On the importance of peer influence for adolescent drug use

Addiction (1996) 91(2), 185-198
On the importance of peer influence for
adolescent drug use: commonly neglected
Department of Health Behavior and Health Education, School of Public Health, University of
North Carolina at Chapel Hill, Chapel Hill, North Carolina & ^Research Triangle Institute,
Research Triangle Park, North Carolina, USA
Peer influence is generally believed to be a major cause of adolescent drug behavior. This paper reviews
research findings on friend selection and projection to suggest that the magnitude of friend influence may be
overestimated. This paper also observes that, although adolescent drug use is assumed to begin in response to
peer group influence, peer groups have rarely been measured in studies of drug behavior. Social network
analysis is identified as a promising method for measuring peer groups. The implications of this review for
research and programs are considered.
The accumulated wisdom of more than two
decades of research on adolescent drug use is
that peer influence is a prominent cause, if not
the most important determinant, among a complicated set of circumstances and risk factors.'"'
The following sampling of quotes, although only
a small portion of the many in the research and
practice literature on adolescent drug use, shows
the importance commonly awarded to peer
...the potential adolescent user of drugs has
many peers who take various substances, and
the formation of friendship circles with such
individuals appears to be a strong causal
influence in drug taking.'"
The best predictors of adolescents' substance
use are the proportion of friends who are users
and their friends' tolerance of use. Adolescent
alcohol and drug use appears to conform to
the behavioral and value structure of the peer
A wide range of influences on initial involvement in substance use has been identifled.
Nevertheless, peer influences (modeling use,
provision of substances, and encouraging use)
are the most consistent and strongest of all
Correspondence to: Karl E. Bauman, Department of Health Behavior and Health Education, School of Public
Health, Rosenau Hall CB#7400, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina,
27599, USA.
0965-2140/96/020185-14 88.00 © Society for the Study of Addiction to Alcohol and other Drugs
Carfax Publishing Company
Karl E. Bauman & Susan T. Ennett
It appears that 'indirect' influence or pressure
is the primary factor in smoking behavior
among adolescents. By associating with friends
who smoke, the adolescent acquires attitudes
favorable to smoking. Thus, it may be more
beneficial for research and health promotion
programs directed toward the prevention of
adolescent cigarette smoking to emphasize the
centrality of the peer group."
Many young people use drugs because their
friends use drugs.'^
The influence of peers has been posited as the
single most important factor in determining
when and how cigarettes are first tried.'^
As the quotes illustrate, peers are believed to
contribute to adolescent drug use both directly
and indirectly through several complex mechanisms: by modeling drug use; by shaping norms,
attitudes, and values; and by providing opportunities and support for drug use.'"*"" The assumption that peers are central to adolescent
drug use is reflected in the social influence
paradigm underlying many drug prevention programs. These programs are founded in part on
the assumed need to buffer peer pressure to use
drugs.'' '»-25
This paper examines critically the importance
attributed to peer influence for adolescent substance use by reviewing contrary evidence.
Specifically, the roles of friendship selection and
behavioral projection are examined. Our review
suggests that peer influence may be less important than commonly believed, although it may
have some limited consequence for adolescent
drug use. Some of the contrary evidence we
consider has been discussed elsewhere,^'' ^' but
that has been rare. The only comprehensive consideration given to this paper's central issues was
a brief comment that referred to little of the
relevant research.^^ Indeed, whether peer
influence might be overestimated has rarely been
mentioned in the research literature"' ^'' ^° and
none of the reports quoted above acknowledged
any of the points that are the focus of this paper.
A related purpose of this paper is to show that
although studies of peer influence typically invoke considerations of peer groups, peer groups
have rarely been measured in this research area.
The implications of not measuring peer groups,
and the potential of social network analysis for
this purpose, are discussed.
