Can School Alcohol Education Programs Contribute to Declines in Drinking Among Students?
Robert E. Mann and Reginald G. Smart, CAMH
Published in Drugs: Education, Prevention and Policy, 1997, 4, 131-137
TABLE OF CONTENTS
Abstract
In Ontario, declines in alcohol use and related measures such
as heavy drinking and drinking-driving among students have
occurred since 1979. Although the reasons for these declines
remain unclear, over the same period the numbers of students
exposed to alcohol education in schools increased substantially.
In an effort to estimate what proportion of the declines in
alcohol-related behaviours among students might be due to the
increase in alcohol education, we used an effect size estimate of
.17 derived from meta-analytic studies to calculate the number of
students beneficially influenced by the increase in alcohol
education. The results suggested that the proportion of the
declines accounted for by the increase in alcohol education was
modest but larger for the more serious indicators of alcohol
problems.
In the past decade in Ontario, Canada a remarkable trend has
been observed in students' use of alcohol, and in their
experience of alcohol-related problems. Important reductions have
been observed in alcohol use and such problems as drinking and
driving in surveys of Grades 7-13 (ages 12-19 years)) students
conducted biennially since 1979 (Adlaf et al., 1993). For
example, the proportion of all students reporting any alcohol use
in the previous 12 months declined from 76.3% in 1979 to 56.5% in
1993. Similarly, the proportion of students with drivers licences
reporting driving after drinking in the previous 12 months
declined from 58.1% in 1979 to 20.4% in 1993. Similar trends have
been described in the U.S. (Johnston et al., 1993). These trends
in Ontario are probably due to many social and cultural changes
but we do not have a complete explanation for them. In the U.S.,
increases in the legal drinking age from 18 to 21 occurred in
most states in the 1980's, and these increases have been linked
to reduced drinking-driving fatalities and reduced alcohol
consumption in affected age groups (O'Malley and Wagennar, 1991).
However, in Ontario, the legal drinking age was raised by only
one year from 18 to 19 in 1978, and evaluations of that change
failed to detect an impact on accident rates (Vingilis and Smart,
1981). The relative price of alcohol in Ontario, an important
determinant of alcohol consumption and problems (Bruun et al.,
1975; Popham et al., 1976), appears to have remained stable over
the period, and the physical availability of alcohol seems to
have increased (Smart and Mann, 1994).
Prevention/health promotion activities, including alcohol
education in the schools, have increased over the past decade. In
Ontario, the proportion of students reporting exposure to alcohol
education in the previous year increased from 54% in 1979 to 78%
in 1991 (Smart et al., 1991). However, the potential impact of
these measures remains unclear. Individual studies of the effects
of alcohol education programs have provided mixed results and
reviews of this literature have been cautious in claiming any
beneficial effects (Schaps et al., 1981; Moskowitz, 1989). More
recently, this literature has been examined using meta-analytic
procedures. Bangert-Drowns (1988) and Tobler (1986) and Rundell
and Bruvold (1988) came to the same conclusion: that alcohol
education programs in schools appear to have statistically
significant, but modest, beneficial effects on drinking
behaviour.
Can these modest effects of alcohol education programs account
for any of the reductions of alcohol use and problems in Ontario
students? The availability of effect sizes derived from
meta-analytic studies and the comprehensive school survey
database (Adlaf et al., 1993) permits some preliminary
estimations of the aggregate or population level impact of the
increases in alcohol education activities observed in
Ontario.
Method
Sources of Data
The data on drinking in this study were derived from the
Ontario Student Drug Use Survey (Smart et al., 1991). That survey
is comprised of nine repeated cross-sectional surveys conducted
biennially since 1977. In this study we use only data for the
years 1979 and 1991. The target population for all surveys is
Ontario students enrolled in Grades 7, 9, 11, and 13 in the
regular public and separate (Catholic) school systems. The 1979
survey employed a stratified multistage cluster design stratified
by grade and geographic region. In 1981, two changes were made to
the study. First, the sample design was modified to employ a more
rigorous stratified (grade by region) single-stage cluster
sample, surveying students from about 200 schools. Second, the
data required weighting to account for variable sampling
fractions and nonresponse by classes and students. The number of
students surveyed by year was 4,794 from 87 schools in 1979 and
3,945 from 179 schools in 1991. Participation rates have remained
high, currently about 85%. Further sampling details are available
from the authors. Data from the survey used here refer to alcohol
use within the past year. For all surveys, the data-gathering
instrument was a self-administered, anonymous questionnaire
completed in class groups in 30 to 40-minute sessions.
