 

X0”)

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

THE EFFECT OF MULTIPLE NORMATIVE MESSAGES ON
PERSUASION

presented by

KATHERINE ANN KLEIN

has been accepted towards fulﬁllment
of the requirements for the

degree In Communication

 

 

Ph/M/ M

Major Professor's Signature
/ 3 7d (1 4007

 

Date

MSU is an afﬁnnative-action, equal-opportunity employer

on--.----.---.-.nn-n-n-u-n-n-a-v--l-.----o--n-n----o-cqncn-u-n-o-n-n-o-n-.-'-s-n-o-.--n--u-a-a-.-.--o-.--o-u-

' J

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

 

 

iii 652813
950510

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DateDue.indd-p.1

THE EFFECT OF MULTIPLE NORMATIVE MESSAGES ON PERSUASION
By

Katherine Ann Klein

A DISSERATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Communication

2007

ABSTRACT
THE EFFECT OF MULTIPLE NORMATIVE MESSAGES ON PERSUASION
By
Katherine Ann Klein
Social norms campaigns have become a popular approach to combat heavy

drinking on college campuses. Despite the popularity of the social norms approach,
studies have failed to consider the consequences of receiving multiple normative
messages. This study, therefore, was an initial investigation into the effects that multiple
normative messages and advertisement believability have on persuasion. A 3x3
experiment with 262 college students was conducted. The results show that both
descriptive and injunctive norms have statistically signiﬁcant effects on attitude, and
attitude change. It was also found that attitudes and behavioral intentions are most
favorable for drinking ﬁve or more drinks when participants receive either two congruent
messages advocating heavy drinking or no messages whatsoever. There was no evidence
that advertisement believability mediated the relationship between message congruency
and attitudes or behavioral intentions. This study not only provides evidence consistent

with the effectiveness of the social norms approach but it also has implications for future

campaigns.

ACKNOWLEDGEMENTS
I would like to thank my committee: Dr. Stan Kaplowitz, Dr. Maria Lapinski, and
Dr. Sandi Smith for their encouragement, guidance, and suggestions. I would especially
like to thank my advisor, Dr. Frank Boster, for his creative ideas, support, direction, and

time.

iii

TABLE OF CONTENTS

LIST OF TABLES ........................................................................... v
LIST OF FIGURES ........................................................................ vi
INTRODUCTION ........................................................................... 1
METHODS ................................................................................. 13
RESULTS ................................................................................... l 8
DISCUSSION .............................................................................. 33
APPENDICES .............................................................................. 42
REFERENCES ............................................................................. 94

iv

LIST OF TABLES
Design of the Study ................................................................................. 44
Study Messages by Condition ..................................................................... 45

Pre-Test Attitude Measure Inter-Item Correlations, Factor Loadings, and Descriptive
Statistics ............................................................................................. 46

Post-Test Attitude Measure Inter-Item Correlations, Factor Loadings, and Descriptive
Statistics .............................................................................................. 47

Pre—test Behavioral Intention Inter-Item Correlations, Factor Loadings, and Descriptive
Statistics .............................................................................................. 48

Post-test Behavioral Intention Inter-Item Correlations, Factor Loadings, and Descriptive
Statistics .............................................................................................. 49

Descriptive Norm Perception Inter-Item Correlations, Factor Loadings, and Descriptive
Statistics .............................................................................................. 50

Injunctive Norm Perception Inter-Item Correlations, Factor Loadings, and Descriptive

Statistics .............................................................................................. 51
Normative Perception Means ...................................................................... 52
Norm Type and Post-Test Attitude Measure Means ............................................ 53

Summary of Regression Findings for Multiple Normative Message Conditions ....54

Norm Type and Post-Test Behavioral Intention Means ....................................... 55
Norm Type and Post Test Attitude Measure Means for all Nine Conditions ............... 56
Norm Type and Attitude Change Means for all Nine Conditions ........................... 57
Norm Type and Behavioral Intention Means for all Nine Conditions ....................... 58

Percentage of Participants Who Find Each Advertisement Believable by Congruency...59
Post-Test Attitude Means based on Message Type and Advertisement Believability. . ...60

Post-Test Behavioral Intention Means based on Message Type and Advertisement
Believability .......................................................................................... 61

LIST OF FIGURES

Condition 1 Advertisement ........................................................................ 84
Condition 2 Advertisement ........................................................................ 85
Condition 3 Advertisement ........................................................................ 86
Condition 4 Advertisement ........................................................................ 87
Condition 6 Advertisement ........................................................................ 88
Condition 7 Advertisement ........................................................................ 89
Condition 8 Advertisement ........................................................................ 90
Condition 9 Advertisement ........................................................................ 91
Donating Blood Advertisement .................................................................. 92
Seatbelt Advertisement ............................................................................ 93

vi

INTRODUCTION

Despite the popularity of the social norms approach, research continues to reveal
inconsistent results, particularly in college students’ drinking. Approximately 80 to 90%
of college students drink alcohol (Haines & Spear, 1996), and a large number of students
are classiﬁed as excessive drinkers. The National College Health Assessment (American
College Health Association [ACHA], 2006) found that in the fall of 2005, 32% of
females and 46% of males at universities across the country reported having ﬁve or more
drinks in one sitting during the previous two weeks. This heavy episodic drinking can
result in consequences such as death, injury, assault, sexual abuse, unsafe sex, drunk
driving, and legal consequences (Hingson, Heeren, Zakocs, Kopstein & Wechsler, 2002).
Furthermore, most consequences of college student drinking do not occur from long term
harm, such as liver damage, but instead from a heavy episode of drinking (N eighbors,
Oster-Aaland, Berstrom & Lewis, 2006). For these reasons, excessive drinking has
become a very important social issue. Many college campuses have adopted a social
norms-based approach (SNA) to combat high drinking rates and reduce injury from
excessive alcohol consumption. The SNA, which was originally formulated in 1986
(Perkins & Berkowitz, 1986), is based on the idea that if people overestimate a norm,
correcting these misperceptions may lead to reduced alcohol consumption. These
campaigns, however, have produced contradictory results. One possible explanation for
these inconsistent results is that students may receive incongruent normative messages.
An example of an incongruent normative message is a descriptive norm which suggests
one behavior and an injunctive norm which suggests a conﬂicting behavior. This study

seeks to clarify the effects of receiving incongruent normative messages.

One of the early studies to report a reduction in excessive drinking following the
implementation of a social norms-based campaign was conducted by Haines and Spear
(1996). They reported an 8.8% decrease in students who were self-identifying as binge
drinkers after exposure to a social norms-based campaign. Not only did this study reduce
reports of heavy episodic drinking, but a greater reduction in the perceived norm of how
many students binge drink after a social norms campaign, compared with a traditional
strategy, was also reported. Since this study, others have also trumpeted the social norms
approach and reported a reduction in excessive drinking at schools implementing such a
program.

It should be noted that since the time of this study the terminology used to refer to
sizeable alcohol consumption has changed. The Haines and Spear (1996) study referred
to heavy drinking as binge drinking. The present study, however, referred to a large
quantity of alcohol in one sitting as heavy drinking and not binge drinking. Although
some researchers (Wechsler, Lee, Kuo, Seibring, Nelson & Lee, 2002) still refer to it as
binge drinking, this term has largely been discontinued. Wechsler and Nelson (2001)
deﬁned binge drinking as ﬁve or more drinks in a row at least once in the past two weeks
for males and four or more drinks in a row in the past two weeks for females. Problems
have been noted with using the term binge drinking (Inter-Association Task Force on
Alcohol and Other Substance Abuse Issues [IATF], 2000; The Higher Education Center,
2000)‘. For these reasons, this study uses the terms heavy drinking or excessive drinking.

Since the Haines and Spear (1996) study, there has been additional evidence of
the effectiveness of the social norms approach. DeJong et al. (2006) conducted a multi-

site randomized trial comparing institutions implementing a social norms-based

campaign with those that did not and found that the perception of student drinking norms,
self-reported alcohol consumption, and alcohol-related consequences increased at all
schools regardless of condition. This increase was much smaller, however, at schools
implementing a social norms-based approach (DeJong et al., 2006).

The social norms approach has also been found to be inﬂuential across different
behavior. Olds, Thombs, and Tomasek (2005) examined seventh through 12th grade
students and found that both injunctive and descriptive norm perceptions that were left
uncorrected were more strongly related to behavioral intentions to initiate cigarette,
marijuana, or alcohol use than were socio-demographic variables. Fabiano, Perkins,
Berkowitz, Linkenbach, and Stark (2004) also found evidence that norms about sexual
violence were misperceived and that correcting these misperceptions may be successful
at reducing sexual violence against women.

Social norms approaches have also been successful in reducing alcohol
consumption in hi gh-risk populations. Perkins and Craig (2006) found that after exposure
to a social norms campaign, frequent alcohol consumption, high-quantity consumption,
and negative consequences from drinking all declined by 30% or more among college
student athletes. Larimer, Turner, Mallett, and Geisner (2004) also found that normative
perceptions inﬂuence behavior among students involved in ﬁatemities and sororities.
This ﬁnding is key as fraternity and sorority members have been labeled a particularly
high-risk population with increased alcohol consumption (Bartholow, Sher & Krull,
2003; Borsari & Carey, 1999; Cashin, Presley & Meilman, 1998). Larimer et al. (2004)
found that descriptive norm perceptions predicted drinking behavior and injunctive norm

perceptions predicted drinking one year after the original survey.

There are, however, other studies which have reported incongruous results. Werch
et al. (2000) found that a social norms campaign targeting ﬁrst-year residential college
students had no signiﬁcant impact on reducing alcohol consumption or alcohol use risk
factors. Clapp, Lange, Russell, Shillington, and Voas (2003) conducted an experiment in
which one residence hall received a descriptive norm campaign and the comparison
residence hall received the policies and laws surrounding alcohol consumption. Clapp et
al. (2003) found that the group receiving the normative campaign decreased their
perception of how much the typical college student drank compared to the comparison
group but that they did not reduce their actual alcohol consumption. There was a slight
increase in alcohol consumption among those who received the normative campaign but
the difference was not statistically signiﬁcant.

Carnpo, Brossard, Frazer, Marchell, Lewis, and Talbot (2003) also report that the
social norms approach is widespread at college campuses but that the results for these
campaigns are obscured by measurement problems. Carnpo et al. (2003) found that
perceptions of friends’ drinking, which are similar to the Theory of Planned Behavior’s
measure of the subjective norm, inﬂuenced drinking behavior but descriptive norms, a
construct central to the social norms approach, did not.

Descriptive and Injunctive Norms

Social norms campaigns feature either or both a descriptive and an injunctive
norm message. Descriptive norms are the perception of what other people do. The
descriptive norm is indicated by the perception of “what is everyone else doing?”
(Cialdini, Reno & Kallgren, 1990) and provide a guide for action, particularly when the

situation is ambiguous (Lapinski & Rimal, 2005). Alternatively, injunctive norms are the

perception of what one should or ought to do (Borsari & Carey, 2003), and are based on
the perception of what is typically approved. Injunctive norms serve as a cue to action as
people are motivated to avoid social sanctions or seek approval (Mead, 1932).

There are a large number of studies that have examined the impact that
descriptive and injunctive norms have on behavioral intentions and behavior. A meta-
analysis by Borsari et al. (2003) that included 23 studies evaluating descriptive and
injunctive norms surrounding college student drinking found that focusing on descriptive
norms was an effective way to inﬂuence perceptions of others’ drinking. The results for
injunctive norms were unclear because only two studies included in the meta-analysis
focused on injunctive norms.

In a series of three studies examining the differential impact of norm type Reno,
Cialdini, and Kallgren (1993) found evidence of a more nuanced relationship. Reno et al.
(1993) claim that injunctive norms have a greater impact across situations than
descriptive norms. In these studies the descriptive norm involved watching someone
refrain from littering and the injunctive norm included observing social disapproval of
another’s littering. It was found that descriptive norms were successful in reducing
littering in a clean environment that had not already been littered. Injunctive norms, on
the other hand, were successful in reducing littering in both a clean and littered
environment (Reno et al., 1993). Trockel, Williams, and Reis’ (2003) results did not
concur with this ﬁnding; they found that focusing on descriptive norms or subjective
norms were statistically signiﬁcant predictors of alcohol consumption among fraternity
men whereas injunctive norms were not.

