J. DRUG EDUCATION, Vol. 34(3) 247-266, 2004
COLLEGE STUDENT BINGE DRINKING AND THE
“PREVENTION PARADOX”: IMPLICATIONS FOR
PREVENTION AND HARM REDUCTION*
ELISSA R. WEITZMAN, SC.D., M.SC.
TOBEN F. NELSON, M.S.
Harvard School of Public Health
ABSTRACT
Considerable attention has been paid to heavy episodic or “binge” drinking
among college youth in the United States. Despite widespread use, the binge
measure is perceived by some as a low intervention threshold. We use data
from the Harvard School of Public Health College Alcohol Study (n = 49,163)
to describe patterns of consumption and harms along a continuum including
the binge measure to demonstrate the validity of the binge threshold and
prevention paradox in college. While the heaviest drinkers are at greatest risk
for harm, they are relatively few and generate proportionately small amounts
of all drinking-harms. The risk of harms is not zero among lower level
drinkers in college. Because they are numerous, they account for the majority
of harms. This paradoxical pattern suggests we moderate consumption among
the majority using environmental approaches, the efficacy of which are
described using case study data from a national prevention demonstration.
Implications for prevention policy, programming, and media advocacy are
discussed.
*Data used in this article were collected with the generous support from the Robert Wood Johnson
Foundation.
247
Ó 2004, Baywood Publishing Co., Inc.
248
/
WEITZMAN AND NELSON
INTRODUCTION
The problem of heavy episodic or “binge” drinking among young adults in college
in the United States has received considerable attention from researchers, college
administrators, substance abuse educators, students, parents, politicians, and the
media [1]. Defined as 5/4 consecutive drinks among men/women at least once in
the two weeks prior to be surveyed, binge drinking occurs among two out of
every five U.S. college students, half of whom did so multiple times [2]. Binge
drinking may typify consumption among other populations of drinkers in the
United States as well [3].
Despite its common use and its stability in repeated national surveys [1, 4], the
binge measure is controversial. Some think that the 5/4 drinking measure is too
low a threshold on which to base preventive interventions [5, 6]. This perspective
may reflect a belief that heavy drinking among youth is a normal expression of
experimentation and development and that chronic effects of long-term alcohol
use and diagnosable alcoholism are more salient consequences, which college
students are likely to avoid when they reduce their consumption after leaving
college [7]. This perspective is not supported by recent longitudinal study [8].
Attempts to address problems associated with the heavy drinking of college
students have employed a variety of methods, including increased punitive
sanctions, expedited judicial process, and counseling and treatment for problem
drinkers [9]. Social marketing strategies that promote a latent healthy drinking
norm in hopes that it will encourage students to reduce their own consumption
have also become popular, although the empirical evidence for these programs
is mixed [10-12].
No major shifts in the prevalence of alcohol-related problems among college
students have been noted during the 1990s [1, 2]. Both traditional and newer
approaches have attempted to intervene among drinkers deemed at “high risk”
for alcohol-related problems based on self-reported consumption patterns or
harms. These high-risk or secondary prevention efforts which often rely on
screening, counseling, and brief interventions, hold a great deal of appeal for
university administrators, health professionals, and interested activists. At the
same time, students who are considered to be normal and not at high-risk are not
targeted by interventions that may potentially limit their consumption. Unfortunately, little evidence exists to support the efficacy of high-risk approaches
[13]. Very few students perceive that they have a problem with alcohol and even
fewer voluntarily present to a passive health care system for alcohol treatment
or help [1]. Increased emphasis on screening and the development of screening
tools to identify problem drinkers have largely been unsuccessful [13], a difficulty
associated with use of a screening approach and not this health problem or
population [14, 15]. Within the college setting and population, high-risk students
typically come to the attention of authorities such as the Office for Student Affairs
or the local police department. Individual motivation to stop drinking is low
COLLEGE STUDENT BINGE DRINKING /
249
among college student drinkers [16, 17], and lasting change is difficult to attain
in the face of underlying environmental conditions that promote alcohol use in
this population, such as a saturation of alcohol outlets in some college environments [18, 19] and targeted marketing practices [20].
