. The definition adopted is the one provided by
the US Department of Justice, describing intimate relationship violence as “homicides,
rapes, robberies and assault committed by intimates [including] current or former
spouses, boyfriends or girlfriends, including same-sex relationships” (US Department
of Justice, 2007).

It would be worth noting that the perpetrator of relationship violence could be
either male or female. However, the present research adopts the perspective where
females are victims and males are perpetrators. This perspective is reflected in the
public service announcement (PSA) used as a stimulus in the study. Choosing to

portray a female victim and a male perpetrator is based on the statistical evidence that

within an intimate violent relationship, females have six times higher chances than
males of being the victims (US Department of Justice, 2007).

After defining the concept of relationship violence as it is used in the present
study, the next section explains that relationship violence deserves more attention
from communication research because this type of research can increase the
effectiveness of prevention communication campaigns.

1.1.3. The need for communication research on relationship violence

Research about relationship violence was mainly conducted in the fields of
sociology, psychology and public health. Reviews of research on relationship
violence and related topics like “family abuse”, report that previous investigations
were mainly conducted from sociological or psychological perspectives (e.g., Gelles &
Conte, 1990, Pirog-Good & Strets, 1989, and Johnson & Ferraro, 2000).

Nevertheless, despite this attention from sociology and psychology research, other
academic disciplines have attributed lower levels of attention to the issue of
relationship violence. The field of communication research is one of them. The result
is a need for communication research that focuses on relationship violence.

It is important to address this need because communication research can increase
the effectiveness of pro-social messages aiming to reduce relationship violence. This
is the advantage from which other social issues have benefited like driving behavior
(e.g., Murry, Jr. et al., 1993; Treise et al., 1999; Wolburg, 2001) and substance
abusc(c.g., see a review of 60 years of research by Crano & Burgoon, 2001). The aim
of the present study therefore is to contribute to the core of communication research

that specifically focuses on relationship violence.

After explaining the need for communication research that specifically focuses on
relationship violence, the next section of this introduction presents an overview on the
present thesis. It explains the objectives of the study, defines the concept of message
response in details, and presents the study’s general research question.

1.2. An overview of the thesis

1.2.1. Purpose of the study and general research question

The present study examines the effect of communicator and recipient’s genders
on the response to relationship violence prevention PSA’s. The findings have the
ambition to shed the light on how the genders of the communicator and the recipient
could affect source credibility, emotional response, cognitive response, attitudes and
behavioral intention. The findings aim to provide useful guidelines for future health
campaigns targeted at relationship-violence bystanders, with the aim of increasing the
effectiveness of such campaigns.

Therefore, the general research question that the present study aims to answer is:

How do communicator and recipient’s genders affect the recipients’ message
response to relationship violence prevention public service announcements
(PSA’s)?

1.2.2. Theoretical framework

As the chapter of Theoretical Framework (chapter 111) explains, the expected main
effects of gender on message response are based on the findings of previous research
that will be reviewed in the first section of chapter III. Additionally, the same chapter

provides two theoretical grounds for the interaction effects between the gender of the

communicator and that of the recipient. These two theoretical grounds are the match—
up hypothesis (Kahle & Homer, 1985) and gender schema theory (Bem, 1981).

Many studies have examined the communicator and recipient genders and their
main influence on message response. As an example, many studies demonstrated that
either the gender of communicator or that of the recipient could affect perceived
source credibility (e.g., Markham, 1988), emotional response (e.g., Levine &
Zimmerman, 1996), cognitive response (e.g., Darley & Smith, 1995), attitudes (e.g.,
Lori & Pradeep, 2003) and behavioral intention (e.g., Nysveen, Pedersen &
Thorbjornsen, 2005).

Additionally, the gender-sensitivity of the message and its gender-salience
produces more chances that the interaction between the communicator and recipients’
gender would affect message response. In fact, Kempf and Palan (2006) maintain that
message response is more likely to be influenced by gender when gender is salient to
the content of the message.

1.2.3. Overview of the study design

The present study was conducted using a 2 between (male or female
communicator) X 2 between (male or female recipient) factorial design. The study
tests the participants’ response to two versions of a relationship violence prevention
PSA where the voice-over is male in one version and female in the other. Participants
were randomly assigned to one of two conditions, depending on whether their genders
match or mismatch with that of the communicator. The participants were
undergraduate Michigan State University students who participated in return for extra-

credit.

Defining message response:

In the present study, message response refers to five dependent variables: source
credibility, cognitive response, emotional response, attitudes and behavioral intention.
The choice of these five variables as message response variables is based on two
reasons. First, these variables appear recurrently in message-response and
information-response models. One example is the dual mediation model of
advertising response (Mitchel & Olson, 1981). This model incorporates two of the
five message response variables tested by the present study: cognitive response and
attitudes.

Second, the five variables in fact constitute sequence of stages that the message
recipient goes through, starting with exposure to the message and ending with
behavioral intention or behavior itself. Upon exposure to the message, the recipient
has a certain degree of perceived source credibility that affects the rest of the response
variables including cognitive and emotional responses, attitudes and behavioral
intention. Also, within the same response sequence, each of the variables exerts
influence on the other.

As a side note, it would be worth mentioning that despite the aforementioned
influence that happens among the five response variables, the present study examines
all of them as dependent variables. There are two reasons to explain this choice. First,
the focus of the present study is the effect of gender on all these five variables, and not
the effects that happen between the variables. Second, the order in which the influence
takes place between variables is still debated among scholars. For example, some

believe that it is attitude that influences behavior (e.g., Lavidge & Steiner,1961), while

others believe that the opposite could also happen (Palda, 1966). Therefore, the
present study considers all the five response variables as dependent variables.

Going back to the importance of including the five response variables in the
present study, more reasons are available to explain the importance of each one of
them. First, source credibility is important because it impacts persuasion (Harmon &
Coney, 1982). On the other hand, source credibility is also influenced by various
variables. Some of these variables are psychographic like the recipients’ initial opinion
prior to exposure to the message (Stemthal et al., 1978). Some other variables are
demographic like age and, of bigger relevance to the present study, gender (e.g.,
Oakley, 2000).

Second, emotional response is important because emotions play an important role
in shaping the response to persuasive messages (Dillard & Peck, 2000) as well as the
subsequent behavioral intention (e.g., Thoits, 1989). Also, according to previous
research, emotional response is influenced by gender (e.g., Levine & Zimmerman,

1996). Therefore, the present study explores whether the participants’ emotional
response to the relationship violence prevention PSA varies according to the
recipient’s gender, the communicator’s gender, and their interaction.

Third, cognitive response is important because it is also shown to influence
attitudes (e.g., Mitchell & Olson, 1981), that in turn influence subsequent behavior.
Also, just like source credibility and emotional response, cognitive response is
impacted by gender (e.g., Darley & Smith, 1995) and is therefore included as a

dependent variable.

Fourth, in addition to source credibility, emotional response, and cognitive

response, this study also measures attitudes as message response variables. In fact,

three types of attitudes are measured: Attitude toward the message itself (attitude pSA);

attitude toward the social issue (attitude issue); and attitude toward the communicator

(attitude voice-over)-

It is important to measure attitudes for two reasons. First, gender was found to
influence attitudes (e.g., Lori & Pradeep, 2003; Byrne, 1961). Second, attitudes shape
the subsequent behavior of message recipients (e.g., Dillard & Peck, 2000). Since the
PSA used in the present study is primarily a persuasion message that aims to change
bystanders’ passive behavior, it follows that it is important to measure their attitudes.

Fifih, measuring behavioral intention is important because it is the precursor of
actual behavior (Parker et al., 1992). Again, previous research found that behavioral
intention can be affected by gender (e.g., Nysveen, Pedersen & Thorbj ornsen, 2005)
which makes it important to measure this response variable within the context of
relationship violence prevention PSA’s.

After defining the concept of message response and explaining the rationale
behind the choice of each of the five response variables, the next section provides
details about the purpose of the present study and the general research question.

1.2.4. Significance of the study

The present study is an important endeavor for five reasons. First, the issue of
relationship violence is serious and persistent. Second it is important to focus on

VOice-over as a communicator because voice-overs are very frequently used in PSA’s

and ads, but still have not received much empirical research attention. Third, research
about bystander behavior is important because the community can play an important
role in eradicating social phenomena like relationship violence. Fourth the focus on
the response of college-aged students is important because, like adults, this age
category is also under the risk of relationship violence. Fifth, PSA’s as a stimulus
deserve research attention because they are a central tool for pro-social campaigns.
The importance of research about relationship violence

Relationship violence has reached alarming levels of spread and frequency, both at
the global and local levels, and therefore deserves the attention of research because it
can increase the effectiveness of relationship violence prevention campaigns. For
example, The World Health Organization (WHO) identified relationship violence in
2002 as the most frequent of all violence forms that women face in the world,
including rape and sexual assault. In the United Stated, a yearly average of about 4
million women experience assault by an intimate partner (American Psychological
Association, 1996). More details regarding the importance if this issue are discussed in
Chapter II — Background, but these statistics can probably already shed light on the
significance of a communication-focused study that deals with relationship violence.

The importance of research focused on PSA voice-overs

In this study, the communicator is the PSA voice-over. Also called an off-camera
commentator or narrator, a voice-over is defined as an announcer who is not on
camera but whom the audience listens to while watching the visual action on a video

(lmber, 2000).

10

Research focus on voice-overs is important for three main reasons. First voice-
overs are very frequently used in commercial adsand social-related PSA’s. However,
not many studies have examined audience response to the voice-overs. In fact, even
when previous studies examined the effects of communicator-recipient gender, the
communicator used in those studies either delivered the message in written form (e.g.,
Bochner & Insko, 1966), or was an individual whom the participants saw on screen
(e.g., Caballero & Pride, 1984) or in person during the experiments (e.g., Berscheid,
1966). When in some studies the communicator was an audio voice, the message was
not an ad or PSA, but rather a long lecture or a speech (e.g., Wood, Gorenflo & Santer,
1994). Therefore, the present study may be among the first to examine the effect of
communicator-recipient similarity, using a voice-over as the communicator.

Since an important portion of PSA’s do use voice-overs, it would be interesting to
examine response to the message delivered by a voice-over. This can help future PSA
and ad designers make the right choice of voice actors.

Second, testing message response when the communicator is a voice-over is
fundamentally different from testing it when the communicator is an individual who
can be seen or interacted with by the participants. For example, many previous source
credibility experiments allowed the participants to see the communicator, either on
screen or directly in the same location (e.g., Berscheid, 1966; Woodside & Davenport,
1978). Other source credibility studies provided the participants with information
about the communicator like their biography or picture for example (e.g., Caballero &

Pride, 1984). While this design was adequate for the purposes of those studies, it also

11

implies a risk of confounding elements such as physical attractiveness, body language
or even dress style.

In real life however, we are daily exposed to a considerable number of ads or
PSA’s while we do not know much about the person who is speaking as the voice-
over. One of the few things we can know however is the gender of the voice-over.
Accordingly, this study tries to test response to the PSA within conditions that are
similar to the audience’s usual exposure to PSA’s. The only thing that will be known
about the voice-over is gender that can be inferred from the voice itself.

Third, from a PSA production point of view, the choice of a voice-over is one of
the important arrangements that PSA and ad designers have to make. Choosing the
right voice-over is important because it is the “voice” that represents the brand or
advocated behavior. This may be as important as the “image” of a message or a brand.
Accordingly, the degree to which a voice-over is likeable or convincing may make a
big difference in the audience’s response to the message. Therefore, studies like the
present one can inform PSA and ad designers during the process of voice-actors’
choice.

The importance of research about bystander behavior

The PSA used as a stimulus in the present study targets relationship violence
bystanders. Therefore, the third argument that explains the significance of the present
study is related to the decision to focus on bystanders as the message recipients.

Public and academic interest in bystander behavior started in the mid-sixties after
the shocking homicide of a New York City resident, Kitty Genovese. Genovese was

murdered in her parking lot within sight and/or hearing of dozens of neighbors, but no

12

one came to her rescue. The incident triggered the interest of psychology and
sociology researchers who tried to understand the factors that would motivate
bystanders to provide help in a distress situation (Laner & Benin, 2001).

Within the context of relationship violence, understanding how to change
bystanders’ passive attitude is of great importance. Unlike the victim and the
perpetrator, bystanders can play a significant role in decreasing the spread of
relationship violence because they have a relatively objective stance towards the
violence situation. In such a situation, the perpetrator and victim are subjective
because they are highly involved in the issue. This usually delays or completely
prevents their action towards change. On the one hand, the perpetrator would not try to
change the situation because he is afraid of legal punishment, wants to maintain
control over the victim, or is simply not aware that his behavior has psychological
causes that can be treated. On the other hand, the victim would not react out of fear,
financial dependence, or if she is unaware of the ways she could get help. This
justifies the importance of changing bystanders’ attitude from passive into active in
order to help the victim, the perpetrator, or both of them.

In fact, bystanders’ participation in eradicating this social phenomenon is probably
indispensible. By choosing either a passive or active behavior, bystanders can
contribute either to the perpetuation or the reduction of this social phenomenon across
generations (Gracia & Herrero, 2006). It is therefore important to conduct
communication-focused research that examines ways to change their attitudes and
behaviors regarding relationship violence.

The importance of research about college students as PSA targets

13

The importance of the present research can also be explained from a fourth
perspective: the age category of the target audience of the PSA used as a stimulus.
The participants of the present survey study are college students, and there are at least
two reasons why this age category deserves attention from research.

First, the age category of college students is at the mid-section between childhood
and adulthood. The issue of relationship violence could be partly solved on the short
run by designing campaigns targeted at current violent couples, who are adults in most
cases. But it is also important to solve the issue of relationship violence in the long
run by preventing it from propagating in future generations. Therefore, it would be
beneficial to make young people aware of the issue and involved in solving it as early
as college-age.

The second reason why focusing on college students is important is that college
students are themselves under the risk of being involved in relationship violence, just
like adults are. In fact, the issue of relationship violence is faced across races, social
strata and ages (e.g., Wilt & Olson, 1996).

It is generally difficult to obtain precise statistics about on-campus relationship
violence due to the “confidential” nature of the problem. But some studies have
revealed alarming numbers. For example, a study conducted in 1981 among college
students revealed that about 60 % of them had personally known someone who had
been involved in relationship violence (Makepeace, 1981). The same study showed
that almost 1 in 4 of the surveyed students had personally experienced a violent
relationship. In 1998, a report from the US Department of Justice stated that women

aged 16-24 experience the highest rate of relationship violence. More recently, Hall

14

Smith et al. (2003) conducted a longitudinal study and concluded that women who
were physically assaulted during adolescence were at greater risk for re-victimization
during college years than those who have not.

Therefore, because the college-age category is exposed to the risk of relationship
violence and because it can play a role in reducing the occurrence of violence in future
generations, it is important to empirically examine the response of college-aged
students to relationship violence prevention PSA’s.

The importance of research focused on testing PSA ’s

Finally, the significance of the present study can also be illustrated by explaining
the importance of communication research that focuses on the response to PSA’s.

It is important to study bystander response to PSA’s because PSA’s are one of the
major tools in social marketing and health communication campaigns. It is through the
use of tools like PSA’s that bystanders can be made aware that their attitudes and
behavior concerning relationship violence can make a big difference. For example, a
PSA can let bystanders know exactly what to do to help reduce relationship violence.
An active bystander could save a couple by directing them to psychological
counseling, or even save a life by reporting a relationship violence case.

In sum, the present study aims to investigate the main and interaction effects of
communicator and recipient genders on persuasive messages within the context of
relationship violence PSA’s targeting bystanders. It examines how the genders of the
message communicator (PSA voice-over) and message recipients (study participants)
individually affect message response, but also how they interact and whether the

match or mismatch of genders affects the audience response to the PSA. What is

15

meant by message response is the participant’s perceived source credibility, emotional
response, cognitive response, attitudes and behavioral change.

The findings of this study aim to respond to a need for communication research
geared toward the improvement of anti-relationship-violence campaigns that target
bystanders. Additionally, this study aims to contribute to other research areas like
“bystander behavior” and “gender similarity” and its effects on message response.

The next section, the last of this chapter, maps out the chapters of the present
thesis and reviews their content.

1.2.5. Organization of the thesis

The present thesis is made up of six chapters. Afier the Introduction (chapter I),
the background of the present study is discussed in chapter 11.

Chapter III explains the theoretical framework of this study, and presents the
hypotheses and research questions. Chapter IV, Method, describes the study design
and procedure, and provides details about the study participants, the stimulus, the
stimulus pretest, the variables and their measures as well as the analysis strategy.

The results of the present study are presented in chapter V, and discussed in
Chapter VI. Chapter VI also discusses the limitations of the present study, provides
suggestions for future research, and explains the study’s practical and academic

implications.

16

CHAPTER II — BACKGROUND

Relationship violence is a social issue faced across generations, races and social
strata (Hotaling & Sugarman, 1986). The present chapter claims that relationship
violence is a serious social phenomenon, a claim based on many facts. First,
relationship violence exists all over the world and takes many forms including
physical, psychological and verbal. Second, violence has multiple physical,
psychological and socioeconomic causes and effects on individuals as well as society.
Also, based on a literature review, this chapter also shows that relationship violence
has not received as much attention from communication research as it did from
psychology and sociology research. The chapter therefore suggests that it deserves
more attention from communication research because such attention is key to more
effective relationship violence prevention campaigns.

2.1. Relationship violence as a serious social issue

2.1.1. Forms, causes and effects

Defining relationship violence: Numerous researchers and theorists have defined
relationship violence in publications pertaining to disciplines as diverse as sociology,
law, health and psychology (Rothenberg, 2002). But across the literature, there is no
consensus about a unified term to describe this social phenomenon. Among other
labels, relationship violence was described as “family violence”, “family abuse”,
“close-relationship abuse”, and “spousal abuse. In some publications, two or more of

these labels are used interchangeably (e.g., Gelles, 1976; Burgess & Draper, 1989),

17

while in others, differences are pointed out. For example, Gelles and Conte (1990)
argue that the term “family violence” is broader than violence in the couple, which
they also call “courtship” or “marital violence”. They explain that the term “family
abuse” is broad because family abuse could also enclose child abuse and elderly
abuse.

For the purposes of the present research, relationship violence is defined as
“homicides, rapes, robberies and assault committed by intimates [including] current or
former spouses, boyfriends or girlfriends, including same-sex relationships” (US
Department of Justice, 2007). Additionally, the present study’s conceptual definition
of relationship violence acknowledges that this phenomenon could take physical as
well as psychological forms. Examples of physical violence are battering and hitting;
psychological violence includes threats, insults and stalking. Sexual abuse is also
another form of relationship violence that creates physical as well as psychological
repercussions. In fact, some researchers believe that sexual abuse in one of the most
common forms of relationship violence (Ellsberg, Heise, Pena, Agurto, & Winkvist,
2001).

Concerning its parties, relationship violence could involve men or women,
whether their relationship is heterosexual or homosexual (Waldner-Haugrud & Gratch,
1997). Also, the perpetrator of relationship violence could either be male or female
(National Coalition against Domestic Violence, 2009). However, statistical evidence
suggests that women have much higher risks than men of being the victims of

relationship violence (Saunders, 1986) Also, research suggests that the most severe

18

forms of violence are perpetrated by males rather than females, and that in many cases
females resort to violence as a form of self defense (Makepeace, 1983).

Relationship violence takes place regardless of age, race or social strata. In a study
covering college students on a national scale, White and Koss (1991) report that 32%
of the surveyed women experienced sexual violence at least once in their lives. These
reported violence experiences spanned from age 14 to college age. Concerning the
prevalence of violence across cultures and races, Heise, Ellsberg and Gottemoeller
(1999) conducted a review of surveys from 36 countries. The review revealed that 10
to 50% of women who previously were in an intimate relationship have been victims
of relationship violence.

Causes of relationship violence: A considerable amount of psychological and
sociological research was conducted to explain the factors behind relationship
violence. The factors leading to relationship violence belong to various categories
including psychological, social, and socioeconomic.

One of the most common psychological causes mentioned by researchers is
previous experience with violence during childhood. Such experiences include either
witnessing violence in the household, or being a victim of violence as a child (e.g.,
Hotaling & Sugarman, 1987) and could lead to future violent behavior during
adulthood.

Another psychological factor is the perpetrator’s desire to maintain economic
and/or relational control over the victim (Telch & Lindquist, 1984). This desire leads
to excessive use of power, in addition to the manipulation of the partner as well as

children (Buel, 2002).

19

Research also confirmed other factors like substance abuse (Hotaling & Sugarman,
1986) and alcohol abuse (Bennett, 1995). Substance and alcohol abuse could increase
the frequency and severity of the perpetrator’s violent behavior. On the other hand,
when substance and alcohol abuse are combined with psychological disorders, they
could become a reason for repeated victimization (Hotaling & Sugarman, 1986).

The victim could also contribute to the perpetuation of the violence situation by
remaining silent and not seeking help or by not leaving the relationship. The victims’
silence also has psychological causes including fear (Mitchell & Hodson, 1983),

shame (Olson et al., 1996) or lack of trust in health care providers who may be a
potential source of help (Buel, 2002). Some victims are also overpowered by a feeling
of guilt, caused by their belief that the failure of their relationship is their own fault
(Buel, 2002). When the victim is a woman, this situation of helplessness and
acceptance is called the “battered woman syndrome” (Walker, 1984). Walker explains
that repeated victimization of women makes them develop a learned habit of accepting
and submitting.

In addition to psychological factors, research also attributed relationship violence
to social factors. These include gender roles for example. Although there is no
consensus about this claim, some researchers suggest that women who were socialized
to accept male dominance are under bigger risks of being victims of relationship
violence than women who were not (Hotaling & Sugarman, 1986). Traditional sex
role expectations are yet another social factor that many researchers identified as a
cause of relationship violence (e.g., Telch & Lindquist, 1984). Additionally,

relationship violence may also be caused by socio-economic factors. In fact, the

20

victims’ inability to speak up may be caused by financial dependence on the abusive
partner (Gelles, 1976).