Multiple Normative Messages

Schultz, Nolan, Cialdini, Goldstein, and Griskevicius (2007) offer an explanation
as to why normative campaigns have produced inconsistent results. Schultz et al. (2007)
posit that when people receive a single normative message, a boomerang effect may
result for those whose behavior is below the actual norm. In the case of alcohol
consumption, those who consume less alcohol than the norm may increase their alcohol
consumption to match the norm. Thus, social norms are powerful for both those whose
behaviors are above and below the norm. Schultz et al. (2007) found that what they
termed a boomerang effect is indeed present when people only receive one descriptive
norm message; however, the boomerang effect disappears when participants receive both
a descriptive and an injunctive norm message that suggest the same behavior.
C onﬂicting Messages

An alternative explanation for the inconsistent results that social norms campaigns
have produced is that people receive conﬂicting messages. A conﬂicting message is one
which is perceived by audiences as advocating two different positions; therefore, it is
two-sided and non-refutational. A conﬂicting normative message is a special type of
conﬂicting message and occurs when the descriptive norm message presents a different
message recommendation than the injunctive norm message. In the realm of alcohol
consumption this situation could occur if a message highlighted social pressure to drink
but claimed that drinking should be avoided, or if it emphasized the beneﬁts of alcohol
consumption but demonstrated no evidence of social pressure to drink. Another type of
conﬂicting message would be one that said that few people drink but encouraged others
to drink. Despite the plethora of studies that examine normative campaigns, there are no

known studies that have explicitly evaluated the relative suasory effect of conﬂicting and

non-conﬂicting normative messages. This deﬁciency in the existing research is somewhat
surprising as numerous studies have documented that people report receiving conﬂicting
health messages (e.g. Brashers, Goldsmith & Hsieh, 2002; Meredith, Jeffe, Mundy &
Fraser, 2001; Yuksel & Corbett, 2005). For instance, conﬂicting health messages have
become such a problem in HIV prevention campaigns that health professionals are
attempting to create an international agency responsible for compiling HIV statistics to
clear up the muddled effects of many agencies supplying mixed messages (Billingham,
2006). Although these mixed messages are not explicitly deﬁned they consist of multiple
agencies offering divergent advice and statistics.

Despite the dearth of studies examining the effects of incongruent normative
messages, many papers point out that people receive incongruent normative messages
quite often. Yanovitzky and Stryker (2001) suggest that students not only receive
normative messages about drinking from university campaigns but also from the media,
and that these messages often conﬂict with each other.

Peralta (2002) also indicates that college students receive incongruent normative
messages. By conducting semi-structured interviews with college students, Peralta (2002)
found that students perceive a body-image norm which encourages thinness but
simultaneously perceive a norm which encourages alcohol consumption. These two
norms directly conﬂict with each other as students believe that if they consume alcohol
they will gain weight. Peralta (2002) concludes that both of these norms inﬂuence
behavior as students still drink, but 40% of students reported concern about the calories
in alcohol and approximately 18% changed their eating habits in response to the number

of calories ingested from consuming alcohol. During the interviews some students

mentioned skipping dinner before drinking, self-induced purging, or extra exercise due to
the increased caloric intake from drinking. Although this study provides evidence of
students receiving conﬂicting messages surrounding alcohol consumption, there is not a
study that controls the content of normative messages so as to look at the effects of
incongruent messages on persuasion.

Without explicitly examining the effects of incongruent messages, some studies
have found that normative messages from different sources are more inﬂuential, thus
raising the possibility that people receive multiple normative messages. Yanovitzky,
Stewart, and Lederrnan (2006) found that perceived alcohol use by close peers, such as
best friends and ﬁ'iends, had a more inﬂuential effect on alcohol use than perceived
alcohol use by students in general. Park, Klein, and Smith (2007a) also found that the
drinking behavior of close friends, as measured by a subjective norm, inﬂuenced
behavioral intention, and that the injunctive and descriptive norms of other university
students did not. Further, Park et al. (2007a) found that both descriptive and injunctive
norms for college students across the United States, as opposed to students at the
participants’ own university, were statistically signiﬁcant predictors of behavioral
intention.

It is somewhat naive to think that people are only receiving normative messages
supplied by university normative campaigns. Further evidence for incongruent normative
messages can be found on college campuses where students are forbidden to drink when
they are underage, but many local liquor stores and bars do not check identiﬁcation, and

alcohol may be allowed on campus or involved in many social activities (Ross, 2004).

Nygaard and Grube (2005) also found evidence of a discrepancy between enacted
norms (i.e., behavior) and normative messages. They conducted 44 semi-structured
interviews with adolescents, who had previously reported driving while intoxicated or
riding with a drunk driver, in order to examine the impact of incongruent messages
surrounding drinking alcoholic beverages. They found that the messages from adult
authorities and the behavior of these same authorities were often discrepant. Adolescents
understood adult messages advising that they should not engage in drunk driving but
reported observing adult behavior which conﬂicted with this message. This discrepancy
left them ambivalent about the dangers of drinking and driving. The ﬁndings indicated
that conﬂict between normative messages and enacted norms leads to ambiguity among
the adolescents and an inability make sense out of the ambiguity surrounding the effects
of drinking and driving.

Elek, Miller-Day, and Hecht (2006) looked at how norms inﬂuence early
adolescent substance use. Although personal anti-drug norms had the largest effect on
overall substance use, this study did ﬁnd that both injunctive and descriptive norms
inﬂuenced early adolescents. The correlation between perceived peer use, or descriptive
norms, and overall substance use was .10, whereas the correlation between parents’
injunctive norms and substance use was -.16 and friends’ injunctive norms and substance
use was -.12. Thus, the more parents and ﬁ’iends disapproved of substance use, the less
likely adolescents were to report substance use; however, the greater the perception of
substance use by peers, the more likely students were to report substance use. This study
shows that adolescents perceive anti-drug injunctive norm messages from their parents

and friends, who inﬂuence behavior, but at the same time perceive that others are using

illegal substances and this factor also inﬂuences their behavior. Thus, adolescents receive
multiple and conﬂicting normative messages and all of these messages inﬂuence their
health outcomes.

Message Sidedness

The previously cited studies provide evidence that people do receive incongruent
normative messages. They do not, however, examine the effects of receiving opposing
messages. One of the possible explanations for how people deal with incongruent
messages may be found in traditional theories of attitude change literature, such as the
message sidedness corpus. In a meta-analysis that examined the persuasiveness of one-
and two-sided messages, Allen (1991) found that a two-sided nonrefutational message
was less persuasive than a one-sided message, and that a two-sided refutational message
was more persuasive than a one-sided message. Two conﬂicting normative messages are
akin to a two-sided nonrefutational message and it can, therefore, be expected that a
conﬂicting message would be less persuasive than a single message.

As the aforementioned review shows, people are likely to receive conﬂicting
messages and to receive them from multiple channels. Despite the dearth of research
examining the effects of incongruent normative messages, predictions can be made about
these effects based on studies that examined refutational messages. Previous research has
found that nonrefutational two-sided messages are ineffective (Allen, 1991) and
approximately cancel each other out (Anderson & Hovland, 1957), yielding no net
attitude change.

It can, therefore, be expected that two normative messages that suggest the same

position, in other words two one-sided messages, are likely to have the greatest effect on

10

attitudes and produce attitude change in the same direction. Examples of two messages
advocating the same position would be both injunctive and descriptive norm messages
either suggesting drinking in moderation or drinking heavy amounts of alcohol. This
reasoning leads to the hypotheses of the study.

The ﬁrst hypothesis predicts that if both normative messages are congruent and
suggest heavy drinking, attitudes will be more favorable toward drinking ﬁve or more
drinks at one sitting. If both normative messages are congruent and suggest moderate
drinking, attitudes will be less favorable. If the messages are incongruent, attitudes will
be intermediate, i.e., between the two extremes. Put differently, repetition, at least up to a
moderate point, facilitates the persuasion process (Cacioppo & Petty, 1979; Weiss, 1971;
Wilson & Miller, 1968), particularly when the repeated messages differ slightly and
emanate from different sources (Harkins & Petty, 1981a, 1981b).

It can also be predicted that normative messages will have a similar affect on
behavioral intentions. Hypothesis two, therefore, predicts that if both normative messages
are congruent and suggest heavy drinking, behavioral intentions will be more favorable
toward drinking ﬁve or more drinks at one sitting. If both normative messages are
congruent and suggest moderate drinking, behavioral intentions will be less favorable. If
the messages are incongruent, behavioral intentions will be intermediate, , i.e., between
the two extremes.

Believability

Attitude change must also be considered in light of advertisement believability.

One possible effect of incongruent normative messages is that university normative

campaigns which supply statistics are often perceived as inaccurate or biased. Polonec,

11

Major, and Atwood (2006) evaluated a normative campaign that featured the message
that most students on campus drink zero to four drinks when they party and found that
72.6% of students did not believe the normative message. This point leads to the third
hypothesis of this study which posits that those who receive two congruent messages are
likely to ﬁnd the advertisement more believable than those who receive two incongruent
messages.

Alternatively, when messages are found to be believable, normative perceptions
are more accurate (Park, Klein & Smith, 2007b). Social judgment theory also points to
the necessity of believability for attitude change. Social judgment theory is based on the
premise that the effect of a persuasive message is determined by how the audience
perceives what the message suggests. (Sherif, Sherif, & Nebergall,1965). Smith, Atkin,
Martell, Allen, and Hembroff (2006) found that messages that fell in the latitude of
acceptance or noncommitrnent were more likely to be effective than messages that fall in
the latitude of rejection. Messages that fall in the latitude of rejection will not be accepted
because they are not believed. Therefore, for a social norms campaign to have the most
effect and change attitudes, the messages must be believable. The fourth hypothesis,
therefore, predicts that believability will mediate the relationship between message
congruency and attitude. The ﬁfth hypothesis predicts that believability will mediate the

relationship between message congruency and behavioral intention.

12

METHODS
Subjects

Subjects in this study were students from communication courses at a large
Midwestern University. A convenience sample of 262 subjects was obtained.
Participation was voluntary and students were offered either extra credit or research
credit in their courses in exchange for their participation. The sample had slightly more
females than the MSU population; the MSU population is 45% male and the sample was
33% male. The ethnic composition of the sample closely approximated the MSU
population: 10.7% of the sample self-identiﬁed as Black or African American, compared
to 8.4% in the general population. Approximately 82.4% of the sample self-identiﬁed as
White or Caucasian as compared to 82.0% of the MSU population. The sample was about
3.1% as Asian or Paciﬁc Islander, compared to 5.7% in the population. About 2.7% of
the sample self-identiﬁed as Hispanic or Latino and 3.2% of the MSU population did. In
the sample the remained 0.4% self-identiﬁed as other and the MSU population is 0.7%
Native American. All participants were undergraduates with 33.2% in their ﬁrst year,
compared to 26% in the MSU population; 26.0% in their second year, compared to 22%
in the MSU population; 20.2% in their third year, compared to 25% in the MSU
population; 14.9% in their fourth year; and 5.3% in their ﬁfth year. The MSU population
is 27% senior but is not broken down by years in college. Participants ranged in age from
18 to 38 with the mean age being 20 (S = 1.88), which is identical to the general
population. About 17.9% of the participants were members of a fraternity or a sorority
compared to about 8.4% of undergraduate students at large.

Design

13

The effectiveness of normative messages about alcohol consumption was
examined using a 3x3 between subjects experiment with an imbedded control condition.
The design of the study is detailed in Table 1. Dependent measurements were obtained
with a pre- and post-test. Immediately following the pre-test, participants viewed a series
of three advertisements that were displayed by an LCD projector. Students viewed a
normative advertisement about alcohol consumption, an advertisement about seatbelts,
and an advertisement about donating blood. There were three different orders of the
advertisements that were selected and assigned at random. All advertisements can be
found in Appendix C. In the control condition, students did not view the normative
advertisement but simply viewed the other two advertisements. The additional two public
service announcements (PSA) were included to conceal the intent of the study. Following
the messages, participants completed the post—test which was identical to the pre-test
except that the behavior items were excluded as actual behavior could not have been
expected to change during the course of the study. The nine different experimental
conditions and messages are detailed in Table 2. The descriptive norm advertisements
featured messages that either suggested drinking in moderation, drinking more heavily, or
were absent. An example of a descriptive norm message suggesting drinking in
moderation is “77% of MSU students drink less than 5 drinks when they party.” An
example of a descriptive norm message suggesting heavy drinking is “23% of MSU
students drink less than 5 drinks when they party.” The control condition was without a
normative advertisement about alcohol consumption. In the other conditions, injunctive
norm advertisements, like the descriptive norm advertisements, either suggested drinking

in moderation, heavy drinking, or were absent. An example of an injunctive norm

14

message suggesting drinking in moderation is “72% of MSU students disapprove of
drinking more than 4 drinks when they party.” An example of an injunctive norm
message suggesting heavy drinking is “28% of MSU students disapprove of drinking
more than 4 drinks when they party.” There were two conditions with incongruent
messages, four with single messages, two with congruent messages, and one devoid of
messages to serve as a control condition. Each participant was only in one condition and
was randomly assigned to that condition.

Instrumentation

In order to gauge the strength of the normative induction, respondents were asked
questions about their injunctive and descriptive norm perceptions after viewing the
normative messages at the post-test. Respondents’ behavioral intention and attitudes
toward drinking ﬁve or more drinks were measured at both pre- and post-test.
Demographic information was measured on the pre-test. Advertisement believability was
measured immediately following the experimental stimulus. A copy of the measures can
be found in the Appendix B.