One strategy that has received less attention in college health prevention is
the population approach, which looks at the entire population rather than focusing solely on the members of that group whose behavior is the most extreme
[21]. This view acknowledges that risk is distributed throughout the entire
population and deviance in the group is directly related to the prevailing normative behavior in the population. Interventions to reduce problem behaviors
thus focus on the community and are directed at producing incremental change
throughout a population, rather than wholesale change among subgroups with
extreme behavior patterns.
Population approaches that address alcohol have demonstrated effectiveness in
reducing the problems that result from alcohol consumption [22-24]. Decreasing
availability of alcohol to underage drinkers, raising alcohol taxes and prices,
instituting mandatory responsible beverage service, and changing the conditions
of availability by limiting outlets are examples of specific population-based
approaches that appear effective at reducing consumption and harms. These interventions influence all members of a population and result in incremental change
at the individual level, yielding large public health benefits.
The observation underlying population approaches to prevention is the “prevention paradox” [21, 25]. This phenomenon is observed for a range of health
areas including alcohol, tobacco, obesity and injury [15, 26]. It is observed when
a large number of people at small risk contribute more cases of a disease or
negative health outcome in a population than do a small number of people at
large risk. Researchers in the United States and internationally have demonstrated
the importance of this perspective for alcohol using large representative data
sets for a variety of negative alcohol related outcomes [27-31].
A major implication of the preventive paradox, when it is used to shape
prevention strategies, is that there may be substantial benefits to the community
from these approaches but relatively small gains to specific individuals. As a
consequence, these types of prevention strategies may be both controversial—by
seeking to incrementally change the behavior of a majority who may correctly
perceive they are at low individual risk of harm—and counterintuitive—by investing prevention resources and attention toward changing the upstream determinants
of behaviors that may appear benign individually while seemingly ignoring the
easily identified problems of persons with more acute symptoms or disease.
For these reasons, it may be important to demonstrate both the validity of the
prevention paradox through examining interrelationships of consumption and
harms, and the impacts of prevention program models that take their cue from
its logic. The “A Matter of Degree” (AMOD) program is a community-based
prevention demonstration that is guided by the prevention paradox and uses a
250
/
WEITZMAN AND NELSON
population prevention approach [32]. Funded by the Robert Wood Johnson
Foundation, the goal of the AMOD program is to produce sustained reductions in
alcohol consumption and harms among college students by addressing features
of the college environment that promote heavy and harmful consumption. These
include factors related to alcohol’s accessibility and availability, marketing,
promotion and pricing.
In this article we use aggregated data from four national cycles of the Harvard
School of Public Health College Alcohol Study (CAS) conducted in 1993, 1997,
1999, and 2001 to examine whether the prevention paradox can be observed for
alcohol consumption and related harms among college students. By describing
the balance of harms according to drinking levels in the population we aim to
identify targets and strategies for intervention efforts that would yield the greatest
population or public health benefit. We also use case study data to describe how
environmental prevention programs that are consistent with strategies suggested
by the prevention paradox may be reducing harms in select college communities
participating in the “A Matter of Degree” (AMOD) program. Findings are discussed in the context of their significance for policy, programming, and media
advocacy efforts.
METHODS
The Survey
The study is based on student reports about drinking behaviors and harms
collected through the Harvard School of Public Health College Alcohol Study
(CAS). The CAS is a national survey program that describes patterns of alcohol
and drug use and other health risk behaviors among a representative sample of
young adults in college. CAS data are collected using a 20-page anonymous
mailed questionnaire administered to students enrolled in a representative sample
of four-year postsecondary institutions in the United States. The CAS has
ongoing approval from the Harvard School of Public Health Human Subject
Committee for analysis of the data. This study used aggregated panel data from
the 1993, 1997, 1999, and 2001 College Alcohol Study. In all four surveys, the
administrators at each college were asked to provide a random sample of approximately 225 undergraduates drawn from the total enrollment of full-time students.
Further details about the sampling, design and characteristics of the CAS have
been published elsewhere [1, 16, 33, 34].
Measures
We selected four measures of alcohol consumption that represent a continuum
of consumption behaviors and five measures of alcohol-related harms that represent salient self-reported problems to the population and setting. We examined the
COLLEGE STUDENT BINGE DRINKING /
251
usual number of drinks consumed on a drinking occasion during the past 30
days (responses ranged from one to nine or more), number of drinking occasions
in the past 30 days (responses ranged from 1 to 20 or more), number of occasions
where the respondent drank enough to become drunk (i.e., unsteady, dizzy, or sick
to your stomach) (responses ranged from 1 to 20 or more), and number of drinks
consumed in the past 30 days. These data were transformed into categories
for analysis.