To assess frequency of drinking alcohol the students were
asked the following question: "In the past 12 months, how often
have you used alcohol-liquor (rum, whiskey, etc.), wine or beer?"
Response categories were (1) no drinking in past 12 months, (2)
drank at special events (e.g. Christmas, weddings), (3) took a
sip to see what it is like, (4) drank once a month or less, (5)
two or three times a month, (6) once a week, (7) two or three
times a week, (8) four or five times a week, or (9) almost every
day. Students were also asked "how often in the past four weeks
have you had five or more drinks on the same occasion".
Prevalence was based on positive responses.
The question on drinking driving asked "How often in the past
12 months have you driven within an hour of drinking two or more
drinks of alcohol". This question was asked only of students who
had a licence to drive. The proportion of licenced students rose
from 26.0% to 39.0% between 1977 and 1991.
Exposure to alcohol education was obtained by answers to the
question "During the past school year how many classes or
lectures did you have that talked about alcohol?" The proportion
answering more than none is the estimate for overall exposure to
alcohol education.
All questions retained the same format between 1979 and 1991.
With the exception of the one on drinking-driving, questions were
asked of all students.
Estimates of the numbers of students in Grades 7 to 13 in 1979
and 1991 were obtained from the Ministry of Education.
Some Estimates and Assumptions
Some estimates of the effectiveness of school based alcohol
education programs are necessary in order to examine their likely
impact on drinking among students. Several meta-analyses have
estimated effect sizes for various types of programs and
outcomes. For example, Tobler (1986) reviewed 143 adolescent drug
prevention programs and found an effect size of .17 for alcohol
when use measures were the outcome variables. Not all of these
programs were school based but most seem to have been. Later
meta-analyses reviews of 33 programs by Bangert-Drowns (1988) and
29 by Rundell and Bruvold (1988) were restricted to school based
programs and published studies with control groups, unlike
Tobler's analysis. The mean effect sizes for alcohol education
programs in Bangert-Drown's study is .22 and Rundell and Bruvold
report .12. If we average the alcohol education effect sizes from
Tobler, Bangert-Drowns and Rundell and Bruvold, we get .17. We
used this effect size in estimating how many of the increased
numbers of students getting alcohol education would have modified
their drinking behaviour.
In choosing the effect size of .17 and using it as shown in
Table 1 we are assuming that:
(i) the effect size of .17 is a reasonable estimate
for alcohol education programs in Ontario schools. In 1989 the
Ministry of Education mandated that all schools should have
alcohol education. Many schools have now introduced them and a
large number of standardized programs are used, many of them
based on programs from the U.S.A., and evaluated programs from
the Addiction Research Foundation and the Alberta Alcohol and
Drug Addiction Commission. There is a wide diversity in these
programs but most try to address changes in knowledge, attitudes
and behaviors around alcohol;
(ii) the effect size is similar for a variety of
drinking-related behaviours, e.g. heavy drinking (5 or more
drinks per occasion), drinking-driving. Programs in Ontario
schools tend to emphasize the risks of heavy drinking such as
drinking driving and do not typically take an abstinence
orientation. Most are innovative and based on theory. Tobler's
analysis showed that alcohol education courses having goals of
wise alcohol use and intervention strategies for safety had
higher effect scores (.27 and .35) than average. Rundell and
Bruvold also showed an effect size of .14 for innovative
programs. Hence, the effect size chosen as .17 may be a
conservative estimate.
Results
Table 1 shows the change in drinking trends, drinking problems
and exposure to alcohol education between 1979 and 1991. There
are large declines in proportions of students drinking alcoholic
beverages, those having 5 or more drinks per occasion and driving
within an hour of drinking alcohol. For the first two measures,
the declines are linear but for the drinking of 5 or more drinks,
the differences are largest between 1979 and 1981 and there
appears to be no overall trend. The proportions of students
receiving some alcohol education increased steadily from a low of
54% in 1979 to 78% in 1991. The average number of reported
classes also increased from 1.3 per student to 2.2 between 1979
and 1991.
Table 1: Trends in Drinking and Alcohol Education Among Ontario Students: 1979 to 1991
Not available on the website at this time.