Effects of relationship violence: Relationship violence has considerable
consequences on individuals as well as society. At the individual level, both the victim
and the perpetrator are affected. The perpetrator faces legal liabilities, as well as
health-related and socioeconomic effects. Health related effects on the batterer include
self-inflicted injuries and psychological distress. Socioeconomic consequences include
decreased levels of concentration and productivity at work or even the loss of
employment (Gerlock, 1999).

On the other hand, effects faced by the victim can take either physical or
psychological forms. Physical consequences of relationship violence span from
bruises to serious or lethal injuries. Whitman and McKnight (1985) collected data
from three American hospitals as part of a study about head injuries and deaths caused
by those injuries. Their study revealed that 30% of the identified cases happened
within an intimate relationship. Out of the total head injuries identified by the study,
15% were caused by a spouse, 14% by a boyfriend, and 1% by a girlfriend.

Physical effects of relationship violence on female victims could go finther than
external injuries. For example, female victims of relationship violence are more likely
than other women to suffer reproductive health problems like unwanted pregnancies,
vaginal and cervical infections and bleeding during pregnancy and low birth weight
after delivery (Curry, Perrin, & Wall, 1998). Substance abuse could also be a physical

consequence due to the psychological instability that is caused by abuse.

21

In addition to physical consequences, victims of relationship violence suffer longer
lasting psychological impact (Berry, 1998). Psychological impacts include mental
disorders and depression (Heise ct al., 1999), as well as low self-esteem and high
levels of psychological distress (e.g., Cascardi & O’Leary, 1992; Levendosky &
Graham-Bermann, 2001). Victims of relationship violence also end up with a low
perception of their ability to control their own lives and the violence situation in which
they are involved (Umberson, Anderson, Glick & Shapiro, 1998).

The physical and psychological impact on the victim bring along social and
professional consequences. Even when physical outcomes are not lethal or physically
impeding, violence is reported to have professional repercussions such as delays and
absence in work places in addition to the loss of concentration, self confidence and
motivation.

At a larger scale, the effects of relationship violence worsen in the presence of
~ children. Children who are exposed to relationship violence have high risks of
suffering social, emotional and academic consequences (e.g., see reviews by J affe,
Wolfe & Wilson, 1990; Holden, Geffner & Jouriles, 1998). In addition to these
immediate effects, research also warns from intergenerational transmission of violence
that perpetuates this social issue from one generation to another (Gelles & Conte,
1990)

2.1.2. Action about relationship violence

In the United States, different types of interpersonal violence, including

relationship violence have been tackled with different approaches. These include

22

reactive measures like punishment, but also preventive measures like education
(Whitman,1988).

Reactive measures: Security and justice systems have designed many reactive
measures to punish relationship violence cases. Opinions diverge about the efficiency
of these measures. An 18-month field experiment conducted by the Minneapolis
Police Department compared the effects of three responses to relationship violence
incidents: I) arresting the perpetrator, ii) expulsing the perpetrator from the household
for 24 hours and iii) trying to restore order (Berk, Campbell, Klup, & Western, 1992).
The experiment showed that arrest was the most effective measure to'avoid future
violence by the same perpetrator.

However, other studies showed that the efficiency of arrest depends on many
factors, including the personal traits of the arrested perpetuator. If the perpetuator is
not attached to his community, he may see the arrest as a challenge that validates his
self-worth, and therefore recommit even more violent acts (Sherman, Smith, Schmidt,
& Rogan, 1992). Similarly, another study by Pate and Hamilton (1992) revealed that
arrest deters violence perpetrators only when they are employed. When they are not
employed, arrest increases the perpetrators’ tendency to commit subsequent abuses.

Preventive measures: Besides reactive measures, research underlines the
importance of early prevention of relationship violence. Hall Smith, White and
Holland (2003) conducted a 4-year longitudinal study among two classes of female
college students. They found that women who experienced physical or sexual violence
in high school were at higher risk of being re-victimized in college than those who

have not. Also, during college years, women who were sexually or physically

23

assaulted during any year were at higher risk of being assaulted during subsequent
years than those who were not.

One of the main preventive measures recommended by researchers is education
(Whitman, 1988). Education efforts could target victims, perpetrators, bystanders and
health care providers. Victims’ education can help them recognize the signs of
violence before it takes place. Since research identified a number of patterns
distinguishing potential or current batterers (e.g., Hotaling & Sugarman, 1986 ),
education may allow potential victims to recognize the patterns of abusers and avoid
falling victims of violence. Education can also inform victims on what to do in case
the violence happened, in order to avoid re-victimization.

On the other hand, perpetrators’ education can help them understand the
psychological and/or social triggers behind their violent behavior. Education can also
inform them of the consequences of violent behavior and lead them to seek the
psychological and/or legal help that they need in order to reduce the severity or
frequency of their violence acts. Many of these educational programs have been
proven useful. For example, Safe Dates Program was launched in public schools in
North Carolina and aimed to raise the awareness of adolescents of the issue of
relationship violence. F our years after the program was launched, an evaluation study
was conducted in order to assess the four-year program’s effects on dating violence
perpetrators (Foshee et al., 2004). The results indicated that, in comparison with a
control group, the program’s participants reported significantly less perpetration and

less victimization over the four years of the program.

24

Additionally, bystanders’ education can be crucial in making communities more
sensitive to any social issue and increasing their willingness to embrace the desirable
behavior (Banyard, Plante, & Moynihan, 2004). Research shows that the degree to
which a particular community is aware of an issue can significantly affect its readiness
to change attitudes and behaviors regarding it (Grimley, Prochaska, Velicer, Blais, &
DiClementc, 1994). This means that the degree of awareness within a particular
community can affect the degree of effectiveness of pro-social messages (Banyard,
Plante, & Moynihan, 2004).

Finally, health care providers also need training in order to help reduce
relationship violence. Educating heath care providers is important because of two
reasons. On the one hand, violence victims are usually reluctant to talk to their health
care providers about their situation out of fear, shame or luck of trust that they can
help (Buel, 2002). On the other hand, physicians are prevented by many factors from
properly helping violence victims. These factors include time constraints, the absence
of training on dealing with such a sensitive issue, fear of offending the victim and a
sense of inability to treat the violence case (Olson et al., 1996). Therefore, medical
care providers need to be trained to recognize the signs, to encourage the victims as
well as perpetrators to talk about their situation, and to provide helpful guidance.

2.2. The need for communication research on relationship violence

The issue of relationship violence has attracted both professional and academic
interest in the shape of governmental and nongovernmental advocacy and policy-

making efforts, as well as communication campaigns. These professional efforts aim

25

to eradicate the problem of relationship violence. Academic interest on the other hand
takes the shape of academic and experimental research.

Research about relationship violence was mainly conducted in the fields of
sociology, psychology and public health. Reviews of research on relationship
violence and related topics like “family abuse”, report that previous investigations
were mainly conducted from sociological or psychological perspectives. Their
primary aim was to understand the reasons behind relationship violence or its effects
on the victims, perpetrators, children or society (for reviews, see e.g., Gelles & Conte,
1990; Pirog-Good & Strets, 1989; Johnson & Ferraro, 2000).

Additionally, a simple search of the phrase “relationship violence” in academic
databases shows the predominance of sociology journals (e. g. Family Relations,
Journal of Marriage and the Family), legal publications (e. g. Law and Human
Behavior), health publications (e.g. The American Journal of Public Health), and
psychology journals (e. g. Current Directions in Psychological Science).

However, despite the considerable attention that the academic disciplines of
sociology and psychology gave to relationship violence, there is a need for more
attention from other disciplines like communication research that can inform future
prevention campaigns. This has created a need for communication-focused research on
relationship violence. It is important to address this need because such research can
boost the effectiveness of the messages, that constitute the core of relationship
violence prevention campaigns. It is true that an increasing number of communication
campaigns have been devoted to relationship violence, but this can be categorized as a

professional interest. Such a professional interest is apparently not matched by a

26

similar degree of academic interest from fields like communication, health
communication and social marketing research.

In fact, a few social issues are most frequently scrutinized by research that
aims at improving communication campaigns. Relationship violence is not one of
them. For instance, a considerable core of research was conducted about ways to best
design effective campaigns about driving behavior (e.g., Murry, Jr. et al., 1993; Treise
et al., 1999; Wolburg, 2001) and addictive substance abuse (e.g., see a review of 60
years of research by Crano & Burgoon, 2001). However, research offering guidelines
from a communication perspective on how to increase the effectiveness of relationship
violence campaigns is hardly available.

One might argue that studies not necessarily need to focus on every single
social issue, since the guidelines provided by previous studies can apply to more than
one issue. For example, someone who adheres to this opinion could maintain that
research findings that confirmed the effectiveness of guilt appeals in changing
adolescents’ hazardous driving behavior could also be useful for campaigns dealing
with substance abuse or relationship violence. This could be correct, but not all the
time because every social issue, including relationship violence, has particular
characteristics. Paying attention to those characteristics can help campaign designers
create effective campaigns.

The characteristics that differentiate relationship violence from other issue can be
discussed on two levels. The first level pertains to the issue itself. Unlike many other
social issues like substance abuse, relationship violence is a highly gender-sensitive

issue. This is due to the fact that, within the context of heterosexual relationships,

27

relationship violence usually involves a man and a woman where the woman is in
most cases the victim (US Department of Justice, 2007). In fact, anthropological
research argues that relationship violence is primarily gender-based due to systems of
belief that legitimize male’s control over female behavior through the use of violence
(Ascensio, 1999).

The characteristic of gender-sensitivity creates the need for communication
research that helps design effective campaigns aiming to eradicate relationship
violence. In fact, many scholars have insisted that the intrinsic characteristics of the
issue affect the content of the message, which in turn affects the response to the
message. For example, Kempf and Palan (2006) maintain that the message recipients
are more likely to be influenced by a communicator of their own sex when gender is
salient to the content of the message. Since gender is salient to the issue of relationship
violence, it follows that the gender of the communicator or the recipient may have an
impact on message response.

Besides, there is a second level of characteristics differentiating social issues from
each other pertaining to the campaigns that are designed to eradicate each social issue.
Let us keep the example of relationship violence in comparison with substance abuse.
One difference between the campaigns aiming to eradicate each of these two issues is
related to the target audience.

On the one hand, anti-substance abuse campaigns usually target the substance
users themselves. To borrow Bagozzi and Moore’s words (1994), the messages used
in such campaigns can be described as “help-self” messages, because they call on the

users themselves to change their own behavior and limit their substance use.

28

On the other hand, relationship violence prevention campaigns could have one of
many target audiences. The target audience could either be the perpetrator, the victim
or bystanders. Therefore, to put it in Bagozzi and Moore’s words (1994), a
relationship violence prevention message would be either help-self, help-other, or a
mixture of both.

One conclusion that may be reached here is that when social issues are different,
campaign messages are different. Therefore, there are no standard guidelines that
campaign designers could use to eradicate any type social issue. On the contrary, in
order for the campaign to be successful, campaign designers may have to use different
guidelines for every social issue, and every type of audience.Thereforc, it is the role
of communication research to find out which guidelines would work best for
relationship violence prevention messages targeting bystanders for example. This is
the aim of the present study.

To recapitulate, this section discussed the need for more communication research
that focuses on relationship violence. It also illustrated the importance to conduct
research that specifically focuses on relationship violence bystanders, instead of
relying on previous studies that examined other social issues. This was explained by
the fact that each social issue has particular characteristics that make it deserve special
attention from communication research.

Two characteristics differentiate relationship violence from other social issues.
Were put forward: its gender-sensitivity and the fact that the target audience could

either be the victim, the perpetrator or bystanders. This makes it necessary to look

29

closely at this issue and conduct research that would better inform relationship
violence prevention PSA’s.

Keeping in mind the gender-sensitivity of the issue of relationship violence, this
study investigates how the genders of the communicator and the recipient affect the
audience’s response to relationship violence prevention PSA’s. What message
response refers to is the recipients’ perceived source credibility, as well as their
emotional and cognitive responses to the message (PSA), their attitudes toward the
issue, the communicator and the message, as well as their behavioral intention.

The next chapter explores the theoretical framework of the present study and puts

forward a series of hypothesis and research questions.

30

CHAPTER III — THEORETICAL FRAMEWORK

The present chapter includes three sections. The first reviews previous literature
about the main effects of communicators and recipients’ genders on message response.
Message response in the present study refers to emotional and cognitive responses,
attitudes toward the issue, the communicator and the message, as well as behavioral
intention.

The second section discusses the match-up hypothesis (e.g., Kahle & Homer,
1985), one of the two theoretical frameworks of the present study. This hypothesis is
used in this study to highlight the importance of gender match-up between the
message communicator and recipient, when the message is gender-salient. Since the
study uses a relationship violence prevention PSA as the stimulus, the message of the
PSA can be described as gender-salient. Therefore, gender match or mismatch
between the PSA communicator (voice-over) and recipient (study participants) may
have an effect on message response.

The third section reviews literature about the gender schema theory (Bern, 1981).
This theory is also used in this study to support the hypothesis that gender match-up
between the communicator and the recipient is important, when gender is salient to the
message. In fact, gender schema theory claims that gender-salient information or
situations can activate in the minds of individuals what Bern (1981) calls a “gender
schema”. A gender schema is a set of cognitive constructs that individuals acquire

since early childhood. When triggered, the gender schema makes individuals evaluate

31

situations and messages based on gender. This gender-based judgment may occur in
the present study since the anti-relationship PSA is gender-salient.

The chapter ends with a section presenting the hypotheses and research questions
that the present study investigates.

3.1. Communicator and recipient genders: Main effects on message response

3.1.1. Effects of the communicator ’s gender on message response

The communicator’s gender is one factor that was shown to influence message
response. This influence was demonstrated on many aspects of message response. The
present section reviews literature showing that the communicator’s gender does affect
source credibility, cognitive response, attitudes and behavioral intention.

Source credibility:

Previous studies examined many dimensions of source credibility. Examples are
“expertise” and “personal attractiveness” (Kenton, 1989), reliability/trustworthiness
(e.g., Hovland, Janis, & Kelley, 1953), physical attractiveness (e.g., Chaiken, 1979),
and dynamism (e.g., Kenton, 1989).

Nevertheless, the present study adopts two dimensions of source credibility:
trustworthiness and expertise (Hovland, Janis, & Kelley, 1953). There are two reasons
for choosing these two dimensions. First, trustworthiness and expertise are the most
recurrent dimensions used to measure source credibility (e.g., Kenton 1989). Second,
within the context of this study, the communicator is a PSA voice-over. It follows that
the PSA audience cannot base its perceived source credibility on the voice-over’s

physical appearance or personal characteristics.

32

Many researchers have examined the effect of the communicator’s gender on
source credibility. Results indicated that male communicators are perceived to be
more credible than female communicators in many situations. This hypothesis was
confirmed on various dimensions of source credibility, even when the content of the
message was exactly the same (e.g., Newcombe & Amkoff, 1979; Markham, 1988). In
Newcombe and Amkoff’ 5 study (1979), the results were attributed to the difference in
speech styles between female and male communicators. Differences in speech style in
turn affected the recipients’ attitudes towards the communicators. In Markharn’s study
(1988), the difference in source credibility was attributed to sex-role stereotypes that
cause males to be perceived to have higher levels of expertise than females in various
situations.

Additionally, Oakley (2000) found that the expertise of the speaker is perceived
higher when the speaker is male than when she is female. The article provides various
explanations including gender-role stereotypes and gender difference in speech style.

Cognitive response:

According to the cognitive response theory (Greenwald, 1968), what determines
the degree of influence of a message is not learning the content of the message, but
rather the recipients’ thoughts about the message, i.e. their cognitive response to it
(Petty, Brinol, & Priester, 2009). Following this conceptual definition, cognitive
response in the present study refers to the thoughts that the participants gnerate during
and after watching the PSA.

The gender of the communicator was also found to affect cognitive response.

Gallois and Callan (1986) examined how communicator and recipient genders impact

33

the processing of messages. The study strongly supported the hypothesis that female
communicators would be more accurately understood than male communicators. This
was confirmed in three different conditions where the message was audiovisual
(picture and sound), visual (picture only) or audio (sound only).

The same study also examined how cognitive response (called message decoding
in that study) differs depending on gender and message type. The study used positive,
negative and neutral messages. The results showed that female communicators were
more accurately understood than male communicators in the maj ority of conditions.
Among other factors, the results were explained based on the difference between the
vocal and physical cues used by male and female communicators.

Accordingly, this last finding would suggest that the recipients’ response to the
PSA in the present study would be more positive when the communicator is female.
However, given the review in the previous section where source credibility was higher
for male than female communicators, and knowing that source credibility affects
cognitive response, the effect of communicator’s gender on cognitive response will be
left open and expressed as a research question.

Attitudes

Another dimension of message response that is affected by the communicator’s

gender is attitudes. The present study examines three types of attitudes: attitude toward

the message itself (attitude pSA), attitude toward the social issue (attitude issue) and

attitude toward the communicator (attitude voice-over)- More specifically, the aim is to

examine whether the genders of the communicator and the recipient have any effects

on the participants’ attitude pSA, attitude issue, and attitude voice-over-

34

Research showed that the communicator’s gender can influence different types of
attitudes. For example, a study conducted by Debevec and Lyer (1986) concluded that
the communicator’s gender can act as a cue that shapes the recipients’ perceived
gender image of a product. This gender image influences the recipients’ attitude
toward the product. This, nevertheless, happened in the study only when the product
already had a gender image, as opposed to neutral products.

Specifically, the effect found by Debevec and Lyer’s study (1986) was that attitude
toward the product as well as purchase intention, were higher when the gender of the
communicator was opposed with the gender image of the product. In other words, the
participants had a better attitude and higher purchase intention when a male
communicator advertised a female-perceived product, and vice versa. Debevec and
Lyer explain these findings by the fact that individuals are increasingly sensitive to
stereotypical gender roles, and are more inclined to accept contemporary gender
portrayals in ads. Therefore, the mismatch between the communicator’s gender and
the product’s gender image represented a “progressive role portrayal”, creating a
better attitude and higher purchase intention.

This last finding may seem contradictory with the rationale of the present study,
but this is not the case. In fact, tn the present study, the message is gender-salient, but
does not have any specific gender-image. It is true that females are most frequently the
victims and males are most frequently the perpetrators of relationship violence
(Saunders, 1986). But this does not negate two facts. First, the issue of relationship
violence can affect both genders (Makepeace, 1983). Second, both genders need to

take part in eradicating it (Flood, 2006). Consequently, the issue of relationship

35

violence is gender-sensitive and gender-salient, but an exclusively female or male
image cannot be attached to it.

Additionally, despite the aforementioned findings by Debevec and Lyer’s study
(1986), the existence of an effect of the communicator’s gender on attitude was all
together refuted by other studies. For instance, Freiden (1984) found that consumers’
attitudes generated by advertising are significantly affected by the type of
spokesperson and the age of the audience, but not the gender of the spokesperson.

Therefore, given the lack on unanimous findings about the effects of the
communicator’s gender on attitudes, these effects will be expressed in research
questions.

Behavioral intention:

According to the theory of reasoned action (Fishbein & Ajzen, 1975), behavioral
intention is a direct precursor of behavior change. Although many studies argue that
behavioral intention does not necessarily lead to actual behavior (e.g., Miniard,
Obenniller and Thomas, 1983), others argue that it can accurately predict actual
behavior (e.g., reviews by Ajzen, 1991; Sheppard, Hartwick & Warshaw, 1988). In
fact, a meta-analysis by Webb and Sheeran (2006) shows that a medium to large
change in behavioral intention leads to a small to medium change in actual behavior.

In the present study, behavioral intention refers to the recipients’ intention to act as
active bystanders of relationship violence by seeking and sharing information about
the issue, but also provide help to the couples who are involved in violent

relationships.

36

The gender of the communicator is one of the factors that have been found to
affect behavioral intention. For instance, many studies have supported the idea that
male communicators have a higher ability than female communicators to produce
behavioral compliance with their advocated message (Eagly, 1983).

However, other researchers maintain that the gender of the communicator is not
the main reason behind compliance. The other explanations provided include social
norms (Milgram, 1974) gender stereotypes (Maccoby & Jacklin, 1974), and
personality differences (Buss, 1981). Therefore, due to these conflicting findings, the
present study does not hypothesize the impact of the communicator’s gender on
behavioral intention, but rather formulates a research question about it.

After reviewing literature about the main effects of the communicator’s gender on
message response, the next section reviews literature about the main effects of the
recipient’s gender.

3.1.2. Eflects of the recipients ’ gender on message response

Just like the communicator’s gender can affect various aspects of message
response, so does the message recipient’s gender. This section reviews literature on
communicator’s gender effects on various aspects of message response, including
source credibility, emotional response and attitudes.

Source credibility:

Most source credibility literature examined the effect of gender on source
credibility with focus on the communicator’s gender rather than the recipient’s gender.
One of the few investigations that mentioned the effect of recipients’ gender on their

perceived source credibility was conducted by Kenton (1989). This study found that

37

even female recipients tend to underestimate female communicators and overestimate
male communicators on some aspects of source credibility like expertise. This
sometimes pushes female communicators to act or look more masculine in order to
gain more credibility.

Given the lack of sufficient previous findings on the effect of the recipient’s
gender on source credibility, this effect is left open as a research question.

Emotional response

A large portion of psychology literature supports the idea that males and females
differ in their degree of emotionality, despite the opposing opinions of some
researchers like Eagly (1978). A PSA test conducted by Levine and Zimmerman
(1996) revealed that female viewers felt stronger emotions than male viewers when
they were exposed to a youth violence PSA. Also, research to a large extent confirms
that women experience stronger emotions than men, or tend to express their emotions
more openly compared with men (see the review of LaFrance & Banaji, 1992).

Attitudes:

The gender of message recipients plays a role in shaping attitude issue- For

instance, Smith (1984) maintains that there is a difference between women and men
both in the way they use violence, and the degree to which they approve of violence.
He explains this difference by a mix of biological and social causes that make men
more inclined to choose forceful options, while women tend to choose middle options.
According to Smith, this statement applies to various types of violence across various
social conditions. These include “foreign affairs, social control and law enforcement,

and interpersonal relations.”