We selected five measures of alcohol-related harms from the 15 measures
collected in the CAS. Measures were selected as exemplary and for salience.
Other measures of harm showed similar distributions of risk throughout the
college student population and are summarized in the results section for brevity
and efficiency.
The harms were self-reported experience of adverse consequences of drinking
at least one time since the beginning of the school year due to student’s own
alcohol use. Subjects were asked to respond to the following question: “Since
the beginning of the school year, how often has your drinking caused you to . . . ?”
A list of consequences from drinking were provided, which are described in
greater detail elsewhere [1]. The selected measures were: 1) fell behind in school
work; 2) did not use protection during sex; 3) damaged property (vandalism); and
4) got hurt or injured. An aggregate measure of five or more harms is also reported
that draws from the full set alcohol-related problems from the CAS survey. In
addition to the above listed harms, the full set includes self-report of: miss a class;
do something later regretted; forget where you were/what you did; argue with
friends; engage in unplanned sexual activity; get into trouble with the campus or
local police; require medical treatment for an alcohol overdose; and among those
who drove a vehicle in the past 30 days, driving after drinking.
Since each of these questions were asked for drinkers only, we identified
all respondents who reported that they abstained from using alcohol using a
separate survey question and assigned a value of zero. We did this to include
those students who did not consume alcohol in the population of all college
students for this analysis.
Students with missing data on any of these variables of interest for this manuscript were deleted from the analysis, resulting in a sample size of 49,163 across
the four national survey samples. No differences in the relationship between
consumption and experience of harm were observed by survey year.
Case Study Data
This report also draws on case study data about the “A Matter of Degree”
(AMOD) program to reduce binge drinking and harms in college. Ten CAS
schools are participating in AMOD, which tests the efficacy of environmental
prevention programs. AMOD schools were selected because of their high
levels (>50%) of binge drinking and their commitment to testing environmental
252
/
WEITZMAN AND NELSON
strategies to reduce heavy drinking and harms. As part of a comprehensive
evaluation of AMOD being conducted by HSPH investigators, all preventive
interventions are described and characterized formally at each site and compared
over time to a time series of that site’s CAS data. Trained on-site evaluators at each
site characterize interventions by: chronology, purpose/objective, strategy and
methods, target groups, resources, barriers and facilitators. Intervention-related
data are collected using participant observation, one-on-one interviewing with
key program personnel, and document review (i.e., quarterly reports, minutes, and
other program communications). All intervention-related data are stored in a
standardized tracking database after they are coded using a uniform observation
protocol developed by the evaluation. The tracking database is then combined
with CAS reports about alcohol-related beliefs and behaviors, including harms.
CAS data are collected from an expanded sample of students (n = 750) at each
AMOD site on an annual basis to generate a complete program time series.
Detailed methods about the evaluation are published elsewhere [32, 35].
Within each AMOD site, intervention and CAS data are compared over time
to assess the impact on student drinking beliefs and behaviors of the range
of prevention activities undertaken. Aggregate and subgroups analyses are
conducted. Tests for trends in each site’s CAS data are conducted using student
data adjusted for age, sex and race. Point estimates and 95% confidence intervals
are generated for each time series measure. This report describes significant
behavioral trends and related harms within given settings that were targets of
specific environmental intervention initiatives at two exemplary sites. This
approach is consistent with methodology in case study research [36, 37].
Data Analysis
Cross-tabulation analyses of experience of each alcohol-related harm with each
of the four measures of alcohol consumption were conducted. The percentage of
students who experienced the harm at each level of alcohol consumption was
considered the risk of experiencing that harm at the population level. The total
percent of students (the number of students per level divided by the total number of
students) at each drinking level was the percentage of students in the population.
We selected the measure being hurt or injured to demonstrate the relationship of harms to several measures of alcohol consumption (Figure 1, parts a-d)
and we selected the measure of usual number of drinks consumed when drinking to examine the relationship of consumption to multiple measures of harms
(Figure 2, parts a-e). These measures were exemplary and representative of
the consumption and harms relationship. Additional analyses are available upon
request from the authors.