Behavioral intention was measured with four items. One item was open ended and
asked participants, “the next time that you party/socialize with friends, how many drinks
do you think you will have.” The remaining three questions employed ﬁve point Likert
Scales with ﬁve being strongly agree and one being strongly disagree. Higher numbers
indicated a greater intention to drink ﬁve or more drinks. Behavioral intention items were
summed to form a behavioral intention index.

Attitudes toward drinking ﬁve or more alcoholic drinks in one sitting were

measured in a semantic differential format that featured 11 pairs of bipolar adjectives.

15

The 11 bipolar adjectives for these items ranged from one to seven where seven indicated
a more favorable attitude toward drinking ﬁve or more drinks. Attitude items were
summed to form an attitude index.

Perceptions of the descriptive norm were measured with four items. Three of the
items were measured with ﬁve point Likert Scales with ﬁve being strongly agree and one
being strongly disagree. Higher numbers reﬂected higher levels of perceived alcohol
consumption. The other item was an open ended question that asked participants “think
of all MSU undergraduate students. What percentage do you think have ﬁve or more
drinks in one sitting at least once during a normal week?” Descriptive norm items were
summed to form an index of descriptive norm perception.

Perceptions of the injunctive norm were measured with ﬁve items. Four of these
items were ﬁve point Likert Scales that were scored in a similar manner to the descriptive
norm perception items. The last item was open ended and asked participants “think of all
MSU undergraduate students. What percentage do you think consider drinking ﬁve or
more drinks in one sitting at least once during a normal week to be an acceptable
activity?” Injunctive norm items were summed to form an index of injunctive norm
perceptions. Higher scores on this index indicated a perception of greater approval for
consuming ﬁve or more alcoholic drinks.

Believability was measured for the advertisement as a whole with one
dichotomous item that asked participants if they found the advertisement believable. A
response of no was coded as zero and yes as one. Thus, a higher number indicated greater

believability.

16

There was also a variable created to test hypotheses four and ﬁve. The conditions
in which participants received two messages, regardless of type, were recoded into three:
(1) the two incongruous message conditions were collapsed, (2) the congruous condition
suggesting heavy drinking, and (3) the congruous condition suggesting moderate
drinking.

Procedure

Participants were randomly assigned to a condition and an order. Three orders
were chosen randomly. The experiment was administered to groups of about 10 at a time,
all in the same condition and order. Upon arriving for the experiment students were told
that they were evaluating potential PSAs. They were asked to give their honest evaluation
of the messages. It took participants approximately 30 minutes to complete the
experiment. At the end of the study they were debriefed about the purpose of the study
and given the actual percentages for both the descriptive and injunctive norm surrounding
alcohol consumption at their university. All study advertisements are located in Appendix

C.

17

RESULTS

Scales

Conﬁrmatory factor analysis (CFA) was performed on all scales to assess content
validity. Reliability analyses were also conducted on all scales. The results of these
analyses are reported in subsequent paragraphs.

Attitudes. Participant attitudes toward consuming ﬁve or more alcoholic drinks
were measured on a set of seven point semantic differential scales. Originally 11 sets of
bipolar adjectives were employed to measure attitudes. Based on the results of CFA, it
was decided that seven items ﬁt best. Items were eliminated if the item produced a
substantial number of errors of a magnitude greater than would be expected by chance
alone. For the pre-test the root mean squared error (RMSE) was .030 and for the post-test
the RMSE was .033. There were no statistically signiﬁcant errors for both the pre-test and
the post-test. Inter-item correlations, factor loadings, and descriptive statistics for the pre-
and post-test attitude items can be found in Tables 3 and 4. The distribution of the pre-
test attitude index was slightly negatively skewed (skew = -.635, standard error = .150,
kurtosis = -.170, standard error = .300) with a mean of 25.29, a standard deviation of
7.93, and an a of .902. The distribution of the post-test attitude index was also slightly
negatively skewed (skew = -.458, standard error = .150, kurtosis = -.311, standard error =
.300) with a mean of 23.96, a standard deviation of 8.05, and an a of .902. The
correlation between the pre- and post-test seven item attitude measure was .88.

Attitude Change. Attitude change was calculated employing the seven attitude
measures and subtracting the pre-test measure from the post-test attitude measure. The

distribution of the attitude change index was slightly negatively skewed (skew = -.274,

18

standard error = .150, kurtosis = .929, standard error = .300) with a mean of -1 .3244, a
standard deviation of 3.569 and an a of .453. A one sample t-test was conducted and
revealed that there was statistically signiﬁcant attitude change as attitudes became less
favorable toward drinking ﬁve or more drinks in one sitting from the pre-test to the post-
test, t(26l) = -6.006, p<.001.

Behavioral Intention. Behavioral intention was measured with ﬁve items. Four of
the items employed a Likert Scale and one was open ended; therefore, all items were
standardized. Based on the results of the CF A, one item was discarded because it lowered
the reliability and produced Signiﬁcant errors. The resulting four items were summed to
form an index which did have one signiﬁcant error for the pre-test and one for the post-
test. For the pre-test the RMSE was .083 with an a of .771. For the post-test the RMSE
was .032 and the a was .801 . Inter-item correlations and descriptive statistics for these
indices can be found in Tables 5 and 6. The distribution for the pre-test was slightly
negatively skewed and platykurtic (skew = -.140, standard error = .150, kurtosis = -.915.
standard error = .300). The distribution for the post-test was symmetrically distributed
but platykurtic (skew = -.020, standard error = .150, kurtosis = -.900, standard error =
.300). Again, because the index was standardized the mean of both the pre- and post-test
distributions was 0. For the pre-test index the standard deviation was 3.673 and for the
post-test the standard deviation was 3.684.

Believability. For every normative advertisement, regardless of the messages
presented, believability was measured with one dichotomous item that asked, “I found

this ad to be believable.” The mean for this measure was .554 and the standard deviation

19

was .498. The distribution of this item was slightly negatively skewed and platykurtic
(skew = -.217, standard error = .159, kurtosis = -1.970, standard error = .318).

Descriptive Norm Perceptions. Descriptive norm perceptions were measured with
four items. Three of the items employed a Likert response scale and one item was open
ended. Due to the differing response formats for these items, all items were standardized.
Based on the results of the CFA, the least number of errors and the smallest RMSE
occurred when all of the original items were retained. The RMSE for the post-test
measure of descriptive norm perceptions was 0.111 and there were no statistically
signiﬁcant errors. Inter-item correlations and descriptive statistics for this index can be
found in Table 7. Because the items were standardized the mean of each item was 0.
Once the items were summed to form an index, the mean was also 0, the standard
deviation 3.261 and or =.832. The distribution of the descriptive norm perception index
was slightly negatively skewed and platykurtic (skew = -.300, standard error = .150,
kurtosis = -1.101, standard error = .300).

Injunctive Norm Perceptions. Injunctive norm perceptions were measured with
ﬁve items. Four of the items featured a Likert Scale and one was open ended; therefore,
all items were standardized. Based on the results of the CF A, one item was thrown out
because it lowered the reliability and produced signiﬁcant errors. The resulting four item
index had no statistically signiﬁcant errors. The RMSE was 0.027 and the 0. was .801.
Inter-item correlations and descriptive statistics for this index can be found in Table 8.
The distribution of the injunctive norm perception index was slightly negatively skewed
and platykurtic (skew = -.383, standard error = .150, kurtosis = -.811, standard error =

.300). Again because the items were standardized, the mean of each item was 0. When

20

the items were summed to form an index, the mean of the distribution was 0 and the
standard deviation was 3.157.
Induction Check

In order to check if students perceived the normative components of each
message, an Analysis of Covariance (ANCOVA) was conducted in which the injunctive
and descriptive norm inductions were independent group factors and the pre-test attitude
index was a covariate. Means and standard deviations of the normative perceptions which
were not adjusted for the covariate are presented in Table 9. There was a main effect for
the injunctive norm condition on the injunctive norm index, F (2,252) = 3.664, p=.027,
such that when the injunctive norm induction suggested drinking ﬁve or more drinks in
one sitting, participants reported that they perceived this injunctive norm suggesting
heavy drinking. There was also a substantial main effect for the descriptive norm
condition on the injunctive norm index, F(2,252) = 6.647, p=.002, such that when the
descriptive norm induction suggested drinking ﬁve or more drinks in one sitting,
participants reported that they perceived an injunctive norm that suggested heavy
drinking as well. There was not a statistically signiﬁcant interaction between descriptive
and injunctive norms, F(4, 252) = .461, p = .765. There was also a main effect for the
injunctive norm condition on the descriptive norm index, F(2,252) = 5.063, p = .007,
such that when the injunctive norm condition suggested drinking ﬁve or more drinks in
one sitting, participants perceived a descriptive norm that also suggested heavy drinking.
There was also a main effect for the descriptive norm condition on the descriptive norm
index, F(2,252) = 7.628, p =.001, such that when the descriptive norm induction

suggested drinking ﬁve or more drinks in one sitting, participants perceived this

21

descriptive norm suggesting heavy drinking. Again, there was not a statistically
signiﬁcant interaction between descriptive and injunctive norms F(4, 252) = .805, p =
.523. Perceptions of the injunctive norm and the descriptive norm were also correlated, r
= .653, p <.001. Thus, the normative messages led participants to infer both descriptive
and injunctive norms for college students about consmning alcohol.

Order Eﬂects

Because three different orders of the advertisements were utilized, an Analysis of
Variance (ANOVA) was conducted to estimate order effects. There was no evidence of
any order effects on the post-test attitude, F(2,25 8) = .358, p =.70, post-test behavioral
intention, F (2,258) = 1.095, p=.336, or attitude change, F(2,259)=.691, p =.502.
ijotheses

Hypothesis One. The ﬁrst hypothesis predicted that if both normative messages
were congruent and suggested heavy drinking, attitudes would be more favorable toward
drinking ﬁve or more drinks. However, if both normative messages were congruent and
suggested moderate drinking, attitudes would be less favorable. Incongruent messages
would produce attitudes that were intermediate. In sum, if only the two message
conditions are considered, two main effects (one for the descriptive norm induction and
one for the injunctive norm induction) are expected.

In order to test this hypothesis, only the conditions in which participants received
two normative messages were considered. An ANCOVA was conducted in which the
pre-test attitude measure was entered into the model as a covariate and the descriptive
and injunctive norm inductions were entered as ﬁxed factors. The means for this analysis,

which are unadjusted for the covariate, are presented in Table 10. The results show a

22

statistically signiﬁcant main effect for descriptive norms on post-test attitude, F (1,113) =
4.786, p = .031, n2 = .041, r = .202, d = .405. There was also a statistically signiﬁcant
effect for injunctive norms on post-test attitude, F (1, 113) = 7.937, p = .006, n2 = .066, r
= .257, d = .507. There was no evidence of an interaction between norm type, F (1, 113)
= 1.179, p = .280, n2 = .010, r=.100, d = .200. Looking at the means presented in Table
10 it is evident that the most favorable attitudes toward drinking ﬁve or more alcoholic
drinks occurred when participants received two congruent messages suggesting heavy
alcohol consumption. The least favorable attitudes toward drinking ﬁve or more drinks
occurred when participants received two congruent messages suggesting moderate
alcohol consumption. Thus, the data were consistent with the hypothesis.

In order to provide a more nuanced assessment of the relationship between
multiple normative messages on attitudes, a regression analysis was conducted, in which
the post-test attitude was regressed onto the pre-test and the two normative conditions.
Regression analysis revealed that the model signiﬁcantly predicted post-test attitudes,
F(3, 114) = 193.842, p<.001. R for the model was .914 and adjusted R2 was .832.
Further, descriptive norms had a substantial positive impact on attitudes, ,6 = .084, t
(114)= 2.204, p =.O30. The impact of injunctive norms on attitude was also statistically
signiﬁcant, ,6 = .107, t (1 14) = 2.816, p =.OO6. Table 11 displays the results of the
regression analysis.

Hypothesis Two. Hypothesis two predicted that if both normative messages were
congruent and suggested heavy drinking, behavioral intentions would be more favorable
toward drinking ﬁve or more drinks at one sitting. If both normative messages were

congruent and suggested moderate drinking, behavioral intentions would be less

23

favorable. If the messages were incongruent, behavioral intentions would be
intermediate.

In order to test this hypothesis, only the conditions in which participants received
two normative messages were considered. An ANCOVA was conducted in which the
pre-test behavioral intention measure was entered into the model as a covariate. The
means for this analysis, which have not been adjusted for the covariate, are presented in
Table 12. The analysis revealed that descriptive norms did not have a substantial impact
on post-test behavioral intentions, F (1,113) = .830, p = .364, n2 = .007, r = .087, d =
.167. Injunctive norms, however, did have a statistically signiﬁcant effect on behavioral
intention, F (1,113) = 5.469, p = .021, n2 = .046, r = .214, d = .429. There was not a
statistically signiﬁcant interaction between descriptive and injunctive norms, F (1,113) =
.085, p = .772, n2 = .001, r = .032, d = .063.

