Download Stepping Back from Social Norms Campaigns

HEALTH COMMUNICATION, 20(3), 277–288
Copyright © 2006, Lawrence Erlbaum Associates, Inc.
CAMERONBACK
STEPPING
AND CAMPO
FROM SOCIAL NORMS CAMPAIGNS
Stepping Back from Social Norms Campaigns:
Comparing Normative Influences to Other Predictors
of Health Behaviors
Kenzie A. Cameron
Center for Communication and Medicine and Division of General Internal Medicine
Northwestern University
Shelly Campo
Department of Community and Behavioral Health and Department of Communication Studies
University of Iowa
Recent health campaigns on college campuses have used a social norms approach, which suggests that one’s perceptions of others’ attitudes and behaviors are the key components in attitude and behavior change. However, the efficacy of social norms campaigns has been mixed.
This study was conducted to assess the relationships among sociodemographics, normative
perceptions, and individual attitudes on 3 health behaviors. Students at 2 universities (N = 393)
completed questionnaires assessing how these variables related to their consumption of alcohol, tobacco use, and exercise behaviors. Regressions indicated that each of these variables was
associated with behavior, but varied independent variables emerged as the salient predictors
among behaviors. In several conditions the effect of normative perceptions on behaviors was
not significant, a finding in direct opposition to social norms marketing. In all 3 behavioral conditions, the variable accounting for the greatest variance was whether or not the individual liked
participating in that particular behavior. Thus, although some social norms marketing may be
meeting with success, it may be the case that predicted attitudinal and behavioral changes will
not be found when applied across diverse health topics.
Social norms marketing campaigns have become a favorite
among many universities attempting to discourage unhealthy behaviors, particularly those related to binge drinking. Recent research suggests that social norms campaigns
have been adopted by nearly half of all colleges and universities in the United States (Wechsler, Seibring, Liu, & Ahl,
2004) and have spread to state health departments (Frauenfelder, 2001). Social norms marketing was named an “idea
of the year” in 2001 by the New York Times Magazine and
has been described as “the science of persuading people to
go along with the crowd” (Frauenfelder, 2001, p. 100).
These campaigns are based on pluralistic ignorance (Allport, 1924; Fields & Schuman, 1976; Prentice & Miller,
Correspondence should be addressed to Kenzie A. Cameron, Center
for Communication & Medicine, Northwestern University, The Feinberg
School of Medicine, Division of General Internal Medicine, 676 North St.
Clair Street, Suite 200, Chicago, Illinois 60611. E-mail: k-cameron@north
western.edu
1993), which suggests that individuals are unable to accurately judge the social norm. The underlying assumption is
that once you correct the perceived norm so that it matches
the actual norm, individuals will alter their behavior accordingly (Perkins & Berkowitz, 1986).
Social norms marketing originates from a Perkins and
Berkowitz (1986) study indicating that students misperceived social norms regarding how much alcohol fellow students consumed. If the misperception were corrected, it was
believed behavior change would follow. The idea first was
used widely at Northern Illinois University in 1990, where
newspaper ads, posters, and handouts conveyed that most
students had five or fewer drinks when they partied. Northern
Illinois University and the University of Arizona, which implemented social norms campaigns along with other harm reduction approaches, reported significant declines in student
drinking behavior (Frauenfelder, 2001; Johannessen, Collins, Mills-Novoa, & Glider, 1999). Researchers differ in
their opinions of the efficacy of social norms campaigns.
278
CAMERON AND CAMPO
Some tout the observed reductions in college drinking (Perkins & Berkowitz, 1986; Perkins, Meilman, Leichlier, Cashin, & Presley, 1999; Stewart et al., 2002), whereas others
question the discovered effects (Campo et al., 2003; Campo
& Cameron, 2006; Campo, Cameron, Brossard, & Fraser,
2004; Keeling, 1999, 2000; Wechsler & Kuo, 2000; Wechsler et al., 2003). Research has begun to explore potential
moderators between normative perceptions and behavior to
better understand why social norms campaigns may succeed
or fail (Rimal & Real, 2003).
Yet, with so much current focus on social norms campaigns, as well as the influx of funds for such campaigns, administrators may have put the cart before the horse. Even
though social norms marketing often is implicitly presented
as being a theory unto itself, it is atheoretical and has not
been rigorously tested, as Keeling (1999, 2000) states in a series of editorials. Rather, common assumptions are advanced
in the social norms literature (e.g., people want to be like others, social norms are more salient than personal attitudes,
normative perceptions are related to actual behavior); these
assumptions are what public health researchers test (Campo
et al., 2003, Perkins & Berkowitz, 1986; Wechsler et al.,
2004). We propose the need to step back from the narrow focus of social norms campaigns to reexamine the influences,
normative and others, on college student health behaviors.
A social norms approach assumes attitude and behavior
change are influenced by one’s perception of others’ attitudes and behavior. Numerous persuasion theories suggest
that one’s own attitudes influence behaviors (cf. Ajzen,
1985; Kim & Hunter, 1993). The theory of reasoned action
(TRA; Ajzen & Fishbein, 1980) and the theory of planned
behavior (TPB; Ajzen, 1985) predict that normative beliefs
affect behavior. These theories consider other influences,
such as personal attitudes, and note that we are not always
motivated to comply with what others think we should do.
Demographic characteristics such as age, sex, and race also
have been identified as related to college student health behaviors (Weitzman, 2004). Other variables shown to be related to drinking behaviors include sex (e.g., Gomberg,
Schneider, & DeJong, 2000), fraternity or sorority membership (e.g., Larimer, Irvine, Kilmer, & Marlatt, 1997; Lo &
Globetti, 1993), race (e.g., Chassin & DeLucia, 1996), and
living arrangement such as residence halls or off-campus
housing (e.g., Olds & Thombs, 2001).
