THESSS

1- llllllllllllHllHllllHllllllllllllIll”lllllllllllllllllll

'9‘” 301820 0067

This is to certify that the

dissertation entitled

THE EFFECTS OF ANGER AND EMOTIONAL FOCUS
ON PERSUASIVE MESSAGE PROCESSING:
A TEST OF THE NEGATIVE STATE RELIEF MODEL

presented by

Monique M. Mitchell

has been accepted towards fulfillment
of the requirements for

Ph . D. degree inCommum' cati on

 

flwA/‘am

fiajor professor

Date May 6, 1999

MS U Ls an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

LIBRARY

Michlgan State
UnIversIty

 

 

 

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

 

09 éUBli

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/98 wCIRCJDatoOmpGS—p.“

THE EFFECTS OF ANGER AND EMOTIONAL FOCUS ON PERSUASIVE
MESSAGE PROCESSING: A TEST OF THE NEGATIVE STATE RELIEF
- MODEL
By

Monique Marie Mitchell

A Dissertation
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Communication

1 999

ABSTRACT
THE EFFECTS OF‘ANGER AND EMOTIONAL FOCUA ON PERSUASIVE
MESSAGE PROCESSING: A TEST OF THE NEGATIVE STATE RELIEF
MODEL
BY

Monique Marie Mitchell

The purpose of this study was to test the Negative State Relief Model’s
(NSR; Cialdini, Baumann, & Kenrick, 1981) ability to explain the research in
persuasion and emotion. An interesting finding in this body of literature is that
persons in a happy mood are more prone to attitude change after reading a
persuasive message, regardless of the quality of the persuasive message.
Persons in a negative mood, on the other hand, are less prone to attitude
change, and are more persuaded by strong persuasive arguments than by weak
persuasive arguments. One theory posited to explain these results is the NSR
model. One limitation of the NSR model, however, is that it discusses emotion
from a valence perspective. Taking a categorical perspective this study
proposes that under certain conditions negative emotions will behave like
positive emotions. One such condition is when individuals are focussing on their
emotion, rather than the task at hand. To this end, a 2 (strong persuasive
message, weak persuasive message) X 3 (control group, focus on emotion,
focus away from emotion) between subjects factorial design is proposed to test
hypotheses. It was predicted that the focus of individuals’ attention will directly
affect their level of distraction when reading a persuasive message. People who

are focused on their emotion will reveal higher levels of distraction. Distraction

will affect recipients’ recall of the persuasive message. Thus, if people are
focusing on their emotion, their level of distraction will be high, and recall will be
low. If people are focusing on the task, their level of distraction will be low and
recall will be high. Persons who are distracted and cannot recall much of a
persuasive message will have a lessened ability to counterargue a persuasive
message. People who cannot counterargue will exhibit more attitude change.
Those who are not distracted by the persuasive message and have better recall
of the message will be better able to counterargue. This effect, however, is
affected by the quality of the persuasive message. Message strength will have a
weak effect for those in the focus toward emotion condition, due to lack of
systematic processing. Message strength will have a strong effect for the
participants in the focus away from emotion condition because of systematic
processing. The data were not consistent with the hypotheses in the main.

Limitations are discussed.

This dissertation is dedicated to my parents, who taught me what hard work
really is.

ACKNOWLEDGEMENTS

This dissertation and the completion of my doctorate degree would not have
been possible without the love, support and encouragement of several people. I
would like to thank each of these people.

To my siblings, Michelle, Misty, and Billy: Thank you for always supporting my
endeavors, and for making my life so enjoyable. I am glad that we are family.

To Dad and Sue: I watch how hard you work. Thank you for giving me the work
ethic and strength it takes to finish graduate school.

To Mom: The love you have given me over the years is incredible. I know that
part of this degree belongs to you. Thank you for pushing me, reminding me
what is really important in life, and for being so proud.

To Marge: Thank you, thank you, thank you for making me stay at M.S.U.l!l Oh,
and for listening to me constantly, for helping me with all of my paperwork, and
for being you.

To Maria and Victoria: I will miss our slumber parties! Thanks for being great co-
authors, good friends, and great listeners.

To the “Gang” (Craig, Cynthia, Jenn, Jen): We had fun, didn’t we?

To Kimo: Thanks for making life an adventurellll I am glad to have a BFF like
you.

To my committee members Ron, Sandi and Stan: Thanks for serving on my
committee! More importantly, thank you for helping me mature as a scholar.
Thank you also for being my friends. My graduate career has been wonderful,
and I will miss it greatly, mainly, because of the friendship and guidance you
have offered.

To Frank: Five years and out! I can never thank you enough for being my
advisor. You have been a counselor, a teacher, a colleague, and most
importantly my true friend. I will take the things you have taught me with me for
the rest of my life. If I can someday be the kind Of advisor you were to me, then I
will have truly accomplished greatness. Thank you for teaching me n_ot what to
think, but to think.

 

Finally, to Stephen: I love you. Thank you for putting up with me through the
dissertation process, and for uprooting your life to move to Oklahoma so that we
could start our life.

TABLE OF'CONTENTS

LIST OF TABLES .............................................................. viii
LIST OF FIGURES ............................................................ viiii
INTRODUCTION ................................................................. 1
LIMITATIONS OF THE NEGATIVE
STATE RELIEF MODEL .............................................. 4
FOCUS TOWARD THE EMOTION ................................ 6
FOCUS AWAY FROM THE EMOTION ........................... 7
COGNITIVE RESPONSES .......................................... 8
METHOD ......................................................................... 11
PARTICIPANTS............................................' ............ 11
DESIGN .................................................................. 11
PROCEDURE .......................................................... 12
INSTRUMENTATION ................................................ 16
RESULTS ......................................................................... 20
DISCUSSION .................................................................... 29

vi

TABLE OF'CONTENTS

LIST OF TABLES .............................................................. viii
LIST OF FIGURES ............................................................ viiii
INTRODUCTION ................................................................. 1
LIMITATIONS OF THE NEGATIVE
STATE RELIEF MODEL .............................................. 4
FOCUS TOWARD THE EMOTION ................................ 6
FOCUS AWAY FROM THE EMOTION ........................... 7
COGNITIVE RESPONSES .......................................... 8
METHOD ......................................................................... 11
PARTICIPANTS............................................' ............ 11
DESIGN .................................................................. 11
PROCEDURE .......................................................... 12
INSTRUMENTATION ................................................ 16
RESULTS ......................................................................... 20
DISCUSSION .................................................................... 29

vi

LIST OF TABLES

TABLE ONE
ATTITUDE CHANGE BROKEN DOWN BY
EXPERIMENTAL CONDITION ................................................... 42

TABLE TWO

CORRELATIONS AMONG THE VARIABLES

FOR FOCUS AWAY FROM EMOTION

CONDITION ........................................................................... 43

TABLE THREE

CORRELATIONS AMONG THE VARIABLES

FOR FOCUS TOWARD EMOTION

CONDITION ............................................................................ 44

vii

LIST OF FIGURES

FIGURE ONE
PREDICTED PATH MODEL FOR THE FOCUS
AWAY FROM EMOTION CONDITION .................................................. 38