The results of the analysis for the effects of alcohol education
are shown in Table 2. Between 1979 and 1991 there was an increase
from 54% to 78% of students getting alcohol education. This
represents an increase of 87,637 students overall. Lipsey (1990)
indicate that an effect size of .17 is equivalent to a difference
of 8.5% between control or baseline numbers and the experimental
group numbers. We have therefore estimated that 7449 students or
8.5% of 87,637 students were affected positively by their alcohol
education under our assumption. Only 56.5% of students drank in
1991 compared to 76.9% in 1979, a decrease of 382,892 students.
Heavy drinking (5 or more drinks) students decreased by 116,534.
Drinking driving decreased dramatically and there were 88,976
fewer drinking drivers. It is clear that the estimated increased
number positively influenced by alcohol education (7,449) in the
1991 data is large enough to account for some of the decreases in
each drinking group. The estimated number beneficially influenced
by alcohol education can account for 6.4% of the decrease in
heavy drinkers and 8.4% of the decrease in drinking drivers.
However, it could potentially account for only 1.9% of the
decrease in numbers of drinkers.
Table 2: Measures of Drinking and Alcohol Education for
Ontario Students, 1979 and 1991; Percents (in brackets) and Numbers
Not available on the website at this time.
Discussion
The question posed in this investigation was whether increases in
alcohol education could account for an important portion of the
declines in students' alcohol use and problems observed in
Ontario between 1979 and 1983. The answer is a qualified yes the
strength of which depends on the measures selected. For the
measure of any alcohol use, only 1.9% of the decreased cases
could be potentially accounted for by the increase in alcohol
education. However, about 6.4% of the decrease in the number of
students drinking five or more drinks on at least one occasion in
the past month and 8.4% of the number of students reporting
drinking and driving, respectively, could potentially be
accounted for by the increase in alcohol education.
Keeping in mind that the size of the reduction in
alcohol-related behavior accounted for by increased education was
larger for behaviors that would be considered more serious, our
results suggest that the increases in alcohol education can
account for a modest proportion of the declines in students'
reported alcohol use and problems. Of course, care must be taken
in interpreting these results. First, many factors such as
broader lifestyle influences, family interactions and media
campaigns, have influenced the drinking behavior in young people
in the past decade. These factors in combination are probably
more influential than school alcohol education. For example,
media and policy level attention to the drinking-driving issue in
North America in the past decade may have influenced students,
but we do not have research-based estimates on how large that
influence might be. A related issue is that the effectiveness of
alcohol education probably is influenced in an interactive
fashion by events occurring in the broader culture, e.g., a
drinking-driving or anti-alcohol campaign in the media may act to
increase the effectiveness of alcohol education in the schools.
Despite the small effect sizes for alcohol education, they are
similar to that found for many types of educational programs but
lower than for most psychotherapy programs as reviewed by Lipsey
and Wilson (1993).
With these reservations in mind, it is worth noting some
convergence of results with other research on population-level
effects of prevention activities. For example, significant
associations between declines in cirrhosis mortality and
morbidity levels, and increases in prevention activities such as
treatment for alcohol abuse and A.A. membership have been noted
(Holder and Parker, 1992; Mann et al., 1988, 1991). However, our
estimates of the proportions of reductions in students' drinking
and related behaviors attributable to alcohol education are lower
than the proportion of the reductions in cirrhosis deaths and
hospitalizations estimated to be due to increased treatment
levels (Smart and Mann, 1993). These findings suggest several
possibilities; (i) the impact of prevention may be less than that
for treatment; (ii) our ability to quantify the factors that
influence youthful drinking and related behaviors is not as good
as our ability to quantify factors influencing cirrhosis
mortality and morbidity; (iii) a broader range of factors
influences the drinking behavior of young people than influences
the drinking behavior of heavy drinkers. Only much more extensive
research could demonstrate which possibilities are most
likely.
Acknowledgments
We are grateful to Edward Adlaf and Robert Bangert-Drowns for
their valuable comments on this work.
References
Adlaf, E.M., Smart, R.G. & Walsh G.W. The Ontario student drug
use survey 1997-1993, Addiction Research Foundation, Toronto,
1993.
Bangert-Drowns, R.L. The effects of school-based substance
abuse education -a meta analysis. Journal of Drug Education, 18,
pp. 243-264, 1988.