Perkins and Berkowitz’s (1986) seminal study examined
the influence of personal attitudes, various sociodemographics (e.g., sex, housing), and perceptions of the campus norm of
student drinking. They found that students’drinking behavior
was significantly related to sex, housing, personal attitudes toward drinking, and the discrepancy between personal attitudes
and perceptions of the campus norm. Their measure of normative perception was operationized by asking participants to
choose the statement they believed best represented the general campus attitude out of five statements that ranged from
very conservative (“Drinking is never a good thing to do”) to
very liberal (“A frequent ‘drunk’is okay if that’s what the individual wants to do”). These items measure perceptions of the
attitudes of fellow college students, not student behavior.
However, social norms campaigns such as those conducted at
Northern Illinois University and Arizona (Frauenfelder, 2001;
Gose, 1997; Haines & Spear, 1996; Johannessen et al., 1999)
present messages related to student behavior (i.e., number of
drinks students drink when they party, percentage of students
who drink less than once a week).
The nationwide College Alcohol Study reported by Perkins and Wechsler (1996) considered sociodemographics,
perceived campus alcohol norm, and personal alcohol attitude in predicting alcohol abuse among college students. The
perceived campus alcohol norm was measured using an index of responses to five Likert-type scale items (e.g., “Students here admire non-drinkers” and “Drinking is an important part of the college experience,” p. 965). These items are
better characterized as general attitude items and do not have
as direct a correspondence to the current social norms campaigns as would items querying the percentage of students
who believe drinking is an important part of the college experience or items that more specifically assess one’s belief
about the drinking behaviors of the general campus population. Perkins and Wechsler measured personal alcohol attitude using an index, the items of which asked students to indicate their response to various situations as to “How much
do you think is appropriate for a college student to drink in
each of the following situations” (p. 964). Nine situations
were provided (e.g., at a party, alone, on a week night), and
response categories were “no alcohol at all,” “only 1 to 2
drinks,” “enough to get high but not drunk,” or “enough to get
drunk” (p. 964). Again, these items do not correspond with
the social norms messages provided in current campaigns.
Fishbein and Ajzen (1975) stress the importance of scale
items being both specific to the context and the situation.
Such specificity increases accuracy and precision of measurement. We advocate the need for measures of multiple
specific normative judgments and attitudes. For any given
behavior, there are numerous normative judgments (e.g., how
many days per week do undergraduates exercise, how many
minutes do undergrads exercise per session, what percentage
of undergrads believe exercising regularly is important). The
variation in the influence of different normative judgments
may account for some of the mixed outcomes in social norms
studies, as they typically target only one normative judgment.
As social norms campaigns are focused on the behavior of
others, attitude measurements should include attitudes toward others’ attitude or behavior (e.g., undergraduates think
drinking to get drunk is a bad idea). In addition, items measuring one’s attitudes toward the behavior itself (e.g., I think
drinking to get drunk is a bad idea), attitudes related to one’s
liking of participating in the specific behavior (e.g., I like to
exercise), and attitudes related to the healthiness of the behavior (e.g., I believe that smoking is bad for health) increase
context specificity of the attitudinal measurement.
STEPPING BACK FROM SOCIAL NORMS CAMPAIGNS
Theories such as the TPB provide guidance to variables
believed to influence behavior. Recently, both the TRA and
the TPB have been used to predict behavioral intentions and
behavior in the health domain (e.g., Coleman, 2002; Greene,
Hale, & Rubin, 1997). Van Ryn, Lytle, and Kirscht (1996)
used the TPB and discovered that different models were predictive for the pro-health behaviors of breast self-exam and
exercise. Campo et al. (2003) used the TPB to examine subjective norms versus general campus norms of alcohol consumption. Yet little has been done to assess the major influences on college students regarding various health behaviors.
Social norms campaigns are being readily adopted and
transferred to other health topics such as smoking and
seatbelt use (Montana State University–Bozeman, 2004).
University administrations and the U.S. Department of Education continue to earmark extensive funds to develop
such campaigns on college campuses, particularly to decrease binge drinking (Higher Education Center for Alcohol and Other Drug Prevention, 2004). However, there is a
lacuna in the literature comparing influences across multiple health behaviors for college students, as well as a lack
of comparison of influences related to unhealthy behaviors
(e.g., smoking) and prohealth behaviors (e.g., exercise). Before more funds are spent on such controversial campaigns,
we must scrutinize the effects of normative judgments as
well as other factors on unhealthy and prohealth behaviors
of college students. We can then ascertain relevant factors
related to health behaviors, as well as learn the relative effects of sociodemographics, normative judgments, and attitudes on health behavior. Ergo, we propose these research
questions.
279
METHOD
Participants
Participants (N = 393) were drawn from populations at two
universities, one in the Southeast (n= 168) and one in the
Northeast (n = 225), and were students in introductory and
advanced communication courses. Communication majors
represented 26% of the sample. Fifty-three percent of the
participants were women and 47% were men. The age of
the participants ranged from 18 to 31 (M = 20.16, SD =
1.61). African American/Blacks comprised 6.6% of the
sample, 80.8% were White, 6.4% Asian/Asian American,
0.5% American Indian/Native American, 1.8% biracial,
2.6% Chicano/Hispanic, and 1.3% reported being of other
racial backgrounds.
Procedure
Participants received either extra credit or course credit for
their involvement in this study. They were administered the
questionnaire in groups. Participants arrived at the classroom
and were instructed that researchers wanted their responses
to a series of health-related messages under consideration for
use at their university. They were advised that the questionnaire would take approximately 30 min to complete and were
given one of six versions of the questionnaire. Following
completion of the questionnaire, participants were debriefed
and were given the researchers’ contact information should
any questions or concerns arise.
Questionnaire
RQ1: What are the sociodemographic, normative judgment, and attitudinal influences on college students’
health behaviors?
RQ2: What is the influence of normative judgments as
compared to other influences on college students’
health behaviors?
RQ3: How do sociodemographic, normative judgment, and
attitudinal influences compare across positive and
negative health behaviors?