FIGURE TWO
PREDICTED PATH MODEL FOR THE FOCUS
TOWARD EMOTION CONDITION ....................................................... 39

FIGURE THREE
PATH MODEL FOR THE FOCUS AWAY
FROM EMOTION CONDITION WITH PATH COEFFICIENTS ................... 40

FIGURE FOUR

PATH MODEL FOR THE FOCUS TOWARD
EMOTION CONDITION WITH PATH COEFFICIENTS ............................. 41

viii

The Effects of Anger and Emotional Focus on Persuasive Message Processing:
A Test of the Negative State Relief Model
Introduction

The purpose of this study is to test the Negative State Relief Model’s
(NSR; Cialdini, Baumann, & Kenrick, 1973) ability to explain the research in
persuasion and emotion. An interesting finding in this body of literature is that
persons in a happy mood are more prone to attitude change after reading a
persuasive message, regardless of'the quality of the persuasive message.
Persons in a negative mood, on the other hand, are less prone to attitude
change, and are more persuaded by strong persuasive arguments than by weak
persuasive arguments. Scholars have posited that this effect exists because
persons in a positive mood lack the motivation and/or ability to process a
persuasive message systematically as proffered by the Elaboration Likelihood
Model (ELM; Petty & Cacioppo, 1986). Persons in a negative mood, on the
other hand, have the motivation and the ability to process persuasive messages
systematically because they desire to relieve themselves of the negative affect
they are experiencing (Cialdini, Baumann, & Kenrick, 1981) as predicted by the
NSR model. There is no literature, however, showing that these findings will be
replicated with negative emotions other than sadness, and if such findings are
dependent on the focus of the respondents’ attention rather than their emotional

state.

Therefore, this study is designed to test the applicability of the NSR model
as a satisfactory explanation for the impact of negative emotions on persuasive
message processing. It is proposed in this study that the valence of an emotion
is not what affects message processing, but that emotional focus is the key
variable. Specifically, this study offers a test of the impact of emotional focus on
message processing, and investigates the extent to which focus affects
distraction, which impacts the effect of message strength on attitude change. To
this end, a 2 (strong persuasive message, weak persuasive message) X 3
(control group, focus on emotion, focus away from emotion) between subjects
factorial design is proposed to test hypotheses.

The NSR model (Cialdini, Baumann, & Kenrick, 1981) has been posited.
as the theoretical underpinning for certain persuasive effects (e.g. see Boster,
Mitchell, Lapinski, Orrego, Cooper & Reinke, 1999) most notably the effects of
guilt on helping behavior. The NSR model is also useful in understanding the
effects of negative emotion on persuasive message processing. Negative
emotions hinder the persuasion process by making attitude and behavior change
more difficult (Bless, Mackie, & Schwarz, 1992; Bohner, Chaiken, & Hunyadi,
1994; Worth & Mackie, 1987). Moreover, participants in a negative mood are
more persuaded by strong arguments than by weak arguments, but the for
persons in a positive mood the effect of message strength on persuasion is trivial
(Bless, Mackie, & Schwarz, 1992; Bohner, Chaiken, & Hunyadi, 1994). Several

studies have produced this interaction effect. In fact, Brentar, Dillard, and Smith
2

(1997) conducted a meta-analysis examining 18 studies (n=3217) which
investigated the effect of mood valence on attitude change. Brentar, Dillard, and
Smith predicted that valence of mood should correlate positively with attitude
change, and their data were consistent with this hypothesis. The average
correlation was p.07, (95% c.i.=+.04 to +.10). Although this correlation is small,
its confidence interval does not contain zero, thus this correlation is statistically _
significant. Although scholars agree that this interaction effect is replicable, the
theoretical underpinning of the effect is arguable and perhaps unknown. One
theory which aids understanding this area is the NSR.

According to Cialdini’s NSR model, negative affect is accompanied by an
intrinsic drive to alleviate the negative feelings (Carlson & Miller, 1987). Persons
who have internalized pro-social standards may alleviate the negative affect by
helping someone in need (Berkowitz, 1987; Carlson & Miller, 1987; Cialdini,
Baumann, & Kenrick, 1981; Cialdini & Fultz, 1990). Individuals may also be
motivated to distract themselves from unpleasant thoughts, and engage in other
activities that divert their attention from the unpleasant affect (Isen & Simmonds,
1978), such as reading a persuasive message or engaging in an alternative task.
In fact, people may be likely to ignore any stimulus that prolongs the negative
affect. Although the NSR model is useful in understanding the relationship
between emotion and persuasive message processing, it is limited in scope.

These limitations will be discussed subsequently.

Limitations of the Negative State Relief Model -

The NSR model discusses negative states from a dimensional
perspective which assumes that emotion can be understood on a continuum
ranging from positive to negative. Dimensional models of affect purport that
emotional states of similar valence share common defining features and suggest
that understanding the underlying features of the dimension will lead to an
understanding of the emotion itself (Dillard & Smith, 1997). In fact, some
dimensional models treat emotions of similar valence as if they are on a bipolar
scale ranging from negative to positive (Russell, 1980). An alternative way to
conceptualize emotion is from a categorical perspective. The categorical
perspective emphasizes the unique qualities of distinct emotional states (e.g.
Ekman, 1992; Izard, 1977). The categorical perspective can gain credibility if
data indicate that negative and positive emotions can behave similarly under
certain conditions. Hence, if data can indicate that negative emotions can
impact persuasive message processing in the same way as positive emotions
do, when persons are focused toward their emotion, such data would also reveal
that a dimensional view of emotion is flawed.