To simplify our task we confine our attention
to the initiation of drug use and we exclude other
behaviors, such as cessation of use by regular,
habitual or addicted users. We proceed by presenting the main evidence often offered for the
thesis that peer influence is a strong determinant
of adolescent drug use. We then present evidence that suggests peer influence for adolescent
drug use may not be as strong as commonly
believed. We next discuss peer groups in the
context of social network analysis. We also identify and evaluate other types of evidence that have
been used to document peer influence. Finally,
we consider the implications of our analysis for
theory, research and drug prevention programs.
Main empirical support offered for the peer
influence axiom
A primary foundation for the belief that peer
influence is a major determinant of adolescent
drug behavior is the strong and consistent correlation between drug use by adolescents and the
drug use that adolescents attribute to their
friends. Research on adolescent drinking and
smoking illustrates this correlation.^' Seventh
graders who said their friends drank alcohol were
much more likely themselves to drink beer
(r=0.48, /)<0.001, n= 1330) and hard liquor
(r = 0.38, p < 0.001, n= 1330) than adolescents
who said their friends did not drink. In a sample
of ninth graders, the correlation between smoking by friends was 0.56 (p<0.001, n=1324).
These correlations are strong by behavioral science standards, are replicated in many studies,
and are found across all drugs considered. Moreover, the statistically significant relationship between the drug behavior of friends is maintained
when other variables are controlled, and friend
behavior typically is stronger than other predictors such as parental behaviors, demographic
characteristics, attitudes and personality attributes.'"' '^^^ The inference typically drawn
from this evidence is that it reflects friend
influence on adolescent drug use initiation.
Qualification of the peer influence axiom
Two types of evidence suggest that peer
influence might be a less important determinant
of adolescent drug behavior than suggested
above: (a) that friendships are determined in
large part by drug use (selection) and (b) that
Peers and adolescent drug use
adolescents attribute their own behavior to
friends (projection).
Friends have similar drug behavior when their
friendships are formed on the basis of common
drug behavior. In these cases, selection rather
than influence produces the association between
friend and adolescent drug use. Selection has
multiple mechanisms: (a) drug users choose
other users to be friends, (b) non-users choose
other non-users to be friends, (c) friendships
dissolve when the drug behavior of friends becomes dissimilar (deselection), and (d) peer
groups restrict membership to people with drug
behaviors like their own. Each mechanism produces an association between the drug behavior
of friends and the adolescent, but it is incorrect
to attribute it to peer influence.
The selection model posits that drug behavior
causes friendships, whereas the influence model
posits that friendships cause drug behavior.
These models assume opposite causal directions,
and knowing which is true or relatively important is fundamental to understanding the etiology
of drug behavior.
Simultaneous examination of selection and
influence effects requires longitudinal data, the
ability to link data provided by friends and adequate numbers of stable and changed friendships and drug behaviors. Given these data
requirements, it is not surprising that few studies
have had the data necessary to determine how
much of the observed association between the
drug behavior of friends is due to peer influence
and how much is due to friend selection.
Five samples of adolescents have been studied
in which peer influence and selection effects
were simultaneously examined. The flndings
suggest that selection contributes substantially to
friend similarity in drug behavior, and therefore
that friend influence is overestimated when selection is not controlled.
Cohen studied 49 friendship groups in a
white, working-class high school during the
1958-59 school year to determine the relative
contributions of influence, selection and deselection to group homogeneity according to 18 individual characteristics.*' The characteristics
studied included smoking, liquor use and beer
drinking. Cohen found that much of the homogeneity of peer group members in these behav-
iors was due to initial selection on the basis of
the behaviors. The contributions of selection to
the homogeneity of boy peer groups were 44%,
55% and 40% for hard liquor drinking, smoking
frequency and beer drinking frequency, respectively. The contributions for girls were 69%, 52%
and 79%, respectively.