Selecting two universities, one of which had an extensive social norms campaign related to binge drinking and
the other, which did not, we collected data to discover the
effects of sociodemographics, normative judgments, and attitudinal factors on personal behavior related to alcohol
consumption, smoking behavior, and exercise behavior. We
measured those normative judgments related to student behavior, which better parallel messages advanced by social
norms campaigns (e.g., Most students drink only four
drinks when they party). We also included multiple measures of attitudes to assess personal behavior as well as
other students’ behavior.
Each student read and responded to two of the three health
topics (i.e., alcohol use, smoking, exercise); we did not present all three topics to all participants to minimize fatigue.
Questionnaires were randomized based on topic order (i.e.,
alcohol use, smoking, exercise). Six versions were developed
so that each topic appeared in a questionnaire version first
and second and was paired with every other topic at least
once; each topic thus was given to two thirds of the participants (n = 260). The six versions were distributed in the same
order; the order was repeated with every seventh student. The
questionnaire included measures of the participants’ own behaviors related to alcohol use, smoking, and/or exercise; normative judgments of other undergraduates’ attitudes and
behaviors; and scales assessing their attitudes toward their
own health behaviors as well as their attitudes toward other
undergraduates’ behaviors. The questionnaire concluded
with sociodemographic questions.
Sociodemographics. In addition to sex, age and race,
participants were asked to indicate if they were a member of a
fraternity or sorority (Greek affiliation) and their living arrangement. Participants were asked if they lived in a resi-
280
CAMERON AND CAMPO
dence hall, off-campus house or apartment, program house
(e.g., thematic living/learning environments), fraternity or
sorority, university or cooperative housing, or other living arrangement. As some researchers have found that living in a
dormitory is related to reduced drinking among college students (Baer, Stacy, & Larimer, 1991; Weitzman, 2004), for
analyses this variable was recoded as dormitory or other. As
the number of minorities was rather small, the race variable
was recoded as White or other.
Health behaviors. All participants were asked to estimate their behaviors regarding the three health topics. Alcohol behaviors were measured by a two-item scale regarding
situational drinking behavior (“On average, how many alcoholic drinks do you consume when you socialize in a setting
with alcohol?” and “Within the last 2 weeks, how many times
have you had five or more alcoholic drinks in a sitting?”; α =
.78), as well as a one-item measure assessing the total number of alcoholic drinks the student reported consuming per
week. Smoking behaviors were measured with two items,
one querying the number of cigarettes or cigars the participant smoked per month, the other asking the number of days
in the last month the participant had smoked. Two items measured exercise behaviors: one asking the number of days per
week the participant exercised, the other seeking an estimate
of the number of minutes per session exercised.
Normative judgments. For each topic, four normative
judgment items were measured. Alcohol normative judgments included (a) how many alcoholic drinks do you think a
typical Cornell University/University of Georgia (UGA)
undergrad consumes when (s)he socializes in a setting with
alcohol (number of drinks), (b) on average, how many alcoholic drinks do you think a typical Cornell/UGA undergrad
consumes in a week, (c) what percentage of Cornell/UGA
students think it is important to avoid drinking to get drunk,
and (d) what percentage of Cornell/UGA students do you
think drink less than 5 alcoholic drinks per week. Measures
of tobacco normative judgments included (a) how many cigarettes/cigars do you think the typical Cornell/UGA undergrad student has used in the last 30 days, (b) how many days
in the last month do you think the typical Cornell/UGA
undergrad student has used tobacco products, (c) what percentage of Cornell/UGA undergraduates believe it is a good
idea not to smoke, and (d) what percentage of Cornell/UGA
undergrad students smoke less than three cigarettes per
month. Exercise items included (a) how much do you think
Cornell/UGA undergrads exercise (days per week), (b) on
average, how much time do you think Cornell/UGA undergrads exercise in one session (in minutes), (c) what percentage of Cornell/UGA undergrads believe exercising regularly
is important, and (4) what percentage of Cornell/UGA undergrads exercise at least twice per week.
Attitude measures. All attitudes were measured on
5-point Likert-type scales from 1 (strongly disagree) to 5
(strongly agree). Four types of attitudes were measured for
each behavior. The degree to which someone liked to engage in
excessive drinking behavior was measured using two items (“I
like to get drunk” and “I enjoy getting drunk”; α = .97). This
value was recoded for analysis, so that a higher score indicated
healthier behavior, that is, not liking to get drunk. One item
measured undergraduate drinking attitudes (“Cornell/UGA
undergrads think that drinking to get drunk is a bad idea”). To
measure attitude toward one’s own drinking behavior, participants were asked to indicate on a 5-point scale the extent of
their agreement to three items (“I don’t have to get drunk to
have a good time,” “I think drinking to get drunk is a bad idea,”
and “I feel better when I do not drink”). These three items fit a
single factor (α = .60). Attitudes toward the harm of drinking to
get drunk were measured with two items: “I believe that drinking to get drunk is harmful to good health” and “Drinking too
much alcohol causes health problems” (α = .66).
Four items measured liking to smoke (e.g., “Smoking is
something I enjoy”; α = .96). This value was recoded for analysis such that a higher score was related to a healthier behavior,
that is, not liking to smoke. Four items were used to measure
participants’ attitudes toward undergraduate smoking behaviors (e.g., “Avoiding smoking is a good practice for Cornell/
UGA undergrads”; α = .88). To measure participants’own attitudes toward smoking, participants were asked to indicate on a
5-point scale the extent of their agreement to three items (e.g.,
“I believe that I should avoid smoking”). These three items fit a
single factor (α = .89). Three items assessed the harm of smoking (e.g., “I believe that smoking is bad for health”; α = .78).