Looking at the means presented in Table 12, the largest intentions to consume
ﬁve or more alcoholic drinks occurred when participants received two congruent
messages suggesting heavy alcohol consumption. The smallest behavioral intention to
consume ﬁve or more alcoholic beverages occurred when participants received two
congruent messages suggesting moderate drinking. As predicted, the means for the
incongruent normative messages were in between the two extremes. In the main, the data
were consistent with the spirit of the second hypothesis. The main effect of the
descriptive norm, however, did not reach conventional levels of signiﬁcance.

In order to examine further the relationship between multiple normative messages
and behavioral intention, behavioral intentions was regressed onto pre-test behavioral

intention, descriptive and injunctive norms. Regression analysis revealed that the model

24

signiﬁcantly predicted post—test behavioral intention, F (3, 114) = 323.740, p<.001. R for
the model was .946 and adjusted R2 was .892. Descriptive norms did not have a
substantial impact on behavioral intention ,6 = .028, t(114)= .908, p =.366. The impact of
injunctive norms on behavioral intention was positive and statistically signiﬁcant ,8 =
.071, t (114) = 2.347, p =.021. The results of this analysis are presented in Table 11.

Because the effect of descriptive norms was within sampling error of zero,
descriptive norms were eliminated from the regression equation and the analysis was
conducted again. Therefore, behavioral intention was regressed onto injunctive norms
and the pre-test measure of behavioral intention. These results are consistent with the
previous analysis. The model signiﬁcantly predicted post-test behavioral intention, F (2,
115) = 484.941, p<.001. R for the model was .946 and adjusted R2 was .892. The impact
of injunctive norms on behavioral intention was positive and statistically signiﬁcant ﬂ =
.070, t(115) = 2.317,p =.022.

Although there were no hypotheses posed about the effects of receiving a single
normative message, the design of the study allows for analysis of all nine conditions.
Further, the data are informative. For these reasons, an ANCOVA including all nine
conditions was conducted with the pre-test was entered into the model as a covariate. The
means from this analysis, which are not adjusted for the covariate, are presented in Table
13. There were two statistically signiﬁcant main effects for norm type on post-test
attitude. Both the descriptive norm, F (2 ,252) = 3.829, p = .023, n2 = .029, r = .170, d =
.341, and for the injunctive norm, F (2, 252) = 5.213, p = .006, n2 = .040, r = .200, d =

.400, produced statistically signiﬁcant effects. The interaction between norm type on

25

post-test attitude was not statistically signiﬁcant, F (4, 252) = 1.702, p = .150, n2 = .026, r
= .161, d= 322.ii

Looking at the means presented in Table 13, attitudes were most favorable for
drinking ﬁve or more drinks when participants received either two congruent messages
which suggested heavy drinking or no messages whatsoever. When participants received
an injunctive norm message which suggested heavy drinking and no descriptive norm
message, attitudes toward drinking ﬁve or more drinks were also higher than the other
conditions.

In order to examine the main effects in more detail a contrast analysis on the main
effects was performed. For the injunctive norm messages, attitudes were substantially
different for the heavy drinking condition compared with the moderate and no message
condition, p = .005. There was, however, no statistically signiﬁcant difference between
the moderate condition and the no message condition, p = .109.

For the descriptive norm conditions, the no message condition and the heavy
drinking message condition were substantially different than the moderate drinking
message condition, p = .007. The heavy message condition was not statistically different,
however, from the no message condition, p = .709.

Behavioral intentions were also examined when participants received a single
message. An ANCOVA was conducted with the pre-test behavioral intention index
entered into the model as a covariate. The means from this analysis are presented in Table
16. The results show that the injunctive norm had a statistically signiﬁcant effect on post-
test behavioral intention, F (2, 252) = 6.510, p = .002, n2 = .049, r = .221, d = .442. The

descriptive norm effect on behavioral intention was not statistically signiﬁcant, F(2, 252)

26

= .914, p = .402, n2 = .007, r = .084, d = .167. There was also not a substantial interaction
between norm type on behavioral intention, F (4, 252) = .450, p = .772, 112 = .007, r =
.084, d = .167. The main effect for injunctive norms is apparent when looking at the
means presented in Table 15, as the greatest intentions to consume alcohol occurred when
the injunctive norm suggested heavy alcohol consumption.

In order to examine the main effects in more detail a contrast analysis on the main
effects was performed. For the injunctive norm messages, behavioral intentions were
substantially different for the moderate drinking condition compared with the no message
and heavy drinking condition, p = .001. There was, however, no statistically signiﬁcant
difference between the no message condition and the heavy drinking condition, p = .133.
There was also a statistically signiﬁcant difference between the behavioral intentions for
the no message condition and the moderate message condition, p = .038. The heavy
drinking condition was also different from the moderate and no message condition, p =
.003.

For the descriptive norm conditions, the moderate drinking condition was not
substantially different from the no message condition and the heavy drinking message
condition, p = .256. The no message condition was also not substantially different from
the heavy drinking condition, p = .462. The no message condition was also not
signiﬁcantly different from the moderate drinking condition, p = .178 and the heavy
drinking condition was not different from the no message condition and the moderate
drinking condition, p = .935.

Hypothesis Three. The third hypothesis predicted that those who received two

congruent messages were likely to ﬁnd the advertisement, as a whole, more believable

27

than those who received two incongruent messages. To test this hypothesis, a message
congruency variable was created where 1 was coded if a participant received two
congruent messages and -1 was coded if a participant received two incongruent
messages. There was a statistically signiﬁcant relationship between message congruency
and advertisement believability, x2 (1, N= 117) = 6.353, p = .012, (p = .233. Table 16
contains the percentages for each of these conditions. Those who received two congruent
messages were more likely to ﬁnd the advertisement believable; whereas, those who
received two incongruent messages were less likely to ﬁnd the advertisement believable.
Thus, the data are consistent with hypothesis three.

Advertisement believability was also examined for each of the types of normative
messages. A correlation analysis was conducted with the dichotomous advertisement
believability variable and the trichotomous injunctive norm condition variable. There was
a statistically signiﬁcant positive relationship between injunctive norms and
advertisement believability, r = .145, p = .027. Those who received an injunctive norm
message which suggested heavy drinking were more likely to ﬁnd the advertisement
believable; whereas, those who received an injunctive norm message which suggested
moderate drinking were more likely to ﬁnd the advertisement unbelievable. There was
not, however, a statistically signiﬁcant relationship between descriptive norm messages
and advertisement believability, r = .077, p =.242.

Hypothesis Four. The fourth hypothesis postulated that advertisement
believability would mediate the relationship between message congruency and attitudes.
To the extent that believability mediates the relationship between the normative messages

and attitude, when messages are believable, congruent, and suggest heavy drinking, then

28

it is expected that students would have a very positive attitude toward consuming ﬁve or
more drinks at one sitting. On the other hand, when messages are believable, congruent,
and suggest moderate drinking, then it is expected that students would have a less
favorable attitude toward consuming ﬁve or more drinks at one sitting. Alternatively,
when messages are believable but incongruent, it is expected that attitudes are
intermediate between these two extremes. When messages are not believable, however,
there is no reason to expect attitudes to vary across these three message conditions. To
test this idea the four, two message conditions were recoded into three: (1) the two
incongruous message conditions were collapsed, (2) the congruous condition suggesting
heavy drinking, and (3) the congruous condition suggesting moderate drinking.

An ANCOVA was conducted with the pre-test attitude measure entered as a
covariate and the new recoded condition variable and advertisement believability as
independent factors. The means for this analysis, which are not adjusted for the covariate,
are presented in Table 17. The results revealed that advertisement believability did not
have a signiﬁcant impact on the post-test attitude measure, F (1, 110) = .170, p = .681, n2
= .002, r = .045, d = .089. Condition did have a substantial impact on post-test attitude, F
(2, 110) = 6.852, p = .002, n2 = .111, r = .333, d = .666, such that when participants
received two congruous messages which suggested heavy alcohol consumption, attitudes
were more favorable toward consuming ﬁve or more drinks in one sitting. The interaction
between the recoded condition variable and advertisement believability was also not
signiﬁcant, F (2, 110) = 1.254, p = .289, n2 = .022, r = .148, d = .297. These results show
that there is no evidence that advertisement believability mediated the relationship

between message congruency and post-test attitudes.

Hypothesis Five. The ﬁfth hypothesis postulated that advertisement believability
would mediate the relationship between message congruency and behavioral intentions.
Similar to the procedure used with hypothesis four, an ANCOVA was also conducted
with the pre-test behavioral intention measure entered as a covariate, the new condition
variable and advertisement believability as independent factors, and the behavioral
intention measure as a dependent variable. The new recoded trichotomous condition
variable was entered as the independent factor. The means for this analysis, which are not
adjusted for the covariate, are presented in Table 18. Advertisement believability did not
have a signiﬁcant impact on behavioral intentions, F (1, 110) = .106, p = .745, 112 = .001,
r = .031, d = .063. The condition variable, however, had a statistically signiﬁcant effect
on behavioral intention, F(1, 110) = 3.192,p = .045, n2 = .055, r = .235, d = .469.
Interestingly, behavioral intentions to consume ﬁve or more alcoholic drinks in one
sitting were greatest when participants received two incongruous messages but did not
believe the advertisement. The interaction between advertisement believability and the
trichotomous condition variable was also not statistically signiﬁcant, F (2, 110) = 1.960,
p = .146, n2 = .034, r = .184, d = .369. Therefore, there was no evidence that
advertisement believability mediated the relationship between message congruency and
behavioral intention. The ﬁfth hypothesis was, therefore, not supported.

The design of this study also allows for analysis of a recent research ﬁnding by
Schultz et al. (2007) who found that in the realm of energy conservation those who were
using less energy increased their energy use when supplied with a descriptive norm
message about energy usage in the neighborhood. Schultz et a1. (2007) found that a single

descriptive norm message served as a point of comparison from which people did not

30

want to deviate. Therefore, after being supplied with a descriptive norm message about
energy consumption, those who consumed more energy than the norm decreased their
energy usage from the baseline and those who used less energy than the norm increased
their energy usage from the baseline to match the norm. Those, however, who had been
low energy consumers and were supplied with a descriptive norm message and an
injunctive norm message in the form of a positively valenced emoticon continued to
consume less energy.

In order to delve further into the ﬁnding that Schultz et a1. (2007) reported,
independent samples t-tests were conducted with the attitude change means. For the
descriptive norm condition which suggested moderate drinking, there was signiﬁcant
negative attitude change from the pretest to the post-test, t(90) = —4.927, p<.05,
indicating that overall participant attitudes became less favorable toward drinking ﬁve or
more drinks. For the no message descriptive norm condition, there was also signiﬁcant
negative attitude change, t(83) =-2.874, p<.05. There was also signiﬁcant negative
attitude change for the descriptive norm condition which suggested heavy drinking, t(96)
= ~2.376, p<.05. Turning to the injunctive norm conditions, there was also signiﬁcant
negative attitude change for the injunctive norm condition which suggested moderate
drinking, t(85) = -2.1279, p<.05. The injunctive norm condition without a message also
produced signiﬁcant negative attitude change, t(87) = -3.296, p<.05. The injunctive norm
condition which suggested heavy drinking also produced negative attitude change which
approached conventional levels of signiﬁcance, t(87) = -1.551, p<.10.

These analyses show that even when presented with normative messages that

encouraged heavy drinking, participant attitudes toward drinking ﬁve or more drinks

31

became less favorable from the pre-test to the post-test. Thus, students were motivated to
hold a less favorable attitude toward consuming large amounts of alcohol and did not try
to match the normative information that they received. This effect held even when
participants were supplied with normative messages that encouraged heavy alcohol
consumption. Thus, Schultz et al.’s (2007) concern of a boomerang effect was not evident

in these data.

32

DISCUSSION

This study was an initial investigation into how receiving multiple normative
messages affected attitudes and attitude change. Although this is the ﬁrst study to do so,
this topic is important to explore as it is highly likely that people do receive conﬂicting
normative messages in their everyday life. Furthermore, the intricate design of this study
allowed for analysis of all possible message conditions and, thus, provided rich data.
There were nine different conditions which included a no message control condition.
Two fake advertisements were also included to conceal the intent of the study.
Additionally, there were three possible orders of these advertisements to minimize any
possible order effects.

The design of study allows for investigation into the relationship between
different types of norms, advertisement believability, and receiving multiple messages.
The ﬁrst and second hypotheses were consistent with the data as attitudes and behavioral
intentions to consume ﬁve or more drinks were greatest when participants received two
congruent messages suggesting heavy alcohol consumption. Attitudes and behavioral
intentions to consume ﬁve or more drinks were smallest when participants received two
congruent messages suggesting moderate drinking, intermediate when they received two
incongruent messages, and largest when they received two congruent messages
suggesting heavy drinking.