Working from a valence perspective, Cialdini notes that what distinguishes
negative emotions from "positive emotions is the focus of the receiver’s thoughts.
Persons who experience negative affect will focus their thoughts away from the
emotion in order to stop experiencing the emotion. In lieu of concentrating on

the negative emotion, they will pay attention to an alternative task (such as
4

scrutinizing a persuasive message). When individuals concentrate on a
persuasive message, they are more likely to be persuaded by a strong
persuasive message, than by a weak persuasive message because it is more
likely that they are systematically processing the message. Persons
experiencing a positive emotion, on the other hand, will focus their attention on
the emotion in order to maintain the good feelings, thus systematic processing is
difficult if at all possible. For example, Worth and Mackie (1987) conducted a
study where they induced persons to be in a positive or neutral mood by having
them either win money in a fake lottery, or not having them win money.
Subsequently, participants were asked to read either a weak or a strong
persuasive message. For persons in the positive mood, the weak message
condition showed mean attitude change of 1.73. Their mean attitude change in
the strong message condition was 1.5, a meager difference. However, for
individuals in the neutral mood, the-weak message condition produced a mean
attitude change of .73, while the mean attitude change was 2.45 in the strong
message condition for those in a neutral mood, a large difference. Although the
NSR model helps us understand such patterns of means, Cialdini does not
address the possibility that persons experiencing any emotion can either focus
their attention toward the emotion or away from the emotion depending on other
characteristics of the emotion induction. Such is the position taken in this

investigation.

Focua Toward the Emotion

Certain emotions motivate us to concentrate on the feelings aroused.
One example of such an emotion is happiness. Research has indicated that
when persons are made to feel happy, the effect of message strength on
persuasion is trivial. Researchers argue that happiness has such an impact
because instead of concentrating on reading the persuasive message, the
persons concentrate on the emotion (e.g. Mackie & Worth, 1989). An alternative
hypothesis is that happy receivers pay attention to emotion related thoughts, not
only because they are happy, but because of other artifacts in the emotion
induction. Happiness is an emotion about which the recipient can do something.
The happy person can continue to think happy thoughts, which will maintain the
happy emotion. Other emotions besides happiness will lead to similar results if
they too include an element of personal efficacy. It is posited that if a negative
emotion induction also includes personal efficacy, then the receivers’ focus will
remain on the emotion, and they will not concentrate on other external stimuli
such as a persuasive message. Hence, the negative emotion would produce
similar message effects as happiness. Under these conditions it would be
expected that the impact of message strength on attitude change typically found
for persons in a negative mood would disappear. If this outcome were observed
it would raise serious questions about the broad application of the NSR model,
and it would introduce important moderators to our understanding of the role of

mood in persuasive message processing.
6

Focus Away From the Emotion.

Cialdini proffers that negative emotions are accompanied by an intrinsic
drive to reduce the negative emotions. Much of the research examining the
effects of negative emotion on persuasive message processing have induced
sadness. Sadness inductions are usually accomplished by asking the participant
to concentrate on the saddest event in their lives, or by having them watch a sad
movie clip. This research has consistently found that those feeling sad pay
attention to the task (i.e., reading the persuasive message), and hence, are more
persuaded by the strong persuasive message than by the weak persuasive
message. Though the affective valence experienced by the person is an
important motivator for seeking relief, what may cause persons to focus on the
task may be features of the emotion induction, rather than the specific emotion.
First, if characteristics of the induction create a situation in which there is nothing
recipients can do to alleviate the source of the negative emotion, they will be
more likely to pay attention to other stimuli to distract themselves from the
emotional state. That is, paying attention to the emotion will prove dysfunctional
in that it would perpetuate the negative feelings. Therefore, the participants will
be likely to pay attention to other tasks (e.g. reading a persuasive message)
because concentrating on the emotion is unproductive.

Perhaps the negative valence of the emotion itself is not what predicts
whether people will focus on issues related to the emotion or focus on a

persuasive message. It is proposed here that certain elements of the negative
7

emotion induction may better predict where the participants’ attentional focus will
be. If participants perceive that they canido something about, or to, the emotion
producing agent they will be more likely to focus on issues related to the emotion
itself (such as how to gain revenge). On the other hand, if participants perceive
that there is no action they can take to impact the emotion producing agent, they
will be more likely to focus on other tasks. The focus of the receiver’s attention
is important because it will help predict the path other cognitive responses, which
directly affect attitude change, will take. These cognitive responses will be
discussed next.
Cognitive Resgonses

The focus of the receivers’ attention will directly affect their cognitive
responses to the persuasive message. Past research has indicated several
variables that are important to the process of attitude change. This study will
investigate three variables that data indicate are important: distraction, recall,
and counterarguing. Scholars have long paid attention to the effects of
distraction on attitude polarity (for review, see Buller, 1986). The research on
distraction has yielded mixed results, however, meta-analytic data have indicated
that distraction facilitates attitude Change. The most common explanation for this
relationship is counterargument disruption (Burgoon, Cohen, Miller, &
Montgomery, 1978; Festinger & Maccoby, 1964). Counterargument disruption
begins with the assumption that when people are in a persuasive situation, they

attempt to counterargue the message. The more counterarguments persons
8

develop, the more likely they will be resistant to attitude change. When
distracted, however, it is difficult to form counterarguments given the difficulty
with comprehending and recalling the persuasive message. Hence, distraction
serves to facilitate attitude change via lessened ability to counterargue. In fact,
Festinger and Maccoby (1964) wrote:
The critical variable would be the extent to which the attention of
the person was distracted from the persuasive communication
while listening to it. If the attention of the listener were distracted
sufficiently to make it quite difficult for him to counterargue, but not
so much as to interfere with his hearing of the speech, this would
represent a maximally effective situation.
Buller’s data indicated that the correlation between distraction and
counterarguing is r = -.17, and the relationship between counterarguing and
attitude change is r = -.50. Given that the correlation between distraction and
attitude change is r = .03, a simple causal chain where distraction leads to
counterarguing, which is antecedent to attitude change, fits with minimal error.
The focus of individuals’ attention will directly affect their level of
distraction when reading a persuasive message. People who are focused on
their emotion will be distracted when reading to the persuasive message.
Distraction will affect recipients’ recall of the persuasive message. Thus, if
people are focusing on their emotion, their level of distraction will be high, and

recall will be low. If people are focusing on the task, their level of distraction will
9

be low and recall will be high; Persons who are distracted and cannot recall
much of a persuasive message will have a lessened ability to counterargue a
persuasive message. People who cannot counterargue will exhibit more attitude
Change. Those who were not distracted by the persuasive message and have
better recall of the message will be better able to counterargue. This effect,
however, is affected by the quality of the persuasive message. Message
strength will have a weak effect for those in the focus toward emotion condition,
clue to lack of systematic processing. Message strength will have a strong effect
for the participants in the focus away from emotion condition because of
systematic processing. The relationships among these variables are depicted in
Figures One and Two.

According to the NSR model people in a negative mood state will desire to
ameliorate their negative feelings, and therefore, distract themselves from
thoughts related to the emotion producing agent. As such, they will focus more
closely on. the task of reading the persuasive message, and hence, be more
persuaded by strong persuasive messages than by weak persuasive messages.
The present study challenges this prediction. It is proffered that the ability to
take action will moderate the impact of message strength on counterarguing and
subsequent attitude change. Using anger as an example of a negative emotion,
this study is designed to determine the NSR model’s ability to predict message

strength’s impact on counterarguing and attitude change under two sets of

IO

conditions: one in which action can be taken (focus toward emotion), and one in
which no action can be taken (focus away from emotion).