38

Additionally, another finding of bigger relevance to the present study confirms that

gender particularly affects attitude issue towards relationship violence. A study by

Locke and Richman (1999) concluded that compared with men, women show a
stronger sympathy with the violence victim, and rank the issue higher in terms of
seriousness.

Concerning Attitude toward the PSA, previous research found mixed differences
between female and male recipients. Struckman-Johnson, Gilliland, Struckman-
Johnson and North (1990) examined the effect of gender on the response to fear-
arousing condom use PSA’s. They concluded that males had a more positive attitude
toward the PSA than females did, by rating it as more effective.

Fleming and Petty (1997, as cited in Maio & Olson, 2000) found that both gender
and gender identity have an impact on attitude toward the message. Their experiment
revealed that both male and female message recipients had a more positive attitude
toward messages when messages matched than when they did not match their gender
identity.

Behavioral intention:

Finally, behavioral intention is yet another message response affected by the
communicator’s gender. Researchers have identified gender as a moderator of
behavioral intention (e.g., Nysveen, Pedersen & Thorbjomsen, 2005). In this study,
the direction of males and females’ behavioral intention differed according to the
nature of the message and of the task.

Other researchers pointed out gender differences in the degree of influenceability

by persuasive messages. Although Eagly (1978) rejected the hypothesis that females

39

are more likely to adhere to advocated behaviors than males, others confirmed it. For
example, Sistrunk and McDavid (1971) found that females tend to conform to
assigned behaviors more than males. They attributed the differences to the nature of
the task or the situation, rather than the traditionally-believed effect of gender roles.

Because of the inconsistency of findings about the role of recipient’s gender in
shaping behavioral intention, this effect is examined trough a research question.

After reviewing literature about the main effects of communicators and recipients’
genders on message response, it would be interesting to examine whether the
interaction of these genders affects message response. More specifically, the
interaction effect concerns what happens when the genders of the communicator and
the recipients match or mismatch, and how this affects message response. The two
following sections present the theoretical framework behind examining gender
match/mismatch. The framework is based on the match-up hypothesis (e.g., Kahle &
Homer, 1985) and gender schema theory (Bem, 1981).

3.2. The match-up hypothesis

One of the theoretical grounds of the present study is the match-up hypothesis
(e.g., Kahle & Homer, 1985). This hypothesis was traditionally used in advertising and
posits that the endorsers of a particular product are more effective when there is a
“match” or a “fit” between the endorser and the product. This “fit” happens through a
particular match-up factor that makes the endorser appropriate for the product that
they endorse. For example, the ad of a sports product would be more credible if the

endorser is a sports professional, rather than an ordinary model or a celebrity who

40

excels in a field different than sports. This match-up makes the endorser more
effective (Kamins, 1990; Till & Busler, 2000).

The match-up hypothesis is relevant to the preset study because in addition to the
main effects of communicator and recipient genders on message response, the present
study aims to examine the interaction of the communicator’s and recipient’s genders.
More specifically, the aim is to investigate whether message response is affected when
the genders of the communicator and the recipient match or mismatch.

The match-up hypothesis originally underlined the importance of matching factors
between the communicator (endorser) and the product they endorse. Therefore, by
using the match-up hypothesis, the present study suggests expending the hypothesis to
underline the importance of gender match-up between the communicator and the
recipient.

In order to illustrate this extension, the next sub-section explains some of the main
mechanisms behind the match-up hypothesis: congruence and gender as a match-up
factor. The subsequent sub-section explains that gender is important as a match-up
factor because the message in this study is gender-salient.

3.2.1. Congruence: a main match-up mechanism

One of the main theoretical concepts explaining the match-up hypothesis is the
mechanism of congruence. Within the context of the match-up hypothesis, congruence
happens when there is a match between the product and the endorser on a particular
characteristic. This characteristic is labeled by many researchers as “the match up

factor” (e.g., Till & Busler, 2000).

41

Traditionally, congruence in relation with the match-up hypothesis was based on
the factor of beauty (Kamins, 1990). For example, the use of attractive models to
advertise beauty-related products was found to positively affect attitude toward the ad
(Baker & Churchill, 1977). Such findings changed the previously held assumption that
any physically attractive endorser would be effective, no matter how well they match
with the product that they endorse (Kamins, 1990).

3. 2. 2. Can gender be a match-up factor?

Considering gender as a match-up factor is important within the context of the
present study because the stimulus is a gender-salient message. Although physical
attractiveness has been traditionally examined as the main match-up factor, previous
research also found gender to be a factor that creates congruence between the product
and its endorsers. For example, even before the “match-up hypothesis” was given this
label, a study by Kanungo and Pang (1973) concluded that there is a need for
“fittingness” between the model used in the ad and the perceived “gender” of the
product. The models they used were either a female, or a male, or a female and male
couple. Among the products they used were a car (perceived as male) and a sofa
(perceived as female). The study measured the participants’ attitude toward the
product. The results were compared with a control condition where no models were
used. Among the study’s findings, the car advertisement featuring a male model
produced positive attitudes among both male and female participants. The sofa
advertisement on the other hand elicited positive attitudes among male participants
when the model was female and negative attitudes among male participants when the

model was male.

42

The fact that gender can be a match-up factor supports the rationale behind the
extension of the match-up hypothesis suggested in the present study. This extension
suggests focusing on gender as a matching factor between the communicator and the
recipient rather than the communicator and the product. The extension is explained in
the next section.

3. 2.3. The match-up hypothesis within the present study

The present study suggests extending the concept of match-up to cover the
communicator and the recipient, instead of the communicator and the product. There
are two reasons behind this suggested extension. First, the act of communication
involves more than just the communicator (endorser) and the object of the
communication (product). In addition to the communicator and the message object, the
message recipient is also one important party within the communication process.
Therefore, the match-up between the communicator and the recipient also deserves
research attention.

In fact, just like many studies supported the importance of matching factors
between the communicator and the product, many others supported the importance of
matching factors between the communicator and the recipient. For example, Ryan and
Giles (1982) examined the effect of matching ethnic origins between the
communicator and the recipient. They found that listener attitudes toward both the
message and the communicator are better when the listeners perceive that their own
ethnic origins matched those of the communicator.

Therefore, by suggesting the extension of the match-up hypothesis, the present

study maintains the core principle of this hypothesis, but changes one of the two

43

 

involved components of the communication act. The core principle it maintains is that
depending on the context of the communication, there are match-up factors that can
make the communication more effective.

The change suggested in this study, however, involves one of the two components
of the message act. Instead of focusing on how the communicator matches the content
of the message, the present study focuses on the effect of the match-up between the
communicator and the recipient.

The second reason for suggesting the expansion of the match-up hypothesis is the
lack of unanimous results concerning the role of the match-up in increasing the
effectiveness of the message (e.g., Kamins, 1990). As reviewed in the previous
section, research only partially showed that the match-up between the communicator
and the product can favorably impact message response.

To sum up, the match up hypothesis is the first theoretical ground of the present
study because the hypothesis can explain why individuals respond to messages
differently depending on whether there is a match/mismatch between the
communicator and the recipient. The second theoretical ground that explains the
importance of gender match-up in a gender-salient message is the gender schema
theory (Bern, 1981). This theory is discussed in the next section.

3.3. Gender schema theory: the role of gender schema in message response

The present section draws on the gender schema theory (Bern, 1981) to provide
theoretical support to the importance of gender match-up between the communicator

and the recipient when the message is gender-salient. This is of high relevance to the

present study because gender is salient to the relationship violence prevention PSA
used as a stimulus.

According to Bem’s gender schema theory (1981), an individual can be described
as gender schematic when they have a general tendency to process information based
on gender-related associations. These gender-related associations serve as a “map”
internalized by individuals since childhood, and used to process gender-related
information. It is this “map” that Bern calls a gender schema. For example, the gender
schema according to Bem (1981) would make sex-typed individuals (feminine females
and masculine males) tend to recall words and concepts in clusters that are consistent
with their socially assigned gender role.

Gender schema is acquired since childhood and shaped by cultural and social
expectations about the attributes of masculinity and femininity (Bem, 1981). Gender
schemata therefore can be present in the mind of any individual. However, it is the
degree of activation of this schema that differs from one individual to another.

Gender schema is activated when the individual is faced with a situation or a
message where gender is salient (Freedman, 1992). In fact, gender schema is just one
of many cognitive schemata internalized by human beings since early childhood.
These schemata serve as cognitive structures that help individuals organize and
process incoming information, evaluate information about the self and others, as well
as later recall of information. These various schemata are triggered depending on the
type of situations or messages that the individual has to process or evaluate.

When gender schema is activated, it guides the individual’s cognitive processing

to encode and organize information in terms of gender. This encoding and

45

organization happens according to the “rules” of the gender schema that are based on
the socially instilled distinction between maleness and femalencss. This way, such
individuals evaluate information based on the cultural definition of what it means to be
a “male” or “female” (Bern, 1981).

This suggests that gender schema theory seems relevant to the present study
because the participants of this study are exposed to a gender-salient message.
Therefore this theory can explain the mechanism through which gender becomes a
factor affecting message response. Since the study participants are asked to watch a
relationship violence prevention PSA, the gender salience of the message may trigger
the gender schemata in their minds. This may cause gender to affect the recipients’
response to the message.

Keeping in mind that gender is suggested in the present study as a match-up factor
between the communicator and the recipient, it is expected that gender match or
mismatch between the communicator and the recipient may give way to different
message responses

3.4. Hypotheses and research questions

As discussed in this chapter, the present study builds its theoretical framework on
three bases. The first is previous findings about the main effects of gender on message
response. The second is the match-up hypothesis (e.g., Kahle & Homer, 1985). The
third is gender schema theory (Bern, 1981). This chapter also justified the potential
impact of gender on message response by the fact that the present study uses a gender

salient message.

46

Based on the theoretical rationale, the present study puts forward the following
hypotheses and research questions:

3.4.1. Main effects: hypotheses and research questions
Gender of the communicator:

H 1: The male communicator will be perceived as more credible than the female
communicator.

H2: The male communicator will produce a stronger emotional response than the
female communicator.

The literature review showed inconsistent findings about the main effects of the
communicator’s gender on cognitive response, attitudes and behavioral intention.
Therefore, these effects are examined through the following research questions:

RQl: What impact does the communicator’s gender have on the recipients’

cognitive response to the relationship violence prevention PSA?

RQZ: What impact does the gender of the communicator have on attitude issues

Attitude voice-over, and attitude pg A in the context of relationship violence prevention

PSA?

RQ3: What impact does the gender of the communicator have on the recipients’
behavioral intention to provide help or seek information about relationship violence?
Gender of the recipient:

H3: Female recipients will have a stronger emotional response to the PSA’s than

male recipients.

H4: Female recipients will have a more positive attitude issue than male recipients.

47

Most previous studies focused on the effects of the communicator’s gender rather

than the recipient’s gender effects on source credibility, cognitive response, attitude

pg A and attitude voice-over and attitude toward the communicator. Therefore, the

impact of the recipient’s gender is examined through the following research questions:
RQ4: What impact does the gender of the recipient have on source credibility?

RQS: What impact does the gender of the recipient have on cognitive response?

RQ6: What impact does the recipient’s gender have on attitude pSA?

RQ7: What impact does the recipient’s gender have on attitude voice-over?

The literature review revealed inconsistent findings concerning the effect of the
recipient’s gender on behavioral intention. Therefore, the following research question
is put forward:

RQ8: What impact does the recipient’s gender have on the recipients’ behavioral
intention to provide help or seek information about relationship violence?

3.4.2. Interaction effects: hypotheses

H5: Perceived source credibility will be higher when the genders of the
communicator and the recipient match, compared to when they do not match.

H6: Emotional response will be weaker when the genders of the communicator
and the recipient match, compared to when they do not match.

H7: Cognitive response will be higher when the genders of the communicator and
the recipient match, compared to when they do not match.

H8: Attitudes will be more positive when the genders of the communicator and the

recipient match, compared to when they do not match.

48

H9: The recipient’s behavioral intention to provide help or seek information about
relationship violence will be higher when the genders of the communicator and the
recipient match, compared to when they do not match.

The next chapter explains the method through which these hypotheses will be tested.

49

CHAPTER IV - METHOD
4.1. Study design:

The present study was conducted using a 2 between (male or female
communicator) X 2 between (Male or female recipient) factorial design. This design is
appropriate for two reasons. First, the study involves two independent variables:
gender of the communicator and gender of the recipient. Second, the study measures
both the main and interaction effects of gender on message response. Additionally, a
between subjects design is appropriate because the analysis involves comparing the
responses of four distinct groups of participants.

4.2. Sample

The participants were undergraduate students from four advertising classes at
MSU. In agreement with their professors, participants received extra-credit for
participation. A total of 243 students completed the survey, including 139 participants
(57.2%) who watched the male version of the PSA, and 104 (42.8%) who watched the
female version of the PSA. Twenty responses were eliminated from the data either
because of their inaccurate recall of the gender of the communicator or their self-
reported duplicate participation. The remaining number of participants was 223.

Among the participants, 33.2% were males and 66.8% females. Knowing that each
participant watched either the male voice-over or the female voice-over version of the
PSA, the participants who watched the male voice-over version included 20.7% males

and 38.3% females. Those who watched the female voice-over version included

50

12.6% males and 28.4% females. (See Table l for a crosstabulation of communicator
and recipients’ genders).

The participants belonged to different races including White/Caucasian (77.1%),
Asian (11.2%), African American (6.7%), Multiracial (2.2%), Hispanic or Latino
(1.3%) and other races (1.3%). Their age ranged between 18 and 28, with an average
of 20 years old.

4.3. Procedure:

Two online surveys were hosted on SurveyGizmo.com and used for data
collection. Each one of the two surveys included a Youtube link to one of two
different versions of the relationship violence prevention PSA used as a stimulus. The
voice-over was male in one version and‘female in the other. Youtube was chosen to
host the PSA’s because it is one of the most popular video sharing websites. This
would make it easy for the study participants to locate and watch the video without
getting distracted by unfamiliar website features.

Each one of the participants was randomly assigned to watch only one of the two
PSA versions. Randomization was based on the last digit of each participant’s student
ID number. According to the university’s enrollment office, student ID numbers are
assigned randomly to the students when they enroll in the university. In the email sent
to the students, they were asked to click the first link (the survey with the male PSA
version) if their ID number ends with an odd digit, and to click on the second link (the
survey with the female PSA version) if their ID number ends with an even digit.

The survey was made of three parts. The first part asked demographic questions

about the participants, in addition to questions measuring gender identity, familiarity

51

with the issue of relationship violence and prior involvement with the issue. The
second part included PSA the link leading to the assigned PSA. In the third and last
part, the participants were asked to report their response to the PSA.

In the end of each survey, the students were redirected to a separate online survey
used to collect the student’s names in order to award them extra credit. Collecting the
names in an independent survey, separately from the data, was meant to guarantee the
unanimity of the participants.

Since four undergraduate classes participated in the survey, duplication of
participants had to be avoided. Participants who belong to more than one of the four
classes were asked not to take the survey twice. Instead, they were invited to take it
only the first time when they receive the invitation email. In case they received the
participation invitation again, they were asked not to respond to the survey and to
simply report their names in order to get the extra-credit.

The survey, stimulus and procedure of the present study were reviewed and
approved by the Institutional Review Board (IRB) at Michigan State University (IRB
# X10-120). The board found that this study meets the criteria for the protection of
human subjects, and therefore deemed it exempt. The consent form approved by [RE
is available in appendix C.

4.4. Stimuli

The PSA’s last 30 seconds each and were produced in a professional studio using
professional equipment provided by government access channel HOM TV. The PSA

features a male actor and female actress who graduated from Michigan State

52

University with a minor in theatre. The female and male voice-overs were played by
two HOM TV anchors.

The PSA shows a scene of physical and psychological violence between a man
(perpetrator) and a woman (victim). The man shakes the woman violently and abuses
her verbally and then slaps her face. The woman looks terrified, cries and begs him not
to hurt her. The man scolds her ordering her to stop crying. Between the short, an on-
screen script reads “Relationship violence should never remain a secret”.

The PSA’s end with a voice-over (communicator) that is female in one version of
the PSA and male in the other. The voice-over addresses the viewers saying “Once
you know about it, it’s no longer just their business. They need your help”. This is
followed by a superscript addressing the viewers: “You can help them before their
time runs out”. Then the phone number of MSU Safe Place is provided on screen (See
the storyboard of the PSA in appendix B).

The only sounds heard in the PSA are those of the victim and the perpetrator as
well as the voice-over. No music was included in order to avoid any confounding
peripheral cues (Park & Young, 1989). Also, one aim of using a voice-over instead of
an on-screen communicator was to avoid the confounding effect of physical likeability
that could impact the participants’ response to the PSA’s (Petty & Cacioppo, 1986).

4.5. Pretest

The pretest aimed to answer four main questions: a) whether viewers could
accurately recall the gender of the PSA communicator, b) whether the message was

accurately understood by the viewers, c) whether the viewers could correctly identify

53

the intended audience as potential relationship violence bystanders, and d) whether the
viewers themselves felt targeted by the PSA’s as potential bystanders.

A total of 10 MSU students took part in the pretest, 40% males and 60% females.
Six students (60%) watched the male version of the PSA, 4 (40%) watched the female
version. The pretest participants were randomly assigned to one of the two PSA
versions. The participants watched the PSA’s individually then were asked to take a
pen-and-paper survey.

Concerning the recall of the communicator’s gender, it was important to find that
most viewers recalled it correctly because the rationale of the present study is based on
both the genders of the communicator and the recipient, and part of it is based on how
these genders interact. The pretest showed that 80% of pretest participants recalled the
gender of the voice-over accurately.

As for the comprehension of the message, 80% of participants provided a very
accurate explanation about the message, stating that it expresses the seriousness of
relationship violence and invites the community to take part in reducing it. The
remaining 20% of participants provided general pro-social interpretations of the
message like “stop relationship violence”, but did not state that it aimed to call on
bystanders’ positive action.

All pretest participants correctly identified the audience to be bystanders, although
only one of them used the term “bystanders”, while the others used terms like
“community” or “witnesses”. Nevertheless, 60% of viewers did not feel personally

targeted as potential bystanders, while 40% did.

54

Overall, the results of the pretest indicated that the message and intended audience
were clearly conveyed through the PSA’s. The results also indicated that the genders
of the communicators were recalled in a generally accurate manner. This meant that
the PSA’s could be used as they are for the main study, as no major modifications
were required based on the pretested results.

4.6. Variables and measures

The next sections provide details about the independent, dependent and control
variables. The sections describe how each one of them was measured and how scale
variables were constructed. For scale variable, factor analyses and reliability test
results are also provided (The survey wording of each variable is provided in appendix
A).

4.6.1. Independent variables:

The present study aims to examine the main and interaction effects of
communicator and recipient genders on message response. Accordingly, the two
independent variables measured in this study are: the gender of the communicator and
the gender of the recipient.

4.6.2. Dependent variables:

i) Source credibility:

Measure: Source credibility was measured using Hovland’s Yale Communication
Model (1953). This model measures source credibility based on two dimensions:
“trustworthiness” and “expertise”. The procedure is derived from studies conducted by
Sternthal et al. (1978) and Harmon and Coney (1982). In order to measure the

participants’ perception of the communicator’s trustworthiness and expertise, four

55

seven-point semantic differential scales were used, ranging from -3 to 3. The items are
“not trustworthy/ trustworthy” and “bad / good” under trustworthiness, and “not expert
in the issue / expert in the issue” and “not experienced / experienced” under expertise.

Factor analysis and Reliability: Exploratory Factor Analysis (EF A) was
conducted using the Principal Components Analysis (PCA) extraction method, with
eigenvalue criteria of >1. The EFA showed clearly one factor with 67.5% of total
variance explained (communalities ranging from .78 to .85). Therefore, the four items
were averaged to construct the variable of source credibility.

The factor loadings of each of the four items were as follows: .62 for “not
trustworthy/ trustworthy”; .72 for “bad / good”; .72 for “not expert in the issue / expert
in the issue” and .62 for “not experienced / experienced”.

In order to assess the internal consistency of the source credibility scale, a
reliability analysis was conducted. The Cronbach’s alpha was .83. This reliability
score is judged good to excellent (e.g., George & Mallery, 2003; Leech, Barrett &
Morgan, 2008). (M: .18, SD= 1.09). (The results of factor analyses are reported in
table 2)

ii) Emotional response:

Measure: The variable of emotional response includes three underlying variables:
Fear, anger and guilt. The survey question measuring emotional response included
nine items on a seven-point Likert scale, ranging from not at all (= -3) to very much (=
3). These included afraid, fearful and scared (under fear); angry, annoyed, irritated,

and aggravated (under anger); and guilty and ashamed (under guilt). The scales of

56

fear and anger were used by Shen and Dillard (2007). The scale of guilt was used by
Dillard and Peck (2000).

Factor analysis and Reliability: Exploratory Factor Analysis (EFA) was
conducted using the Principal Components Analysis (PCA) extraction method, with
eigenvalue criteria of >1. The EF A showed that the first three factors accounted for
the strongest levels of variance. The factor “afraid” accounted for 53.6% of the total
variance (factor loading = .93), “fearful” accounted for 19.73% of the variance (factor
loading = .92), and “scared” accounted for 14% of the variance (factor loading = .93).
The factors’ communalities ranged from .94 to .95. These three items were therefore
averaged into one separate variable labeled “Fear”.

The EF A analysis showed a second cluster containing four components: “Angry”
accounted for 4.53% of the total variance (factor loading = .7), “annoyed” accounted
for 3.15% of the total variance (factor loading = .89), “irritated” accounted for 2.17%
of the total variance (factor loading = .91), and “aggravated” accounted for 1.46% of
total variance (factor loading = .81). The communalities of the four items ranged from
.73 to .87. Based on these results, the four factors were averaged into one variable
labeled “anger”.

Finally, the EFA revealed a third cluster of factors that included two components:
“Guilty” accounted for .72% of total variance (factor loading = .93) and “ashamed”
accounted for .06% of total variance (factor loading = .89). The communalities of the
two items ranged from .89 to .91. These two factors were averaged into one variable

labeled “guilt”.