To measure participants’ attitudes toward liking to exercise, four items were used (e.g., “I like to exercise”; α =
.85). Four items were used to measure participants’ attitudes toward undergraduate exercising behaviors (e.g.,
“Cornell/UGA undergrads need to exercise to maintain
good health.”; α = .90). Four items assessed attitudes toward participants’ own exercising (e.g., “I believe that I
should exercise regularly”). These four items fit a single
factor (α = .89). To assess attitudes toward the benefit (i.e.,
lack of harm) of exercising, four items fitting a single factor
were used (“I believe that exercising is important to maintain good health”; α = .83).
Table 1 provides the means and standard deviations for all
of the measures. Data were analyzed using hierarchical regression, entering the sociodemographic variables in the first
block, normative judgments in the second block, and attitudes in the final block.
RESULTS
We found that, although students at the Southeastern university smoked and exercised more, predictors of their drinking,
smoking, and exercise behavior did not vary from those of
STEPPING BACK FROM SOCIAL NORMS CAMPAIGNS
281
TABLE 1
Means and Standard Deviations of Measures
Measure
Normative judgment of how many drinks typical undergraduate consumes in a week
Normative judgment of how many drinks typical undergraduate consumes in a setting
Normative judgment of percentage of students who think it is important to avoid drinking to get drunk
Normative judgment of percentage of students who drink less than 5 drinks a week
Normative judgment of how many cigarettes/cigars typical undergraduate has used in last 30 days
Normative judgment of how many days in the last month the typical undergraduate has smoked
Normative judgment of percentage of undergraduates who think it is a good idea not to smoke
Normative judgment of percentage of undergraduates who smoke less than 3 cigarettes a month
Normative judgment of how many days per week undergraduates exercise
Normative judgment of how many minutes per session undergraduates exercise
Normative judgment of percentage of undergraduates who think exercising regularly is important
Normative judgment of percentage of undergraduates who exercise at least twice per week
Attitude toward liking to get drunk (recoded)
Attitude toward undergraduates’ drinking behavior
Attitude toward own drinking behavior
Attitude toward harm of drinking to get drunk
Attitude toward liking to smoke (recoded)
Attitudes toward undergraduates’ smoking behavior
Attitude toward own smoking behavior
Attitude toward harm of smoking
Attitude toward liking to exercise
Attitude toward undergraduates exercising behavior
Attitude toward own exercising behavior
Attitude toward healthiness of exercising
Situational drinking behavior
Drinks consumed per week
Number of cigarettes/cigars smoked in past month
Number of days smoked in past 30 days
Days per week exercised
Minutes per exercise session
the Northeastern university; the pattern of significant predictors was identical when separate regression analyses were
run. University affiliation was not a significant predictor of
behavior within the regression models. We found no topic order effects based on presentation order. Thus, the reported regression analyses were collapsed across universities and
questionnaire versions.
Hierarchical Regression Results
Alcohol. Hierarchical regression indicated that sex,
Greek affiliation, the participant’s normative judgment of the
amount it was believed other students drink per setting, his or
her liking to drink to get drunk, and his or her attitude toward
the harm of drinking to get drunk predicted both the student’s
binge drinking behavior and the number of drinks he or she
consumed per week. Men and fraternity and sorority members were heavier binge drinkers and consumed more drinks
per week. The more a participant believed that other students
drank per setting and the more he or she liked to drink, the
more the student binge drank or consumed per week,
whereas the more the student believed that drinking to get
M
SD
9.77
4.73
36.57
46.16
61.58
13.80
73.65
51.03
2.53
42.55
73.17
50.36
3.04
2.15
3.03
4.15
4.17
4.57
4.73
4.81
4.32
4.33
4.58
4.58
3.03
8.73
27.54
4.41
3.28
57.55
7.28
2.05
21.13
21.59
81.89
9.13
22.55
23.96
1.12
17.54
18.32
18.98
1.09
1.12
1.08
0.69
1.05
0.62
0.52
0.39
0.69
0.71
0.58
0.53
2.54
10.77
89.25
8.99
1.89
33.19
drunk is harmful to health, the less the student binge drank or
consumed per week. Tables 2 and 3 provide the beta weights,
partial correlations, R2, and R2 change for each block.
Smoking. Results indicated that sex, the participant’s
normative judgment of the number of days he or she believed
others smoke per month, his or her attitude toward liking to
smoke, his or her attitude toward undergraduate smoking behavior, and the participant’s attitude toward his or her own
smoking were predictive of the number of cigarettes students
smoked per month. Men smoked more than women, and the
more a participant believed that others smoked per month,
the more he or she smoked. In addition, the more students
smoked, the more they believed that other undergraduates
should not smoke, but the less they felt they personally
should avoid smoking. Table 4 provides the beta weights,
partial correlations, R2, and R2 change for each block.
Sex, liking to smoke, and attitude toward the participant’s
own smoking were significant predictors for number of days
students smoked per month. Those who smoked on more
days were more likely to be men, like to smoke more, and believe less strongly that they should avoid smoking. Table 5
282
CAMERON AND CAMPO
TABLE 2
Hierarchical Regression Predicting Binge Drinking
β
Predictor
Step 1—Sociodemographics
Sex
Greek
White
Dorm
Step 2—Normative judgments
How many drinks do you think a typical undergraduate consumes in a week?
How many drinks do you think a typical undergraduate consumes in a setting?
What percentage of students do you think believe it is important to avoid
drinking to get drunk?
What percentage of undergraduates drink less than 5 drinks per week?
Step 3—Attitudes
Liking to drink to get drunk (recoded)
Undergraduates think that drinking to get drunk is a bad idea
Attitudes toward own drinking behavior
Attitudes toward harm of drinking to get drunk
Partial
Correlation
–.24**
.15*
–.06
–.07
–.31
.19
–.08
–.09
.09
.22*
.05
.10
.22
.06
–.01
–.01
–.26**
.05
.05
–.22**
.29
.06
.07
–.26
R2
∆R2
F
.23
.23**
19.05**
.34
.10**
15.53**
.49
.15**
19.20**
R2
∆R2
F
.24
.24**
20.06**
.31
.07**
13.85**
.38
.08**
12.65**
Note. Betas and partial correlations are taken from the final regression equation. F is F from analysis of variance at step of entry.