It is hypothesized that in the emotion focus condition participants will be
distracted and have low recall of the persuasive message. Low recall, in turn will
lower their ability to counterargue and the effect of message strength (via
counterarguing) on attitude change will be trivial. In the task focus condition,
participants will not be distracted by their emotion and will have better recall.
Thus, they will have a heightened ability to counterargue and will be more
persuaded by the strong persuasive message than by the weak persuasive
message (see Figures One and Two). Finally, it is also hypothesized that there
will be a positive, linear relationship between attitude change and post message
behavior.

Method
Participants

In exchange for course credit, 120 undergraduates enrolled in a
communication course at Michigan State University agreed to participate in the
experiment. Seventy-seven percent of the participants were female, and 73%
were Caucasian. The mean age was 21 years (§ = 2.20 years).

Design

A 2 (strong persuasive message, weak persuasive message) X 3 (control,

focus toward emotion, focus away from emotion) independent groups design

was employed. Participants were assigned randomly to experimental groups
11

yielding 20 participants in the control, weak message condition; 18 participants in
the control, strong message condition; 19 participants in the foCus away from
emotion, weak message condition; 20 participants in the focus away from
emotion, strong message condition; 19 participants in the focus toward emotion,
weak message condition; and 24 participants in the focus toward emotion, strong
message condition.
Procedure
One week before the experiment participants were asked to fill out a
questionnaire tapping into pre-message attitude toward wearing bicycle helmets.
The participants were then asked to Sign up for a time to enter the laboratory.
Participants were told they would be reading‘ a public service

announcement about wearing bicycle helmets, and would be asked to answer a
series of questions afterward. Next, all participants, except those in the control
group, were instructed, “Since I also work for Michigan State administration I was
asked if I could also get some reactions from you about another university
related topic. Michigan State will be changing its graduation requirements, and
will be officially releasing a statement regarding this late next week. Before the
memo is released the board of trustees would like me to get student reactions.”
In the focus away from emotion condition, participants then read the following:

Michigan State University’s Board of Trustees recently released a

memo to all units on campus stating that beginning the fall of 2000

all students will berequired to take a‘comprehensive exam prior to
12

graduation. This exam will test students’ general knowledge, and
knowledge specific to their major(s) similar to the Graduate
Entrance Exam (GRE). If students fail this exam they will be
required to take a capstone course (at the students’ cost) provided
by their college, which will prepare them to take the exam a
second, and last, time. Students will not be allowed to graduate if
they fail the exam a second time. It is important to note that MSU is
changing its graduation requirement, and no student is omitted by a
“grandfather clause.” That is, all seniors must take this exam
beginning September 2000. Reportedly, this is a move toward an
initiative that is being called “Academic MSU.” Although some
students are angry about this decision, the MSU administration has
already begun implementing this requirement. In fact, an unnamed
administrator was quoted stating “Michigan State students are an
embarrassment. They score lower than other students nation-wide
and are unprepared for the “real world;” we no longer can put up
with such non-academic attitudes.” Another MSU trustee stated,
“Students here (at MSU) don’t care about getting an education.
They are lazy. It’s about time we demand that they learn
something.” lnforrnation regarding this new requirement can be

retrieved off of the lntemet at Http:llwww.academicmsu.com. This

information will be released officially to students, communities, and
13

the press late next week. Preaident McPheraon and hi§ colleagues
have alreadv con§ulted aamall goup of atudentaand are no lo_ng_e_r_
accepting any feeMk. At this point there is nothing MSU
students can do about thLS deciaion.

Participants in the focus toward emotion condition read the same memo,
except the last paragraph was changed such that an action could be taken:

This information will be released officially to students, communities,
and the press late next week. Although President McPherson is
convinced the comprehensive exam is the answer to a slew of
academic problems at Michigan State, he is accepting feedback.
Students can reach him via e-mail. If you would like to reach
President McPherson and tell him how you feel about this decision
an anonymous e-mail system has been set up for you in the
communication department. You can reach MSU administrators
and tell them what you think of their system, after you finish the rest
of this survey.

After participants read the memo they were asked to fill out a
questionnaire which served as the first emotion induction check. Next, all
participants read either a strong or weak persuasive message about wearing
bicycle helmets. The strong message stated:

There are over 100 million people who ride a bicycle in America.

While bike riding is great fun, precautions must be taken, such as
14

wearing a bicycle helmet. It is known by experts that wearing a
bicycle helmet while riding is a must, and the fact is that only 10% of
adults do so, and only 2% of children wear a bicycle helmet. Each
year 50,000 bicyclists suffer serious head injuries due to not
wearing a helmet. In fact, biking is the most common cause of
recreational injury, accounting for almost 600,000 emergency room
visits each year. Specifically, in 1987, 561,765 people were treated
In emergency rooms for injuries suffered due to biking accidents.
Head trauma, which can be prevented by wearing a bicycle helmet,
results in approximately 85% of bicycle fatalities. In Eastern
Ontario, out of 880 head injuries, the largest group were due to
bicycle accidents.
Simply wearing a bicycle helmet can put an end to these staggering
statistics. Research has indicated that wearing a bicycle helmet can
reduce the risk of head iniugy by 85%. In addition, federal
researchers have stated that “one death every day, and one head
injury every four minutes could be prevented if every cyclist wore a
helmet.” Be smart, wear a bicycle helmet whenever you ride a bike.
The weak message stated:
There are many people who ride a bicycle in America. While bike
riding is great fun, precautions must be taken, such as wearing a

bicycle helmet. While it is known by locals that wearing a bicycle
15

helmet while riding may be necessary, the fact is that not many
people wear a bicycle helmet. Each year some bicyclists suffer
mild head injuries, though the cause of the head injuries is still not
clear. Biking has been cited to be a common cause of recreational
injury, accounting for a few emergency room visits each year. A
couple years ago, many people were treated in the hospital for
injuries suffered due to biking accidents. Head trauma results in a
few of bicycle fatalities. Some scholars believe that wearing a bike
helmet might reduce these injuries, but the evidence is unclear. In
Eastern Ontario, out of hundreds of head injuries, a large group
were due to bicycle accidents. Wearing a bicycle helmet might put
an end to these injuries. Research has indicated that wearing a
bicycle helmet might reduce the risk of head injury by 3%. The fact
is, wearing a bicycle helmet can’t hurt you and it is fashionable. Be

smart, wear a bicycle helmet whenever you ride a bike.