Kandel studied similarity of marijuana use in
best friend dyads in New York State high schools
at the beginning and end of the 1971-72 school
year."" The (j) coefficients calculated to express
the strengths of influence and selection effects
were 0.43 and 0.45 (n = 783), respectively, suggesting that they contributed equally to the similarity of marijuana use by friends. One
commentator on Kandel's findings wrote that
attributing '...the similarity of marijuana use between friends to the mutual influence of friends
on each other overestimates influence of friends
by approximately 100 percent by failing to take
into account the process of friend selection."*'
Fisher & Bauman studied beer drinking and
cigarette smoking in two separate studies of seventh and ninth graders, respectively. The subjects were enrolled in North Carolina schools
specifically chosen to be typical of most other
schools.** Subjects completed questionnaires in
their homes in 1980 and again 1 year later in
1981. Information was collected that enabled
linkage of data provided by friends. For both
smoking and alcohol, selection effects were
stronger than influence effects. Across all drug
behaviors, 4> coefficients computed to estimate
friend influence for initiation were SO. 16,
whereas four of six (/> coefficients calculated to
express friend selection based on drug behaviors
were s 0.35. Ennett & Bauman reanalysed these
data after creating peer groups with social network techniques.'" They concluded that selection and influence make equal contributions to
drug behavior homogeneity of peer groups.
Aseltine studied marijuana use by 435 friendship pairs enrolled in public high schools in three
Boston area communities.'" Data were collected
from the predominantly white and Catholic subjects in three waves from 1988-90. Reciprocal
associations between marijuana use by friends
were examined using the panel feature of the
research design and structural modeling. The
findings suggested that failure to control for
selection effects when examining the association
between drug behaviors of friends may overestimate peer influence by nearly 60%.
Karl E. Bauman & Susan T. Ennett
These studies suggest that selection plays a
substantial role in creating similar drug behaviors
among friends. Nevertheless, once friendships
are formed they may reinforce common behaviors and, in that sense, operate as an influence
among users. Dealing with influence for users
extends beyond our focus on drug use initiation,
but it is an important topic for future research.
We conclude from the studies that have been
able to consider both influence and selection
effects that selection may make a substantial
contribution to the association between drug behaviors of friends and that failure to control for
selection may overestimate the contribution of
ban adolescents in grades 8 and 11 in a large
metropolitan area." Perceived reports of friend
smoking were more strongly associated vmh the
adolescents' reports of their own smoking
(kappa = 0.49) than were actual reports of friend
smoking (kappa = 0.24). Similarly, Iannotti &
Bush surveyed 3160 fourth and flfth graders
about their use of alcohol, cigarettes, marijuana
and crack cocaine.'^ For each drug behavior,
perceived reports were more strongly correlated
than actual reports with adolescent drug use.
It might be argued that perceived reports are
more valid indicators of friend behavior than
actual reports. However, the opposite was suggested by two studies that used biochemical indicators of friend cigarette smoking and found the
indicators to be more strongly correlated with
actual than with perceived reports."' "
In nearly all studies of adolescent drug behavior,
A possible problem with the use of perceived
the behavior of friends is measured by asking reports to study peer influence derives from the
adolescents to describe the behavior of their well-established principle that people project
friends rather than by asking the friends to report their own attributes, including their own behavtheir own behavior. We refer to these two mea- iors, to others.'^" This principle can be exsures of friend behavior as 'perceived and actual tended to the present concern: when asked to
reports', respectively. If perceived and actual re- describe the drug behavior of friends, some adoports of friend behavior are strongly correlated, lescents may project their own behavior to their
and if they are equally correlated with adolescent friends. When projection occurs and the behavdrug behavior, then the distinction between iors of the friend and adolescent actually differ, a
them would be unimportant. However, in the misclassification of friend behavior always infew studies in which both measures were used, creases the positive association between the beperceived reports were much more strongly re- havior of friends. In every such case, friend and
lated than actual reports to adolescent drug use. subject behavior are considered to be the same
The North Carolina study of alcohol use de- when they actually differ. Moreover, in contrascribed above obtained both perceived and ac- diction to the peer influence model, which posits
tual reports of friend behavior. The correlation friends as a cause of adolescent behavior, when
between perceived reports of friend beer drinking projection occurs the behavior considered to be
and the adolescents' own beer drinking was the friend's is the consequence rather the cause
r— 0.48; the correlation of 0.07, when substitut- of the adolescent's behavior. The assumed coning actual for perceived reports, was signiflcantly tribution of peer influence to adolescent drug
different and much smaller (p< 0.001, n = behavior, therefore, may be artiflcially inflated
1330).^' The corresponding correlations for hard when perceived reports of friend behavior are
liquor drinking in that study were 0.38 for used." That adolescents overestimate drug use
perceived reports and 0.11 for actual reports prevalence, and that this is most pronounced
(p<0.001, M=1321). In the North Carolina among users, may reflect and magnify this source
study of cigarette smoking, the correlation of of bias.'*
0.56 between perceived reports of friend
Adolescents do appear to attribute their own
behavior and adolescents' own behaviors was cigarette smoking to the behavior of their friends.