Interestingly when taking into account the single message conditions, behavioral
intentions to consume ﬁve or more drinks were greatest when injunctive norms suggested
heavy drinking regardless of the descriptive norm message or when participants received

no messages at all. The power that injunctive norms have on behavioral intentions was

33

also illustrated by the statistically signiﬁcant main effect for injunctive norms on
behavioral intention. Descriptive norms did not produce a statistically signiﬁcant main
effect on behavioral intention. This ﬁnding mirrors previous ﬁndings which have found
that injunctive norms, more so than descriptive norms, have a strong association with
behavior across different types of behavior (e. g. Cialdini, Kallgren, & Reno, 1991; Klein
& Boster, 2005).

The injunction check data also point to the relationship between norm types. For
the injunctive norm condition, participants perceived both injunctive and descriptive
norms. For the descriptive norm condition, participants also perceived both injunctive
and descriptive norms. There was, however, no evidence of an interaction. Each
induction, therefore, bolstered the other. This is an important ﬁnding because normative
campaigns often only focus on descriptive norms. This data show that even if normative
campaigns only focus on one type of norm, the audience is likely to perceive both types
of norms. This is perhaps because what is approved of is often what is done (Cialdini et
al., 1990).

Although there are no known previous quantitative studies that have looked at the
effects of receiving multiple normative messages on persuasion, the message sidedness
literature has looked at the effects of multiple messages on persuasion. Allen (1991)
found that a two-sided nonrefutational message was less persuasive than a one-sided
message, and that a two-sided refutational message was more persuasive than a one sided
message. Two incongruent normative messages are akin to a two-sided nonrefutational
message and a single message is equivalent to a one-sided message. The adjusted posttest

means are consistent with Allen’s (1991) ﬁnding. When incongruous, or two-sided

34

nonrefutational, messages were administered, attitudes were intermediate. When one-
sided messages suggested heavy drinking, attitudes were most positive toward drinking.
And, when one-sided messages suggested moderate drinking, attitudes were least positive
toward drinking.

These ﬁndings provide additional evidence that the social norms approach can
have powerful effects on college alcohol consumption. Further support for the social
norms approach was evidenced as both descriptive and injunctive norms had signiﬁcant
main effects on the post-test attitude measure and also attitude change. Early support for
the power of social norms came from Perkins and Berkowitz (1986) and Cialdini, Reno,
and Kallgren (1990) and has continued with a recent study by DeJong et al. (2006) who
found that campuses that implemented a social norms-based campaign had lower
perceptions of student drinking levels and lower alcohol consumption compared with
schools that did not implement a norm-based campaign. However, others have doubted
the effect that the social norms approach has on drinking attitudes, behaviors, and
intentions (Wechsler et al., 2003). This study provides evidence that norms as part of
persuasive advertisements have strong associations with attitudes toward heavy drinking
and can be inﬂuential when employed in social norms campaigns.

The results of this study should be considered in light of the environment in which
it was conducted. This study was conducted at a university where a social norms
campaign has been in place for six years. All participants in this study have been exposed
to an ongoing normative campaign throughout their college careers. Therefore, any effect
on attitudes and behavioral intentions that occurred was above and beyond any changes

that have already taken place due to this pre-existing campaign.

35

Linkenbach (2003) has developed a model for social norms campaigns based on
reducing alcohol related crashes in Montana. This model consists of the steps necessary
to take a campaign from the strategic planning steps to evaluation. Linkenbach (2003)
notes that during this process it is very important that the messages are delivered often
enough to achieve saturation and create audience awareness. It takes approximately eight
months to one year to carry out a normative campaign and evaluate its effectiveness
(Linkenbach, 2003).

This ﬁ'amework for a successful social norms campaign provides interesting
insight into the current study. Participants were only exposed to the campaign message
once and their attitudes and behavioral intentions were measured immediately following.
The experimental normative messages were able to affect both attitudes and behavioral
intentions after only one exposure. Therefore, if the same campaign messages were
delivered repeatedly in an extensive social norms campaign then the ﬁndings of this
study might be even more pronounced.

Further, a stronger relationship between advertisement believability and attitudes
and behavioral intention might be evident in an extensive campaign. Even though this
study did not ﬁnd evidence that advertisement believability mediated the relationship
between message congruency and attitudes or behavioral intentions, other studies have
found that advertisement believability plays a crucial role in estimating peer alcohol
consumption (Park et al. 2007b). Therefore, a longer social norms campaign might ﬁnd
that advertisement believability does mediate the relationship between message

congruency and behavioral intentions and between message congruency and attitudes.

36

As expected, there was a signiﬁcant relationship between message congruency
and advertisement believability. Those who received two congruent messages found the
advertisements more believable than two incongruent messages. Although this result may
seem intuitive, it is an important ﬁnding. Other studies have pointed out that college
students receive incongruent messages about alcohol consumption (e. g. Yanovitzky &
Stryker, 2001; Peralta, 2002; Ross, 2004; Nygaard & Grube, 2005). Therefore, if students
receive incongruent normative messages about alcohol consumption, this message
environment is going to have detrimental affects on the believability of normative
campaigns produced by universities.

An interesting ﬁnding of this study was the association between norm type and
believability. The relationship between descriptive norms and advertisement believability
was not statistically signiﬁcant but the relationship between injunctive norms and
advertisement believability was. This result corroborates studies that have found that the
types of norms are not only conceptually different but that they are also related to other
variables differently (Klein & Boster, 2006; Park & Smith, 2007; Park et al. 2007a).

The design of this study also allowed for an additional investigation into Schultz
et al.’s (2007) ﬁnding. Although the results of this study did not replicate Schultz et al.’s
(2007) ﬁnding, one possible explanation for this may be explained by the type of
behavior studied. Energy consumption is a very ambiguous behavior. It is difﬁcult to
gauge, not only the energy usage of neighbors, but also how much energy consumption is
approved. Further, the behavior is enacted in a private setting. Alcohol consumption,
however, is a much more public and a less ambiguous behavior (Lapinski & Rimal,

2005). It is a lot easier to observe normative behavior and approval around alcohol

37

consumption compared to energy consumption. Lapinski and Rimal (2005) assert that
behavioral attributes like privacy and ambiguity effect normative inﬂuence. When a
situation is public and unambiguous, norms have a much more powerful effect as people
are more susceptible to descriptive normative forces than when a situation is private and
ambiguous (Lapinski & Rimal, 2005). Therefore, the difference in the behavioral
attributes of alcohol consumption and energy consumption may yield a difference in
ﬁndings.

Although the behavior of drinking is more public and ambiguous, the messages
that participants received in the present study were more ambiguous than the messages in
the Schultz et al. (2007) study. This may be an alternate explanation for the difference in
ﬁndings. Participants were given an actual point estimate of the energy consumption
norm in the Schultz et al. (2007) study. The present study, however, was not as precise
and offered participants a range of normative behavior. Therefore, the message was more
ambiguous and might not have produced what Schultz et al. (2007) termed a boomerang
effect as participants did not have an exact ﬁgure to strive to reach. An interesting
replication of the present study would be to supply participants with exact ﬁgures about
their peers’ alcohol consumption. For instance, messages could be designed so that they
give participants the mean number of drinks that students at their university consume in
one sitting. Or a replication of the Schultz et al. (2007) study could be designed where
people are given ranges of normative energy use to see if this still produces what Schultz
et al. (2007) termed a boomerang effect. Therefore, the differences in ﬁndings between
the present study and the Schultz et al. (2007) may be due to the variation in either the

ambiguity of the behavior or the messages.

38

Limitations

Overall this study shows that the different norm types can have effects on
attitudes and that receiving incongruent messages can obfuscate these effects. Although
this study was an important initial investigation into how incongruent messages affect
believability and attitudes, it is not without limitations. One possible limitation of this
study is that the sample was a convenience sample. A larger, random sample might make
the results more generalizable to college students across the United States. It should be
noted, however, that unlike many studies in which college students are simply studied out
of ease, this study sampled college students because they were the most relevant
population for this area of study. Normative campaigns have largely focused on reducing
heavy episodic alcohol consumption among college students; therefore, a study looking at
the effects of incongruent normative messages about alcohol consumption makes sense
with a college student sample.

Another possible limitation of this study is the immediate measure of attitudes and
attitude change following the viewing of the normative advertisements. A future study
might want to look at multiple exposures to incongruent messages over time. The
measure of believability might also be considered a limitation. Although there was only
one indicator for this construct, this type of dichotomous measure has been used
effectively in past studies (e.g. Park et al., 2007b). Many researchers face the dilemma
that a one item measures creates because not only is it difﬁcult to create more than one
measure for certain constructs but survey length must also be considered in order to avoid

survey fatigue. Therefore, a replication of this study that includes more indicators for this

39

construct might further illuminate the relationship between believability and normative
messages.
Implications and Future Directions

Although this study is not without limitation it does have some important
implications for future normative campaigns. This study shows that when participants
receive two congruent normative messages, attitude change can be achieved. Many
normative campaigns and studies only employ one type of normative message, usually a
descriptive norm message. This study shows that not only do injunctive norms have
effects but that employing these messages together can be powerful. Future social norms-
based campaigns should look to employ both descriptive and injunctive norms jointly to
maximize results.

Because this study was an initial investigation into the effect of conﬂicting
normative messages, it can be used as a foundation for continued research on this topic.
Future studies might seek to replicate the ﬁndings of this study but with messages that are
framed differently. This study used the phrasing “ﬁve or more drinks” when delivering
injunctive norm messages and “four or fewer” when delivering descriptive norm
messages. A future study could look to change this wording to see if the effects still hold
or if the effects can be attributed to framing.

In addition to changing the wording of the messages, the wording of the attitude
measure could also be altered in a future study. The attitude measure in this study asked
participants about their attitude toward drinking ﬁve or more drinks in one sitting. Future

studies may want to also ask other attitude measures to see if the results still hold.

40

This study could be replicated, not only with differently worded messages and
attitude measures, but also with a different topic. Previous studies (e. g. Klein & Boster,
2006) have found that norms may operate differently with different types of behavior.
Therefore, replicating this study with other behaviors that norms have been applied to
such as organ donation (Park & Smith, 2007), reusing hotel towels (Goldstein,
Griskevicius & Cialdini, 2007), water conservation (Corral-Verdugo & F rias-Armenta,
2006), smoking (Gunther, Bolt, Borzekowski, Liebhart, & Dillard (2006), and rape
(Bohner, Siebler & Schmelcher, 2006) might be beneﬁcial and demonstrate the extent to

which the results of this study can be generalized across topic.

41

APPENDICES

42

APPENDIX A

Tables and Figures

43

Table 1: Design of the Study

Descriptive Norm

 

Injunctive Norm

Suggesting
drinking large

quantities

Null

Suggesting

drinking in

moderation

Suggesting
drinking in

moderation

Incongruent

Message

Single Message

Congruent

Message

Null

Single Message

No Message

Condition

Single Message

Suggesting
drinking in large

quantities

Congruent

Message

Single Message

Incongruent

Message

 

44

Table 2: Study Messages by Condition

 

 

Descriptive Norm
Suggesting Null Suggesting
drinking in drinking large
moderation quantities
Injunctive
Norm Suggesting “77% of XXX “23% of XXX
drinking students drink 4 or students drink 4 or
large fewer drinks when fewer drinks when
quantities they party.” they party”
“28% of XXX “28% of XXX “28% of XXX
students disapprove students students disapprove
of drinking 5 or disapprove of of drinking 5 or
more drinks when drinking 5 or more drinks when
they party” more drinks they party”
when they party”
1 2 3
Null “77% of XXX No Message “23% of XXX
students drink 4 or Condition students drink 4 or
fewer drinks when fewer drinks when
they party.” they party”
4 5 6
Suggesting “77% of XXX “72% of XXX “23% of XXX
drinking in students drink 4 or students students drink 4 or
Moderation fewer drinks when disapprove of fewer drinks when
they party.” drinking 5 or they party”
more drinks
when they party.”
“72% of XXX “72% of XXX
students disapprove students disapprove
of drinking 5 or of drinking 5 or
more drinks when more drinks when
they party.” they party.”
7 8 9

 

45

Table 3: Pre-Test Attitude Measure Inter-Item Correlations, Factor Loadings, and
Descriptive Statistics

 

 

l 2 3 4 5 6 7 FL Mean St.
Dev
.82 3.402 1.360
.666" .77 4.234 1.510
.612“ .620” .77 4.701 1.692
.623" .568" .618" .77 3.309 1.253
.661 ** .543" .573" .663“ .77 3.399 1.250
.573" .599" .582'MI .517“ .541" .73 3.651 1.429
.548" .500M .503" .480"”'I .525“ .513"”ll .66 2.674 1.318

 