57

Cronbach’s alpha reliability coefficients indicated a high internal consistency of
the three constructed emotional response variables: (1 =97 for fear (M= -.35, SD=
1.68), (1 =89 for anger (M= .60, SD= 1.37), and (1 =89 for guilt (M= -1.25, SD=] .62).

iii) Cognitive response:

Measure: Cognitive response measure was based on the Thought-Listing
Procedure of Petty and Cacioppo (1979). After viewing the PSA, respondents were
asked to generate up to 10 thoughts that crossed their minds while watching it.

After data collection, the thoughts were coded by 2 independent coders, both blind
to the study’s conditions. The coders coded each thought into one of the following
three categories: “This thought opposes the behavior advocated by the PSA” (coded as
-1) ; “This though is neutral or irrelevant to the behavior advocated by the PSA”
(coded as 0); and “This thought supports the behavior advocated by the PSA” (coded
as 1). Intercoder reliability was measured using Perreault and Leigh’s Index (1989)
(P/L Index). The intercoder reliability coefficient was .80. Any discrepancies were
resolved by discussion.

Because the stimulus is a pro-social PSA, the cognitive responses retained for
analysis were the positive thoughts expressed by the participants. This refers to the
thoughts that support the advocated behavior. The count technique was used in order
to construct the cognitive response variable. (M= 3.63, SD= 3.19).

iv) Attitudes:

Attitude toward the issue

Measure: The participants were asked to express their attitude toward the issue of

relationship violence. Attitude issue was measured using a seven-point semantic

58

differential scale that ranged from -3 to 3 and included four items. These items are

trivial / serious, unimportant / important, not worth much concern / worth a lot of

concern, and irrelevant / relevant. This scale was used by Dillard and Peck (2000).
Factor analysis and Reliability: The BF A showed clearly one factor with 85.20%

of total variance explained (communalities ranged from .83 to .86). Therefore, the four

items were averaged into one variable: Attitude issue- (M= .28, SD= 1.50).

The factor loadings of the four items were as follows: .93 for “trivial / serious”, .92
for “unimportant / important”, .93 for “not worth much concern / worth a lot of
concern”, and .91 for “irrelevant / relevant”.

The Cronbach’s alpha reliability coefficient also indicated a high internal
consistency ((1 =93). This reliability coefficient is deemed excellent (e.g., George &
Mallery, 2003; Leech, Barrett & Morgan, 2008).

Attitude toward the communicator

Measure: In the present study, the communicator is the PSA voiceover. The
participants’ attitude toward the voice-over was measured using the scale of Stem,
Mullenix, Dyson and Wilson] (1999). This scale was used in an experiment to measure
attitudes towards a tape recorded voice. This makes it closer to the present study
where the communicator is a voice-over.

Participants were asked to “evaluate the speaker in this PSA” on a seven-point
Likert scale ranging from -3 to 3. The scale included 8 items chosen from the scale of
Stem, Mullenix, Dyson and Wilsonl (1999): not competent/ competent, not
honest/honest, not assertive/ assertive, not informed/informed, evasive/

straightforward, not qualified/qualified, not sincere/ sincere, and meek/forceful.

59

Factor analysis and Reliability: The EFA showed clearly one factor with 53.61%

of total variance explained (communalities ranging from .35 to .64). Therefore, the 8

items were averaged to construct the variable of attitude voicgover. (M= .80, SD= .93).

The factor loadings of each of the items were as follows: .75 for “not competent/
competent”, .79 for “not honest/honest”, .66 for “not assertive/ assertive”, .80 for “not
informed/informed”, .73 for “evasive / straightforward”, .75 for “not
qualified/qualified”, .72 for “not sincere/ sincere”, and .59 for “meek/forccf ”.

The Cronbach’s alpha reliability coefficient of a =.87also indicated a high internal
consistency. This reliability coefficient is judged as good to excellent (e.g., George &
Mallery, 2003; Leech, Barrett & Morgan, 2008).

Attitude toward the PSA:

Measure: In order to measure the participants’ attitude pSA, a seven-point

semantic differential scale was used. This scale was previously used by Dillard and Ye
(2008). This scale included 8 items ranging from -3 to 3. The eight items accounted
for the two dimensions of perceived PSA effectiveness as suggested by Dillard and Ye
(2008): impact (4 items) and attributes (4 items). The scale includes the following
items: persuasive/not persuasive, effective/not effective, convincing/not convincing,
compelling/not compelling (impact), reasonable/not reasonable, logical/not logical,
rational/not rational, and true to life/not true to life (attributes).

Factor analysis and Reliability: The EFA showed that the first two factors
accounted for the strongest levels of variance: 58.75% and 18.58% of the variance

explained respectively. Therefore, the first cluster of factors was averaged into one

variable: attitude p3 A (communalities ranging from .80 to .86). The items

60

corresponding to this cluster are: “Not persuasive / persuasive” (factor loading = .87),
“not effective/ effective” (factor loading = .89), “not convincing / convincing” (factor
loading = .84) and “not completing / compelling” (factor loading = .9).

Cronbach’s alpha reliability coefficient of a =.934 indicated a high internal
consistency of the constructed variable (M= 2.33, SD= .97).

v) Behavioral intention:

Measure: On a seven-point Likert scale, respondents reported how likely or
unlikely they are to perform a number of acts that indicate positive behavioral
intention concerning relationship violence. These acts indicate their willingness to
actively help relationship violence victims or perpetrators, or seek or spread
information about relationship violence. The scale includes six items. Examples are
“how likely are you to provide help to relationship violence victims?” and “How likely
are you to seek information on how to help reduce relationship violence”. The
participants’ answers ranged from “not at all likely” (= -3) to “very likely” (= 3). This
scale was previously used by Bagozzi and Moore (1994). The wording was modified
in order to fit the context of relationship violence.

Factor analysis and Reliability: The BF A showed clearly one factor with 57.88%
of total variance explained (communalities ranging from .33 to .78). Therefore, the six
items were averaged to construct the variable of behavioral intention (M= -.71, SD=
1.31). The factor loadings of the six items were .58, .64, .88, .87, .77 and . 75

respectively.

61

The Cronbach’s alpha reliability coefficient also indicated a high internal
consistency (or =.84). This reliability coefficient is judged as good to excellent (e.g.,
George & Mallery, 2003; Leech, Barrett & Morgan, 2008).

4.6.3. Control variables:

The study includes five control variables: a) identification with PSA characters, b)
prior involvement in relationship violence, c) empathy with the victim, (1) gender
identity, and e) familiarity with the issue. The next sections briefly present the
rationale behind choosing each one of the control variables, as well as their measures.
Factor analysis and reliability test results are reported for the scale variables: empathy,
gender identity and familiarity with the issue. The sections also report the results of
one-way AN OVA tests and correlation tests conducted in order to choose the control
variables that are useful while testing each hypothesis or research question. The list of
control variables and the dependent variables that they affect is provided in table 3.
The results of the one-way ANOVA analyses that determined the choice of nominal
covariates are reported in tables 4 to 6. The results of correlations that determined the
choice of scale covariates are reported in table 7.

a) Identification with PSA characters

Rationale: Identification is a mechanism that happens when the viewer “puts
himself in the place of a character and momentarily feels that what is happening to that
character is happening to himself” (Maccoby & Wilson, 1957). In this process, the
viewer assumes the emotions felt by the character he identifies with, and shares the
same emotions. For example, if the character is in danger, the viewer would feel fear

(Maccoby & Wilson, 1957).

62

In the present survey study, the PSA shows a male perpetrator and a female
victim. When male and female participants watch the PSA, some of them may identify
with one of the PSA characters. This may affect their response to the message.

It is important to include identification as a control variable because research
shows that it can affect the five dependent variables measured in this study: source
credibility (e.g., Basil, 1996), emotional response, cognitive response (e.g., Zillmann
et al., 1996), attitudes (e.g., Kelman, 1958) and behavioral intention (e.g., Ansager,
Austin & Pinkleton, 2001 ).

Measure: The respondents were asked to report their perceived similarity with and
likeability of the two PSA characters: the victim and the perpetrator. Answers to these
questions were collected using dichotomous (yes or no) items. This measure was used
by Maccoby and Wilson (1957).

The use of identification as a control variable: Identification is made up of two
variables: likeability of characters and perceived similarity with characters. The one-
way AN OVA results for likeability showed that it had a significant impact on two

emotional response variables, including fear (F (2, 219) =3.61, p < .05) and anger (F
(2, 219) =3.64, p < .05), in addition to a significant impact on attitude issue (F (2, 219)
=12.39, p < .01) and attitude pSA (F (2, 219) =4.67, p < .05).

Likability did not have a significant impact on the rest of the variables. Therefore,

it was used as a covariate only in the analysis of gender effects on fear, anger, attitude

issue and perceived PSA persuasiveness.

63

Concerning the second identification variable, perceived similarity with the PSA
characters, the one-way ANOVA showed that it had a significant impact only on one
emotional response variable : fear (F (2, 219) =5.03, p < .01). The impact of
perceived similarity on the rest of the variables was not statistically significant.
Therefore, perceived similarity was used as a covariate only while analyzing the
effects of gender on fear.

b) Prior involvement in relationship violence:

Rationale: Participants could have prior involvement with relationship violence as
witnesses, victims or perpetrators. It is important to measure prior experience as a
control variable because it can increase the participants’ issue involvement. This was
found to enhance cognitive responses, that in turn can either increase or decrease
persuasion (Petty & Cacioppo, 1979). Additionally, Ramon and Severance (1970)
found that issue involvement (they labeled ego-involvement), affects attitude change,
source credibility, and message evaluation. It is therefore important to control for prior
involvement in relationship violence in order to avoid any confounding effects on
dependent variables.

Measure: This variable was measured using a dichotomous (yes or no) question.
The question asked the participants whether they know anybody (including
themselves) who was involved in a violent relationship.

The use of prior involvement as a control variable: The one-way AN OVA
indicated that prior personal or acquaintance involvement had a significant impact on

guilt (F (1. 220) = 8.21, p < .01) and on behavioral intention (F (1, 220) = 5.07, p <

64

.05). Accordingly, this variable was used as a control variable when the hypothesis
concerning guilt and behavioral intention was analyzed.

c) Empathy

Rationale: Empathy is usually considered an emotional response variable (e.g.,
Batson et al., 1981). However, the present study uses empathy as a control variable,
and this decision is based on the following reasons. Because the topic of the PSA is
relationship violence, and because the PSA shows a scene of violence against a victim,
empathy can influence the recipients’ other responses. This happens because empathy
makes PSA viewers understand the feelings of the characters that they watch and even
assume their roles (Mehrabian & Epstein, 1972). This in turn affects their responses to
the PSA.

This assumption was supported when one-way ANOVA analyses were conducted
in order to check whether empathy has significant impacts on the response variables.
The AN OVA showed that all dependent variables were significantly impacted by
empathy. Therefore, given this effect and given the type of PSA message used in the
present study, I decided to use empathy as a control variable in order to avoid spurious
results.

Measure: Empathy was measured using four statements that were provided in the
survey and respondents were asked to rate their agreement with them on a seven-point
scale (-3 = not at all agree; 3 = strongly agree). This scale was used by Bagozzi and
Moore (1994).

Factor analysis and Reliability: The BF A showed clearly one factor with 73.8%

of total variance explained (communalities ranging from .55 to .83). Therefore, the

65

four items were averaged to construct the variable of empathy (M= -.77, SD=1.54).
The factor loadings of the four items were .88, .91, .88 and .74 respectively.

Cronbach’s alpha reliability coefficient also indicated a high internal consistency
(or =.87).

The use of empathy as a control variable: Since empathy is a scale variable,
Pearson’s correlation was used in order to determine whether it has significant
relationship with any of the dependent variable. The results showed that empathy is
significantly correlated with all dependent variables. Therefore, it was included as a
control variable while testing all hypothesis and research questions. (Table 7
summarizes the correlation indexes of scale covariates with the corresponding
dependent variables).

(1) Gender identity:

Rationale: Because response within this study will be based on gender, it would
be important to take into account the participants’ gender identity — that is how
feminine or masculine they perceive themselves to be. In fact, psychologists believe
that varying degrees of femininity and masculinity exist within every one of us (Gill et
a1. 1987, quoted in Palan 2001). This could affect the degree of identification felt by
every individual participant, and therefore act as a confounding variable that impacts
the participants’ response to the PSA.

Measure: Gender identity was measured using a scale of 8 items chosen from
Bem’s gender identity scale (1981). These items were used by many researchers

including Burke and Tully (1977).

66

On a seven-point Likert scale (-3 to 3), each participant was asked to rate how
much each item applies to themselves, being men or women. The scale included the
following eight items: Emotional / unemotional, Timid / bold, Masculine / feminine,
soft / hard, powerful / powerless, loud / quiet, aggressive / nonaggressive, and
sensitive / insensitive.

Factor analysis and Reliability: The BF A showed that the first two factors
accounted for the majority of the total variance with respectively 35.9% and 23% of
the total variance explained (communalities ranging from .53 to .64). The EFA
revealed that four items out of eight had a high loading, and therefore were averaged
to construct the gender identity variable (M= .19, SD= .72). The corresponding
question items are “timid/bold” (factor loading = .80), “powerful/ powerless” (factor
loading = .70), loud / quiet” (factor loading = .82), and aggressive / nonaggressive
(factor loading = .72).

The Cronbach’s alpha reliability coefficient also indicated a high internal
consistency (o. =.77).

The use of gender identity as a control variable: Pearson’s correlation showed
that gender identity was significantly correlated with only one dependent variable:
fear. Therefore, gender identity was used as a covariate only while testing the
hypotheses and research questions related to fear.

e) Familiarity with the issue:

Rationale: Prior knowledge about the issue could bias the response to it. Since the
participants are MSU students, some of them might have attended the series of

awareness workshops organized on campus by MSU Safe Place and the Sexual

67

Assault Program. They also may have received information about the issue of
relationship violence from other communication or educational sources.

If they were previously exposed to any awareness or educational messages about
relationship violence, some of the participants may easily identify positive bystander
behavior as the pro-social behavior. Being aware of the pro-social nature of a message
can affect response to it. In fact, Dillard and Ye (2008) maintain that when message
recipients are aware that the message is recommending a pro-social behavior, they are
more likely to perceive higher effects on themselves than on others. This is what they
call a reverse third-person effect. This effect may bias survey responses concerning
emotional response, cognitive response, attitudes and behavioral intention.

Measure: This variable was measured using a scale measuring two items:
familiarity and knowledge. The participants reported the degree of their familiarity
and knowledge of the issue on a seven-point Likert scale ranging from -3 to 3.

Factor analysis and Reliability: The EFA showed clearly one factor with 88.72%
of total variance explained (communalities = .887 for both factors). Therefore, the two
items were averaged to construct one variable: Familiarity with the issue of
relationship violence (M= -.04, SD=1.42). The factor loadings were .94 for both items.
The Cronbach’s alpha reliability coefficient was (1 =87, reflecting a good to excellent
internal consistency.

The use of familiarity as a control variable: Pearson’s correlation test showed

that familiarity with the issue was significantly correlated with anger, attitude issue and

behavioral intention. It was therefore used while testing hypotheses and research

questions related to these three variables.

68

4.7. Analysis strategy:

To examine the main and interaction effects of gender on the five dependent
variables, a series of Univariate Analysis of Variance (ANOVA) and Univariate
Analysis of Covariance (ANCOVA) was conducted. The ANCOVAs allowed
integrating both the dependent variables of interest and the covariates. When the
analysis yielded statistically significant impacts, the means were analyzed in order to
determine the direction of the effect. The analytic tool used for this study is SPSS and

results were tested at a level of significance of p value less than .05.

69

CHAPTER V - RESULTS

5.1. Descriptive statistics:

The results of this study showed many differences between the message responses
of male and female recipients, depending on whether the communicator was female or
male. Not all these differences were statistically significant. Nevertheless, the
following section reports all significant as well as non-significant differences in order
to provide a sense of the direction of each message response depending on the genders
of the communicator and the recipients.

First, concerning source credibility, the female communicator was significantly
perceived as more credible than the male communicator. The female communicator
was perceived as more credible both by male recipients (M=.57 for the female
communicator Vs. M=-.08 for the male communicator) and by female recipients
(M=.40 for the female communicator Vs M=.05 for the male communicator). (Table 8
provides the means and standard deviations of dependent variables by gender).

Second, emotional response includes the three variables of fear, anger and guilt.

F emalc recipients expressed more fear than males, both when the communicator was a
female (M= -.21 for females Vs. M= -.51 for males) and when the communicator was
a male (M= -.1 for females Vs. M= -.89 for males). This difference between female
and male reported fear was statistically significant.

Additionally, female recipients expressed more anger than males only when the

communicator was a male (M= .67 for female recipients Vs. M = .64 for male

70

recipients) compared with when the communicator was female (M= .51 for female
recipients Vs. M = .57 for male recipients). As for guilt, female recipients reported
more guilt compared with male recipients, both when the communicator was male
(M= 1.14 for female recipients Vs. M = -1.58 for male recipients) and when the
communicator was female (M= -1.45 for female recipients Vs. M = -1.23 for male
recipients).

Third, concerning cognitive response, the recipients reported more positive
thoughts in total than negative or neutral thoughts. The average reported positive
thoughts was about 3.5 compared with .77 on average for negative thoughts and 2.25
on average of neutral thoughts.

Fourth, differences were also noticed between female and male recipients when it

came to attitudes. Female recipients expressed a significantly higher attitude issue than

male recipients, both when the communicator was male (M=2.52 for females Vs.

M=l .83 for males) and when the communicator was female (M=2.62 for females Vs.

M=1.91 for males). Attitude p5 A was clearly different between female and male

recipients when the communicator was male (M=.36 for females vs. M=.06 for males).

However, attitude pg A was almost the same between female and male recipients when

the communicator was female (M=.333 for females vs. M=.339 for males).

As for attitude voice-overs the male communicator generated a higher attitude voice-

over from female recipients (M=.91) than from male recipients (M=.6). On the other

71

hand the female communicator generated almost the same degree of attitude voice-over

from both female recipients (M=.8) and male recipients (M=.81).

Finally, the behavioral intention reported by both female and male recipients was
generally low. When the communicator was male, female recipients expressed a
slightly higher behavioral intention than males did (M=-.73 for females Vs. M=-.74
for males). When the communicator was female, female recipients expressed a lower
behavioral intention than males did (M=-.67 for females Vs. M=-.64 for males). (See
table 8 for the means by genders).

5.2. Hypotheses and research questions testing:

5.2.1. Testing the main effects:
5.2.1.1. Effects of the communicator ’s gender:

In what follows, I report the results of the analyses conducted to test the main
effects of the communicator’s gender on message response. Table 9 summarizes the
results of two-way AN OVAs and Table 10 summarizes the results of AN COVAs of
gender’s effects on all dependent variables.

Source credibility

H 1: The male communicator will be perceived as more credible than the
female communicator.

This hypothesis was tested using a two-way ANOVA as well as an analysis of
covariance (AN COVA).

The two-way ANOVA showed that the communicator’s gender had a significant
impact on source credibility (F (1, 218) = 10.19, p < .01). However, while the

hypothesis predicted that the male communicator will be perceived as more credible

72

than the female communicator, the means revealed an opposite-direction effect. It was
the female communicator who was perceived to be more credible than the male
communicator, by both male and female participants (table 8). Therefore, hypothesis 1
was not supported.

The ANCOVA confirmed that the gender of the communicator has a significant
effect on source credibility (F (1, 217) = 12.39, p < .01). Also, the direction of the
effect was opposite to what the hypothesis predicted. Both male and female
participants perceived the female communicator to be more credible than the male
communicator (table 8). Therefore hypothesis 1 was not supported.

Emotional response

H2: The male communicator will produce a stronger emotional response than
the female communicator.

The AN OVAs showed that the communicator’s gender had no significant effect on
fear (F (1, 218) = .34, p > .05), no significant effect on anger (F (1, 218) = .32, p >
.05) and no significant effect on guilt (F (1, 218) = .47, p > .05).

A series of ANCOVAs was also conducted in order to test this hypothesis. The
covariates included to test the effect on fear were empathy with the victim, likeability
of the PSA characters, perceived similarity with PSA characters, and gender identity.
The covariates used to test the effect on anger were empathy with the victim,
likeability of the PSA characters and familiarity with the issue of relationship
violence. The covariates used in testing the effect on guilt were empathy with the

victim and prior involvement in relationship violence.

73

The AN COVAs confirmed that the gender of the communicator had no significant
effect on fear (F (1, 214) = 1.51, p > .05), no significant effect on anger (F (1, 215) =
.15, p > .05), and no significant effect on guilt (F ( 1, 216) = 1.10, p > .05). This means
that the recipients’ emotional response would not be significantly different, whether
the communicator was male or female. Therefore, hypothesis H2 was not supported.

Cognitive response

RQl: What impact does the communicator’s gender have on the recipients’
cognitive response to the relationship violence prevention PSA?

The two-way ANOVA revealed no significant impact (F (1 , 218) = 1.34, p > .05).
The AN COVA was conducted using empathy as a covariate, and also revealed no
significant impact of the communicator’s gender on cognitive response (F (1,217) =
2.07, p > .05). This means that the recipients’ cognitive response would not be
significantly different depending on whether the PSA voice-over was male or female.

Attitudes

RQZ: What impact does the gender of the communicator have on attitude
issue, attitude voicwver, and attitude pg A in the context of relationship violence

prevention PSA’s?

First, the two-way ANOVA revealed that the gender of the communicator had no

significant effect neither on attitude issue (F (1, 218) = .44, p > .05), nor on attitude

voice-over (F (1, 218) = .11, p > .05), nor on attitude pSA (F (1, 218) = .31, p > .05).
Concerning the AN COVA, the covariates included in analyzing the effect on

attitude issue were empathy with PSA characters and familiarity with the issue; the

74

covariate included in analyzing the effect on attitude voice-over was empathy; and the

covariates included in analyzing attitude p3 A were empathy and the likeability of PSA

characters.