Logistic regression models were fit to examine the consumption-harm relationship using the aggregate national survey data. We modeled all of the 12 alcoholrelated harm measures, plus the measure driving after five or more drinks and an
COLLEGE STUDENT BINGE DRINKING /
253
Figure 1. Individual risk of being hurt or injured from one’s own drinking
overlaid on population distribution of this harm by four consumption measures.
aggregate measure of five or more of these harms with each of the three alcohol
consumption measures. Additionally a spearman rho correlation coefficient was
calculated to examine the relationship between each of the alcohol consumption
measures. The change in alcohol related harms at the exemplary AMOD sites
was examined using logistic regression.
254
/
WEITZMAN AND NELSON
Figure 1. (Cont’d.)
All analyses were conducted using SAS version 8.2 [38] on the Unix platform.
Consistent with other analyses of multiple year CAS data [1, 35], the data sets
were directly standardized for age, sex, and race to a single survey year to ensure
valid aggregation and longitudinal assessment.
COLLEGE STUDENT BINGE DRINKING /
255
Figure 2. Percentage of students reporting specific drinking-related harms
by usual number of drinks per drinking occasion in the past 30 days.
256
/
WEITZMAN AND NELSON
Figure 2. (Cont’d.)
COLLEGE STUDENT BINGE DRINKING /
257
RESULTS
Consumption and Harms
Individual reports of drinking-related harms increased as level of consumption
increased for multiple measures of drinking style.
Figure 1.a displays the relationship between self-report of being hurt or
injured across usual number of drinks consumed when drinking. Figure 1.b shows
this same relationship for being hurt or injured and the number of drinking
occasions in the past 30 days. Figure 1.c shows the same relationship between
being hurt or injured and the number of drunken occasions in the past 30 days.
Figure 1.d shows the same relationship for the number of drinks consumed in
the past 30 days.
These results are exemplary of the consumption-harms relationship. For all 12
of the harms, driving after five or more drinks and the five or more aggregate
harm measure, we found a significant linear relationship (p < .0001 for every
tested relationship) with each of the alcohol consumption measures using logistic
regression. For the usual number of drinks per occasion (modeled as a continuous
variable with 10 categories, min = 0, max = 9 or more) the slope for each of the
harms ranged from 0.28 to 0.75 (0.40 for being hurt or injured). For number of
drinking occasions in the past 30 days (modeled as a continuous variable with six
categories, min = 0 times, max = 20 or more times) the slope for each of the harms
ranged from 0.51 to 1.26 (0.73 for being hurt or injured). For the number of
drunken occasions in the past 30 days (modeled as a continuous variable with
six categories, min = 0 times, max = 20 or more times) the slope for each of the
harms ranged from 0.68 to 1.91 (0.85 for being hurt or injured). Finally, for the
total number of drinks consumed in the past 30 days (modeled as a continuous
variable, min = 0 drinks, max = 360 drinks) the slope for each of the harms ranged
from 0.01 to 0.08 (0.02 for being hurt or injured).
The similarity of results by consumption measure is not surprising given
the very high correlations between the measures of alcohol consumption. The
usual number of drinks on a drinking occasion was strongly correlated with
drinking occasions (rho = 0.71; p < .0001), drunken occasions (rho = 0.72;
p < .0001), and number of drinks in the past 30 days (rho = 0.88; p < .0001).
Similarly, drinking occasions was correlated with drunken occasions (rho = 0.70;
p < .0001) and number of drinks in the past 30 days (rho = 0.94; p < .0001), as
was drunken occasions with number of drinks in the past 30 days (rho = 0.77;
p < .0001).
Despite this relationship of increasing risk with increasing alcohol consumption, there were very few students who consumed alcohol at the highest levels
in these distributions. The majority of harms at the population level occurred
among drinkers consuming at a less extreme level, as evident by the population patterns of consumption. This results from the larger percentage of students
258
/
WEITZMAN AND NELSON
overall who drank at a level associated with a lower, but non-zero, level of
risk. The percentage of students who experienced a given alcohol-related harm
at each level of alcohol consumption is illustrated by the pie charts shown
in Figure 2 parts a-e.