Note: Factor Loading is abbreviated as FL
** Correlation is statistically signiﬁcant at the .01 level (2-tailed)

46

Table 4: Post-Test Attitude Measure Inter-Item Correlations, Factor Loadings, and
Descriptive Statistics

 

 

l 2 3 4 5 6 7 FL Mean St.
Dev
1 .87 3.172 1.361
2 .758M .85 3.812 1.593
3 .564" .591" .68 4.565 1.686
4 .658""'I .615“ .551” .76 3.100 1.252
5 .786“ .705" .563M .696“ .87 3 .222 1.242
6 .552" .641 ** .502” .455" .567" .68 3.512 1.445
7 .592" .516" .411" .492" .574M .456" .65 2.686 1.307

 

Note: Factor Loading is abbreviated as FL
** Correlation is statistically signiﬁcant at the .01 level (2-tailed)

47

Table 5: Pre-test Behavioral Intention Inter-Item Correlations, Factor Loadings, and
Descriptive Statistics

 

 

l 2 3 4 FL Mean St. Dev
1 .76 4.997 3.042
2 .700" .96 3.015 1.241
3 .762“ .876“ .94 2.973 1.230
4 .670“ .920" .832“ .91 3.077 1.187

 

Note: Factor Loading is abbreviated as F L
All means and standard deviations are shown prior to standardization
** Correlation is statistically signiﬁcant at the .01 level (2-tailed)

48

Table 6: Post-test Behavioral Intention Inter-Item Correlations, Factor Loadings, and
Descriptive Statistics

 

 

l 2 3 4 FL Mean St. Dev
1 .78 4.916 3.049
2 .710" .94 2.981 1.192
3 .781" .870" .94 2.874 1.195
4 .686" .906" .840“ .92 3.019 1.143

 

Note: Factor Loading is abbreviated as FL
All means and standard deviations are shown prior to standardization
** Correlation is statistically signiﬁcant at the .01 level (2-tailed)

49

Table 7: Descriptive Norm Perception Inter-Item Correlations, Factor Loadings, and
Descriptive Statistics

 

 

1 2 3 4 FL Mean St. Dev
1 .67 47.218 22.131
2 .558M .89 3.309 .975
3 .504" .647" .71 3.157 .944
4 .489M .652“ .467“ .71 3.031 .998

 

Note: Factor Loading is abbreviated as FL
All means and standard deviations are shown prior to standardization
** Correlation is statistically signiﬁcant at the .01 level (2-tailed)

50

Table 8: Injunctive Norm Perception Inter—Item Correlations, Factor Loadings, and
Descriptive Statistics

 

 

1 2 3 4 FL Mean St. Dev
1 .60 46.235 23.1 17
2 .410“ .74 3.302 .908
3 .491** .593” .78 3.321 .917
4 .444" .544" .513" .70 3.092 .866

 

Note: Factor Loading is abbreviated as FL
All means and standard deviations are shown prior to standardization
** Correlation is statistically signiﬁcant at the .01 level (2-tai1ed)

51

Table 9: Normative Perception Means

 

Descriptive Norm

 

Moderate
Drinking
Moderate Drinking
DN Perception -1.428 (2.651)
IN Perception -.622 (2.856)
No Message
Injunctive DN Perception —.221(3.344)
Norm IN Perception -.512 (2.783)
Heavy Drinking
DN Perception .307 (4.052)

IN Perception -3.29 (3.825)

No Message Heavy Drinking

-1473 (3.81) .097 (2.892)

-.769 (3.620) -.040 (2.966)

.195 (2.948) .536 (3.160)

.138 (3.174) .112 (3.290)

1.420 (2.877) .536 (3.160)

1.267 (2.854) .837 (2.699)

 

Note: Numbers in parentheses are standard deviations.

52

Table 10: Norm Type and Post-Test Attitude Measure Means

 

 

 

Descriptive Norm Mean
Moderate Heavy
Drinking Drinking

Injunctive Moderate 23.607 23.633 23.140
Norm Drinking (8.029) (8.771) (8.363)
Heavy 24.032 27.103 25.517
Drinking (9.871) (7.451) (8.848)
Mean 23.356 25.339 24.348
(8.996) (8.266) (8.659)

 

Note: Numbers in parentheses are standard deviations.

53

Table 11: Summary of Regression Findings for Multiple Normative Message Conditions

 

 

B t p R Adjusted F p
R2
Attitude .914 .832 193.842 <.001
IN .107 2.816 .006
DN .084 2.204 .030
Behavioral .946 .892 323.740 <.001
Intention
IN .071 2.347 .021
DN .028 .908 .366
Behavioral .946 .892 484.941 <.001
Intention
(without DN)
IN .070. 2.317 .022

 

54

Table 12: Norm Type and Post-Test Behavioral Intention Means

 

Descriptive Norm Mean

 

Moderate Heavy

Drinking Drinking

 

Moderate -0.643 0.727 0.066

Injunctive Drinking (3.506) (3.644) (3.613)
Norm Heavy 0.618 0.955 0.781
Drinking (3.785) (3.318) (3.541)

Mean 0.020 0.839 0.429

(3.679) (3.459) (3.579)

 

Note: Numbers in parentheses are standard deviations.

55

Table 13: Norm Type and Post Test Attitude Measure Means for all Nine Conditions

 

 

 

Injunctive Norm Mean
Moderate None Heavy
Drinking
Moderate 22.607 22.594 24.032 23.088
Descriptive Drinking (8.029) (7.910) (9.871) (8.593)
Norm None 22.964 25.179 24.571 24.238
(7.584) (8.287) (6.17) (7.369)
Heavy 23.633 23.071 27.103 24.609
Drinking (8.771) (7.702) (7.451) (8.1 15)
Mean 23.081 23.568 25.216 23.962
(23.568) (7.955) (8.068) (8.053)

 

Note: Numbers in parentheses are standard deviations.

56

Table 14: Norm Type and Attitude Chang Means for all Nine Conditions

 

Injunctive Norm

 

 

Moderate None Heavy Mean
Drinking

Moderate -2.500 -2.438 -1.387 -2. 100
Descriptive Drinking (4.308) (4.247) (3.667) (4.069)
Norm None -2.107 -0.036 -0.857 -1.000
(3.563) (3.574) (1.860) (3.185)
Heavy -1.800 -1.357 0.690 -0.828
Drinking (2.696) (3.058) (3.496) (3.250)
Mean -2.128 -1.330 -0.534 -1.324
(3.534) (3.784) (3.230) (3.569)

 

Note: Numbers in parentheses are standard deviations.

57

Table 15: Norm Type and Behavioral Intention Means for all Nine Conditions

 

Injunctive Norm Mean

 

Moderate None Heavy

Drinking
Moderate -0.643 -0.901 0.61 8 -0.304
Descriptive Drinking (3.506) (3.612) (3.785) (3.662)
Norm None -1.012 0.390 0.252 -0.123

(3.234) (3.946) (3.779) (3.677)

 

Heavy 0.727 -0410 0.955 0.437
Drinking (3.644) (4.139) (3.318) (3.714)
Mean -O.285 -0.334 0.612 0.000

(3.512) (3.884) (3.605) (3.684)

 

Note: Numbers in parentheses are standard deviations.

58

Table 16: Percentage of Participants Who Find Each Advertisement Believable by
Congruency

 

 

Message Type
Incongruent Congruent Message
Message
Advertisement No 59.02 35.71
Believability Yes 40.98 64.29

 

59

Table 17: Post-Test Attitude Means based on Message Type and Advertisement

 

 

 

Believability
Message Condition Mean
Congruous Incongruous Congruous
Moderate Heavy
Advertisement No 23.769 25.61 1 26.857 25.339
Believability (5.718) (9.421) (6.594) (25.339)
Yes 21.600 21.280 26.905 23.295
(9.694) (8.595) (7.930) (8.911)
Mean 22.607 23.836 26.893 24.274
(8.029) (9.270) (7.500) (8.659)

 

Note: Numbers in parentheses are standard deviations.

60

Table 18: Post-Test Behavioral Intention Means based on Message Type and

Advertisement Believability

 

 

 

Message Condition Mean
Congruous Incongruous Congruous
Moderate Heavy
Advertisement No -0.160 1.840 0.375 1.193
Believability (3.300) (-1.012) (.956) (3.527)
Yes -1.061 -1.012 0.956 -0.347
(3.737) (3.300) (3.265) (3.475)
Mean -0.643 0.671 0.811 0.390
(3.506) (3.685) (3.285) (3.569)

 

Note: Numbers in parentheses are standard deviations.

61

APPENDIX B

Experimental Measures

62

Main Study
Pre-Test

63

Note: Items marked with a * were used in the analysis

We are in the process of evaluating a series of public service announcements (PSAs)
about topics that are pertinent to college students. We are interested in your opinion
about potential messages for these PSAs. You will be supplied with sample
advertisements and then answer questions about them. Please answer every
question honestly as your input is valued. There is a front and back to each page.

Before you evaluate advertisements, we’d like to collect some background information.
Demographics

*1. Sex: D Male [:l Female

*2. Age: 1:118 D19 D20 [:1 21 1:122 D23 D24 [:IOther
*3. Racial Identiﬁcation: [:1 Black 1:] White [:| Asian or Paciﬁc Islander 1:]
Hispanic or Latino [:1 American Indian or Alaskan Native [:1 Other
*4. Year in School: 1:] 1St year undergraduate 1:] 2“d year undergraduate 1:] 3rd year

 

 

undergraduate D 4th year undergraduate [:1 5th year or more undergraduate 1:] Graduate
or professional [3 Other

 

*5. Are you a member of a social fraternity or sorority? (National Interfratemity
Conference, National Panhellenic Conference, or National Pan-Hellenic Council) 1:]
Yes D No

For the next few questions, please consider one drink as approximately a 4 ounce
glass of wine, a 12 ounce bottle or can of beer, a 12 ounce bottle or can of wine
cooler or a shot of liquor straight or in a mixed drink.

Behavior

6. In the last month, on how many different occasions did you drink alcohol? -

 

7. On a normal night out, how many drinks do you consume?

 

8. Thinking back to the last time that you partied/socialized, how many drinks did you
have, if any

9. On average how long do you usually party when you go out, i.e., how many hours is it
from the time that you start drinking until the time you stop drinking for the night?

 

64

10. I often pace my alcohol consumption when I go out partying (1 per hour).

[:1 Strongly agree 1:] Agree 1:] Don’t agree or disagree [:1 Disagree 1:] Strongly disagree
1 1. I seldom drink 4 or fewer drinks when I go out partying. (Recoded)

[:1 Strongly agree E] Agree [:1 Don’t agree or disagree [:1 Disagree [3 Strongly disagree
12. I drink moderately or not at all when I go out partying.

[:1 Strongly agree 1:) Agree [I Don’t agree or disagree 1:] Disagree 1:] Strongly disagree
13. I often drink excessive amounts of alcohol (Recoded)

[:1 Strongly agree 1:] Agree 1:] Don’t agree or disagree [:1 Disagree 1:] Strongly disagree
14. I usually drink 4 or fewer drinks when I socialize with friends.

[:1 Strongly agree [:1 Agree [:1 Don’t agree or disagree 1:] Disagree [:1 Strongly disagree
15. When I go out, I usually drink heavy amounts of alcohol. (Recoded)

[:1 Strongly agree [:1 Agree [:1 Don’t agree or disagree [:1 Disagree El Strongly disagree
Descriptive Norm Perceptions

16. Think of all MSU undergraduate students. What percentage do you think have 5 or
more drinks in one sitting at least once during a normal week?

 

Frequency of Drinkﬂg

17. Within the last 30 days, how‘often do you think the typical MSU student drank
alcohol?

Descriptive Norm Perceptions

18. Most MSU students have 5 or more drinks when they party. (Recoded)

1:] Strongly agree [I Agree [:1 Don’t agree or disagree [:1 Disagree 1:] Strongly disagree
19. Most MSU students have 4 or fewer drinks on one occasion when they party.

1:] Strongly agree [I Agree [:1 Don’t agree or disagree [:I Disagree 1:] Strongly disagree

65

20. On a normal night out, most MSU students have 5 or more drinks when they
socialize. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
Personal Approval

21. I ﬁnd heavy drinking to be a acceptable activity. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
22. I think that excessive drinking is an activity that one should avoid.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
23. I have ethical objections to binge drinking.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
Iniunctive Norm Perceptiogg

24. Think of all MSU undergraduate students. What percentage do you think consider

drinking 5 or more drinks in one sitting at least once during a normal week to be an
acceptable activity?

 

25. Most MSU students would approve of me limiting my alcohol consumption to 4 or
fewer drinks when I party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

26. Most MSU students would endorse my decision to have 5 or more drinks when I

party. (Recoded)
D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

27. Most MSU students would approve of drinking 5 or more drinks in one sitting.
(Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
28. Most MSU students think that I should have 4 or fewer drinks when I party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

66

Attitudes

Please place a check mark (/) on the line that bests corresponds to your opinion.