*p < .01. **p < .001.
TABLE 3
Hierarchical Regression Predicting Drinks per Week
Predictor
Step 1—Sociodemographics
Sex
Greek
White
Dorm
Step 2—Normative judgments
How many drinks do you think a typical undergraduate consumes in a week?
How many drinks do you think a typical undergraduate consumes in a setting?
What percentage of students do you think believe it is important to avoid
drinking to get drunk?
What percentage of undergraduates drink less than 5 drinks per week?
Step 3—Attitudes
Liking to drink to get drunk (recoded)
Undergraduates think that drinking to get drunk is a bad idea
Attitudes toward own drinking behavior
Attitudes toward harm of drinking to get drunk
β
Partial
Correlation
–.28**
.19*
–.09
–.04
–.32
.21
–.11
–.05
.23*
.00
.00
.21
.00
.00
–.06
–.06
–.17*
.09
.02
–.17*
.17
.10
.02
–.18
Note. Betas and partial correlations are taken from the final regression equation. F is F from analysis of variance at step of entry.
*p < .01. **p < .001.
provides the beta weights, partial correlations, R2, and R2
change for each block.
Exercise. Hierarchical regression indicated that race, a
participant’s normative judgments related to the number of
days per week and minutes per session he or she believed
other students exercised, the participant’s normative judgments of the percentage of students believing exercise is
important, his or her attitude toward liking to exercise, and
attitude toward his or her own exercise behavior were pre-
dictive of the number of days participants exercised per
week. Whites were less likely to exercise than students of
other races. The more days per week participants believed
that other students exercised, the more days per week they
exercised. However, the less time a participant believed other
students exercised per session and the less he or she believed
other students thought regular exercise was important, the
more days per week he or she exercised. The more students
liked to exercise and the more students believed that they
should exercise, the more days per week they exercised. Ta-
STEPPING BACK FROM SOCIAL NORMS CAMPAIGNS
283
TABLE 4
Hierarchical Regression Predicting Number of Cigarettes Used in Last 30 Days
β
Predictor
Step 1—Sociodemographics
Sex
Greek
White
Dorm
Step 2—Normative judgments
How many cigarettes/cigars do you think typical undergraduate has used in last 30 days?
How many days in last month has typical undergraduate smoked?
What percentage of undergraduates believe it is a good idea not to smoke?
What percentage of undergraduates smoke less than 3 cigarettes per month?
Step 3—Attitudes
Liking to smoke (recoded)
Attitudes toward undergraduates’ smoking behavior
Attitudes toward own smoking behavior
Attitudes toward harm of smoking
Partial
Correlation
–.17**
–.04
–.07
.00
–.20
–.04
–.08
.01
–.06
.20**
.07
.03
–.06
.21
.08
.03
–.47***
.19*
–.20*
–.05
.43
.15
–.16
–.04
R2
∆R2
.06
.06**
3.73**
.10
.04*
3.17**
.33
.26***
10.90***
∆R2
F
F
Note. Betas and partial correlations are taken from the final regression equation. F is F from analysis of variance at step of entry.
*p < .05. **p < .01. ***p < .001.
TABLE 5
Hierarchical Regression Predicting Number Days Smoked in the Past Month
β
Predictor
Step 1–Sociodemographics
Sex
Greek
White
Dorm
Step 2–Normative judgments
How many cigarettes/cigars do you think typical undergraduate has used in last 30 days?
How many days in last month has typical undergraduate smoked?
What percentage of undergraduates believe it is a good idea not to smoke?
What percentage of undergraduates smoke less than 3 cigarettes per month?
Step 3—Attitudes
Liking to smoke (recoded)
Attitudes toward undergraduates’ smoking behavior
Attitudes toward own smoking behavior
Attitudes toward harm of smoking
Partial
Correlation
–.14*
.05
–.04
.05
–.21
.07
–.06
.07
–.03
.09
.08
.00
–.04
.11
.10
.00
–.65**
.10
–.20*
.04
.62
.09
–.18
.04
R2
.07
.07*
4.50*
.10
.034
3.41*
.56
.45**
23.94**
Note. Betas and partial correlations are taken from the final regression equation. F is F from analysis of variance at step of entry.
*p < .01. **p < .001.
ble 6 provides the beta weights, partial correlations, R2, and
R2 change for each block.
Sex, participants’ normative judgments related to minutes per session other students’ exercise, their normative
judgments of the percentage of students believing exercise
is important, their attitudes toward liking to exercise, and
attitudes toward their own exercise behavior predicted the
number of minutes students exercised per session. Men exercised longer than women. The more minutes per session
participants believed that other students spent exercising,
the longer they themselves exercised. The less participants
believed other students thought exercise was important, the
more they exercised. The more students liked to exercise as
well as the more that students believed they should exercise, the longer they spent during each exercise session. Table 7 provides the beta weights, partial correlations, R2, and
R2 change for each block.
Relation of Normative Judgments to Behavior
Normative judgments were related to participants’ behavior
in five out of the six studied behaviors (see Tables 2–7).
When observing the beta weights, it is evident that the type
of normative judgment associated with each behavior dif-
284
CAMERON AND CAMPO
TABLE 6
Hierarchical Regression Predicting Number of Days per Week of Exercise
Partial
Correlation
β
Predictor
Step 1—Sociodemographics
Sex
Greek
White
Dorm
Step 2—Normative judgments
How many days per week do you think undergraduates exercise?
How many minutes/session do undergraduates exercise?
What percentage of undergraduates do you think believe exercising regularly is important?
What percentage of undergraduates exercise at least twice per week?