Participants were then asked to complete packet of questions which

included all dependent measures. These measures will be discussed

subsequently.

Instrumentation

Emgtion In;l_uc_:tion Check. The extent to which the participants

experienced anger was measured via four items asking them to indicate how

angry, mad, enraged or furious they (1 = none of this emotion, 7 = a great deal of

16

this feeling) felt. Participants were also asked to indicate how sad, happy and
fearful they felt in order to assess participants experience of anger, in lieu of
other emotions. A pilot test indicated that the experimental induction did induce
anger. Mean anger scores were M=15.95 (§=14.85). Scores could range from
four to 28, hence it was concluded that the participants experienced an average
amount of anger as a result of this induction.

Persuasive Message Strength Induction Check. Participants were also
asked to fill out five short Likert type items ranging from one to seven (1 = not
persuasive, 7 = very persuasive) as an induction check for message strength.
Items indicated how compelling, persuasive, dumb, weak, and convincing the
message was. Negatively worded Items were reflected such that higher scores
indicated more perceived message strength. A pilot test indicated that the
strong message was perceived to be more persuasive (M=18.53, §=3.86) than
the weak persuasive message (M=17.66, §=3.16). Moreover, a 2 (toward
emotion, away from emotion) X 2 (strong message, weak message) analysis of
variance (ANOVA) produced a substantial main effect for message quality in this
pilot test (F (1,70)=7.33, p<.01, eta2=.10, E32). hence, this induction was
concluded to be successful.

Attitude Change. Attitude change was measured via five Likert type
questions tapping into the individuals’ attitude about wearing bicycle helmets
(1=wearing a bicycle helmet is a bad idea, 7=wearing a bicycle helmet is a good

idea). Example items are, “I think wearing a bicycle helmet is important” and
17

“Wearing a bicycle helmet is a smart choice.” Attitude change was assessed by
calculating the difference between the participant’s post-message attitude and
their pre-message attitude.

[Legal Recall was measured by counting the amount of persuasive
arguments the participant correctly recalled. Both the weak and strong
persuasive messages included 14 persuasive arguments. A zero to two rating
was given for each of the 14 indicators. Zero indicated the persuasive message
was not mentioned, one indicated that the gist of the message was recalled, and
a two indicated that the message was recalled exactly. Two trained coders
coded 10 percent of the thought listing tasks. Their coding responses were
correlated and used to calculated inter-coder reliability, yielding an alpha of .94.

Counterarguing. Participants were asked to think back about their attitude
toward bicycle helmets. Next, they were given the instruction, “Think about what
this message was advocating. Now think of every point you can that is counter
to the message. What counter points or counter arguments can you come up
with?” Each counterargument was counted.

Distraction. Participants were also asked to list every thought they
had while reading the bicycle helmet message. This thought listing task was
coded for emotional thoughts (thoughts indicating how angry they were about the
comprehensive exams, or any other explicit indication of their anger) and number
of thoughts which were off of the bicycle helmet topic. Emotional thoughts and

off topic thoughts were used as indicators of distraction. A pilot test indicated
‘ l8

that participants in the focus toward emotion condition had more emotional
thoughts (_M=.33, _s_=.63) than did participants in the focus away from emotion
condition (M=.11, _s_=.31). A 2 X 2 ANOVA also indicated a main effect for focus
on emotional thoughts (F (1, 72)=3.94, p=.05). There were no substantial effects
for off topic thoughts. Two trained coders coded 10 percent of the thought listing
tasks. Their coding responses were correlated and used to calCulated inter-
coder reliability, yielding a reliability of .98 for off topic thoughts and 1.00 for
emotional thoughts. These measures were correlated to assess whether they
were alternate indicators of distraction, and could be summed to form an index.
These indicators were not significantly correlated (_r=-.10, p>.05), and therefore
were not considered alternate indicators. Moreover, in the main study, a 2 X 3
ANOVA was also performed on the dependent variable, emotional thoughts, as a
measure of distraction. There was a substantial main effect for focus (F(2,114)=
3.02, p=.05, eta2 =.05, _r=.22) indicating that individuals in the experimental
groups had more emotional thoughts (focus toward emotion: M=.33, a: .80,
focus away from emotion: M=.31, §=.66) than the participants in the control
group (M=0.00, §=0.00). The experimental groups did not have different means,
however, as predicted. This main effect was examined with caution, however,
because there was also a significant message by focus interaction (F(2,114)=
3.27, p<.05, eta2=.05, 5.22) indicating that for participants in the focus toward
emotion condition strong messages elicited more emotional thoughts than did

weak messages. For participants in the focus away from emotion condition,
19

weak messages elicited more emotional thoughts than strong messages did.
Because the dependent variable, emotional thoughts, was not effective it was
not used as a dependent variable. Henceforth, distraction will be referring to
number of off topic thoughts.

Behavior. Two weeks after the laboratory session, participants were
called on the telephone and asked about their bicycle helmets wearing behavior.
Behavior was measured using four Likert type items ranging from one to seven
(seven indicating mOre of the recommended behavior). Example items are, “I
have priced bicycle helmets” and “l have worn a bicycle helmet since my
participation in the study.”

Results

Confirmatory factor analysis procedures were performed on the scales
yielding the expected seven factor solution for anger, fear, sadness, happiness,
attitude change, message strength, and recall. Analyses revealed that the
scales formed unidimensional solutions.

First, the emotion induction was assessed for effectiveness. The four
items measuring anger were summed to form an index. This index was skewed
positively with an overall mean of 14.96 and a standard deviation of 8.96. The
reliability of this index was estimated by standardized item alpha and was found
to be .98. A 2 X 3 message type by focus ANOVA performed on these data
produced a substantial main effect for focus (F(2,1 13)=15.82, p<.01, etaz =.22,

_r=.47) with no other substantial effects emerging. Inspection of the means
20

indicated that the participants receiving the emotion induction (i.e. both
experimental groups) experienced more anger (focus away from emotion:
M=15.92, _s_=8.95; focus toward emotion: M= 14.00, §=8.86) than the control
group experienced (M=6.61, §=4.08). Because research has indicated
individuals may experience more than one emotion at a time, the participants’
level of fear, sadness, and happiness was also assessed.