Shifting Drinking Distributions and Reducing
Harms at AMOD Sites: Two Case Studies
Consumption shifted downward incrementally at several AMOD sites as indicated by time series analyses of two sites’ student samples adjusted for age,
sex, and race. Larger proportions of students consumed fewer drinks per drinking
occasion and/or on fewer occasions when we looked at student self-reports about
consumption in off-campus and Greek-house party settings. Sites with substantial
downward shifts in consumption among students attending these events reported
declines in harms among these same groups, as predicted by the prevention
paradox (Figures 3.a–b., and 4.a–b.). Patterns of declining consumption and harms
are consistent with preventive interventions undertaken at these sites that target
the upstream determinants of drinking. At site A, where we described moderation
of off-campus consumption and declines in related harms, interventions included:
increased enforcement and sanctions in judicial system; new standards for
Greek systems students; parental notification practices to inform parents of
students’ illegal behaviors; an increase in substance-free residence halls; mandatory beverage server training; making minor-in-possession a civil not criminal
offense; and an off-campus accountability project with compliance and keg checks
at liquor licensees. At site B, where we described moderation in Greek-party
consumption and declines in related harms, interventions included: institutionalization of a uniform police reporting form; prohibitions of alcohol advertisements
in campus publications; landlord ordinances to promote responsible rental practices to youth; institutionalization of a two-strikes and suspension behavioral
code; off-campus transportation events policy; parental notification; party host
rules (to discourage supply of alcohol to underage students); and a freshman
sanction policy to strengthen policy enforcement. All of these changes focused
on restricting access and availability of alcohol to youth, limiting advertising
about alcohol to youth, and increasing the linkages among authorities that
supervise youth.
DISCUSSION
Our analyses of national survey data show that the bulk of drinking-related
harms reported by college students accrue to drinkers who consume at nonextreme levels, the groups some might consider to be at less than “high-risk."
While we found that a given individual’s risk for experiencing harms increases
with consumption, most drinkers do not consume extreme amounts of alcohol.
COLLEGE STUDENT BINGE DRINKING /
259
Figure 3. Moderation of consumption at off-campus parties and concomitant
moderation in harms among drinking youth at those events over time.
260
/
WEITZMAN AND NELSON
Figure 4. Moderation of consumption at Greek parties and concomitant
moderation in harms among drinking youth at those events over time.
COLLEGE STUDENT BINGE DRINKING /
261
Rather, the abundance of drinking-related harms in college results from a combination of low to moderate individual risk multiplied by the large number of
non-extreme drinkers. This pattern held in national analyses of CAS data across
multiple consumption measures and in analyses of survey data from each of
ten high binge schools participating in the “A Matter of Degree” program. The
reported consumption-harms relationship is consistent with predictions of the
prevention paradox.
Our finding suggests that, from a public health perspective, targeting the
entire population, including non-extreme drinkers, using environmental prevention strategies is both logical and likely to yield substantial community-wide
reductions in harm. In addition to having the greatest public health benefit,
encouraging small reductions in consumption among the majority of drinkers
may be more feasible for prevention practitioners in college than the goal of
dramatically changing consumption among frequent heavy drinkers. The latter
may have longstanding patterns of frequent heavy drinking that require intensive
individual efforts and/or treatment from clinicians.
The Promise of Environmental Strategies
Case study analyses of the AMOD demonstration data indicate that: a) designng
and implementing environmental interventions even in high prevalence communities is possible; and b) downward shifts in consumption appear to be
associated with downward shifts in harms, as predicted by theory. In particular, it appears that environmental strategies to alter access and availability and
increase linkages among supervisory agents hold promise for harm reduction
in the college setting, as suggested by other studies [19, 39, 40].
These analyses are not experimental, and do not completely rule out alternative hypotheses about a causal relationship between intervention and outcome.
However, they do provide a longitudinal perspective in the context of known
intervention activity. Further, they are consistent with theory and suggest
a plausible mechanism through which change in patterns of harm can occur.
Additional research using quasi-experimental study designs of both naturally
occurring and planned interventions can provide evidence about the utility of
environmental approaches.
Implications and Challenges
This study lends support to the use of primary prevention efforts targeting
low to moderate drinkers to achieve maximum alcohol-related harm reduction.
Drinkers consuming at or below a binge level may best be reached by prevention
that targets for change environmental influences (such as outlet density, price,
and advertising) while more extreme drinkers may best be reached using individually based secondary or tertiary (treatment-related) approaches. In all likelihood, shifting the distribution of drinking and harms will require a range
262
/
WEITZMAN AND NELSON
of strategies—including ones characteristic of environmental prevention. Ultimately, we need to become more sophisticated in crafting and targeting prevention efforts so that the field moves away from the misconception that “one size
fits all” when designing prevention programs.