The AN COVAs confirmed that there is no significant effect for the
communicator’s gender on any of the three attitude variables (F (1, 215) = .30, p
<.05), (F (1, 217) = .19, p > .05) and (F (l, 216) = 1.02, p > .05) respectively.
Therefore, all three attitudes would not be significantly different depending on
whether the PSA voice-over is male or female.

Behavioral intention

RQ3: What impact does the gender of the communicator have on the
recipients’ behavioral intention to provide help or seek information about
relationship violence?

The AN OVA showed no significant effect of the communicator’s gender on the
participant’s behavioral intention (F (1, 218) = .18, p > .05). Concerning the
AN COVA, the covariates included were empathy and prior involvement in
relationship violence. The ANCOVA also showed no significant effect of the
communicator’s gender on behavioral intention (F (1,215) = .51, p > .05), suggesting
that the recipients’ behavioral intention would not significantly change depending on
whether the PSA voice-over is male or female.

5.2.1.2. Effects of the recipient’s gender:

In what follows, I report the results of the hypotheses and research questions

concerning the main effects of the recipient’s gender on message response. Table 9

75

summarizes the results of two-way ANOVA tests and table 10 summarizes the results
of ANCOVA tests.

Source credibility

RQ4: What impact does the gender of the recipient have on source
credibility?

The two-way AN OVA showed that there was no significant impact of the
recipient’s gender on source credibility (F (1, 218 = .01), p > .05). Similarly, the
ANCOVA confirmed the null finding (F (1 , 217) = .01, p > .05), suggesting that the
degree of the communicator’s perceived credibility would not significantly change
depending on the gender of the recipient.

Emotional response

H3: Female recipients will have a stronger emotional response to the PSA’s
than male recipients.

Two-way ANOVAs showed no significant effect neither on anger (F (1, 218) =
.006, p > .05) or on guilt (F (1, 218) = 1.15, p > .05). However, the effect on fear was
statistically significant (F (1, 218) = 4.99, p < .05), such that females felt higher
degrees of fear than males, both when the communicators were male and female.

Similarly, the ANCOVAs showed no significant effect of the recipients’ gender on
anger (F (1, 215) = .13, p > .05) and guilt (F (1, 216) = 1.07, p > .05). However, it
showed a significant effect on the emotion of fear (F (l, 214) = 7.15, p < .05). The
direction of the effects was also the same as the one revealed by the two-way

ANOVA: female recipients expressed more fear than males. Therefore, hypothesis H3

76

was partly supported, suggesting that only fear, but not anger or guilt, changes
significantly depending on whether the PSA recipient is male or female.

Cognitive response

RQS: What impact does the gender of the recipient have on cognitive
response?

The two-way ANOVA revealed no significant impact of the recipient’s gender on
cognitive response (F (1, 208) = 2.19, p > .05). Similarly, the ANCOVA confirmed the
null finding (F (1,217) = 3.07, p > .05). This suggests that whether the recipient is
male or female, the cognitive response to the PSA would not be significantly different.

Attitudes
H4: Female recipients will have a more positive attitude issue than male

recipients.

The two-way AN OVA showed a significant impact of the recipient’s gender on

attitude issue (F (1, 218) = 26.91, p < .01), such that female recipients reported a higher

attitude toward the importance of the issue of relationship violence compared with
male recipients.
The ANCOVA confirmed the null finding (F (1 , 215) = 29.44, p < .01), and the

direction of the effect was the same. Therefore, hypothesis H4 was supported.
RQ6: What impact does the recipient’s gender have on attitude pSA?
The two-way ANOVA showed that the recipient’s gender had no significant

impact on attitude pg A (F (1 , 218) = .43, p > .05). Similarly, the ANCOVA showed no

significant impact on attitude pg A (F (1 , 216) = .32, p > .05). This suggests that the

77

recipients’ attitude toward the PSA would not significantly change depending of their

gender.
RQ7: What impact does the recipient’s gender have on attitude voicwver?

The two-way ANOVA showed that the recipient’s gender had no significant
impact on attitude toward voiccmfl (F (l , 218) = 1.25, p > .05). This result corresponds

with the ANCOVA showing no significant effect of the recipient’s gender on attitude
voice-over (F (1, 217) = 1.49, p > .05). This also suggests that the recipients’ attitude

toward the voice-over would not significantly change, whether they are male or
female.

Behavioral intention

RQ8: What impact does the recipient’s gender have on the recipients’
intention to positively participate in reducing the spread of relationship violence?

The two-way AN OVA revealed no significant effect of the recipient’s gender on
the recipients’ gender on their intention to behave as positive bystander and contribute
to the reduction of relationship violence (F (1, 218) = .002, p > .05). Similarly the
ANCOVA revealed no significant effect on behavioral intention (F (1, 215) = .01, p >
.05), suggesting that the recipients’ attitude toward the PSA would not be significantly
different, whether they are male or female.

5.2.2. Testing the interaction effects:

In what follows, I report the results of the hypotheses concerning the effect of the
interaction between the communicator and recipient’s genders on message response.
Table 9 summarizes the results of two-way ANOVAs and TablelO summarizes the

results of ANCOVAs.

78

Source credibility

H5: Perceived source credibility will be higher when the genders of the
communicator and the recipient match, compared with when they do not match.

The AN OVA showed no significant impact of the gender interaction on source
credibility (F (1, 218) = .94, p > .05). Similarly, the ANCOVA revealed no significant
impact (F (1, 217) = .71, p > .05). Therefore, this hypothesis was not supported. This
suggests that the perceived credibility of the communicator would not be significantly
different, whether the genders of the communicator and the recipient match or
mismatch.

Emotional response

H6: Emotional response will be weaker when the genders of the
communicator and the recipient match, compared to when they do not match.

The two-way AN OVA showed no significant impact of gender interaction on fear
(F (1, 218) = 1.00, p > .05), anger (F (1, 218) = .04, p > .05) and guilt (F (1, 218) =
.63, p > .05). Similarly, the ANCOVAs revealed no significant impact of the genders
interaction on either one of the emotions, including fear (F (1, 214) = .27, p > .05),
anger (F (1, 215) = .01, p > .05) and guilt (F (1, 216) = .92, p > .05). Therefore,
hypothesis H6 was not supported, suggesting that the recipients’ emotional response
would not be significantly different, whether their genders matched or mismatched
with that of the communicator.

Cognitive response

H7: Cognitive response will be higher when the genders of the communicator

and the recipient match, compared to when they do not match.

79

The two-way ANOVA conducted to test the gender interaction effect on cognitive
response revealed no significant impact (F (1, 218) = .70, p > .05). The ANCOVA
confirmed the null finding (F (1,217) = 1.41, p > .05). Therefore, hypothesis H7 was
not supported, suggesting that the recipients’ cognitive response to the PSA would not
be significantly different, whether their genders matched or mismatched with that of
the communicator.

Attitudes

H8: Attitudes will be more positive when the genders of the communicator
and the recipient match, compared to when they do not match.

The two-way AN OVA showed that that the interaction of the communicator and

recipient’s genders had no significant impact on attitude issue (F (l , 218) = .003, p >

.05), no significant impact on attitude voice-over (F (1, 218) = 1.33, p > .05), and no

significant impact on attitude pSA (F (1, 218) = .47, p > .05).
The ANCOVAs also showed that the gender interaction had no significant impact

on attitude issue (F (1, 215) = .22, p > .05), no significant impact on attitude voice-0ver

(F (1, 217) = 1.10, p > .05), and no significant impact on attitude pSA (F (1, 216) = .25,

p > .05). Therefore, hypothesis H8 was not supported, suggesting that the three
attitudes would not be significantly different depending on whether the genders of the
communicator and the recipient match or mismatch.

Behavioral intention

80

H9: Behavioral intention to provide help or seek information about
relationship violence will be higher when the genders of the communicator and
the recipient match, compared to when they do not match.

Finally, the two-way AN OVA showed no significant impact of the interaction of
the communicator and recipient’s genders on behavioral intention (F (1 , 218) = .006, p
> .05). The ANCOVA confirmed the null finding (F (1, 215) = .76, p > .05).
Therefore, hypothesis H9 was not supported either, suggesting that the recipients’
behavioral intention would not be significantly different depending on whether the
genders of the communicator and the recipient match or mismatch.

After presenting the results of hypotheses and research questions testing, the next
chapter discusses the finding, as well as the limitations, implications and future

research recommendations.

81

CHAPTER VI — DISCUSSION

6.1. Overview of the findings

The present study investigated the main and interaction effects of gender on the
response to relationship violence prevention PSA’s. Message response refers to source
credibility, emotional response, cognitive response, attitudes toward the issue of
relationship violence, toward the communicator and the PSA, and behavioral
intention.

The study yielded three significant findings. First, the gender of the communicator
(PSA voice-over) has a significant impact only on source credibility, but not on
emotional response, cognitive response, attitudes and behavioral intention.
Particularly, the results indicate that the female communicator has been perceived to
be more credible than the male communicator both by male and female recipients. In
fact, among the participants who watched the female version PSA, male recipients
reported a higher perceived source credibility than female recipients.

Second, the study found that the gender of the message recipients affects only one
aspect of emotional response: fear. Particularly, female recipients reported higher
degrees of fear than the male recipients. The message recipient’s gender did not have
any significant effects on other aspects of emotional response, including anger and
guilt.

Third, the study found that the gender of the recipients significantly affects attitude

toward the issue, such that females had a higher attitude issue than males. The gender

82

 

of the recipients did not affect any other type of attitudes, including attitude voicemm

and attitude pSA, nor did it significantly affect behavioral intention.

Concerning the interaction effects of the communicator and recipients’ genders,
the results showed no significant effect on any of the response variables. This
indicates that the response to a relationship violence prevention PSA would not
significantly differ whether the genders of the communicator and the recipient
matched or mismatch.

6. 2. Discussion

6.2.]. Discussing the signifrcantfindings:

The effects of communicator ’s gender on source credibility

The present study found that the female communicator was perceived to be more
credible than the male communicator. While this supports previous findings that the
gender of the communicator does affect source credibility, it opposes the usually-
found direction of gender’s effect. In fact, the bulk of previous source credibility
research found that male communicators tend to be perceived more credible than
female communicators (e.g., Newcombe & Amkoff, 1979; Markham, 1988). The
present study found the opposite.

Many factors could be approached to explain this finding, but some of them have
been refuted by the method of the present study, namely through the use of control
variables. First, one of those factors could be empathy with the victim. Since the study
stimulus (the PSA) depicts a female victim and a male perpetrator, one may consider
empathy with the female victim as a factor that must have spurred a higher perceived

credibility in favor of the female communicator. However, because of the significant

83

impact that empathy had on source credibility, it was used in the present study as
control variable. The results were still in favor of the female communicator in terms of
sourse credibility.

Second, one may think of identification with the female victim as a factor that
worked in favor of the female communicator. But this may not be the case either. In
fact, previous research about identification showed that audiences tend to identify with
characters that are perceived as good or exhibiting positive behavior rather than those
who are perceived as bad or exhibiting negative behavior. This process is called
wishful identification (F eilitzen & Linné, 1975).

Following this logic, we would expect the participants to identify with the female
victim rather than the male perpetrator. This was true in the present study where two
aspects of identification were measured: perceived similarity with the characters and
likeability of the characters. More participants expressed likeability toward the victim
(26.9%) rather than the perpetrator (3.6%), and more perceived similarity with the
victim (14.3%) rather than the perpetrator (2.3%).

However, those who did report identification with either the victim or the
perpetrator represented only 30.5% and 16.6% of the participants respectively. On the
other hand, most participants did not identify at all with any of the two PSA
characters. While 69.1% answered that they did not like any of the two characters,
83% answered that they did not perceive themselves to be similar to either of the two
characters. The reason may be that even the victim’s behavior was not perceived as
“good”. The fact that she does not fight back may have created an impression of

submissiveness or weakness.

84

Consequently, since most participants did not express identification with any of
the two characters, both identification variables —- likeability and perceived similarity -
did not have a significant effect on source credibility. This means that identification
could not have significantly biased source credibility in favor of the female
communicator.

One potential explanation for the superior credibility of the female
communicator would be the topic itself of the PSA. Because the PSA deals with
relationship violence, and because women are more often the victims of relationship
violence, the study participants may have felt more credibility toward the female than
male communicator. This effect may have been boosted by the significant positive
correlation between attitude toward the issue of relationship violence and source
credibility (.19).

Another explanation that the present study did not measure is the mental
association that the participants may have established between the male actor
(perpetrator) and the male voice-over. This probably explains why female
respondents, but also male respondents had a negative attitude toward the male
communicator (M = .6 and SD = .87 for male participants; M = .91 and SD = .93 for
females).

The effect of recipient’s gender on fear

The results indicated that the gender of the recipient significantly affects the
degree of fear elicited by the PSA, especially among female recipients who reported

higher degrees of fear compared with male recipients. As is the case with source

85

credibility, identification with the victim or empathy could have been potential factors
to explain this finding, but this may not be the case.

The reason why empathy and identification may not have caused the significant
impact of gender on fear is that they both were used as control variables. First, the
effect of the recipients’ gender on emotional response (including fear) was tested
using a factorial ANOVA. The effects turned out to be significant only on fear.
Second, the same effect was tested again but this time using an ANCOVA test,
including both empathy and likeability as control variables. The result remained the
same such that gender had a significant effect only on fear. Consequently, it may be
concluded that neither empathy nor identification played a role in creating the effect of
gender on the emotional response of fear.

One explanation for the result would be the significant impact of the recipient’s

gender on attitude issues which in turn has a significant positive correlation with fear.

In fact, the present study found that the gender of the recipients had a significant

impact on attitude issue: such that females had a positive attitude toward the issue than

males did. This means that females, more than males, perceived the issue of
relationship violence as important and worthy of concern. Since there was a significant
positive correlation between attitude toward the issue and fear (.181), this can
probably explain why the gender of the recipient significantly impacted the emotional
response of fear, and why it was females who reported more fear than males.

The finding that it was females who expressed more fear than males may also be

explained by previous findings that females either feel stronger emotions compared

86

with males, or tend to be more expressive than males when it comes to emotions (see
the review of LaFrance & Banaji, 1992).

The effect of recipient ’s gender an attitude toward the issue

The recipient’s gender was found to significantly affect attitude issue, such that

females perceived the issue to be important and worthy of concern more than males
did. One explanation for this finding could be the gender-sensitivity of the issue
(Ascensio, 1999). This gender-sensitivity makes it logical that the gender of the
recipient may impact their attitude toward the issue.

Concerning the direction of the effect, where females had a more positive attitude
toward the issue of relationship violence than males, this could be explained by the
gender-sensitivity of the issue, but also the fact that females are usually the victims of
relationship violence. This widespread fact and belief could have made females hold a
higher attitude concerning the importance of the issue of relationship violence.

6. 2. 2. Discussing the null findings

The results showed a lack of significant gender influence on emotions like anger
and guilt, as well as attitude voice-over and attitude pg A and behavioral intention. Many

arguments could be suggested to explain these null findings. They include a)
mediating effects between some variables, b) the nature of the communicator (voice-
over), c) sample size and sample attributes, d) the technical evaluation of the PSA, and
e) audience characteristics.

a) Mediating effects

The present study treated all message response variables as dependent variables,

including source credibility, emotional response, cognitive response, attitudes and

87

behavioral intention. Treating them all as dependent variables aims to keep the focus
on the influence of gender on all the response variables, rather than the influence of
the variables on each other. However, previous research demonstrated that there are
mediating effects that take place between the response messages themselves. Not
taking these mediating effects into consideration in the present study may have
affected the results.

One example of previous findings about mediating effects is that cognitive
response mediates the relationship between advertising and attitudes toward the brand
(e.g., Mitchell & Olson, 1981; Wright, 1973). Also, emotional response mediates the
relationship between message content and the attitudes of message recipients toward
the message or brand (e.g., Holbrook & Batra, 1987). Kim, Allen and Kardes (1996)
also identified mediating effects of affective and cognitive processes on the effects of
advertising on attitudes toward the brand.

These previous findings therefore suggest that the relationship between gender and
the response variables measured in the present study may not be direct. It rather
involves mediating processes that may affect the results. While the findings of the
present study constitute a preliminary effort toward understanding the effect of gender
on message response within the context of relationship violence PSA’s, it would be
recommended to take the mediating effects into consideration in future research.

b) The nature of the PSA communicator:

One reason for the insignificant effects of gender on attitude toward the
communicator may be the fact that the communicator is a voice-over. The fact that the

recipients did not physically see or interact with the communicator, or the fact that

88

they did not listen to the communicator for a long-enough duration may have
weakened the degree of response.

In fact, in most previous studies where attitude toward the communicator was
examined, the communicator was either a person who directly talked with the
participants (Woodside & Davenport, 1974) or someone they saw on screen, or a
combination of on-screen and audio communication (e.g., Kelman & Eagly, 1965).
This implies that the ability to see and sometimes personally interact with the
communicator may create stronger attitudes communicator.

It is true that in some previous experiments the communicator was just heard on
tape (e.g., Whittler, 1991). However, these experiments made the participants listen to
interviews (e.g., Kelman, 1958) or speeches (e.g., Stern, Mullenix, Dyson & Wilson,
1999), the length of which may have allowed the recipients to form an attitude toward
the communicator. Conversely, the voice-over used in the present study speaks only in
the end of the 30-second PSA. Therefore, the brief exposure to the communicator’s
message may have weakened the degree of generated response.

c) Sample considerations

Two factors related to the study sample may have affected the strength or direction
of the findings. First, the sample size was limited. Although the survey was sent to a
total of 525 students, 243 responses were received, and only 223 participations were
finally used for the data analysis. The number of participants dropped for many
reasons including some students’ decision not to participate (participation was

voluntary), incomplete surveys, or duplicate participations.

89

 

Additionally, the analysis of the present study required dividing the participants
into four cells, thus creating a less-than-desired number of participants in each cell,
such that the number in each cell ranged from 85 at the high end down to only 28. In
fact, although there are no rules concerning the ideal sample size, researchers usually
use 50 to 100 participants in each cell (Wimmer & Dominick, 2006).

Also, since the present study involves five dependent variables, some researchers
would recommend using about 100 participants for every dependent variable
(Gorsuch, 1983, as cited in Wimmer & Sominick, 2006). Consequently, the effects
may have been weaker than desired, which may explain why many findings were null.

Second, the skewness of some results toward female participants may have been
due to the inequality of the cells both in terms of size and gender. In fact, because the
two conditions (male and female communicator) were randomly assigned to male and
female participants, some cells ended up receiving more participants than the others.
Thus, for the female communicator condition, the cell of females included 63
participants (28.4% of total participants), while the cell of males included only 28
participants (12.6%). Similarly for the male communicator condition, the cell of
females included 85 participants (38.3%) while the cell of males included only 46
participants (20.7%).

d) PSA technical evaluation

At least three dependent variables may have been affected by the audience’s

evaluation of the quality of the PSA’s production. In fact, although the present study

measured attitude pSA, this scale focused on perceived persuasiveness, rather than

what the recipients thought about the quality of the PSA or how much they liked or

90

disliked it. Also, some researchers have drawn a distinction between perceived
effectiveness of a PSA (measured in the present study) and the liking of a message
(Dillard & Peck, 2000). PSA evaluation, based on the liking of the PSA, could also be
able to affect message response. For example, a positive ad evaluation was found to
positively impact behavioral intention (e.g., Mackenzie & Spreng, 1992).

In order to get a sense of what the participants of the present study thought about
the quality of the PSA, it would be interesting to read through the thoughts generated
for the cognitive response variable. In fact, one of the categories used for coding the
thoughts generated by participants was “irrelevant or neutral thoughts”. Since this
study focuses on the thoughts that supported the advocated behavior, all the thoughts
that included an evaluation of the PSA’s technical production were coded as “neutral
or irrelevant”.

At this stage, assrunption cannot be statistically based on the content of those
thoughts. Nevertheless, reading through the thoughts could suggest that the
participants’ evaluation of the PSA may have affected their response to it. Examples
of those thoughts are “the black and white made it seem much more intense”, “the
black and white was effective”, or “the lighting is really weird”. Comments were also
made about the actors. Examples are “Male was a bad actor” or “need better acting”.
A qualitative analysis of the cognitive response thoughts may reveal more precise
trends of how the participants’ evaluation of the PSA production may have affected
message response.

e) Community considerations

One last explanation for the absence of significant effects on most message response

91

variables, including behavioral intention, can be the degree of the participants’
readiness to positively respond to such pro-social messages about relationship
violence. Research showed that communities do not always have the same degree of
responsiveness to messages that call them to change their social behaviors and beliefs
(Banyatd, Plante & Moynihan, 2004).

Edwards et al. (2000) constructed a community-readiness model, suggesting that
the effectiveness of prevention strategies, including pro-social messages, depends on
the level of the community’s awareness of the issue. The more aware the community
the more ready it will be to listen to the message, and to actively respond to it and deal
with it (Banyard, Plante & Moynian, 2004).

In fact, this model may explain the low levels of behavioral intention in the present
study, since the measured familiarity and prior involvement with the issue were also
low. First, the participants had a low level of familiarity with the issue of relationship
violence. Their average familiarity was negative (M= -. 14) with a standard deviation
of 1.61. Second, when asked whether they had ever received a help request from a
relationship violence victim or perpetrator, 91% of the participants answered “no”.
When asked whether they had ever interfered to help solve a relationship violence
situation, 72.6% answered “no”. Third, when asked whether they had received any
training related to relationship violence, 68.2% of the participants answered “no”.
This, according to the community-readiness model (Edwards et al., 2000) indicates
that this cohort of students will need higher levels of awareness before they can have a
stronger response to relationship violence PSA’s.

6.3. Limitations of the study

92

Despite its potential theoretical and professional contributions, the present study
has a number of limitations that need to be addressed by future research. First, in this
study, source credibility, emotional response, cognitive response, attitudes and
behavioral intention are all treated as dependent variables. However, previous research
used source credibility, emotional response and cognitive response as mediators (e.g.,
Mitchell & Olson, 1981; Wright, 1973). This may have affected the way the variables
interfered and therefore affected the results.