Four items measuring fear were summed to form an index. This index
was skewed positively with a mean of 10.90 and a standard deviation of 7.98.
The reliability of this scale was estimated by standardized item alpha, and was
found to be .98. A 2 X 3 ANOVA performed on these data produced a
substantial main effect for focus (F(2,1 13)=10.35, p<.01, eta2=.16, p.40) with no
other substantial effects emerging. Inspection of the means indicated that the
participants in the experimental groups experienced more fear (focus away from
emotion: M=13.34, §=8.27, focus toward emotion: I_VI_=12.69, §=8.80) than the
control group experienced (Mr-6.52, §=4.13). Four items measuring sadness
were summed to form an index. This index had a mean of 11.00 and a standard
deviation of 6.04. The reliability of this scale as estimated by standardized item
alpha was .90. A 2 X 3 ANOVA performed on these data produced a main effect
for focus (F (2,1 13)= 4.9, p<.01, eta2=.08, [=.28) with no other substantial effects
emerging. Inspection of the means indicated that participants in the
experimental conditions experienced more sadness (focus away from emotion:

M=12.31, §=5.80; focus toward emotion: _Il_ll=12.04, §=6.83) than the participants
21

in the control group experienced (M=8.54, §= 4.40). The four happiness items
were summed to form an index. The mean of this index was 16.00 with a
standard deviation of 6.82. The reliability of this index, as estimated by
standardized item alpha was .95. A 2 X 3 ANOVA performed on these data
indicated a main effect for focus (F(2,112)= 10.46, p<.01, eta2 =.16, _r=.40) with
no other significant or substantial effects emerging. Inspection of the means
indicated that participants in the control group experienced more happiness
(M_=9.92, §=2.59) than did the participants in the experimental groups (focus
away from emotion: M=6.79, §=3.70; focus toward emotion: _M=8.26, _s_=3.44).

Three items were summed to form the message strength induction check.
This scale was skewed negatively with a mean of 13.83 and a standard deviation
of 4.30. The reliability of this index was estimated by standardized item alpha
and was found to be .85. A 2 X 3 focus by message type ANOVA performed on
these data produced a substantial main effect for message type (F( 1, 114)=
23.24, p<.01, eta2 =.17, [=.41) with no other substantial or significant effects
emerging. Inspection of the means indicated that participants receiving the
strong persuasive message reported it being more persuasive (_M=15.53, §=3.31)
than those receiving the weak persuasive message (M=12.02, §=4.52).

The distribution of responses to the distraction measure (i.e., number of
Off topic thoughts) was skewed positively with a mean of 1.5 and a standard
deviation of 2.04. A 2 X 3 message type by focus ANOVA performed on these

data produced no substantial effects. The means, however, did indicate that
22

participants in the focus toward emotion condition had more off topic thoughts
(M=1.77, §=2.6) than did the participants in the focus away from emotion
condition (_M=1.11, §=1.6) though the difference was trivial, and statistically
insignificant.

The 14 recall items were summed to form an index. This index was
skewed positively with a mean of 4.49 and a standard deviation of 3.01. The
reliability of this scale was estimated by standardized item alpha and was found
to be .36.

The number of counter arguments each participant developed was
summed to form an index. This index was skewed positively with a mean of 2.76
and a standard deviation of 1.60.

The four attitude change items were summed to form an index. This
index was skewed positively with a mean of .73 and a standard deviation of 3.77.
The reliability of this index was estimated by standardized item alpha and was
found to be .61.

The first hypothesis proposed that emotional focus would impact
distraction such that those participants in the focus toward emotion condition
would be more distracted than either the participants in the focus away from
emotion condition and the control group. As previously mentioned, a 2 X 3 focus
by message type ANOVA performed on these data produced no substantial
effects. The means, however, did indicate that participants in the focus toward

emotion condition had more off topic thoughts (I\_ll=1.77, §=2.6) than did the
23

participants in the focus away from emotion condition (M=1.11, §=1.6) but the
difference was trivial and statistically insignificant.

The second hypothesis predicted that focus would impact recall of the
persuasive message. Specifically, participants in the focus away from emotion
condition were predicted to recall more of the persuasive message than
participants in the focus toward emotion condition given that they were proposed
to be less distracted. This hypothesis was not consistent with these data. A 2 X
3 ANOVA performed on these data did not produce a main effect for focus, or a
substantial interaction effect. A main effect for message type, however, was
found (F(1,98)= 11.29, p<.01, eta2=.10, p.32). Inspection of the means
indicated that participants were better able to recall the strong persuasive
message (M=5.42, §=3.50) than the weak persuasive message (_M=3.43, §=1.96)
regardless of emotional focus.

The third hypothesis predicted that focus would impact counterarguing. It
was proposed that participants in the focus toward emotion condition would be
less able to counterargue the persuasive message given their level of predicted
distraction. Moreover, it was predicted that it would be easier to counterargue a
weak persuasive message than a strong persuasive message. A 2 X 3 ANOVA
performed on these data were not consistent with this hypothesis. No
substantial main effects or interaction effects emerged. Participants were

equally able to counterargue in all conditions.

24

The fourth hypothesis predicted a message by focus interaction for
attitude change. It was predicted that for participants in the focus toward
emotion condition, message type would have a trivial impact on attitude change.
For participants in the focus away from emotion condition, however, message
type was predicted to affect attitude change substantially. A 2 X 3 ANOVA
performed on these data was not consistent with this hypothesis. First, a main
effect for message type was found (F(1,114)= 7.53, p<.01, eta2=.06, [=.24).
Inspection of the means indicated that more change was found when
participants read the strong persuasive message. These means are presented
in Table One. There was also a main effect for focus, although it did not reach
conventional levels of statistical significance (F (2,1 14)= 2.61, p=.07, eta’=.04,
[=.20). The means indicated that persons in the focus toward emotion condition
had less attitude change than did participants in the focus away from emotion
condition, and participants in the control group (see Table One). Moreover, to
provide a more rigorous test a contrast analysis was also employed on the
attitude change data. The contrast —1 was applied to the weak message, focus
toward emotion condition, 0 was applied to the weak message, focus away from
emotion condition, 0 was applied to the strong message, focus toward emotion
condition, and 1 was applied to the strong message, focus away from emotion

condition. No substantial or significant effects emerged.

25

The final hypothesis predicted a positive, linear relationship between
attitude change and post message behavior. These data were inconsistent with
this hypothesis (_r=-.06, p>.05).