There are several challenges to creating more rational and environmentallybased prevention programs. The first challenge reflects the need for data
resources. A rational approach generally requires good surveillance data in
order to match problems to strategies and to monitor change. Investment in
epidemiological surveillance may be important for addressing these problems. A
second challenge relates to changing public perceptions about the nature of the
alcohol abuse problem and its likely solutions. Our findings about the relationship
between consumption and harms at the community level may be contrary to
conceptions of the problem and its solution held by the lay public and policy
makers. If communities think that the bulk of alcohol abuse problems comes
from the small number of extreme drinkers (with their highly visible drinking
problems), then education about the actual distribution of harms by consumption
may an be important step in building support for prevention efforts that focus
on the majority. Targets for education would include both the lay public and
policymakers since the support of both groups is required to design, enact and
sustain environmental prevention efforts. This type of media advocacy effort
around alcohol [41-43] is different from the strategy of the social norms approach
that has gained popularity in the prevention community. Efforts to change social
norms about drinking recently have been based on efforts to correct students’
misperceptions about the drinking levels of other students, to encourage the
adoption of moderate drinking norms. While popular, these efforts have not been
effective in changing behaviors [10]. Rather, our analyses suggest a two-pronged
approach to prevention. First, use media advocacy to explain the links between
low levels of consumption and harms and promote harm reduction. To the best of
our knowledge, no comprehensive effort has been made to educate the public or
policy makers about the relationship between consumption and harms from the
perspective of the prevention paradox. Second, use policy controls to limit access
to and availability of alcohol to incrementally shift consumption toward lower
average levels.
A third challenge relates to generating the political will and wherewithal to
implement environmental changes to reduce harms. Within AMOD, efforts to
target individual drinkers using traditional information-based approaches met with
little resistance but appeared to produce few or no results [35]. However enacting
environmental changes at the demonstration sites, including changes to patterns
of pricing and promotion of alcohol, met with considerable resistance, including
aggressive efforts from the alcohol industry [18, 32]. Being able to generate and
sustain the political will be necessary to change the environment to reduce harms
may require both a well educated public and policy sector, as well as time and
other resources.
COLLEGE STUDENT BINGE DRINKING /
263
ACKNOWLEDGMENTS
Dr. Henry Wechsler at the Harvard School of Public Health provided a critical
reading of this manuscript and Lisa Irish and Amanda Rudman at the HSPH
helped with manuscript preparation.
REFERENCES
1. H. Wechsler, J. E. Lee, M. Kuo, M. Seibring, T. F. Nelson, and H. Lee, Trends in
Alcohol Use, Related Problems and Experience of Prevention Efforts among U.S.
College Students 1993-2001: Results from the 2001 Harvard School of Public Health
College Alcohol Study, Journal of American College Health, 50, pp. 203-217, 2002.
2. P. M. O’Malley and L. D. Johnston, Epidemiology of Alcohol and other Drug Use
among American College Students, Journal of Studies on Alcohol, S14, pp. 23-39,
2002.
3. T. S. Naimi, R. D. Brewer, A. Mokdad, C. Denny, M. K. Serdula, and J. S. Marks,
Binge Drinking among U.S. Adults, Journal of the American Medical Association,
289, pp. 70-75, 2003.
4. Centers for Disease Prevention and Control (CDC), Youth Risk Behavior Surveillance:
National College Health Risk Behavior Survey—United States, 1995, Morbidity and
Mortality Weekly Report, 46, pp. 1-54, 1997.
5. W. DeJong, Finding Common Ground for Effective Campus-Based Prevention,
Psychology of Addictive Behaviors, 15:4, pp. 292-296, December 2001.
6. W. DeJong, Definitions of Binge Drinking [Letter to the Editor]. Journal of the
American Medical Association, 289, p. 1635, 2003.
7. K. M. Jackson, K. J. Sher, H. J. Gotham, and P. K. Wood, Transitioning Into and Out
of Large-Effect Drinking in Young Adulthood, Journal of Abnormal Psychology,
110, pp. 378-391, 2001.