Second, some potentially confounding factors were not taken into consideration
because the survey was conducted online. One example of these elements is the mood
of the audience at the time of the survey response, which was shown to affect
evaluation and information processing (Bagozzi, Gopinath & Nyer, 1999). Future
research could attempt to control for this variable.

Third, this study measured behavioral intention, but not behavioral change itself.
A measured degree of behavioral intention does not necessarily mean that the behavior
change will happen in reality (Kelman, 1958). Some researchers also concluded that
attitudinal changes often last only a couple of months (Brecklin & Ford, 2001).

Fourth, a number of issues related to the study sample constitute other limitations
that could be avoided by future research. These issues include the less-than desired
number of participants (N=223), the inequality of cell sizes (the smallest included 28
and the biggest 85 participants), as well as the lack of gender balance in the sample
(148 females versus only 75 males).

Finally, since the PSA used in this study was produced by students, the production

quality may be another limitation to consider. Negative evaluations of the PSA by

93

participants may have affected their response to the message itself. Future studies may
avoid this limitation by using more professionally advanced videos.

Also, although the PSA of the present study was pretested, the focus was on
detecting accurate recall of the gender of the voice-over and the comprehension of the
message. Future research may improve this aspect by including PSA evaluation
questions in the pretest. This may allow the researchers to improve the PSA before
conducting the study.

6.4. Suggestions for future research

Based on the limitation that I have just discussed, 1 put forward the following
suggestions for future research. First, since the present study did not treat response
variables as mediators, future research could take the presentendeavor firrther by
considering the mediation effects of those variables.

Second, given the difference between behavioral intention and actual behavior,
and that the PSA used in the present study will be aired on government access channel
HOM TV, it would be interesting to conduct a subsequent study in cooperation with
MSU Safe Place in order to check the existence of any real changes that might have
resulted from watching the PSA on TV. The behavioral change to be measured is
whether viewers have responded to the message of the PSA by being active bystanders
in case they knew of any relationship violence case.

Third, other potential research paths could be of interest. The main finding in the
present study is the significant impact of the communicator’s gender on source
credibility, and that the female communicator was perceived as more credible than the

male communicator. This indicates that while designing relationship violence PSA’s,

94

it is important to consider the gender of the voice-over. Nevertheless, it would be
interesting to examine whether this finding applies solely to relationship violence
PSA’s or to other PSA’s treating other gender sensitive issues like abortion or gender-
based discrimination. Therefore, future studies could check whether female
communicators will still be more credible than the male when the message is gender
sensitive.

Also, in relation with this source credibility finding, future research could further
examine the effects of communicator’s gender on message response by comparing the
response to gender-salient messages (both male and female-oriented messages) with
the response to gender-neutral messages.

Finally, another potential alley in relation with this finding is to consider whether
the results remain the same when the communicator is a direct speaker or writer, rather
than a voice-over. In fact, previous studies that examined gender and source credibility
used a communicator who is a direct speaker or debater, either present in the same
room or seen on TV.

6.5. Implications

6.5.1. Practical implications

The findings suggest a set of guidelines for relationship violence prevention
PSA’s. First, this study confirms that it is important to take into account the specific
characteristics of each particular pro-social issue while designing PSA’s about it. In
the context of the present study, the main characteristic of the issue of relationship

violence is gender sensitivity and gender-salience of the messages. It seems that these

95

characteristics have made the female communicator gain more credibility than the
male communicator.

Since most previous source credibility research found that, in many contexts, male
communicators have more credibility than female communicators, the finding of the
present study emphasizes the effect of the nature of the message itself on source
credibility. Therefore, future anti-relationship PSA’s may be more effective if the
communicator is female than if the communicator is male. This implication is
important because within the context of persuasive messages, choosing the right
communicator can be crucial to the success of the message (e.g.,Stemthal, Dholakia &

Leavitt, 1978; Harmon & Coney, 1982).

Second, this study found that the recipient’s gender affects attitude issue , such that

females have a higher attitude toward the issue of relationship violence than males.
This means that females, more than males, perceive the issue of relationship violence
to be important and deserves concern.

However, since most relationship violence perpetrators are males, this indicates
the need to orient future awareness efforts toward male audiences. The aim is to
increase their attitude toward the issue, and therefore make them more responsive to
relationship violence prevention campaigns. In fact, some researchers have supported
the importance of targeting men in order to decrease the spread of relationship
violence (Flood, 2006).

Third, the findings suggest another implication concerning the use of fear appeals
in relationship violence prevention PSA’s. The present study found that the recipient’s

gender significantly affects fear, such that females expressed more fear than males.

96

This may suggest that fear appeal might be more effective in addressing female
audiences than male audiences. This may be useful for future campaigns that primarily
target female audiences. Examples are those addressed to female victims of
relationship violence aiming to encourage then to speak up, or even those dealing with
other health issues like breast cancer, encouraging women to make early medical
check-ups.
6.5.1. Theoretical implications

The present study aims to contribute to many fields of research. First, the findings
contribute to source credibility literature, by stating that female communicator can be
more credible than male communicators in the context of relationship violence PSA’s.
This challenges the general belief that male communicators are more credible than
female communicators across contexts.

Second, the findings of this study contribute to the literature of bystander behavior
(see Howard & Crano, 1974; Laner & Benin, 2001 for reviews of previous research).
As stated in the discussion section, the findings suggest that the response generated by

this PSA did not prompt recipients to express a high intention to behave as active

bystanders. Among other implications, this means that a high attitude issue does not

necessarily lead to a positive behavioral intention.

Finally, the findings also contribute to the literature of emotional appeals and their
effects on persuasion. On the one hand, the findings suggest that strong fear does not
necessarily mean that behavioral intention will be high. This contributes to the
findings of various previous studies about whether fear appeals can make a message

more persuasive. While some researchers report a negative relationship between the

97

degree of fear and persuasion (e.g., Janis & F eshbach, 1953; Snider, 1962 as cited in
Stemthal & Craig, 1974), others found that fear and persuasion have a positive
relationship (e.g., Dabbs & Leventhal, 1966; Komzweig, 1967).

On the other hand, the null findings about anger and guilt suggest that gender is
probably not the only variable that researchers need to focus on while trying to
identify the variables that affect emotional response. In fact, the null finding
concerning gender effects of anger and guilt seems to contribute to the inconsistent
findings about gender differences in emotional response. Fisher and Dubé (2005)
report that in only half the studies that examined gender differences in emotional
response, females and males report the same degree of intensity of many emotions

including as anger and guilt.

98

CONCLUSION

In sum, relationship violence is an alarming social issue that deserves attention
from various academic fields including communication research. Communication
research about relationship violence is important because it can contribute to the
efiectiveness of relationship violence prevention campaigns.

The present study contributes to the communication research aiming to improve
PSA’s targeting relationship violence bystanders. More specifically, this study
examines the effects of the communicator’s and recipient’s genders on message
response. Message response refers to source credibility, emotional response, cognitive
response, attitudes and behavioral intention.

The study found three main findings. First, the gender of the communicator has a
significant effect on source credibility, such that the female communicator was
perceived to be more credible than the male communicator. Second, the recipient’s
gender has a significant effect on only one aspect of emotional response - fear — but
not other aspects like anger and guilt. Female recipients expressed higher degrees of

fear than male recipients. Third, the recipients’ gender also has a significant impact on

only one aspect of attitudes — attitude issue — but not on attitude voice-over or attitude

pSA. Female recipients expressed higher attitudes toward the issue of relationship

violence than male communicators.
The findings of this study have many theoretical implications. One of these

implications is the contribution to the literatures of source credibility, since the finding

99

challenges the previously established beliefs in favor of male communicators. Second,
the findings also contribute to the literature of bystander behavior, since the study
stimulus was a PSA targeting potential relationship violence bystanders. Third, the
findings also contribute to the literature of emotional appeals, both by confirming the
effect of the recipient’s gender on fear, and by refuting gender effects on anger and
guilt.

The main practical implications suggest that it is important for future PSA
designers to consider the gender salience of their message while choosing the PSA
voice-over. The implications also suggest that PSA designers need to know the
characteristics of their audience and design the PSA’s accordingly. There are two
reasons behind this implication. First, the gender of the PSA recipients affects their
attitudes toward the issue of the PSA, and therefore may affect their subsequent
behavior. Second, the gender of the recipients also affects some emotional responses

like fear. This in turn may have an impact on the response to the PSA’s.

100

fiat—e _

0383:5922. e». 08:55:83.. 2:— woom 3:8. 0232a,.

 

 

 

 

max
3:? 12:»? deg.
3»? 02:: .3 «m 3.
955558.. s. cm age 8.3. and: web:
mafia—a 02:: mm mm 3
.x. cm Hon: 56$ Nmbtx. t .o..\..
He?— OoE: E Em NNN

$03.08— wumfo gages _oo.oo\o

 

101

flea—e N

My...» :2. mow—e Define—.2 5.1.5.8

 

2:. 22.24 \ 862..

23 26222.8: \ 26222.8:
2:. 2.53.2.3: 2.53.5.2...
we: \ nee:

>353" 23 a. e: \ <2... 5:2.
finale." 2:» a: a: \ «65. 5:2.
manna“ 2:. a. a: \ «65. 5:2.
.553." 23 a. e: \ <2... 5:2.
.3532.“ 23 a: a: \ <2... 5:2.
3.1589 2:. a: a: \ <2... 5:2.

>mm5<e89 23 a. a: \ <2...
5:2.
0:25." 23 a. a: \ <2... 5:2.

>255e9 Zen .2 a: \ «.2.! 5:2.

@55558
e\.. :2. 5.29:8 86552—
: 5.. mam—e

 

mo:..ne m2... .352. O5: 0:555:23.
2.33:5
.qm 6...
.mn .3.
.mm gm
.3 .3
.3 .8
.3 be
.3 .8
.3. .3
.mm .3
.8 .3
.m0 .3
.8 .8
.3 .8
New qu Woe fmo
3.3 09$ 3. _ o co. _ q
.3 .3 .m0 .m9

 

Siege N 00:33:02

102

32.5 N 835:2:

 

H.222 \ .255.
c:.5..:..5:. \ .5—.:...::.

2:. $2.... 5:2. 852.: \ 3.2....
a 5. 2. 8:22.:
5.29.5. \ 3.9.9:.

2:. 85:25. \ 85:25.
2:. 5:8. \ .258.

2:. 3.247.: \ 5,324.2.
2:. 55.52. \ 55.52.
@525 \ 9.25.3522...
2 :. .5252. \ 0:252.
2:. 2:02.: \ 2:22.:

7.52» \ $5.22.:—

2:. 62.5522. \ 5.5525
2 2 2.527.: \ 2.52....

2:. o::<.:2.5 \ n::<.:2:5
2 :. 85.2.55 \ 2.5.2.55

25:525.
:2. 2. 5.2.52. 2.22:2.
: 5.. «2....

 

>352: >355 >355 0:55:52...
WM...» <:.In?:<2. .5.»
.3... .mm
.8 .2.
.8 .mo
.3. .2
.3 a.
.3 .3
do .3
.mo .2.
.3 .mu
.3 .3
.3 .mm
.mo .3
.3. .2
.3... .mo
.3 .mo
.0. .3
who gum um:
mmbo 3.3 3.2.
.3 .m. .8

 

3.2.... M 0:55:02

103

Adm—=8 N 8:52.83

 

5.8582. 8 5.358 .8... 3 8:33:55 $52.8 <85...
9... a. 9.. 58.! \ <2! 58.5

5.85.2... 8 5.358 .8... 8 8.3.2556 $52.8
cow—.3333 :5. a. a: 58.! \ <2! 58.5

5.85.2... 8 88.13.3533: v.3... 8.3.2.3... $22.88
.5 «8.83.. A5... a. a: 58.! \ <81! 58.5

5.8552. 8 88.13.258.35 2. .33. 8 .8... 89.8
8.»..cnm5e $22.88 33 a. a: 58.! \ <2! 58.5
5.85.2... 8 was; <58 .5... .12.... 3... 5:5! :5. a. a:
58.! \ <2! 58.5

5.85.2... 8 2... :8 5.5.8.. $.35. E :8 <58 9... a. a:
58.! \ <2! 58.5

Minna—.8
e\e 35.19.88 8.6.9.8..
n 3.. was?

 

.8552... 9.5.35.5?!
5.2.5...

.mm .3
.3 L.
.3 9m
.3 go
.3 do
.!m .3

93

Sam

.3

104

H.958 w

 

r... 9.. 0399.989 99.. >388; .8 8989. <9...9...8

 

 

0391.98

92.8.89 5895...! 9.. 9m.»
2.9.988... .589....89.39 .5

9.9.8.8.. 9.5591! 4...... 9m.»
2.9.9883 3.89.58839 N5

5999...! i... =8 <3...9

0898.. 589...!

....9.9...9....! 1.... =8 .88 9..
8.9.3.59... <3898

09.489999...“ 8.89.89. <9...9...8
5.89.. 8.99.399. 82.998 .5
>9m8.. 8.99.399. .89—.998 N5

>amfiggfl mmwgfl
>235? mum.»

189.. 8.99.399. 8.9998 .5

3.9.8 2.8.5.5.!

.89.. 8.99.399. .89—.998 .5
>9m8.. 8.99.399. .89—.998 N5
9.... 8.99.399. 82.998 8

>nn==fla .899
.92....38 3.898..

>225? .999
.w8..9<3..9. 389.39

.89..

>9m8 89.9..999. .89—.998 N5
wa..9<3..9. 389.39

105

.958 A

 

O98 -¢<9 >99. m... 9.. <9...998 3.. 5.488.. 9.. .5899...

 

99 .u 89.. >9 8.. 99.. >388 .988

 

 

 

$5998 99.. 89:88 E. MM Elm m
£89..
.8282. @393 N No.5 .93 .99. .
$3.59 0893 NE 89.8 N...
>9M8..
.8288: 0.99.8 N 3.3 9.59 .9...
€599 9.99.8 N . o .5999 ..mm
>235? .898
.8288: 0.95.8 N N . .N. .93 fog...
15...... @3998 N... 55.8 .mm

 

.. 8 A .8. :mA .2. ...._.A .8.

106

 

 

 

 

 

<9_..9...8 999 8988 E MM KM m
189..
.8283 999.8 N 3.99 3.3 9.8:
$3.59 09.99.98 39 9o. .9. NE
AMA .S.

107

H.958 9

098 -<<9 >99. 8... 9.. <9...998 3.. 5488.8 9.. 1.3.. .9<9.<89.89. .9 ”8.9.398...

 

$9398 99 Q9... 99.. .w8..9<3..9. 3.89.39

 

 

 

$9393 999 89988 H. Mm Km m
09...
.8288: @898 . N...~ N. .5 wk...
$.33 0.89.8 .91.... 999.... N9.
.w8..9<3..9. 389.39
.8889 0.99.8 . W99. m9“. .89.
55.3 9.99.8 Be 39.3 .5...

 

.mA .8. :8 .2.

108

9.9.9.9 9

09.99.9993 9.. 7.9.9.. <9...9...99

 

99.99 999. .9

 

 

 

.9 9 9 9 9 9 9 9.9::
9 9.99

9 .99: 9.99

9 .9... .99: 9.99

9 .9. .99: .99: 9.99

9 .9... .9... .99: .99 9.99

9 .99: .99: .9... .9. .99.. 9.99

9 .9... .99 .99 .99 .99. .99: 9.99

9 .99: .99: .99: 9... .99: .3: .99: 9.99

9 .9... .9... .3: .99 .99.. 99: .99.. .9... .99

.9 .99: .99: .99: .9... .99: .99. .99: .99: .99: .99

: -99 -99: -99 -99 -.99 .99 .9. -99 -.99 -99 9.99

.9 .99 .9. 99. .99 .99 .99: .99 .99 .99: 99.. -99 9.99
29.9.

 

<9...9.9.9 99.999. .. 999.99 9999......3.9 N. ..99... w. 99m9... 9. m5... m. 99m9...<9 99999999.

9. 93.999 .9999. 4. 93.999 5.8.9993 m. 9....999 99>. 9. 999999999. .9.99..99. .9.
9599...... 3. «9999199993.. 3. ..99....9...Q 3.... 2.9 .9999 9.. 99.999999... <.9.9999.

109

 

H.929 «

09.9.9.9..999 9.. 2.9.9.. <9...9.9.99 2.9.9 .99....9. 99.9

 

 

 

9 9 999999999 99 99
. 9.99

9 .99: 9.99

9 .99. .3: 9.99

9 .99 .99: .99 9.99

9 .99: .99: .99: .99 9.99

9 .99: .9... .99: -99 .9... .99

9 .99: .99 .99 29.9 .99. .99: 9.99

9 .99. .99.. .99: .99 .99. .99. .99.. 9.99

9 .99: .99: .99.. .99: .99: .99. .99: .99: 9.99

99 .99. .99: .99: .99: .3: .99: .99 .99: .3: 9.99

.. 3...”. J...” .5. 3...... :99 Jam... .3 :99. -..m. 39.9. .9...

.N 39. b9 .3 -99 .N9 -99 .9 km .99 .99 a... .99.
29.9.

<9...9.9.9 99.9999 .. 999.99 9.99.9.5... N. ..99... u. 99«9... A. «9.... m. 99«9...<9 .99—.9999.
9. 9....999 .9999. q. 9....999 5.8.39... «. 9....999 99>. 9. .99—9.9.9.9. .9.99..99. .9.

9.9.9.3.. Z. «9999.. 599...... .N. ..9.9...9....v. 1.... ...9 .9999 9.. ..9.9..999....9 «99.9999.

 

110

 

9.9.9.9 9

09.99.9993 9.. 39.9.. «99199.99 ....9.99.9 9999.

 

 

. N w A m 9 ... m 9 .9 .. .N

9 ..99
9 .99: ..99

9 ..9 ...... ..99

.9 -.9. .99.. .99 ..99

9 .99.. .99: .99 ..9 ..99

9 .99 .9. .... -.9. ..9 ..99

9 .99: ... .9. .99 .9. ..9 ..99

9 .99.. ..99: ..9. ..9. .99.. ..9 ...9: ..99

9 .99.. .99.. .99.. .99: .99.. .... .99.. .99: ..99

.9 .99: .99: .99: .99: ...9: .99 .9... .99: .99: ..99

.99 -.N9. -.9N .9N 9.. ..u .9. .99 .9“. .99 ..99
.N .94 .9N .99 J99 J99 ..u -.9. -.9N ..u .99 .9N ..99

29.9.
<9...9.9.9 99.999. .. 999.99 9.99.9.5... N. ..99... m. 99«9... ... «9.... m. 99«9...<9 99999999.

9. 9....999 .99..... q. 9....999 3.8.22. 9. 9....999 99>. 9. .99..9<.9..9. .9.99..99. .9.
2999...... ... «9999.. 599...... .N. ..99....9...... 9...... ...9 .9999 9.. ..9.9..999...9 <.9.9999.

~
~

 

111

 

...9...9 .9

39999 999 m.9999..9 59.999999 9m .99 99999. <9...9.9.99 .9 09999..

 

 

 

 

..99 9
<9...9.._9 0933939939 ”99.9.99. 3.9.393 ”9.99.393 3 Mb
m99..99 39.9 39.9 -..99 N99 39.. ... .
9.99.9.9... 9939.9 -....99 N99 .99 . . . 9
9939.9 39.9 .. .N9 ..99 . 9. .99
9939.9 -..99 9.99 .99 .99
999.. 39.9 39.9 -....99 9.99 -..... ..m.
.939..999. 9939.9 -..99 . .99 -. .9 ..99
3.99.... .. 9.29% a... -999 9.99 -.99 .99
9939.9 -..99 9.99 -.N. ..99
>9«9.. 39.9 39.9 .N.. .99 ...... .99. ..N9
.3998... 9.2.9.9.... -999 9.99 .9. ......
.. 9999999 N. 9 939.9 39.9 -..99 9.99 .9. ......
9939.9 -..99 9.99 .9. ...9
G9... 39.9 39.9 -9 .99 .9599 -. .99 ..9.
.3598... 9.39.... -999 9.99 -. .... 9.99
..9999999 9.
9939.9 39.9 -..99 9.99 -. .N. . .99
9 939.9 -999 9.99 -. . . 9 ..99

 

99.9.9 .9 99999999

112

 

 

 

9999.9 .9 999.3999.
.w99M9

<9...99_9 093.999.99.99 ”99.9.99. 3.9.393 399.393 R 9'6
6999...: 39.9 39.9 .99 999 9.. u 9 .99
.99—.9999 9939.9 .99 999 9.3 9....
9939.9 39.9 .99 .999 9.99 9....
9939.9 .99 999 9.99 9.99
>3399 .939 39.9 39.9 -. .99 9.99 .99 ...9-9
9939.9 -.99 9.99 9.99 .99
9939.9 39.9 -.99 999 .9. ...N
9939.9 -99 9.99 9.99 .99
>....999 3.99.3.2. 39.9 39.9 -999 9.99 .99 9..
9939.9 -. .99 N99 9. .99
9939.9 39.9 -.99 9.99 .9. .99
9939.9 -999 9.99 .99 99

 

99.9.9 .9 999.3999

113

 

 

 

 

9999.9 .9 999.3999.
”99m9

<99.9_._9 6933939939 ”99.9.99. 3.9.393 399.393 E ml...
>....999 99> 39.9 39.9 -999 9.99 99 ..99
9939.9 -999 9.99 .99 ..99
9 939.9 39.9 -999 9.99 .9 9 .. .9.
9939.9 -999 9.99 .9 9 ..99
.99—92.999. 39.9 39.9 -999 999 -.99 . .99
.9.99..99 9 939.9 -999 9 .99 -9 9 .. .N..
9939.9 39.9 -999 9 .99 -.99 . . 99

9939.9 -999 9.99 -.99 ..99

114

...9—9.9 ..