Step 3—Attitudes
Liking to exercise
Attitudes toward undergraduate exercise behavior
Attitudes toward own exercise behavior
Attitudes toward healthiness of exercising
–.07
–.04
–.14*
.06
–.08
–.05
–.17
.07
.17*
–.14*
–.14*
.07
.16
–.16
–.15
.07
.47***
–.01
.26**
–.16
.38
–.01
.20
–.12
∆R2
R2
F
.02
.02
1.01
.06
.04*
1.85
.34
.29***
10.51***
Note. Betas and partial correlations are taken from the final regression equation. F is F from analysis of variance at step of entry.
*p < .05. **p < .01. ***p < .001.
TABLE 7
Hierarchical Regression Predicting Number of Minutes of Exercise per Session
Predictor
Step 1—Sociodemographics
Sex
Greek
White
Dorm
Step 2—Normative judgments
How many days per week do you think undergraduates exercise?
How many minutes/session do undergraduates exercise?
What percentage of undergraduates do you think believe exercising regularly
is important?
What percentage of undergraduates exercise at least twice per week?
Step 3—Attitudes
Liking to exercise
Attitudes toward undergraduate exercise behavior
Attitudes toward own exercise behavior
Attitudes toward healthiness of exercising
β
Partial
Correlation
–.11*
–.03
–.07
.02
–.13
–.03
–.07
.02
–.08
.29**
–.14*
–.08
.31
–.14
.06
.06
.40**
–.10
.17*
–.10
.39
–.08
.13
–.07
R2
∆R2
F
.02
.02
1.40
.14
.12**
4.88**
.30
.17**
8.74**
Note. Betas and partial correlations are taken from the final regression equation. F is F from analysis of variance at step of entry.
*p < .05. **p < .001.
fers among behaviors; moreover, normative judgments are
not remarkably strong predictors. In some cases, one’s sex
was a stronger predictor of behavior than was normative
judgment. The normative judgment of “behavior in a setting” (e.g., drinks per setting, days smoked per month, minutes exercised per session) was the best predictor from the
four normative judgment variables tested; it was predictive
for four out of the six health behaviors. However, in all
cases in which normative judgments were predictive of
behavior, other sociodemographic or attitudinal variables
were as predictive if not more so. Normative judgments as
to the percentage of undergraduate students who engaged in
the particular behavior (e.g., percentage of students who
drink less than five alcoholic drinks per week) were never a
significant predictor of behavior. Further, normative judgments as to the percentage of undergraduates who thought
engaging in a particular behavior was a good idea were
only predictive for the positive behavior of exercising;
those who exercised more believed that other students
found the behavior to be less important. Table 8 provides a
STEPPING BACK FROM SOCIAL NORMS CAMPAIGNS
285
TABLE 8
Comparison of Predictors
Predictor
Sociodemographics
Sex
Greek affiliation
Whites compared to others
Dorm vs. other housing
Normative judgments
Drinks per week, exercise day per week, cigarettes per month
Drinks per setting, days smoked per month, exercise time per session
Avoid drinking to get drunk, good idea not to smoke, exercise regularly
important (health behavior)
Attitudes
Liking to engage in the behavior (coded such that higher score was related to
healthier behavior)
Attitudes toward undergraduates’ behavior
Attitudes toward own behavior
Attitude toward healthiness of behavior
Binge
Drinking
Scale
Drinks
per
Week
–.24***
.15**
–.28***
.19**
No. of
Cigarettes
in Last 30
Days
–.17**
No. of
Days
Smoked in
Past
Month
No. of
Days per
Week
Exercise
–.14**
No. of
Minutes per
Exercise
Session
–.11*
–.14*
.23**
.22**
–.26***
–.22***
.17*
–.14*
–.14*
.20**
–.17**
.29***
–.16*
–.47***
–.65***
.47***
.40***
.19*
–.20*
–.20**
.26**
.17*
–.17**
*p < .05. **p < .01. ***p < .001.
comparative view of the significant predictors for each of
the six behaviors.
Predictors Across Positive
and Negative Health Behaviors
The only predictor consistently significant across both positive and negative health behaviors was the participants’ liking
to engage in that particular behavior. Sex was predictive for
five of the six behaviors, both positive and negative: Men engaged in the behavior more often than women in all instances
except the number of days they exercised per week (for
which there was no sex difference). Greek affiliation was associated only with alcohol consumption. Living arrangement
(e.g., dormitory or other housing) was not predictive for any
of the behaviors under scrutiny. The effect of normative judgments varied across health behaviors. Participants’ attitudes
toward their own behavior were predictive for all smoking
and exercising behaviors but were not predictive for alcohol
consumption. Attitudes toward the healthiness or harm of the
behavior were predictive only for the negative behavior of
excessive alcohol consumption.
DISCUSSION
This study took a step back from social norms campaigns to
examine the relative effects of sociodemographic variables,
normative judgments, and individual attitudes on personal
behavior related to alcohol consumption, smoking, and
exercising. We thus were able to assess predictors for both
healthy and unhealthy behaviors. The data were collected at
two universities, with no significant difference in predictors
of behaviors found between locations. This lack of difference
suggests that the results are generalizable to college students.
Results indicated that different variables were predictive
among behaviors, which is consistent with other research on
prohealth behaviors (van Ryn et al., 1996). Normative judgments were not consistently predictive for the behaviors under scrutiny. Liking to engage in a particular behavior and
participant sex were the only predictors that remained consistent across positive and negative health behaviors. Otherwise, no clear pattern of predictors, whether sociodemographic, normative judgment, or attitudinal, emerged in these
data.
Perkins and Wechsler (1996) note that personal alcohol
attitude “holds the lion’s share of effect on personal alcohol
abuse” (p. 968) in their sample. They found the perceived
campus alcohol norm to be a significant predictor independent of other sociodemographic variables, yet the majority of
the effect was accounted for by personal alcohol attitude. Our
results parallel theirs in all cases, across both positive and
negative health behaviors, as attitudes were a more significant predictor than normative judgments. The participant’s
attitude toward liking to engage in the particular activity
(drinking, smoking, exercising) was always a significant predictor of behavior and often was by far the most statistically
significant predictor. Future research should explore the relationship between normative judgments and liking to engage
in a behavior.