F urtherrnore, this study predicted specific causal relations among these
variables (see Figures One and Two). Hence, these path models were tested
using ordinary least squares to estimate parameters. In order to test a path
analytic model the correlations among the predicted variables must be
examined. The correlations for these models are presented in Tables Two and
Three. The first model predicted that for the focus away from emotion condition,
recall and perceived message strength would be causally antecedent to
counterarguing, which would predict attitude change. The path coefficients
generated by this model are presented in Figure Three. In order to test the fit of
this model several analyses were employed. First, the size of the path
coefficients was considered. The link between recall and counterarguing was -
.05, a trivial link. This small link indicates lack of fit. Next, confidence intervals
were drawn around each of the links. If the confidence interval includes zero,
then the link is considered trivial. The confidence interval for the link between
recall and counterarguing was (95%=-.58 to .48), the interval includes zero. The
confidence interval for the link between perceived message strength and
counterarguing was (95%=-.13 to .53). Finally, the confidence interval for the
link between counterarguing and attitude change was (95%=-.59 to .19). Each

interval includes zero, hence this model is not consistent with these data.
26

Finally, the errors generated by this model were considered. Errors are
calculated by the difference between the obtained and predicted correlations for
the unspecified links. Large errors indicate lack of fit of a path model. The error
generated for the link between recall and attitude change was -.22, and the error
for the line between message strength and attitude change link was .37. These
errors appear to be large, however, the chi-square analysis indicates that they
are within sampling error of zero (x2=1.62 (2), p>.05). These results indicate that
important paths are weak, in the incorrect direction, and errors generated by this
model are large. Thus, these data indicate that it is very unlikely that the causal
mechanisms predicted in this study actually predict attitude change.

The second model predicted that for the focus toward emotion condition,
distraction would be causally antecedent to recall which would be causally
antecedent to counterarguing, which would predict attitude change directly.
Perceived message strength was also predicted to be antecedent to
counterarguing. The main difference between this model and the previous
model is the perceived message strength is proposed to have a weak link to
counterarguing. The path coefficients generated by this model are presented in
Figure Four. In order to test the fit of this model, several analyses were
employed. First, the size of the path coefficients was considered. The smallest
link in this model is the link between counterarguing and attitude change (-.18).
The size of this path coefficient is considered to be small which indicates lack of

fit. Next, confidence intervals were drawn around each of the links. The
27 '

confidence interval for the link between distraction and recall was (95%=-.06 to
.92). The confidence interval for the link between recall and counterarguing was
(95%=—.22 to .92). The confidence interval for the link between counterarguing
and attitude change was (95%=-.57 to .21). Finally, the confidence interval for
the link between perceived message strength and counterarguing was (95%=-
.66 to .20). Each interval includes zero, hence this model is not consistent with
these data. The errors generated by this model were also considered. The error
generated for the link between memory and perceived message strength was
.21, the error for the link between perceived message strength and attitude
change was .32, the error for distraction and counterarguing was .05, the error
for distraction and attitude change was .02, and the error for memory and
attitude change was .85. Most of these errors are large, however, the chi-square
analysis indicates that they are within sampling error of 'zero (x2=4.48 (5), p>.05).
Again, important paths are weak and in the wrong direction and errors generated
by this model are large. Distraction was positively related to recall such that the
more off topic thoughts individuals had, the more they recalled about the
persuasive message. It was predicted that distraction (i.e. off topic thoughts)
would decrease recall. Moreover, these results indicate that the more
participants recalled the more they counterargued (regardless of message
strength). These results indicate that the causal mechanisms hypothesized are

not consistent with these data.

28

Discussion

The purpose of this study was to challenge the NSR model. It was
proposed that under conditions of distraction, such as when individuals are
focussed on their emotional state, people will not be able to distinguish weak and
strong persuasive messages, hence they will be equally persuaded by both.
Because research has indicated a trivial message strength effect with persons
who are happy, by obtaining data'which showed this can also be the case with
persons who are angry (when attention is focussed toward their emotion), the
NSR model would be challenged. Moreover, this study predicted specific causal
relations among variables thought to affect persuasion, such as recall and
counterarguing. Specifically, it was hypothesized that in the focus toward
emotion condition participants would be distracted and have low recall of the
persuasive message. Low recall, in turn, would lower their ability to
counterargue, and the effect of message strength on attitude change would be
trivial. In the focus away from emotion condition participants would not be
distracted and have more recall. Thus, they would have a heightened ability to
counterargue and would be more persuaded by the strong persuasive message
than by the weak persuasive message. These data were not consistent with
these hypotheses. In fact, there is no substantial evidence that emotional focus
affects any of the specified dependent variables. Moreover, the specified causal

models obtained egregious errors, and did not fit.

29

One may conclude that the NSR model was consistent with these data. In
fact, these data indicate. similar patterns of means for the dependent variable
attitude change, as was found in past studies examining negative emotion. As
mentioned previously, past research has indicated that individuals in a negative
mood are able to distinguish between strong and weak persuasive messages,
and are more persuaded by strong messages. The participants in this study
exhibited more attitude change after reading a strong persuasive message about
bicycle helmets regardless of emotional focus. This study purported that this
finding would be moderated by how actionable the angering message was
(hence causing distraction), but these data are not consistent with such an
assertion. One noteworthy flaw, however, is that there was no induction check
for actionability. Hence, although the students were told that they could e-mail
the president of the university, they might not have felt any sense of personal
efficacy or revenge. Therefore, they might not have been sufficiently distracted.
If the participants were not sufficiently distracted, then the hypotheses proffered
by this study might be worthy of future study, but with a more effective emotional
focus induction.

Moreover, there are other possible reasons why the hypotheses were
inconsistent with these data. Several factors may have been influential in this
process. First, there was some measurement error in this study. Specifically,
the recall measure was particularly unreliable (a=.36) although it did produce a

large effect. It is known that error of measurement can flatten slopes, and hence
3o

attenuate correlations. Given that recall is the key endogenous variable in the
path model for the focus away from emotion condition, it is not surprising that the
model failed. Such lack of reliability could create unpredictable relationships with
key variables in the path model.

The distraction measure also caused problems. Individuals had as many
off topic thoughts in the focus away from emotion condition as they did in focus .
toward emotion conditions. The counterarguing measure also posed serious
questions. It is particularly intriguing that although the participants were able to
distinguish a weak persuasive message from a strong persuasive message, and
rate the strong message higher in persuasiveness, they were able to
counterargue both messages equally well. One would posit that it should be
more difficult to counterargue a strong persuasive message than a weak
persuasive message. Perhaps the participants in this study were developing
counterarguments that neither message addressed. In fact, upon reading the
counterarguments, several participants developed counterarguments such as
“Bicycle helmets are ugly”, and “I would look like a dork if I wore a bike helmet”.
Because neither message address fully how fashionable bike helmets are, the
fact that one message was of higher quality than the other would not stop
participants from developing such counterarguments. Future research would be
well served to code counterarguments and investigate differences in attitude

change dependent on the types of counterarguments one develops.