 

>99. 9.9 9.. <99.9999 ..99 .99 99999. <99.9.._99

 

 

 

 

 

 

<99.9.._9 999 999999 E. mm 39 9
999999 9999.995. .
0933939939... «99999 ..99 . ...: ...-3 .9...
”99.9.99..m 999999 A”. . .9. 9 9.9 .9.
> a .9 . .99 .99 99
9999
09339399319 «99999 .3 . 99 99 .99
92.9.2.9... 92.92 9.. . .98 5.8 $9.
> x .9 . 9.99 9.99 .99
>9m99
0933939989; «99999 Cw. . .9. .9. .99
”99.9.9999 «99999 A”. . .9. .9. .99
> 9 .9 . 99 99 99
09...
0933939989... 999999 .3 . ..99 ..99 .39
”99.9.9939 «99999 A”. . 999 9.99 ...9
> x .9 . ..99 ..99 .99

 

.799. ..9A.o..

...-99.9 . . 999.3999.

115

...-9.9.9 .. 999.3999.

 

<99.9...9 999 999999
0999...: 99999999
093399.99.99.9 «99999 3.
”99.9.99; M99999 :9.
> a .w

>gmn=gn mmm—ufl
093399.99.99.9 «99999 3.

”99.9.9999 «99999 A...
> x .w

992...”; <9.99-9<99
0933939939.. «99999 9».

”99.9.99; 999999 A...
> x .9

>....999 99>
0933939989; 999999 99.

”99.9.9999 «99999 A...
> x .9

”9993999. .9.99..99
0933939989.... «99999 99.
”99.9.9999 «99999 A”.
> u .w

9.. mm 39 9
_ 99.9 .9999 9.9
_ a... B... ......
_ .N.. .N.. ..o
. .9. .9. .99
_ B93 ...9. 99.8...
_ .oo .oo .99
. ..o ..99 ...
. row 999 999
. .9.9 9.9 999
_ ..o ..o .9.
_ .99 .99 .99
_ re. re. .9.
_ .9. ...9 .9.
. .oo .99. .99
. .9. .9__ .9.

 

 

-:9A9x:.

116

€9.99 5

>99. 99 99. 0999999999 9.99 U9 99999. «999.9299

 

 

 

Um

mm

7am

m99999 99999993~

93999.9 _ 9 _ .99 9 _ .99 9 _ .99.....
09999999999819 «99999 3. _ .N.: .N.: 599
”999—9999.9 “99999 E. . .oo .oo be
> x w . .3 .3. .3
@9999 m: 89.3 ..o.
.9999. NNN 3. .3

1999
95999.9 _ 39.: 39.: :9 .99.:
5.99935! 9m 9m.» 9999999999 . Hum Sum 9.9.
99999799 9999—993. . .3 .3 .3
92.92 9.293 _ 5.3 3.3 9.99..
695.999.99.919 «99999 c». . ~99 N. 3 ..m.
92.38% 92.92. a. 9 3.: 3.: 9. _ 9:
> x w . .mm .99 .3
@9999 NE 393 T:
99.9. NNN 93.2.

 

. 9 A .8. ..9A .2. ER .92.

3.939 5 99999999.

117

 

 

118

399.9 3 99999999.
9.. mm 7.5 m.
>9M99
95999.9 _ $3 $3 3.99.:
9:39.93 3 99> 32.2.29 _ 9.3 9.3 9.8..
...9—99:995. 3.... ...9 .9999 . m... . w... No»
095.999.99.919 «99999 3. . .8 .3 ..m
”99.9.9999 «99999 a. . .No .N.. ..w
> a .w . .8 .8 .o.
@9999 N; ummb. ..mm
99.9. NNN 99. .w.
99..»
933.3 _ :3 :3 99.99.:
99.99 .9<9.<9.9999 . True .999 9:99..
0939.99.88.99 «99999 3. . ...o ...o .99
”99.9.95; «99999 E. . who Mao ..ou
> u w .. Now Now .3
H9999 89 £3.99 Pm»
H999. NNN 89.99
]

 

9 A .8. .9A .98.
399.9 .N 99999999.

 

..99—9.9 3 99999999.

 

9.. mm 79m m.

O9«9...<9 99999999

95999.9 9 $9.99 .3999 89.89.:
095.999.99.919 «99999 tr. . .939 .939 NS
”99.9.9999 «99999 E. . N99. N99. 93
>uw . ...99 2.99 ..9.
@9999 N3 3%.: ......
99.9. NNN MN . m be

>2..999 .92...
93999.9 _ 9.99 9.99 9.99..
5.89.9.5~ 9.. 9m.» 999999.999 . PNN mum PM...
m.99....99..w t...— .—.9 .9999 . .939 :39 3.99.:
099.599.99.919 «99999 9?. . .wo .N.... ...9
9999.899 92.9919. _ 99.8 99.8 99.99.:
> x w . .3 .3 .mm
@9999 N . m .993 39
99.9. NNN .9. 9.99

119

 

. 9 A .8. ...9A .92.
399.9 3 99999999.

.99—9.9 3 999.3999.

 

9..

 

>3..999 <9.99.9<99

.99—99.5.

0939.99.99.99; «99999 .>.
”99.9.9999 «99999 ..w.

> 9 .w

@9999

99.9.

>...E99 99>

5999.5.

5.999.995. 9.. 9m> 999999.999
0959.99.99.99; «99999 ...».
”99.9.9999 «99999 ..w.

> u .w

H9999

99.9.

..9A .2 .....9A .92.

N
tQu—In—bn—tu—ty—t

N

N
\l

#__#

N.9
NNN

mm

.99...
.2.
..No

.3
39.9w

33.9.9.
3.9..
..99
.90
.uo
39......
m . Wm.

25

3.3
..9
..No
.90
..3

33.9.9.
5.90
..9m
.99
.wo
..99

.99...
..8
...9
...9

8.8...
9.99..
..99
.99
.8

3.99.9 3 99999999.

120

3.929 5 999999990

 

 

w999<m999_ .9.99..99
.99—99.5.
.9199 99.57.2999.
9.995.993. s9... ...9 .9999
095.999.99.919 «99999 .3
”999.999.; «99999 .5
> x .w
M9999
.9.9—

.....cm .92.

9..

HHHHH

N99
NNN

mm

3%.:
..90
NS

.9.
.00
.o«

N399

h.0909

79m

39.:
9.90
No.

.9.
.oo
.o«
.00

39.99...
98
9.2

.3
.8
.8

121

APPRENDIX A

Wordin of the Surve uestions b Variables Used
Demograghics:
I. THE FIRST SET OF QUESTIONS ASK ABOUT YOURSELF
1. What is your sex? (Please choose from the drop box)
1) Male
2) Female

2. Which year are you in university? (Please choose from the drop box)
1) Sophomore
2) Freshman
3) Junior
4) Senior

3. What was your age on your last birthday? years old (e.g., 23
years old)

 

4. Which one of these racial groups BEST describes you? (CHOOSE ONLY 93$
ANSWER) (Please choose from the drop box)
‘ 1) African American

2) Hispanic or Latino

3) Asian

4) White/Caucasiasn

5) Multi-racial

6) Other (e. g., American Indian, Asian, Native Hawaiian or Other Pacific
Islander):

Gender identgz’

5. As a female/ male, you usually are...

1) Unemotional -3 -2 -1 O l 2 3 Emotional
2) Timid -3 -2 -l O l 2 3 Bold

3) Masculine -3 -2 -1 0 l 2 3 Feminine
4) Soft -3 -2 -1 O 1 2 3 Hard

5) Powerless -3 -2 -1 O 1 2 3 Powerful
6) Quiet -3 -2 -1 O 1 2 3 Loud

7) Nonaggressive -3 -2 -1 O l 2 3 Aggressive
8) Insensitive -3 -2 -1 O l 2 3 Sensitive

122

’ I

Prior involvement in relationship violence
8. Do you know any person (man or woman, including yourself) who has been

involved in a violent intimate relationship?
0) No
1) Yes
9. Have you ever interfered to stop a relationship violence situation?
0. No
1. Yes

F amiliargy' with the issue at relationshgg’ violence:

12. What do you think you are with the issue of relationship violence?

1) Not at all familiar -3 -2 -l 0 l 2 3 Very familiar

2) Not at all - Very
Knowledgeable '3 '2 -1 0 1 2 3 Knowledgeable

Source credibilgg’

15. Please evaluate the narrator you have just listened to, based on the following
scale:

1) The narrator is NOT The narrator is an

-3-2—10123

an expert expert
2) The narrator is _3 _2 _1 O 1 2 3 The narrator is
NOT experienced experienced

3) The narrator is NOT _3 _2 _1 O 1 2 3 The narrator is
trustworthy trustworthy

4) The narrator is bad 3 -2 -l O l 2 3 The narrator is good

123

Mivoice-over
16. Please evaluate the narrator (voice-over) you have just listened to, based on
the following scale:

1) The narrator is _3 _2 _1 0 The narrator is
NOT competent competent

2) The narrator is 3 2 1 O The narrator is
NOT honest - ' ' honest

3) The narrator is _3 _2 _1 O The narrator is
NOT assertive assertive

4) The narrator is _3 _2 _1 0 The narrator is
NOT informed informed

5) The narrator is The narrator is
evasive '3 ‘2 '1 0 straightforward

6) The narrator is _3 _2 _1 O The narrator is
NOT qualified qualified

7) The narrator is _3 _2 -l O The narrator is
NOT sincere sincere

8) The narrator is _3 _2 _1 O The narrator is
meek forceful

Identi motion with PSA characters
17. Which of the two main characters did you like most?

1) Man 2) Woman 0) none of the two

18. Which of the two main characters do you feel is most like you?

1)Man

Attitude ES A

19. What do you think about the video? Please select the number that represents

your answer:

1) The video was NOT _3 _2 _1 3 The video was
persuasive persuasive

2) The video was NOT The video was

. -3 -2 -1 3 .

effective effective

3) The video was NOT _3 _2 _1 3 The video was
convincing convincing

4) The video was NOT The video was

. -3 -2 -1 3 .

compellmg compelling

5) The video was NOT The video was

-3 -2 -1 3

reasonable reasonable

6) The video was NOT _3 _2 -l 3 The video was
logical logical

7) The video was NOT The video was

. -3 -2 -1 3 .

rational rational

8) The video was NOT _3 _2 _1 3 The video was
true to life true to life

2) Woman 0) none of the two

Attitllde issue;

21. What do you think about the issue of relationship violence?

4)

Trivial -3
Unimportant -3
Not worth much _3
concern

Irrelevant -3

Emotional response:

-2
-2

-2
-2

-1
-1

-1
-1

O
O

O
0

1
1

1
1

2 3

2 3
2 3
2 3

Serious
Important
Worth a lot of
concern
Relevant

22. What word would describe your emotional reaction to the video?

1)

Afraid Not at all
Fearful Not at all
Scared Not at all
Angry Not at all
Annoyed Not at all
Irritated Not at all

Aggravated Not at all
Guilty Not at all
Ashamed Not at all

Cogitive response
23. Please write up to 10 thoughts that crossed your mind while watching the

PSA. Then for each thought, choose which statement describes it most

-3
-3
-3
-3
-3
-3
-3
-3
-3

-2
-2
-2
-2
-2
-2
-2
-2
-2

-l
-1
-l
-l
-l
-1
-l
—1
-1

OOOOOOOOO

125

pap-AH—A—np—sp—ay—sp‘

NNNNNNNNN

3

wwwwwwww

Very much
Very much
Very much
Very much
Very much
Very much
Very much
Very much
Very much

Empathy with the victim at relationshgg' violence

24. To what degree do you agree with each of the following statements? Please
indicate your degree of agreement, 1 being "do not agree at all", and 7 being
"strongly agree"

1) I felt as though I
was right there in
the PSA Not at
experiencing what all
the woman was
experiencing

2) As I watched the Very much
video, I could Not at
actually feel the -3 -2 -1 O 1 2 3

all

fear the woman was
experiencing

3) At the end, I felt the
pain as if I was
struck

4) The video ended to
evoke within me a Not at Very much
desire to offer help all
to the woman

-3 -2 -1 0 1 2 3 Verymuch

Very much
all

126

I

 

Behavioral intention
25. After viewing this video and listening to the narrator’s recommendation, how
likely are you ...?

1) to provide help to Not at
relationship . -3 -2 -1 O 1 2 3 Very likely
. . . all lrkely
Violence vrctrms ?

2) to provide help to Not at
relationship all likely
violence
perpetrators?

3) to seek information Not at
about relationship all likely -3 -2 -1 0 1 2 3 Very likely
violence in general

-3 -2 -1 0 l 2 3 Verylikely

4) To seek Not at
information on how all likely
to help reduce -3 -2 -1 0 1 2 3 Very likely
relationship
violence

5) to share the video Not at
with your friends all likely -3 -2 -1 0 l 2 3 Very likely
and family?

6) to call the number Not at

shown in the video? all likely '3 '2 '1 0 1 2 3 Very likely

127

APPENDIX B

Sto board of the 30-second PSA

 

Video

Perpetrator
shakes the victim
violently as she
tries to break free
from him, and
begs him not to
hurt her.

 

On-screen Relationship HUICIKL‘
graphic:
“Relationship
violence”

shmrlrl nmvr rmnrnn

On—screen
graphic: “Should
never remain”

 

128

Audio

Perpetrator: I told
you a hundred
times how crazy
this makes me!
Victim: Let go,
please! You’re
hurting me!
Perpetrator: I’ll
show you what hurt
feels like for what
you’ve done!

Perpetrator
suddenly slaps the

victim. Sound of the slap.

On-screen a S e c re t
graphic “A
secret”

Victim: (crying)

 

Victim sitting on please don’t
the floor, crying
and trying to Voice-over: Once
protect herself, as you know about it,
the perpetrator it’s no longer just
stands with a tight their business
threatening fist.
Perpetrator:
(scolding the
victim) I said stop
crying!

129

On-screen
graphic: “You can
help them now
before their time
runs out”.

On-screen
graphic: “Call
MSU Safe Place:
517-355-1100”

 

Voice-over: They
need your help.

You can help them

l)L‘l()lC lllUl' lllllt‘ lllllS Ulll

Call MSU Safe Plarc

317 3531100

 

130

APPRENDIX C

[RB-Approved Consent Form
Title of Study: Message Evaluation Study (MES)

You are invited to participate in a research study on persuasive media message
effectiveness. We ask that you read this form carefully. After reading this page, if you
agree to take part in this study, please press “I agree”. If you do not agree to take part,
please press “cancel”. Participation in this study is not an obligation and you will not
incur any loss or risk if you decide not to participate.

Background Information:

You are being asked to participate in this study on this single occasion for extra credit
in your class. This survey is about evaluating persuasive media messages and aims to
examine whether these messages affect your beliefs and behaviors. You will be asked
to watch a Public Service Announcement (PSA) about intimate relationship violence,
and then answer a survey which will take 20 to 30 minutes to complete. You are not
expected to participate in any treatments that would incur the risk of physical or
mental injury during your participation in this study. Please note that you need to be
18 years or older to be eligible to participate in this study. If you do not wish to
participate or if you are younger than 18, please press “cancel” below then talk with
your instructor in class. Your instructor will give you an alternative task you can
complete.

Benefits and Risk:

You are not expected to participate in any treatment that would incur the risk of
physical or mental injury during your participation in this study.

The immediate benefit which participants are expected to yield by taking part in the
study is extra-credit in the course during which this research has been presented.
Also, the information you provide will be used for designing more effective media
messages and advancing social scientific knowledge.

Confidentiality:

Your answers to this survey will be kept private. Your privacy will be protected to
the maximum extent as allowable by law. No one will know what your answers
were. You can feel free to answer each question based on your actual knowledge
and behavior. The questions that ask about your background will be used only to

131

describe the types of students completing this survey. The information will not be
used to find out your name. In the end of this survey, you will be asked to provide
your last name in order to allow the instructor to award you the extra-credit.
However, NO names will ever be reported as part of the results of this study.

Voluntary Nature of the Study:

Completing this survey is voluntary. You may choose not to participate, and you
may withdraw from the study at any time without penalty. Moreover, you can
withdraw or refuse to answer any particular question without penalty.

Contacts and Questions:

You have the right to contact the investigators if you have any objections to or
concerns with any aspect of this study. The contact information is as follows: Dr. Hye-
J in Paek (5 1 7-432-83 77/ paekh@msu.edu), Kaoutar Tbatou (517-599-2452 /
tbatouka@msu.edu).

If you have any questions or concerns regarding your rights as a study participant, or if
you are dissatisfied at any time with any aspect of this study, you may contact—
anonymously, if you wish—the Michigan State University Human Subject Protection
Programs at Michigan State University, by phone: (517) 355-2180, fax: (517) 432-
4503, email: irb@msu.edu, or regular mail: 202 Olds Hall, East Lansing, MI 48824.

Pressing “I agree” below and typing the date indicates your voluntary agreement to
participate in this study.

(I agree) (Date)

132

BIBLIOGRAPHY
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and
Human Decision Processes, 50, 179-211.

American Psychological Association. (1996). Violence and the family: Report of the
American psychological association presidential task force on violence and the
family.

Andsager, J. L., Austin, E. W., & Pinkleton, B. E. (2001). Questioning the value of
realism: Young adults’ processing of messages in alcohol-related public service
announcements and advertising. Journal of Communication, 51, 121—142.

Bagozzi, R. P., Gopinath, M., & Nyer, P. U., (1999). The role of emotions in
marketing. Academy of Marketing Science Journal, 27 (2), 184-206.

Bagozzi, R. P., & Moore, D. J. (1994). Public service advertisement: Emotions and
Empathy Guide prosocial behavior. The Journal of Marketing, 58 (1), 56-70.

Baker, M. J ., & Churchill, G. A., Jr. (1977). The impact of physically attractive
models on advertising evaluations Journal of Marketing Research, 14(4).

Banyard, V. L., Plante, E. G., & Moynihan, M. M. (2004). Bystander education:
Bringing a broader community perspective to sexual violence prevention.

Basil, M. D. (1996). Identification as a mediator of celebrity effects. Journal of
Broadcasting and Electronic Media, 40 (4), 478—495.

Batra, R., Ray, M. L. (1986). Affective responses mediating acceptance of
advertising. Journal of Consumer Research, 13, 234-249.

Batra, R., Ray, M.J (1986). Situational effects of advertising repetition: the
moderating influence of motivation, ability and opportunity to respond. Journal of
Consumer Research, 12, 432-445.

Batra, R., Stephens, D. (1989). Attitudinal effects of ad-evoked moods and emotions:
The moderating role of motivation. Psychology and Marketing, 11 (3), 199-215.

Batson, C. D., Duncan, B. D., Ackerrnan, P, Buckley, T., & Birch, K. (1981). Is

empathic emotion a source of altruistic motivation? Journal of Personal and
Social Psychology, 40(2), 290-302.

133

 

Berry, D. B. (1998). The Domestic Violence Sourcebook: Everything You Need to
Know. Los Angeles: NTC Contemporary.

Buel, S. M. (2002). Treatment guidelines for healthcare providers’ interventions with
domestic violence victims: experience from the USA. International Journal of
Gynecology and Obstetrics. 78(1), 849-854. -

Byrne, D. (1961). Interpersonal attraction and attitude similarity. Journal of
Abnormal and Social Psychology, 62 (3), 713-715.

Bern, S. L. (1981). Gender schema theory: A cognitive account of sex typing
Psychological Review, 88(4), 354-364.

Bennett, L. W. (1995). Substance abuse and the domestic assault of women. Social
Work, 40 (6), 760-71.

Berk, R. A., Campbell, A., Klap, R., & Western, B. (1992). The deterrent effect of
arrest in incidents of domestic violence: A Bayesian analysis of four field
experiments. American Sociological Review, 57(5), 698-708.

Berscheid, E. (1966). Opinion change and communicator-communicatee similarity
and dissimilarity. Journal of Personality and Social Psychology, 4 (6), 670-680.

Bochner, S., & Insko, C. A. (1966). Communicator discrepancy, source credibility,
and opinion change. Journal of Personality and Social Psychology, 4 (6), 614-
421.

Borzekowski, D. L. G., & Poussaint, A. F. (1999). Public service announcement

perceptions : A quantitative examination of anti-violence messages. American
Journal of Preventive Medicine, 17 (3), 181-188.

Brecklin, L. R., & Forde, D. R. (2001). A meta-analysis of rape education programs.
Violence and Victims. 16 (3), 303-321.

Brock, T. C. (1965). Communicator-recipient similarity and decision change. Journal
of Personality and Social Psychology, 1 (6), 650-654.

Brovennan, Donald M., Edward L. Klaiber, Yutaka Kobayashi, and William Vogel

(1968), Roles of activation and inhibition in sex differences in cognitive abilities,
Psychological Review, 75 (1), 23-50.

134

Browne, N. (1980). Film form / voice-over: Bresson’s The Diary of a Country Priest.
Yale French Studies, 60, 233-240.

Burgess, R. L., & Draper, P. (1989). The explanation of family violence: The role of
Biological, behavioral, and cultural selection. Crime and Justice, 11, 59-116.

Buss, D. M. (1981). Sex differences in the evaluation and performance of dominant
acts. Journal of Personality and Social Psychology, 40, 147-154.

Caballero, M. J ., & Pride, W. M. (1984). Selected effects of salesperson sex and
attractiveness in direct mail advertisements, The Journal of Marketing, 48 (l ),
94-100.

Cascardi, M., & O’Leary, K. D. (1992). Depressive symptomatology, self-esteem, and
selfblame in battered women. Journal of Family Violence, 7 (4), 249—259.

Chaiken, S. (1979). Communicator physical attractiveness and persuasion. Journal of
Personality and Social Psychology, 37 (8), 1387-1397.

Cohen, J. (2001). Defining Identification: A theoretical look at the identification of

audiences with media characters. Mass Communication and Society, 4 (3), 245-
264.

Crano, W, D., &Burgoon, M. (2001). Mass Media and Drug Prevention: classic and
contemporary theories and research. Routledge.

Curry, M. A., Perrin, N., & Wall, E. (1998). Effects of abuse on maternal
complications and birth weight in adult and adolescent women. Obstetrics &
Gynecology, 92(4), 531-534.

Darley, W. K., & Smith, R. E. (1995). Gender differences in information processing

strategies: An empirical test of the selectivity model in advertising response.
Journal of Advertising, 42 (1), 41-56.