29. I think that drinking 5 or more alcoholic drinks in one sitting is

* A Good Thing
*Acceptable
*Not Fun

*Something that
people should do

Responsible
Not Relaxing
Necessary
*Good
Unwise
*Unfavorable

*Beneﬁcial

 

 

 

 

 

 

 

 

 

 

 

67

A Bad Thing (Recoded)
Unacceptable (Recoded)
Fun

Something that people
should not do (Recoded)

Irresponsible
Relaxing
Not Necessary

Bad (Recoded)

Wise

Favorable (Recoded)

Not Beneﬁcial (Recoded)

30. I think that drinking 4 or fewer alcoholic drinks in one sitting is

 

 

 

 

 

 

 

 

 

 

* A Good Thing A Bad Thing (Recoded)
*Acceptable Unacceptable (Recoded)
*Not Fun Fun

*Something that Something that people
people should do should not do (Recoded)
Responsible Irresponsible

Not Relaxing Relaxing

Necessary Not Necessary

*Good Bad (Recoded)

Unwise Wise

*Unfavorable Favorable (Recoded)
*Beneﬁcial Not Beneﬁcial (Recoded)

 

Please place a check mark (3’) next to your response.
Behavioral Intention

*31. The next time that you party/socialize with friends, how many drinks do you think
you will have

* 32. I intend to limit my alcohol consumption to 4 or fewer drinks the next time that 1

party.
D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
* 33. I plan to drink 5 or more drinks the next time that 1 party. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
*34. I will limit my alcohol consumption to 4 or fewer drinks the next time that 1 party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

68

Perceived Behavioral Control

35. It will be easy to limit my alcohol consumption to 4 or fewer drinks the next time that
1 party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

36. I am conﬁdent that I will drink 5 or more drinks the next time that 1 party. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

37. I can limit my alcohol consumption to 4 or fewer drinks the next time that 1 party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

38. Even if I wanted to, I do not think that I could limit my alcohol consumption to 4 or
fewer drinks the next time that 1 party. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
39. Sometimes I go out and drink more than I really want to. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
40. I often feel pressure to drink when I go out partying. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
41. If my friends are drinking, I feel that I should keep up with them.(Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
42. I feel pressured to drink more than I ought to drink. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
Fake Questions

43. I plan to donate blood in the next few months. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
44. I think that donating blood is something that everyone should do.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

69

45. Donating blood is not very important.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
46. I have a lot of respect for people who donate blood.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

47. I think that donating blood is

 

 

 

 

A Good Thing A Bad Thing (Recoded)
Acceptable Unacceptable (Recoded)
Not Fun Fun

Something that Something that people
people should do Should not do (Recoded)
Responsible Irresponsible

 

48. I plan to wear my seatbelt the next time that I ride in a car.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
49. I think that wearing your seatbelt is something that everyone should do.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
50. Wearing a seatbelt is not very important.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
51. I have a lot of respect for people who always wear their seatbelt.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

70

52. I think that wearing a seatbelt is

A Good Thing

 

Acceptable

 

Not Fun

 

Something that
people should do

 

Responsible

 

71

A Bad Thing (Recoded)
Unacceptable (Recoded)
Fun

Something that people
should not do (Recoded)

Irresponsible

Experimental Stimulus Questions

72

You are now going to see a few advertisements on the screen in the front of the room.
After each ad is placed on the screen please answer the following questions.

Advertisement 1: Please consider the advertisement on the screen.
Advertisement Believability

*I found this ad to be believable D yes D no

Messaga Characteristics

This advertisement presents new information D yes D no

The message that said “23% of MSU students drink 4 or fewer drinks when they party”
seems believable.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

The message that said “72% of MSU students disapprove of drinking 5 or more drinks
when they party” seems believable.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
It seems reasonable that 23% of MSU students drink 4 or fewer drinks when they party.
D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

It seems reasonable that 72% of MSU students disapprove of drinking 5 or more drinks
when they party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
Advertisement 2: Please consider the advertisement on the screen.

Fake Questions

I found this ad to be believable D yes D no

This advertisement presents new information D yes D no

It is believable that one pint of blood can save up to 3 lives.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

73

I think donating blood is something that everyone Should do.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
I donate blood as often as I can remember.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
How many times did you donate blood in the last year?

Advertisement 3: Please consider the advertisement on the screen.

I found this ad to be believable D yes D no

This advertisement presents new information D yes D no

The message made me think that I should wear my seatbelt.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
I always wear my seatbelt.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

It is believable that if you do not wear your seatbelt then there may be dangerous
consequences.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

74

Post — Test

75

Descriptive Norm Perceptions

*1. Think of all MSU undergraduate students. What percentage do you think have 5 or
more drinks in one sitting at least once during a normal week?

 

Frequency of Drinking

2. Within the last 30 days, how often do you think the typical MSU student drank
alcohol?

Descriptive Norm Perceptions

*3. Most MSU students have more than 5 drinks in when they party. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
*4. Most MSU students have 4 or fewer drinks in one occasion when they party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

*5. On a normal night out, most MSU students have 5 or more drinks when they
socialize. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
Personal Approval

6. I ﬁnd heavy drinking to be a acceptable activity.(Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
7. I think that excessive drinking is an activity that one should avoid.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
8. I have ethical objections to binge drinking.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
Iniunctive Norm Perceptions

*9. Think of all MSU undergraduate students. What percentage do you think consider

drinking 5 or more drinks in one sitting at least once during a normal week to be an
acceptable activity?

 

76

10. Most MSU students would approve of me limiting my alcohol consumption to 4 or
fewer drinks when I party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

*1]. Most MSU students would endorse my decision to have 5 or more drinks when I
party. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

*12. Most MSU students would approve of drinking 5 or more drinks in one sitting.
(Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
*13. Most MSU students think that I should have 4 or fewer drinks when I party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

77

Attitudes
Please place a check mark (V) on the line that bests corresponds to your opinion.

14. I think that drinking 5 or more alcoholic drinks in one sitting is

 

 

 

 

 

 

 

 

 

 

* A Good Thing A Bad Thing (Recoded)
*Acceptable Unacceptable (Recoded)
*Not Fun Fun

*Something that Something that people
people should do Should not do (Recoded)
Responsible Irresponsible

Not Relaxing Relaxing

Necessary Not Necessary

*Good Bad (Recoded)

Unwise Wise

*Unfavorable Favorable (Recoded)
*Beneﬁcial Not Beneﬁcial (Recoded)

 

78

15. I think that drinking 4 or fewer alcoholic drinks in one sitting is

* A Good Thing
*Acceptable
*Not Fun

*Something that

people should do _ __

Responsible
Not Relaxing
Necessary
*Good
Unwise
*Unfavorable

*Beneﬁcial

 

 

 

 

 

 

 

 

 

79

A Bad Thing (Recoded)
Unacceptable (Recoded)
Fun

Something that people
should not do (Recoded)

Irresponsible
Relaxing
Not Necessary

Bad (Recoded)

Wise

Favorable (Recoded)

Not Beneﬁcial (Recoded)

Behavioral Intention
Please place a check mark (/) next to your response.

*16. The next time that you party/socialize with friends, how many drinks do you think
you will have

*17. I intend to limit my alcohol consumption to 4 or fewer drinks the next time that 1

party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
* 18. I plan to drink 5 or more drinks the next time that 1 party. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
*19. I will limit my alcohol consumption to 4 or fewer drinks the next time that 1 party.
D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
Perceived Behaviorai Control

20. It will be easy to limit my alcohol consumption to 4 or fewer drinks the next time that

I party.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
21. I am conﬁdent that I will drink 5 or more drinks the next time that 1 party. (Recoded)
D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
22. I can limit my alcohol consumption to 4 or fewer drinks the next time that I party.
D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

23. Even if I wanted to, I do not think that I could limit my alcohol consumption to 4 or
fewer drinks the next time that 1 party. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
24. Sometimes I go out and drink more than I really want to. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

80

25. I often feel pressure to drink when I go out partying. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
26. If my friends are drinking, I feel that I should keep up with them. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
27. I feel pressured to drink more than I ought to drink. (Recoded)

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
Fake Questions

28. I plan to donate blood in the next few months.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
29. I think that donating blood is something that everyone should do.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
30. Donating blood is not very important.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
31. I have a lot of respect for people who donate blood.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

32. I think that donating blood is

A Good Thing _______ A Bad Thing (Recoded)
Acceptable _______ Unacceptable (Recoded)
Not Fun _______ Fun

Something that Something that people
people should do _______ should not do (Recoded)
Responsible _______ Irresponsible

81

 

33. I plan to wear my seatbelt the next time that I ride in a car.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
34. I think that wearing your seatbelt is something that everyone should do.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
35. Wearing a seatbelt is not very important.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree
36. I have a lot of respect for people who always wear their seatbelt.

D Strongly agree D Agree D Don’t agree or disagree D Disagree D Strongly disagree

37. I think that wearing a seatbelt is

 

 

 

 

A Good Thing A Bad Thing (Recoded)
Acceptable Unacceptable (Recoded)
Not Fun Fun

Something that Something that people
people should do should not do (Recoded)
Responsible Irresponsible

 

82

 

APPENDIX C

Study Advertisements

83

Figure 1: Condition 1 Advertisement

B [C A S PA R rlI\ N 1

g’:». 1 .
i ‘3

° 77% (11' \lSlr 51111101114 (Irink -+ (11' l'mu‘r drinks

“11011 Hwy purl).

° Zi'il’fnnl'NlHlalrlilt‘nls(lisilmn‘nw‘111(11‘111k111‘2‘

3111' 111(11‘1Hll'1111x'h’ “11111 Hwy purl}:

All}: ”1945‘ $15.11-
u h. 1 \ r n a 1 1 V

 

Note: The ﬁne print above states, “National Health Assessment; MSU 2006. N =
1005, margin of error 3%.”

84

Figure 2: Condition 2 Advertisement

° 21:‘i‘,"01)1‘.\1317sllldt‘lllndisujilirtn't-til-drinking

3(11‘1111t1'tt111'i11kn “11111111113 [)211'1}.

 

Note: The ﬁne print above states, “National Health Assessment; MSU 2006. N = 1005,
margin of error 3%.”

85

 

Figure 3: Condition 3 Advertisement

BE A SPAR'I‘AN!

A

° 2390 111' XIHIi students drink -t (11‘ fewer drinks

when they party.
0 28111) ()1 MSU students disapprove ul' drinking
3 or 111111‘e drinks when they party.

mung.“ sun.
l V I \‘l R S I I Y

 

Note: The ﬁne print above states, “National Health Assessment; MSU 2006. N = 1005,
margin of error 3%.”

86

Figure 4: Condition 4 Advertisement

B IS A Sl’.‘\li71‘:\ Ni

° 77%. ()1 \1518 students drink -t (11' fewer drinks

\\'11e11111e_\ purl);

 

Note: The ﬁne print above states, “National Health Assessment; MSU 2006. N = 1005,
margin of error 3%.”

87

Figure 5: Condition 6 Advertisement

when they party.

MICHIGAN sl.\_ll
t- s r \ r R \111

 

Note: The ﬁne print above states, “National Health Assessment; MSU 2006. N = 1005,
margin of error 3%.”

88

' 77% (>1 \1518 students drink -1 (11' fewer drinks
when they p211‘t_\‘.

( I

7}. In ()1 \IHl' students disupprnve 111' drinking

3111'111111'edi'i11ks when they purl);

MICHIGAN STATE
11 N 1 v t a s 1 r v

 

Note: The ﬁne print above states, “National Health Assessment; MSU 2006. N = 1005,
margin of error 3%.”

89

Figure 7: Condition 8 Advertisement

° 7255111110151studentsdisupprtwenl‘drinkin‘:

3(1r111nredri11ks when they party.

 

“KNOW SIAII
l . l l 5 I I

 

Note: The ﬁne print above states, “National Health Assessment; MSU 2006. N = 1005,
margin of error 3%.”

90

Figure 8: Condition 9 Advertisement

L

' 23% (11' \151' students drink -1 (11' fewer drinks

when the} party.

° 72% (11' \IHL students disupprtn'e ul' drinking

 

3 Hr 111111'edri11ks when they pin‘t}.

 

Note: The ﬁne print above states, “National Health Assessment; MSU 2006. N = 1005,
margin of error 3%.”

91

Figure 9: Donating Blood Advertisement

 

92

Figure 10: Seatbelt Advertisement

BUCKLE UP ’W’

' 0151 PAY 1HE PRICE

7 , é m Sara ofM‘rchigon
Trafﬁc Sabw Cemmiorion

 

93

REFERENCES

Allen, M. (1991). Meta-Analysis comparing the persuasiveness of one-Sided and two-
sided messages. Western Journal of Speech Communication, 55, 390-404.