The messages most advocated by pundits of social norms
campaigns are those that provide receivers with an “accu-
286
CAMERON AND CAMPO
rate” view of peers’ behaviors, for example, “2/3 of Rutgers
students drink 3 or fewer drinks, and 1 in 5 don’t drink at all”
(Higher Education Center for Alcohol and Other Drug Prevention, 2004). However, in no conditions, neither positive
nor negative health behaviors, was the normative judgment
of other students’ behaviors a significant predictor of own behavior. Another common formulation in social norms campaigns is the use of messages relating to the “accurate” perceptions of peers’ attitudes, for example a University of
Oregon message that proclaims “87.6% of first year students
report that drinking alcohol does not make them feel sexier”
(Higher Education Center for Alcohol and Other Drug Prevention, 2004). In our study, this variable was predictive only
for exercising behaviors. Therefore, although normative
judgments may have some effect on college students’ health
behaviors, they are neither consistent nor the optimal predictors of such behaviors.
Measures created for this study assessed multiple normative judgments and attitudes, providing a more extensive
view of the effects of normative judgments and attitudes on
behavior than past research. Our measurements of normative
judgments (querying amount of engaging in behavior as well
as percentage of students holding a particular attitude or
engaging in a particular behavior) are in greater correspondence to current social norms messages. These current messages tend to present statements about specific student behavior, for example, “Most of us at UW-Whitewater have 5 or
fewer drinks per week” (Higher Education Center for Alcohol and Other Drug Prevention, 2004). Our measures elicited
participant judgments of such behaviors among peers,
whereas past measurement failed to do so. Yet even with
these more specific normative judgment measures, the data
indicate that normative measures are not consistent predictors of student behavior as it relates to alcohol consumption,
smoking, or exercising.
Limitations
The majority of attitudinal scales had reliabilities greater
than .80. The lowest scale reliabilities were related to alcohol attitudes. The measure of participants’ attitudes toward
undergraduate alcohol use was a one-item measure; future
studies may seek to enhance this measure as well as increase the reliability of the measurements of alcohol attitudes. Based on this study, we do not know if there is an interaction between normative judgments and liking to
engage in a behavior on likelihood of engaging in a behavior. Although results pointed to significant effects on individual health behavior, in all cases there remains a large
portion of variance unaccounted for. Future studies need to
ascertain other significant predictors of positive and negative health behaviors of college students and determine if
interaction effects exist.
Implications
This research extended past studies of influences on college
student behavior by considering positive and negative behaviors and furthered our understanding of behavioral influences
by including sociodemographic variables and multiple measures of normative judgments and personal attitudes. As different variables were found to be predictive for the varying
health behaviors, these results bring into question the effectiveness of social norms campaigns, not only for alcohol consumption but for other behaviors such as smoking, seatbelt
use, and so forth. The results indicate that one’s liking to engage in a particular behavior is consistently a significant predictor of both addictive and nonaddictive behaviors.
Smokers in this study conceivably recognized the harmful
effects of smoking, as the more they smoked the more they
believed others should not smoke. This result may seem
counterintuitive; however, it is well known that smoking
cigarettes is hazardous to one’s health. Although smokers
may be unwilling to give up their own habit, they may recognize the dangers and addictive properties of smoking and believe that others should avoid use. The findings related to exercise (the more a participant exercised, the fewer minutes he
or she thought other students exercised and the less he or she
thought other students thought exercise is important) may be
a result of participants regularly being in an exercise facility
and seeing others exercising for a short time.
For behavioral intervention campaigns to meet with success, it appears we must learn how best to influence individuals’ predilection for behaviors, both healthy behaviors we
wish to encourage (e.g., exercise) and unhealthy behaviors
we wish to diminish (e.g., excessive alcohol use). We caution
against practitioners indiscriminately transferring campaigns
from one topic (e.g., from alcohol to tobacco) or type of message (attitudinal or behavioral) to another without proper
evaluation. Using untested campaigns, even with the best intentions, can have positive, no, or negative effects on audience behaviors (Fishbein, Hall-Jamieson, Zimmer, von
Haeften, & Nabi, 2002).
CONCLUSION
The premise of social norms campaigns is that if you alter
your normative judgments of others’ behavior, then you will
alter your own attitudes and behavior. Although some normative judgments were statistically significant predictors of individual behavior, no one judgment (of the four types measured) was a consistent predictor. Rather, when normative
judgments were significant, as in the case of binge drinking
behavior, it was the judgment most closely related to the specific behavior (e.g., drinks per setting, drinks per week) that
attained significance. Ergo, health campaigns, whether social
norms or other persuasive campaigns, must be targeted at rel-
STEPPING BACK FROM SOCIAL NORMS CAMPAIGNS
evant attitudes and normative judgments. Simply using a
message derived from social norms marketing is unlikely to
consistently affect college student behavior, as the relevant
attitudes and normative judgments change both within and
across positive and negative health behaviors.
ACKNOWLEDGMENTS
The authors contributed equally to this article. This research
was supported in part by the Cornell University Agricultural
Experiment Station federal formula funds Project NYC131401, received from Cooperative State Research, Education and Extension Service, U.S. Department of Agriculture.
Any opinions, findings, conclusions, or recommendations
expressed in this publication are those of the authors and do
not necessarily reflect the view of the U.S. Department of
Agriculture. We are grateful to Somjen Frazer, Dominique
Brossard, Esther Baker, Baseema Banoo, Deanna Caputo,
Sarah Doherty, Erin Payne, and Tommy Wood for their assistance. We also wish to thank Teri Thompson and two anonymous reviewers for their comments. A version of this manuscript was presented at the annual American Public Health
Association Conference, San Francisco, November 2003.
REFERENCES
Ajzen, I. (1985). From intentions to actions: A theory of planned behavior.