Dillard, J. P., & Peck, E. (2000). Affect and persuasion: Emotional responses to public
service announcements. Communication Research, 27, 461- 495.

Dillard, J. P., & Ye, S. (2008). The perceived effectiveness of persuasive messages:

Questions, structure, referent and bias. Journal of Health Communication, 13
(2), 149-168.

135

Eagly, A. H. (1978). Sex differences and influenceability. Psychological Bulletin, 85
(1), 86-116.

Edwards, R. W., Jumper-Thurman, P., Plested, B. A., Oetting, E. R., & Swanson, L.
(2000). Community readiness: research to practice, Journal of Community
Psychology, 28(3), 291-307.

Feilitzen, C., V., & Linné, O. (1975). The effects of television on children and
adolescents: Identifying with television characters. Journal of Communication,
25 (4), 51-55.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An
introduction to theory and research. Reading, MA: Addison-Wesley.

Fischer, E., & Arnold, S. J. (1997). Sex, gender identity, gender role attitudes, and
consumer behavior. Psychology and Marketing, 11 (2), 163-182.

Fisher, R. J ., & Dubé, L. (2005). Gender differences in response to emotional
advertising: A social desirability perspective. Journal of Consumer Research,
31(4), 850-858.

Frieze, I. H., & Browne, A. (1989). Violence in marriage. In Ohlin, L. E., & Tonry,
M. H. (eds.) Family Violence. Chicago, IL: University of Chicago Press.

Frijda, N. H., Kuipers, P., & ter Schure, E. (1989). Relations among emotion,

appraisals, and emotional action readiness. Journal of Personality and Social
Psychology, 57, 212-228.

Fleming, M. A., & Petty, R. E. (2000). Attitudes, Behavior and Social Context. Terry,
D.J., & Hogg, M. A (ed.). Routledge.

Flood, M. (2006). Changing men: Best practice in sexual violence education. Women
against Violence: An Australian Feminist Journal. 18, 26-36.

Foshee, V. A., Bauman, K. E., Ennett, S. T., Linder, F., Benefield, T., & Suchindran,
C. (2004). American Journal of Public health, 94(4), 619-624.

Freedman, S. A. (1992). Triggering the gender schema: A theoretical proposition.
Information Analyses, Educational Resources Information Center (070).

Freiden, J. B. (1884). Advertising spokesperson effects: An examination of endorser
type and gender on two audiences. Journal of Advertising Research, 24 (5), 33-
40.

Gelles, R. J. (1976). Abused Wives: Why do they stay? Journal of marriage and the
family. 38 (4), 659-668.

136

Gelles, R. J ., & Conte, J. R. (1990). Domestic violence and sexual abuse of children :
A review of research in the Eighties. Journal of Marriage and the Family, 52
(4), 1045-1058.

George, D., & Mallery, P. (2003). SPSS for Windows step by step: A simple guide and
reference. 11.0 update (4“ ed. ). Boston: Allyn & Bacon.

Gerlock, A. A. (1999). Health impact of domestic violence. Issues in mental health
nursing, 20, 373-385.

Greenwald, A. G. (1968). Cognitive learning, cognitive response to persuasion, and
attitude change. In A. G. Greenwald, T. C. Brock, and T. M. Ostrom (Eds),
Psychological foundations of attitudes (pp. 147-170). New York: Academic
Press.

Gracia, E., & Herrero, J. Public attitude toward reporting partner violence against
women and reporting behavior. Journal of Marriage and Family, 68 (3), 759-
768.

Grimley, D., Prochaska, J. 0., Velicer, W. F., Blais, L. M., & DiClemente, C. C.
(1994). The transtheoretical model of change. In Brinthaupt, T. M., & Lipka, R.
P. (Eds.), Changing the self: philosophies, techniques and experiences (201-227).
Albany, NY: State University of New York Press.

Harmon, R. R., & Coney, K. (1982). The persuasive effects of source credibility in
buy and lease situations. Journal of Marketing Research, 19 (2), 255-260.

Hall Smith, P., White, J. W., & Holland, L. J. (2000). A longitudinal perspective on
dating violence among adolescent and college-age women. American Journal of
Public Health, 93 (7), 1104 — 1109.

Heilman, M. E., & Stopeck, M. H. (1985). Attractiveness and corporate success:
Different causal attributions for males and females. Journal of Applied
Psychology, 70(2), 379-388.

Heise L. L., Ellsberg M., & Gottemoeller M. (1999). Ending violence against women.
Population Reports. Sseries L (11). Baltimore: Johns Hopkins University School
of Public Health, Center for Communications Programs.

Hoffner, C., Buchanan, M. (2005). Young adults‘ wishful identification with
television characters: The role of perceived similarity and character attributes.
Media Psychology, 7, 325-351.

137

 

Holbrook, M. B., & Batra, R. (1987). Assessing the role of emotions as mediators of

consumer responses to advertising. The Journal of Consumer Research, 14 (3),
404-420.

Holden, G. W., Geffner, R. A., & Jouriles, E. N. (1998). Children Exposed to Marital
Violence: Theory, Research, and Applied Issues. Washington DC: American
Psychological Association.

Hong, L. (2000). Towards a transformed approach to prevention: Breaking the link
between masculinity and violence. Journal of American College Health, 48,
269-279.

Hotaling, G. T., & Sugarman, D. B. (1986). An analysis of risk markers in husband
and wife violence: The cmrent state of knowledge. Violence and Victims, 1(2),
101-124.

Hovland, C. I., Janis, I. L., & Kelley, H. H (1953). Communication and persuasion.
New Haven: Yale University Press.

Huhmann, B. A., Brotherton, T. P. (1997). A content analysis of guilt appeals in
magazine advertisements. Journal of Advertising. 26(2), 35-46.

Huang, C. J ., Gunn, T. (2001). An examination of domestic violence in an African
American community in North Carolina: Causes and consequences. Journal of
Black Studies, 31 (6), 790-811.

Hunt, S. C., Martin, A. M. (2001). Pregnant women, violent men: What midwives
need to know. Elsevier Health Sciences.

Imber, J ., & Tofiler, B. A. (2000). Dictionary of Marketing Terms. Hauppaugue, NY:
Barron’s Educational Series.

Jaffe, P.G., Wolfe, D. A., & Wilson, S. K. (1990). Children of Battered Women.
Newbury Park: Sage.

Johnson, M. P. (2008). A Typology of Domestic Violence: Intimate Terrorism,

Violent Resistance, and Situational Couple Violence. Lebanon, University Press
of New England.

Johnson, M. P., & Ferraro, K. J. (2000). Research on domestic violence in the 19903:
Making distinctions. Journal of Marriage and Family, 62 (4), 948-963.

138

Joseph, W. B. (1982). The credibility of physically attractive communicators: A
review. Journal of Advertising, 11 (3), 15-24.

Kahle, L. R., & Homer, P. M. (1985). Physical attractiveness of the celebrity endorser:
A social adaptation perspective The Journal of Consumer Research 1 1(4), 954-
961.

Kamins, M. A. (1990). An investigation into the "match-up" hypothesis in celebrity
advertising: When beauty may be only skin deep Journal of Advertising, 19(1),
4-13.

Kelman, H. C. (1958). Compliance, identification and internalization: Three processes
of attitude change. Journal of Conflict Resolution,2, 51-60.

Kelman, H. C., & Eagly, A. H. (1965). Attitude toward the communicator, perception
of communication content, and attitude change. Journal of Personality and Social
Psychology, 1(1), 63-78.

Kenton, S. B. (1989). Speaker credibility in persuasive business communication: A
model which explains gender differences. The Journal of Business
Communication, 26 (2), 143-157.

Kim, J ., Allen, C. T., & Kardes, F. R. (1996). An investigation of the meditational
mechanisms underlying attitudinal conditioning. Journal of Marketing Research,
33(3), 318-328.

Kempf, D. S., & Palan, K. M. (2006). The effects of gender and argument strength on
the processing of word-of-mouth communication. Academy of Marketing
Studies Journal, 10(1), 1-18.

Komzweig, N. (1967). Behavior change as a function of fear arousal and personality,
unpublished doctoral dissertation, Yale University.

Kozloff, S. (1983). Humanizing “the voice of God”: Narration in “The Naked City”.
Cinema Journal, 23 (4), 41-53.

Lafrance, M., Banaji, M. (1992). Toward a reconciliation of the gender-emotion
relationship. In Clark, M. M. (Ed.), Emotion and social behavior (178-201).
Newbury Park: Sage.

139

 

Landy, D., and H. Sigall. (1974). Beauty is talent: Task evaluation as a function of the

performer's physical attractiveness. Journal of Personality and Social
Psychology, 29, 299-304.

Laner, M. R., & Benin, M. H. (2001). Bystander attitudes toward victims of violence:

who’s worth helping? Deviant Behavior: An Interdisciplinary Journal, 22, 23-
42.

Leech, N. L., Barrett, K. C., & Morgan, A. M. (2008). SPSS for intermediate
statistics: Use and interpretation. (3rd ed.). New York, NY: Taylor & Francis
Group.

Levendosky, A. A., & Graham-Bermann, S. A. (200]). Parenting in battered women:

The effects of domestic violence on women and their children. Journal of Family
Violence, 16 (2), 171-192.

Levine, I.S., Zimmernam, J, D. (1996). Using qualitative data to inform public policy:
Evaluating “Choose to De-F use”. American Journal of Orth0psychiatry. 66 (3),
363-3 77).

Locke, L. M., Richman, C. L. (1999). Attitudes toward domestic violence: Race and
gender issues. Sex Roles, 40 (3/4), 227-247.

Lori, D. W., Pradeep, K. (2003). Web advertising: Gender differences in beliefs,
attitudes and behavior. Internet Research, 13 (5), 375-385.

Lowry, D., T. (1973). Demographic similarity, attitudinal similarity and behavior
change. The Public Opinion Quarterly, 37 (2), 192-208.

Maccoby, E. E., & J acklin, C. N. (1974). The Psychology of Sex Differences. Stanford,
California: Stanford University Press.

Maccoby, E. E., Wilson, W. C. (1957). Identification and observational learning from
films. Journal of Abnormal and Social Psychology, 55 ( 1), 76-87.

Mackenzie, S. B., & Spreng, R. A. (1992). How does motivation moderate the impact
of central and peripheral processing on brand attitudes and intentions? The
Journal of Consumer Research, 18(4), 519-529.

Mackie, D., M., & Worth, L. T. (1990). Processing of persuasive in-group messages.
Journal of Personality and Social Psychology, 58 (5), 812-822.

140

 

Maio, G. R., & Olson, J. M. (2000). Why We Judge: Functions of Attitudes. Laurence
Erlbaum Associates: Mahwah, New Jersey.

Makepeace, J. M. (1981). Courtship violence among college students. Family
Relations, 30(1), 97-102.

Makepeace, J. M. (1983). Life events stress and courtship violence. Family Relations,
32(1), 101-109.

Markham, P. L. (1988). Gender and the Perceived Expertness of the Speaker as
Factors in ESL Listening Recall. TESOL Quarterly, 22 (3), 397-406

Mehrabian, A., & Epstein, N. (1972). A measure of emotional empathy. Journal of
Personality, 40, 525-543.

Milgram, S. (1974). Obedience to Authority: An Experimental View. New York:
Harper & Row.

Mills, J ., Jellison, J. M. (1968). Effects on opinion change of similarity between the
communicator and the audience he addressed. Journal of Personality and Social
Psychology, 9 (2), 153-156.

Miniard, P. W., Oberrniller, C., Page, T, J, Jr. (1983). A further assessment of
measurement influences on the intention-behavior relationship. Journal of
Marketing Research, 20 (2), 206-212.

Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator
of advertising effects on brand attitudes? Journal of Marketing Research, 18 (3),
318-332.

Mitchell, R. E., & Hodson, C. A. (1983). Coping with domestic violence: social
support and psychological health among battered women. American Journal of
Community Psychology, 11(6), 626-654.

Mitchell, M. M., Brown, K. M., Morris-Villagram, M., & Villagram, P. D. (2001).
The effects of anger, sadness, and happiness on persuasive message processing:

A test of the negative state relief model. Communication Monographs. 28 (4),
347-359.

Murry, J. P. Jr., Stam, A., Lastovicka, J. L. (1993). Evaluating an Anti-drinking and
driving advertising campaign with a sample survey and time series intervention
analysis. Journal of the American Statistical Association, 88 (421). 50-56.

141

National Coalition against Domestic Violence (2009). Domestic violence facts.
Retrieved fiom
http://wwwncadvorg/files/DomesticViolenceFactSheet%28National%29.pdf

Nabi, R. L. (2002). Anger, fear, uncertainty and attitudes: a test of the cognitive
functional model. Communication Monographs. Annandale: 69 (3 ), 204-217.

Newcombe, N., & Amkoff, D. B. (1979). Effects of speech style and sex of speaker
on person perception. Journal of Personality and Social Psychology, 37 (8), 1293-
1303.

Oakley, J. G. (2000). Gender-based barrier to senior management positions:
Understanding the scarcity of female CEOs. Journal of Business Ethics, 27 (4),
321-334.

Olson, L., Anctil, C., Fullerton, L., Brillman, J ., Arbuckle, J ., & Sklar, D. (1996).
Increasing emergency physician recognition of domestic violence. Analysis of
Emergency Medicine, 27(6), 741-746.

Pate, A. M., & Hamilton, E. E. (1992). F ormal and informal deterrents to domestic
violence: the Dade County Spouse Assault Experiment. American Sociological
Review, 57(5), 691-697.

Petty, R. E., Brinol, P., & Priester, J. P. (2009). Mass media attitude change:
Implications of elaboration likelihood model of persuasion. In Bryant, J ., &
Oliver, M. B. (Eds), Media Effects: Advances in Theory and Rsearch (125-164).
New York: Routledge.

Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and
peripheral routes to attitude change. New York: Springer-Verlag.

Petty, R. E., & Cacioppo, J. T. (1979). Issue involvement can increase or decrease
persuasion by enhancing message-relevant cognitive responses. Journal of
Personality and Social Psychology, 37 (10), 1915-1926.

Pirog-Good, M. A., & Stets, J. (Eds). ( 1989). Violence in dating relationships:
Emerging social issues. New York: Praeger.

Ramon, J. R., & Severance, L. J. (1970). Ego-involvement, discrepancy, source
credibility and attitude change. Journal of Personality and Social Psychology, 16
(2), 175-190.

Reeves, B., Miller, M. (1976). A multidirnentional measure of children’s
identification with television characters. Journal of Broadcasting, 22, 71-86.

142

Rogers, E. M., & Bhowmik, D. K. (1970). Homophily-heterophily: Relational
concepts for communication research. The Public opinion Quarterly, 34 (4),
523-538.

Rositter, J. R., & Bellman, S. (2005). Marketing communications: Theory and
applications. F renchs Forest, Australia: Pearson Education.

Runkel, P. J. (1956). Cognitive similarity in facilitating communication. Sociometry,
19 (3), 178-191.

Ryan, E. B., & Giles, H. (1982). Attitudes toward language variation: Social and
applied contexts. London: Arnold.

Saunders, D. G. (1986). When battered women use violence: Husband-abuse or self-
defense? Victims and Violence, 1(1), 47-60.

Shen, L., Dillard, J .P. (2007). The influence of behavioral inhibition/approach
systems and message framing on the processing of persuasive health messages.
Communication Research. 34 (4), 433-467.

Sheppard, B. H., Hartwick, J ., & Warshaw, P. (1988). The theory of reasoned action:
A meta-Analysis of past research with recommendations for modifications and
firture research. Journal of Consumer Research, 15 (3), 325-343.

Sherman, L. W., Smith, D. A., Schmidt, J. D., & Rogan, D. P. (1992). Crime,
punishment, and stake in conformity: Legal and informal control of domestic
violence. American Sociological Review, 57(5), 680-690.

Simons, H. W., Berkowitz, N. N., & Moyer, R. J. (1970). Similarity, credibility and
attitude change. Psychological Bulletin, 73 (1), 1-14.

Sistrunk, F., & McDavid, J. W. (1971). Sex variable in conforming behavior. Journal
of Personality and Social Psychology, 17 (2), 200-207.

Slater, M. D., Rouner, D. (2002). Entertainment education and elaboration likelihood:
Understanding the processing of narrative persuasion. Communication Theory.

12 (2), 173-191.

Smith, R. E. (1993). Integrating information fiom advertising and trial: Processes and
effects on consumer response to product information.

143

 

Smith, R. E & Swinyard, W. R. (1982). Information response models: An integrated
approach. The Journal of Marketing, 46 (1), 81-93.

Smith, T. W (1984). The Polls: Gender and attitudes toward violence. The Public
Opinion Quarterly, 48 (1), 384-396.

Sood, E. (2002). Audience Involvement and Entertainment Education.
Communication Theory. 12 (2), 153-172.

Stemthal, B., & Craig, C. S. (1974). Fear appeals: Revisited and revised. The Journal
of Consumer Research, 1 (3), 22-34.

Stemthal, B., Dholakia, R., & Leavitt, C. (1978). The persuasive effect of source
credibility: Tests of cognitive response. The Journal of Consumer Research, 4
(4), 252-260.

Stoller, R. J. (1964). A contribution to the study of gender identity. International
Journal of Psycho-analysis, 45, 220-226.

Stout, P. A & Rust, R. T. (1993). Emotional feelings and evaluative dimensions of
advertising: Are they related? Journal of Advertising, 22 (1), 61-71.

Struckman-Johnson, C. J ., Gilliland, R. C., Struckman—Johnson, D. L., & North, T. C.
(1990). The effects of fear of AIDS and gender on responses to fear-arousing

condom use advertisements. Journal of Applied Social Psychology, 20 (17),
1396-1410.

Telch, C. F., & Lindquist, C. U. (1984). Violent versus nonviolent couples : A
comparison of patterns. Psychotherapy, 21(2), 242-248.

Thoits, P. A. (1989). The sociology of emotions. Annual Review of Psychology, 15,
317-342.

Till, B. D., & Busler, M. (2000). The match-up hypothesis: Physical attractiveness,
expertise, and the role of fit on brand attitude, purchase intent and brand beliefs.
Journal of Advertising, 29(3 ), 1-13.

Tjaden, P., & Thoennes, N. (2000). Extent, nature and consequences of intimate
partner violence: Findings from the National Violence Against Women Survey.
Washington DC: Department of Justice (US). Retrieved fi'om
http://wwwncjrsgov/pdffiles1/nij/181867.pdf.

144

Treise, D., Wolburg, J. M., Otnes, C. C. (1999). Understanding the "Social Gifts" of
drinking rituals: An alternative framework for PSA developers. Journal of
Advertising, 28 (2). 17-31.

Triandis, H. C. (1960). Cognitive similarity and communication in dyad. Human
Relations, 13, 175-183.

Turner, J. C., Hogg, M. A.. (1987). Rediscovering the social group: a self-
categorization theory. Oxford & New York: Basil Blackwell.

Umberson, D., Anderson, K., Glick, J ., & Sharpio, A. (1998). Domestic violence,
personal control and gender. Journal of Marriage and Family, 60(2), 442-452.

US Department of Justice (2007). Intimate partner violence in the United States.
Retrieved at: http://www.oip.gov/bis/pub/pdf/ipvus.pdf.

US Department of Justice (1998). Violence by intimates: Analysis of data on crimes
by current or former spouses, boyfriends, and girlfiiends. Retrieved at:
http://www.oip.gov/bjs/Dub/pdf/invus.Ddf.

Waldner-Haugrud, L. K., & Gratch, L. V. (1997). Victimization and perpetration rates

of violence in gay and lesbian relationships: Gender issues explored. Violence
and Victims, 12 (2), 173-184.

Walker, L. (1984). The Battered Woman Syndrome. New York: Springer.

Ward, D. A., Seccombe, K., Bendel, R., Carter, L. F. (1985) Cross-sex context as a
factor in persuability sex differences. Social Psychology Quarterly. 48 (3), 269-
276.

Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions engender
behavior change? A meta—analysis of the experimental evidence. Psychological

Bulletin, 132 (2), 249-268.

White, J. W., & Koss, M. P. (1991). Courtship violence: incidence and prevalence in a
national sample of higher education students. Violence Victims, 6, 247-256.

Whitman, S. (1988). Ideology and violence prevention. Journal of the National
Medical Association 80(7), 73 7-743.

145

 

Whitman, S., & McKnight, J. L. (1988). Ideology and injury prevention. International
Journal of Health Servives, 15(1), 35-46.

Whittler, T. E. (1991). The effects of actors’ race in commercial advertising: Review
and extension. Journal of Advertising, 20(1), 59-60.

Wimmer, R. D., & Dominick, J. R. (2006). Mass Media research: An Introduction (8th
edition). Belmont, CA: Thomson Wadsworth.

Wilt, S., Olson, S. (1996). Prevalence of domestic violence in the United States.
Journal of the American Medical Women’s Association, 51 (3), 77-82.

Witte, K., Allen, M. (2000). A meta-analysis of fear appeals: Implications for health
communication campaigns. Health Education and Behavior, 27, 591-615.

Wolburg, J. M. (2001). The “risky business” of binge driving among college students:
Using risk models for PSA’s and anti-drinking campaigns. Journal of
Advertising, 30 (4), 23-39.

Wolfe, D. A., Jaffe, P. G. (1999, Winter). Emerging strategies in the prevention of

domestic violence. The Future of Children, 9, 3, Domestic violence and
children, 133-144.

Woodside, A. G., Davenport, J. W. Jr. (1974). The effect of salesman similarity and
expertise on consumer purchasing behavior.

World Health Organization. (2002). World health report on violence and health.
Retrieved from: http://whqlibdocwho.int/hq/2002/9241 54561 5.pdf.

Wright, P. (1973). The cognitive processes mediating acceptance of advertising.
Journal of Marketing Research, 10 (1), 53-62.

Wright, P. (1975). Factors effecting cognitive resistance to advertising. The Journal
of Consumer Research, 2 (1), 1-9.

Zillmann, D., Gibson, R., Sundar, S. S, & Perkins, W. Jr. (1996). Effects of

exemplification in news reports on the perception of social issues. Journalism
and Mass Communication Quarterly, 73 (2), 427-444.

146