31

Although the hypotheses posed by this study were not consistent with
these data several lessons were learned that future research should address.
First, whether or not emotional focus affects persuasive message processing is
still an empirical question that needs to be investigated. Second, the impact that
negative emotion has on persuasive message processing is an interesting
substantive area worthy of investigation regardless of focus. Research is still
Short on examinations of negative emotional states other than sadness.
Scholars, like Cialdini, assume that negative states will behave similarly, but this
assertion is largely unproven. Future research should examine message
irrelevant fear, guilt, anger, and distress among others. If such emotional states
do behave similarly then the valence perspective will gain considerable
credibility. On the other hand, if future data indicate differences in attitude
change, recall, and counterarguing among similarly valenced negative emotions,
the categorical perspective will gain more credibility. This study did indicate that
anger had an effect on message processing, however, emotional focus did not.
It would be an important addition to the body of literature on this topic to replicate
this study, and compare anger to happiness.

Notes:

1This study was replicated with an audio taped message. The results

were equivalent, in the main, to the results reported here.

32

REFERENCES

33

References

Berkowitz, L. (1987). Mood, self-awareness, and willingness to help.
Journal of Peraonality and Social Paychology, Q, 721-729.

Bless, H., Mackie, D. M., & Schwarz, N. (1992). Mood effects on attitude
judgments: Independent effects of mood before and after message elaboration.

Journal of Personality and Social Psychology, fl, 585-595.

 

Bohner, G., Chaiken, S., & Hunyadi, P. (1994). The role of mood and
message ambiguity in the interplay of heuristic and systematic processing.
Europeaa Journal of Social Psychology, 25, 207-221.

Boster, F. J., Mitchell, M. M., Lapinski, M. K., Orrego, V. 0., Cooper, H., &
Reinke, R. (in press). The impact of guilt and type of compliance-gaining
message on compliance. Communication Mariam

Brentar, J. E., Dillard, J. P., & Smith, B. A. (May, 1997). Message-
irrelevant affect and persuasion: A meta-analysis. Paper presented at the
annual meeting of the lntemational Communication Association, Montreal.

Buller, D. B. (1986). Distraction during persuasive communication: A
meta-analytic review. Communication Monographs, 5_3, 91-114.

Burgoon, M., Cohen, M., Miller, M. D., & Montgomery, C. L. (1978). An
empirical test of a model Of resistance to persuasion. Human Communication

Research 5, 27-39.

 

34

Carlson, M., & Miller, N. (1987). Explanations of the relation between
negative mood and helping. Psychological Bulletin. 102. 91-108.

Cialdini, R. B., Baumann, D. J., & Kenrick, D. T. (1981). Altruism as
hedonism: Helping and self-gratification as equivalent responses. Journal of

Personality flld Social Psychology, 4_0, 1039-1046.

 

Cialdini, R. B., & Fultz, J. (1990). Interpreting the negative mood-helping
behavior literature via “mega”-analysis: A contrary view. Psychological Bulletin,
191, 210-214.

Dillard, J. P., & Smith, B. A. (1997). Affect and persuasion: Evidence for
cognitive coloration of message processing. Paper presented at the National
Communication Association’s annual conference, Chicago, Ill.

Ekman, P. (1992). An argument for basic emotions. Cognition and
Emotions, 6, 169-200.

Festinger, L., & Maccoby, N. (1964). On resistance to persuasive
communication. Journal of Abnormal and Social Psycholpgy, §_8_, 359-366.

lsen, A. M., & Simmonds, S. (1978). The effect of feeling good on a task
that is incompatible with good mood. Social Psychology Quarterly, 51, 346-349.

Izard, C. E. (1977). Human Emotions. New York: Plenum Press.

Mackie, D. M., & Worth, L. T. (1989). Processing deficits and the

mediation of positive affect in persuasion. Journal of Personalig and Social

Psychology, fl, 27-40.

35

Petty, R. E., & Cacioppo, J. T. (1986). Communication and Per§ya_siar_1;
Central and Peripheral Routea to Attitade Chaaga New York: Springer.

Russell, J. A. (1980). A circumplex model of affect. Jourfll of Peracmlity
and Social stcholagy, _3_9_, 1161-1178.

Worth, L. T., & Mackie, D. M. (1987). Cognitive mediation of positive affect

in persuasion. Social Cognitiona a, 76-94.

36

FIGURES AND TABLES

37

Figure One

Predicted path model for the focus away from emotion condition.

+ —— ATTITUDE
RECALL ———> COUNTERARGUTNG —> CHANGE

PERCEIVED
MESSAGE
STRENGTH

38

Figure Two

Predicted path model for the focus toward emotion condition.

+

_ ATTITUDE
DISTRACTION ___+RECALL—-> COUNTER —> CHANGE

ARGUING

PERCEIVED
MESSAGE
STRENGTH

 

..................... -> Denotes a weak link

39

Figure Three

Path model for the focus away from emotion condition with path coefficients.

-.05
-.20
RECALL —>COUNTER —> ACIIENI UGDEE
ARGUING
PERCEIVED , .20
MESSAGE

STRENGTH

4O

Figure Four

Path model for the focus toward emotion condition with path coefficients.

’43 '35 COUNTER- "‘8 ATTITUDE

DISTRACTION ———>RECALL—>ARGUING —> CHANGE

-.23

PERCEIVED
MESSAGE
STRENGTH

41

Table One

Attitude change broken down by experimental condition.

CONTROL FOCUS ‘ FOCUS
TOWARD AWAY FROM

EMOTION EMOTION
WEAK STRONG WEAK STRONG WEAK STRONG
M .15 1.00 -1.26 .96 .47 2.90

S 3.07 3.74 3.62 3.30 4.75 3.28

42

Table Two

Correlations among the variables for the focus away from emotion condition.
Uncorrected correlations are found in the bottom triangle, correlations corrected
for attenuation due to error of measurement are in the top triangle,

communalities are in the diagonal.

RECALL COUNTER- ATTITUDE MESSAGE

ARGUING CHANGE STRENGTH

RECALL .36 -.03 -.21 .07
COUNTERARGUING -.02 1.00 -.20 .19
ATTITUDE CHANGE -.1o -.16 .61 .33
MESSAGE .04 .16 .24 .85

STRENGTH

43

Table Three

Correlations among the variables for the focus toward emotion condition.
Uncorrected correlations are found in the bottom triangle, correlations corrected
for attenuation due to error of measurement are in the top triangle, reliabilities

are in the diagonal.

RECALL COUNTER ATTITUDE PERCEIVED DISTRACTION

-ARGU|NG CHANGE MESSAGE

STRENGTH
RECALL .36 .28 .78 .29 .43
COUNTER- .17 1.00 -.18 -.15 .16
ARGUING
ATTITUDE .37 -.14 .61 .36 .01
CHANGE
MESSAGE .16 -.12 .26 .85 .18
STRENGTH
DISTRACT. .26 .16 .O1 .17 1.00

"IIIIIIIIIII11111111111“