In J. Kuhl & J. Beckmann (Eds.), Action control: From cognition to behavior (pp. 11–39). New York: Springer-Verlag.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice Hall.
Allport, F.H. (1924). Social psychology. Boston: Houghton Mifflin.
Baer, J. S., Stacy, A., & Larimer, M. (1991). Biases in the perception of
drinking norms among college students. Journal of Studies on Alcohol,
52, 580–586.
Campo, S., Brossard, D., Frazer, M. S., 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.
Campo, S., & Cameron, K. A. (2006). Differential effects of exposure to social norms campaigns: A cause for concern. Health Communication, 19,
209–219.
Campo, S., Cameron, K. A., Brossard, D., & Frazer, M. S. (2004). Social
norms and expectancy violation theories: Assessing the effectiveness of
health communication campaigns. Communication Monographs, 71,
127–149.
Chassin, L., & DeLucia, C. (1996). Drinking during adolescence. Alcohol
Health and Research World, 20, 175–180.
Coleman, C. L. (2002). Examining influences of pharmacists’ communication with consumers about antibiotics. Health Communication, 15,
79–99.
Fields, J. M., & Schuman, H. (1976). Public beliefs about the beliefs of the
public. Public Opinion Quarterly, 40, 427–448.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior:
An introduction to theory and research. Reading, MA: Addison-Wesley.
Fishbein, M., Hall-Jamieson, K., Zimmer, E., von Haeften, I., & Nabi, R.,
(2002). Avoiding the boomerang: Testing the relative effectiveness of
287
antidrug public service announcements before a national campaign.
American Journal of Public Health, 92, 238–245.
Frauenfelder, M. (2001, December 9). Social-norms marketing. New York
Times Magazine, 100.
Gomberg, L., Schneider, S. K., & DeJong, W. (2000). Evaluation of a social
norms marketing campaign to reduce high-risk drinking at the University
of Mississippi. American Journal of Drug and Alcohol Abuse, 27,
375–389.
Gose, B. (1997, October 24). Colleges try to curb excessive drinking by saying moderation is okay. Chronicle of Higher Education, 44 (9), A61–A62.
Greene, K., Hale, J. L., & Rubin, D. L. (1997). A test of the theory of reasoned action in the context of condom use and AIDS. Communication Reports, 10, 21–33.
Haines, M., & Spear, S. F. (1996). Changing the perception of the norm: A
strategy to decrease binge drinking among college students. Journal of
American College Health, 45, 134–140.
Higher Education Center for Alcohol and Other Drug Prevention. (2004).
Social norms and social marketing. Retrieved May 4, 2004, from http://
www.edc.org/hec/socialnorms/
Johannessen, K., Collins, C., Mills-Novoa, B., & Glider, P. (1999). A practical guide to alcohol abuse prevention: A campus case study in implementing social norms and environmental management approaches. Tucson:
Campus Health Services, The University of Arizona.
Keeling, R. P. (1999). Proceed with caution: Understanding and changing
norms. Journal of American College Health, 47, 243–246.
Keeling, R. P. (2000). Social norms research in college health. American
Journal of College Health, 49, 53–56.
Kim, M. S., & Hunter, J. E. (1993). Relationships among attitudes, behavioral intentions, and behavior: A meta-analytic view of past research, Part
2. Communication Research, 20, 331–364.
Larimer, M. E., Irvine, D. L., Kilmer, J. R., & Marlatt, G. A. (1997). College
drinking and the Greek system: Examining the role of perceived norms for
high-risk behavior. Journal of College Student Development, 38, 587–598.
Lo, C. C., & Globetti, G. (1993). A partial analysis of the campus influence
on drinking behavior: Students who enter college as non-drinkers. The
Journal of Drug Issues, 23, 715–725.
Montana State University-Bozeman, (2004). Most of us campaign. Retrieved May 4, 2004, from http://www.mostofus.org/
Olds, R. S., & Thombs, D. L. (2001). The relationship of adolescent perceptions of peer norms and parent involvement to cigarette and alcohol use.
Journal of School Health, 71, 223–228.
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., Meilman, P. W., Leichlier, J. S., Cashin, J. R., & Presley, C.
A. (1999). Misperceptions of the norms for the frequency of alcohol and
other drug use on college campuses. Journal of American College Health,
47, 253–258.
Perkins, H. W., & Wechsler, H. (1996). Variation in perceived college drinking norms and its impact on alcohol abuse: A nationwide study. Journal of
Drug Issues, 26, 961–974.
Prentice, D. A., & Miller, D. T. (1993). Pluralistic ignorance and alcohol use
on campus: Some consequences of misperceiving the social norm. Journal of Personality and Social Psychology, 64, 243–256.
Rimal, R. N., & Real, K. (2003). Understanding the influence of perceived
norms on behavior. Communication Theory, 13, 184–203.
Stewart, L. P., Lederman, L. C., Golubow, M., Cattafesta, J. L., Goodhart, F.
W., Powell, R .L., et al. (2002). Applying communication theories to prevent dangerous drinking among college students: The RU SURE campaign. Communication Studies, 53, 381–399.
van Ryn, M., Lytle, L. A., & Kirscht, J. P. (1996). A test of the theory of
planned behavior for two health-related practices. Journal of Applied Social Psychology, 26, 871–883.
288
CAMERON AND CAMPO
Wechsler, H., & Kuo, M. (2000). College students define binge drinking and
estimate its prevalence: Results of a national survey. Journal of American
College Health, 49, 57–64.
Wechsler, H., Nelson, T. F., Lee, J. E., Seibring, M., Lewis, C., & Keeling, R.
P. (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.
Wechsler, H., Seibring, M., Liu, I. C., & Ahl, M. (2004). Colleges respond to
student binge drinking: Reducing student demand or limiting access.
Journal of American College Health, 52, 159–168.
Weitzman, E. R. (2004, February 1). Social developmental overview of
heavy episodic or binge drinking among college students. Psychiatric
Times, 21(2), 57–59, 80.