8. A. C. Merline, P. M. O’Malley, J. E. Schulenberg, J. E. Bachman, and L. D. Johnston,
Substance Use Among Adults 35 Years of Age: Prevalence, Adulthood Predictors,
and Impact of Adolescent Substance Use, American Journal of Public Health, 94,
pp. 96-102, 2004.
9. H. Wechsler, K. Kelly, E. Weitzman, J. P. SanGiovanni, and M. Seibring, What
Colleges are Doing about Binge Drinking: A Survey of College Presidents. Journal of
American College Health, 48, pp. 219-226, 2000.
10. H. Wechsler, T. F. Nelson, J .E. Lee, M. Seibring, C. Lewis, and R. P. Keeling,
Perception and Reality: A National Evaluation of Social Norms Marketing
Interventions to Reduce College Students’ Heavy Alcohol Use, Journal of Studies on
Alcohol, 64:4, pp. 484-494, 2003.
11. S. Campo, D. Brossard, M. S. Frazer, T. Marchell, D. Lewis, and J. Talbot, Are Social
Norms Campaigns Really Magic Bullets? Assessing the Effects of Students’ Misperceptions on Drinking Behavior, Health Communication, 15:4, pp. 481-497 2003.
12. J. D. Clapp, J. E. Lange, C. Russell, A. Shillington, R. B. Voas, A Failed Norms Social
Marketing Campaign, Journal of Studies on Alcohol, 64:3, pp. 409-414, May 2003.
13. M. E. Larimer and J. M. Cronce, Identification, Prevention and Treatment: A Review
of Individual-Focused Strategies to Reduce Problematic Alcohol Consumption by
College Students, Journal of Studies on Alcohol, S14, pp. 148-163, 2002.
264
/
WEITZMAN AND NELSON
14. G. Rose, The Strategy of Preventive Medicine, Oxford University Press, Oxford, 1992.
15. B. Lewis and G. Rose, Prevention of Coronary Heart Disease: Putting Theory into
Practice, Journal of the Royal College of Physicians of London, 25, pp. 21-26, 1991.
16. H. Wechsler, G. W. Dowdall, G. Maenner, J. Gledhill-Hoyt, and H. Lee, Changes in
Binge Drinking and Related Problems among American College Students between
1993 and 1997, Journal of American College Health, 47, pp. 57-68, 1998.
17. J. R. Knight, H. Wechsler, M. Kuo, M. Seibring, E. R. Weitzman, and M. A. Schuckit,
Alcohol Abuse and Dependence among U.S. College Students, Journal of Studies on
Alcohol, 63:3, pp. 263-270, 2002.
18. H. Wechsler and B. Wuethrich, Dying to Drink: Confronting Binge Drinking on
College Campuses, Rodale Books, Emmaus, Pennsylvania, 2002.
19. E. R. Weitzman, A. Folkman, K. L. Folkman, and H. Wechsler, The Relationship
of Alcohol Outlet Density to Heavy and Frequent Drinking and Drinking-Related
Problems among College Students at Eight Universities, Health and Place, 9, pp. 1-6,
2003.
20. Center on Alcohol Marketing and Youth, Overexposed: Youth as a Target of Alcohol
Advertising in Magazines, October 3, 2002, Available at: http://campy.org/research/
21. G. Rose, Sick Individuals and Sick Populations, International Journal of Epidemiology, 14, pp. 32-38, 1985.
22. Substance Abuse and Mental Health Services Administration (SAMHSA), Center
for Substance Abuse Prevention, Division of State and Community Systems
Development, Preventing Problems Related to Alcohol Availability: Environmental
Approaches, Reference Guide, Third in the PEPS Series, DHHS Publication No.:
(SMA) 99-3298, Rockville, MD, Department of Health and Human Services, 1999.
Available at http://www.health.org/govpubs/PHD822/aar.htm)
23. National Institute on Alcohol Abuse and Alcoholism (NIAAA) Task Force on
College Drinking. A Call to Action: Changing the Culture of Drinking at U.S. Colleges,
NIH Publication No: 02-5010, National Institute on Alcohol Abuse and Alcoholism,
Bethesda, Maryland, 2002. Available at
http://www.collegedrinkingprevention.gov/Reports/TaskForce/TaskForce_TOC.aspx
24. R. J. Bonnie and M. E. O’Connell (eds.), Reducing Underage Drinking, National
Academies Press,Washington, D.C., 2003.