Bruun, K., Edwards, G., Lumio, M., Makela, K., Pan, L.,
Popham, R.E. Room, R, Schmidt, W., Skog, O.J., Sulkunen, P., &
Osterberg, E. Alcohol control policies in public health
perspective. (Finnish Foundation for Alcohol Studies), Helsinki,
1975.
Holder, H.D. & Parker, R.N. Effect of alcoholism treatment on
cirrhosis mortality: A twenty-year multivariate time series
analysis. British Journal of Addiction, 87, pp. 1263-1274,
1992.
Johnston, L., Bachman, J. & O'Malley, P. Drug use rises among
American teenagers as fewer see dangers and as peer norms begin
to change. (Ann Arbor, Institute for Social Research), 1993.
Lipsey, M.W. Design Sensitivity: Statistical Power for
Experimental Research. Sage Publications, Newbury Park, 1990.
Lipsey, M.W. and Wilson, D.B. The efficacy of psychological,
educational and behavioural treatment. American Psychologist, 48,
pp. 1181-1209, 1993.
Mann, R.E., Smart, R.G., Anglin, L. & Rush B. Are decreases in
liver cirrhosis rates a result of increased treatment for
alcoholism? British Journal of Addiction, 83, pp. 683-688,
1988.
Mann, R.E., Smart, R.G., Anglin L. & Adlaf, E.M. Reductions in
cirrhosis deaths in the United States: Associations with per
capita consumption and AA membership. Journal of Studies on
Alcohol, 52, pp. 361-365, 1991.
Moskowitz, J.M. The primary prevention of alcohol problems: a
critical review of the research literature. Journal of Studies on
Alcohol, 50, pp. 54-88, 1989.
O'Malley, P. & Wagenaar, A. C. Effects of minimum drinking age
laws on alcohol use, related behaviors, and traffic crash
involvement among American youth: 1976-1987. Journal of Studies
on Alcohol, 52, 478-491, 1991.
Popham, R.E., Schmidt, W. & deLint, J. The effects of legal
restraint on drinking, in Kissin, B. and Begleiter, H. (Eds) The
Biology of Alcoholism, Volume 4: Social Aspects of Alcoholism.
Plenum, New York, 1976.
Rundell, T.G. and Bruvold, W.H. A meta analysis of
school-based smoking and alcohol use prevention programs. Health
Education Quarterly, 15, pp. 317-334, 1988.
Schaps, E., DiBartolo, R., Moskowitz, J., Palley, C.S. &
Churgin S. A review of 127 drug abuse prevention program
evaluations. Journal of Drug Issues, 11, pp. 17-43, 1981.
Smart, R.G. & Mann, R.E. The relative importance of treatment,
health promotion and alcohol controls in the decrease of alcohol
consumption and problems in Ontario: 1975-1993. Unpublished
manuscript. Addiction Research Foundation, Toronto, 1994.
Smart, R.G., Adlaf, E.M. & Walsh, G.W. The Ontario Student
Drug Use Survey: Trends Between 1977 and 1991. Addiction Research
Foundation, Toronto, 1991.
Smart, R.G. & Mann, R.E. Recent liver cirrhosis declines:
Estimates of the impact of alcohol abuse treatment and alcoholics
anonymous. Addiction, 88, pp. 193-198, 1993.
Tobler, N.S. Meta-analysis of 143 adolescent drug prevention
programs: Quantitative outcome results of program participants
compared to a control or comparison group. Journal of Drug
Issues, 16, pp. 537-567, 1986.
Vingilis, E. & Smart, R.G. Effects of raising the legal
drinking age in Ontario. British Journal of Addiction, 76,
415-424, 1981.
The views expressed in this paper are those of the authors and
do not necessarily reflect those of the Centre for Addiction and
Mental Health.
Direct reprint requests to
Robert E. Mann
Centre for Addiction and Mental Health
(previously the Addiction Research Foundation)
33 Russell St.
Toronto, ON, M5S 2S1
Tel. 416-595-6111
Web Site: http://www.camh.net
Publisher
Carfax Publishing Limited
Contact: Margaret Hutchinson
Tel: +44 (0) 1235 401 000
Fax: +44 (0) 1235 401 550
E-mail: margh@CARFAX.CO.UK
Web Page: http://www.carfax.co.uk
Address:
P.O. Box 25, Abingdon, Oxfordshire
OX14 3UE, United Kingdom