American College Health Association (2006). American College Health Association —
National College Health Assessment (ACHA-NCHA) Web Summary. Retrieved
January 11, 2007 from
http://www.acha.org/projects_programs/ncha_sampledata.cﬁn

Anderson, N.H. & Hovland, CL (1957). The representation of order effects in
communication research. In C.l. Hovland (Ed.), The order of presentation in
persuasion (pp. 158-169). New Haven: Yale University Press.

Bartholow, B. D., Sher, K. J ., & Krull, J. L. (2003). Changes in heavy drinking over the
third decade of life as a function of collegiate fraternity and sorority involvement:
A prospective, multilevel analysis. Health Psychology, 22, 616—626.

Billingham, D. (Nov. 11- Nov 17, 2006). Does the world need another AIDS authority?
The Lancet, 368, 1639-1641.

Bohner, G., Siebler, F., & Schmelcher, J. (2006). Social norms and the likelihood of
raping: Perceived rape myth acceptance of others affects men's rape proclivity.
Personality and Social Psychology Bulletin, 32, 286-297.

Borsari, B. E., & Carey, K. B. (1999). Understanding fraternity drinking: Five recurring
themes in the literature, 1980—1998. Journal of American College Health, 48, 30—
37.

Borsari, B. & Carey, KB. (2003). Descriptive and injunctive norms in college drinking:
A meta-analytic integration. Journal of Studies on Alcohol, 64, 331-341.

Brashers, D.E., Goldsmith, D.J., & Hsieh, E. (2002). Information seeking and avoiding in
health contexts. Human Communication Research, 28, 258 - 271.

Cacioppo, J ., & Petty, R. (1979). The effects of message repetition and position on
cognitive response, recall, and persuasion. Journal of Personality and Social
Psychology, 37, 97-109.

Campo, S., Brossard, D., Frazer, S.M., Marchell, T., Lewis, D. & Talbot, J. (2003). Are
social norms campaigns really magic bullets? Assessing the effects of students’
misperceptions on drinking behavior. Health Communication, 15, 481-497.

Cashin, J. R., Presley, C. A., & Meilman, P. W. (1998). Alcohol use in the Greek system:
Follow the leader? Journal of Studies on Alcohol, 59, 63—70.

94

Cialdini, R.B., Reno, R.R. & Kallgren, CA. (1990). A focus theory of normative
conduct: Recycling the concept of norms to reduce littering in public places.
Journal of Personality and Social Psychology, 58, 1015-1026.

Clapp, J .D., Lange, J.E., Russell, C., Shillington, A. & Voad, RB. (2003). A failed norms
social marketing campaign. Journal of Studies on Alcohol, 64, 409 — 414.

Corral-Verdugo, V. & Frias-Armenta, M. (2006). Personal normative beliefs, antisocial
behavior, and residential water conservation. Environment and Behavior, 38, 406-
421.

DeJong, W., Schneider, S.K., Towvim, L.G., Murphy, M.J., Doerr, E.E., Simonsen, N.R.
et al. (2006). A multisite randomized trial of social norms marketing campaigns to
reduce college student drinking. Journal of Studies on Alcohol, 67, 868-879.

Elek, E., Miller-Day, M. & Hecht, M. (2006). Inﬂuences of personal, injunctive and
descriptive norms on early adolescent substance use. The Journal of Drug Issues,
36, 147-172.

Fabiano, P., Perkins, H.W., Berkowitz, A.B., Linkenbach, J. & Stark, C. (2004).
Engaging men as social justice allies in ending violence against women: Evidence
for a social norms approach. Journal of American College Health, 52 (3), 105-
112.

Haines, M. & Spear, SF. (1996). Changing the perception of the norm: A strategy to
decrease binge drinking among college students. Journal of A merican College
Health, 45, 134-140.

Harkins, S.G., & Petty, R.E. (1981a). Effects of source magniﬁcation of cognitive effort
on attitudes: An information-processing view. Journal of Personality and Social
Psychology, 40, 401-413.

Harkins, S.G., & Petty, R.E. (1981b). The multiple source effect in persuasion: The
effects of distraction. Personality and Social Psychology Bulletin, 7, 627-633.

Hingson, R.W., Heeren, T., Zakocs, R.C., Kopstein, A. & Wechsler, H. (2002).
Magnitude of alcohol-related mortality and morbidity among US. college
students ages 18 — 24. Journal of Studies on Alcohol, 63, 136-144.

Goldstein, N.J., Griskevicius, V. & Cialdini, RB. (2007). Invoking social norms: A
social psychology perspective on improving hotels' linen-reuse programs. Cornell
Hotel and Restaurant Administration Quarterly, 48, 145-150.

Gunther, A., Bolt, D., Borzekowski, G., Liebhart, J. & Dillard, J. P. (2006). Presumed

inﬂuence on peer norms: How mass media indirectly affect adolescent smoking.
Journal of Communication, 56, 52-68.

95

 

The Higher Education Center (2000). A note to the ﬁeld: On ‘binge drinking.’ Retrieved
June 5, 2007 from http://www.higheredcenter.org/press-releases/OO1020.html.

Inter-Association Task Force on Alcohol and Other Substance Abuse Issues (2000).
Alcohol and other substance abuse issues’ proclamation. Retrieved June 1, 2007
from http://www.iatf.org/proc1a.htm

Klein, K. A., & Boster, F. J. (2006). Subjective, descriptive, and injunctive norms: Three
separate constructs. Paper presented at the annual conference of the International
Communication Association. Dresden, Germany.

Lapinski, M.K. & Rimal, RN. (2005). An explication of social norms. Communication
Theory, 15, 127-147.

Larimer, M., Turner, A., Mallett, K., and Geisner, I. (2004). Predicting drinking behavior
and alcohol-related problems among fraternity and sorority members: Examining
the role of descriptive and injunctive norms, Psychology of Addictive Behaviors.
18, 203-212.

Linkenbach, J. (2003). The Montana model: Development and overview of a seven-step
process for implementing macro-level social norms campaigns. In H. W. Perkins
(Ed.), The social norms approach to preventing school and college age substance
abuse(pp. 182—205). San Francisco: Jossey-Bass.

Mead, OH. (1932). Mind Self and Society ﬁom the Standpoint of a Social Behaviorist
Chicago: University of Chicago.

Meredith, K.L., Jeffe, D.B., Mundy, L.M. & Fraser, V.J. (2001). Source inﬂuencing
patients in their HIV medication decisions. Health Education and Behavior, 28,

40-50.

Neighbors, C., Oster-Aaland, L., Berstrom, R.L. & Lewis, MA. (2006). Event and
context-speciﬁc normative misperception and high-risk drinking: 21St birthday
celebrations and football tailgating. Journal of Studies on Alcohol, 67, 282 - 289.

Nygaard, P. & Grube, J .W. (2005). Mixed messages: Contributions to adolescent
drinking and driving. Addiction Research and Theory, 13, 411-426.

Olds, R.S., Thombs, D.L. & Tomasek, J .R. (2005). Relations between normative beliefs
and initiation intentions toward cigarette, alcohol and marijuana. Journal of
Adolescent Health, 37, 7 -— 13.

Park, H.S. & Smith, SW. (2007). Distinctiveness and inﬂuence of subjective norms,
personal descriptive and injunctive norms, and societal descriptive and injunctive
norms on behavioral intent: A case of two behaviors critical to organ donation.
Human Communication Research, 33, 194-218.

96

Park, H.S., Klein, K.A. & Smith, S. (2007a). The power of subjective norms, university
descriptive and injunctive norms, and US. descriptive and injunctive norms on
drinking behavior intentions. Paper presented at the International Communication
Association Conference in San Francisco, CA.

Park, H.S., Klein, K.A. & Smith, S. (2007b). College students’ estimation and accuracy
of other students’ drinking: Believability of advertisements featured in a social
norms campaign. Paper to be presented at the National Communication
Association Conference in Chicago, IL.

Peralta, R.L. (2002). Alcohol use and the fear of weight gain in college: Reconciling two
college norms. Gender Issues, 20, 23-42.

Perkins, H. W. & Berkowitz, A. D. (1986). Perceiving the community norms of alcohol
use among students: Some research implications for campus alcohol education
programming. International Journal of the Addictions, 21, 961-976

Perkins, H.W. and Craig, D.W. (2006). A successful social norms campaign to reduce
alcohol misuse among college student-athletes. Journal of Studies on Alcohol, 67,
868-879.

Polonec, L.D., Major, A.M, & Atwood, LE. (2006). Evaluating the believability and
effectiveness of the social norms message “Most Students Drink 0 to 4 Drinks
When They Party.” Health Communication, 20, 23-34.

Reno, R.R., Cialdini, R. & Kallgren, CA. (1993). The transituational inﬂuence of social
norms. Journal of Personality and Social Psychology, 64(1), 104 -112.

Ross, V. (2004). Depression, anxiety, and alcohol and other drug user among college
students. Retrieved November 30, 2006 ﬁ'om
http://www.higheredcenter.org/pubs/prev-updates/depression.html.

Schultz, P.W., Nolan, J.M., Cialdini, R.B., Goldstein, N.J. & Griskevicius, V. (2007).
The constructive, destructive, and reconstructive power of social norms.
Psychological Science, 18, 429 — 434. '

Sherif, C. W., Sherif, M., & Nebergall, R. E . ( 1965). Attitudes and attitude change: The
social judgment-involvement approach. Philadelphia: W. B. Saunders.

Smith, S. W., Atkin, C.K., Martell, D., Allen, R. & Hembroff, L. (2006). A social
judgment theory approach to conducting formative research in a social norms
campaign. Communication Theory, 16, 141-152.

97

Trockel, M., Williams, S.S., Reis, J. (2003). Considerations for more effective social
norms based alcohol education on campus: An analysis of different theoretical
conceptualizations in predicting drinking among fraternity men. Journal of
Studies on Alcohol, 64, 50-59.

Wechsler, H. & Nelson, T.F. (2001). Binge drinking and the American college student:
What’s ﬁve drinks? Psychology of Addictive Behaviors, 15, 287-291.

Wechsler, H., Lee, J.E., Kuo, M., Seibring, M., Nelson, T.F. & Lee, H. (2002). Trends in
college binge drinking during a period of increased prevention efforts. Findings
from 4 Harvard School of Public Health College alcohol study surveys: 1993-
2001. Journal of American College Health, 50, 203-217.

Wechsler, H., Nelson, T.F., Lee, J.E., Seibring, M., Lewis, G, & Keeling, RP. (2003).
Perception and reality: A national evaluation of social norms marketing
interventions to reduce college students’ heavy alcohol use. Journal of Studies on
Alcohol, 64, 484-494.

Weiss, RP. (1971). Role playing and repetition effects on opinion strength. Journal of
Social Psychology, 85, 29-35.

Werch, C.E., Pappas, D.M., Carlson, J .M., DiClemente, C.C., Chally, P.S., & Sinder,
J.A.. (2000). Results of a social norm intervention to prevent binge drinking
among ﬁrst-year residential college students. Journal of American College
Health, 49, 85-92.

Wilson, L.R., & Miller, H. (1968). Repetition, order of presentation, and timing of
arguments and measures as determinants of opinion change. Journal of
Personality and Social Psychology, 9, 184-188.

Yanovitzky, 1., Stewart, L.P. & Lederman, LC. (2006). Social distance, perceived
drinking by peers, and alcohol use by college students. Health Communication,
19, 1-10.

Yanovitzky, l. & Stryker, J. (2001). Mass media, social norms, and health promotion
efforts: A longitudinal study of media effects of youth binge drinking.
Communication Research, 28, 208-239.

Yuksel, H. & Corbett, K.K. (2005). Mixed messages: A qualitative study of the meaning

and context of high school students’ tobacco use in turkey. Health Promotion
International, 20, 360-366

98

 

‘ The Journal of Studies on Alcohol and Drugs has adopted a policy in which the
term binge can only be used to describe a prolonged period of alcohol consumption
(usually two or more days) and a withdrawal from usual activities and obligations in
order to continue alcohol consumption (http://www.j sad.com/j sad/static/binge.html,

2007).

H An ANOVA was also conducted with attitude change as the dependent variable.
Again, the two main effects were statistically signiﬁcant. The results show a statistically
signiﬁcant effect of normative messages on attitude change, F (8, 253) = 2.825, p = .005,
n2 = .082, r - .286, d = .572. Both the descriptive norm, F (2 ,253) = 3.583, p = .029, n2 =
.028, r = .170, d = .341 and the injunctive norm messages, F (2, 253) = 4.713, p = .010,
112 = .036, r = .190, d = .379 affected attitude change. The interaction between descriptive
and injunctive norms on attitude change was not statistically signiﬁcant, F (4, 253) =
1.505, p = .201, n2 = .023, r = .152, d = .303. These results Show that the type of norm
message received has an impact on attitude change but that the normative messages do

not interact with each other. The means for this analysis are presented in Table 14.

99

 

   

1tittijjtittttjtjj111111