25. G. Rose, The Strategy of Preventive Medicine, Oxford University Press, Oxford, 1992.
26. T. Stockwell, J. Toumbourou, P. Letcher, D. Smart, A. Sanson, and L. Bond, Risk
and Protection Factors for Different Intensities of Adolescent Substance Use: When
Does the Prevention Paradox Apply? Drug and Alcohol Review, 23:1, pp. 67-77,
March 2004.
27. R. Caetano and C. Cunradi, Alcohol Dependence: A Public Health Perspective,
Addiction, 97, pp. 633-645, 2002.
28. D. A. Cohen, K. Mason, and R. Scribner, The Population Consumption Model, Alcohol
Control Practices and Alcohol-Related Traffic Fatalities, Preventive Medicine, 34,
pp. 187-198, 2001.
29. G. Gmel, S. Klingemann, R. Muller, and D. Brenner, Revising the Preventive Paradox:
The Swiss Case, Addiction, 96, pp. 273-284, 2001.
30. L. T. Midanik, T. W. Tam, T. K. Greenfield, and R. Caetano, Risk Functions
for Alcohol-Related Problems in a 1988 U.S. National Sample, Addiction, 91,
pp. 1427-1437, 1996.
COLLEGE STUDENT BINGE DRINKING /
265
31. O.-J. Skog, The Prevention Paradox Revisited, Addiction, 94, pp. 751-757, 1999.
32. E. R. Weitzman, T. F. Nelson, and H. Wechsler, Assessing Success in a CoalitionBased Environmental Preventionn Programme Targeting Alcohol Abuse and Harms,
Nordic Journal of Substance Use, 20, pp. 1-9, 2003.
33. H. Wechsler, A. Davenport, G. Dowdall, B. Moeykens, and S. Castillo, Health
and Behavioral Consequences of Binge Drinking in College: A National Survey
of Students at 140 Campuses, Journal of the American Medical Association, 272,
pp. 1672–1677, 1994.
34. H. Wechsler, J. E. Lee, M. Kuo, and H. Lee, College Binge Drinking in the 1990s: A
Continuing Problem Results of the Harvard School of Public Health 1999 College
Alcohol Study, Journal of American College Health, 48, pp. 199-210, 2000.
35. E. R. Weitzman, T. N. Nelson, H. Lee, and H. Wechsler, Reducing Drinking and
Related Harms in College: Evaluation of the “A Matter of Degree” Program, American
Journal of Preventive Medicine, 27, pp. 187-196, 2004.
36. P. H. Rossi and H. E. Freeman, Evaluation: A Systematic Approach, Sage Publications,
Inc., Newbury Park, California, 1989.
37. R. K. Yin, Case Study Research Design and Methods, Applied Social Research
Methods Series, Vol. 5, Sage Publications, Newbury Park, California, 1989.
38. SAS Institute, SAS/STAT User’s Guide, Version 6 (4th ed.), SAS Institute, Cary,
North Carolina, 1994.
39. E. R. Weitzman, T. F. Nelson, and H. Wechsler, Taking Up Binge Drinking in College:
The Influence of Personal, Social and Environmental Factors, Journal of Adolescent
Health, 32, pp. 26-35, 2003.
40. H. Wechsler, J. E. Lee, T. F. Nelson, and H. Lee, Drinking Levels, Alcohol Problems,
and Secondhand Effects in Substance-Free College Residences: Results of a National
Study, Journal of Studies on Alcohol, 62, pp. 23-31, 2001.
41. L. Wallack, Two Approaches to Health Promotion in the Mass Media, World Health
Forum, 11, pp. 143-154; discussion, pp. 155-164, 1990.
42. L. Wallack and L. Dorfman, Media Advocacy: A Strategy for Advancing Policy
and Promoting Health, Health Education Quarterly, 23, pp. 293-317, 1996.
43. J. M. Moskowitz, The Primary Prevention of Alcohol Problems: A Critical Review
of the Research Literature, Journal of Studies on Alcohol, 50, pp. 54-88, 1989.
Direct reprint requests to:
Elissa R. Weitzman, ScD, MSc
Dept. of Society, Human Development and Health
Harvard School of Public Health
Landmark Center
401 Park Drive
P.O. Box 15678
Boston, MA 02215
e-mail: eweitzma@hsph.harvard.edu