signiflcantly larger (p<0.001, n=1324) than We reanalysed data from the study of smoking
the correlation of 0.42 between actual reports among North Carolina ninth graders discussed
of friend behavior and adolescents' own earlier in this review.'"' ""^ These analyses are
limited to the s 843 friendship pairs with inforThe above flndings were replicated by Urberg mation for all of the necessary measures. Suband her colleagues in their study of 2334 subur- jects were classified as current cigarette smokers
Peers and adolescent drug use
who said they were current smokers and had
smoked at least a pack of cigarettes in their
lifetime or who said they were not current smokers but had carbon monoxide levels 2:9 parts
per million (p.p.m.) in their alveolar breath.'' All
other subjects were classified as non-smokers.
The findings were consistent with the conclusion
that there may be a strong tendency for adolescents to use their own smoking behavior to describe the smoking behavior of their friends.
Specifically, when the best friend was not a
smoker 22.2% of the adolescents who smoked
said the best friend smoked, whereas 2.3% of
the adolescents who did not smoke said their
best friend smoked (x^ = 60.9, contingency
coefficient = 0.267, p = 0.000, n = 791). Among
adolescents whose best friends were current
smokers, 88.7% of adolescents who were smokers said their best friend was a smoker, whereas
37.5% of the non-smokers said that their best
friend was a smoker (;(^ = 31.8, contingency
coefficient = 0.462, /) = 0.000, n=117). The
same strong patterns were observed when considering second and third best friends.
The findings in the preceding paragraph could
be attributable to factors other than projection.
They could refiect a tendency of friends to give
the impression to others that they are more like
them than they really are, and it is the false
impression by friends rather than smoking behavior that is being reported by the adolescent.
In this case, however, the association between
perceived reports of friend behavior and adolescent behavior is not due to the infiuence of the
friend's smoking behavior. Another possibility
could be that adolescents smoked in the presence of fiiends, subsequently quit smoking, and
accurately reported themselves as non-smokers
while their friends continued to think of them as
smokers. Because smokers are likely to smoke
with other smokers, our data could make it appear that behavior is overestimated when it is
not. Or adolescents who try smoking (i.e. experiment with the substance) could consider themselves to be non-smokers but be considered
smokers because of the experimentation by their
smoking friends. There has been no research on
how projection might infiuence reports of friend
behavior, perhaps because the possibility is seldom recognized. These alternative explanations,
therefore, are offered as suggestions for consideration by future research and not as evidence.
The use of perceived reports of friend behavior
is sometimes justified by the reasoning that what
adolescents think their fiiends do is more
infiuential than what the friends actually do.
Note that in this reasoning the cause of adolescent behavior is assigned to the adolescent
and not to the friend; the peer infiuence model
assigns cause to the friend. Another reason for
the use of perceived reports is that they are much
less expensive to obtain because they do not
require data collection from the fiiends. Regardless of the reason, fnend infiuence will be overestimated by any part of the association between
adolescent behavior and perceived reports that is
due to inaccurate projection of adolescent behavior to the friend.
We conclude that in studies relying on adolescent reports of friend behavior, the association
between the drug behavior of friends may exaggerate the role of peer infiuence.
Control for selection and projection effects
With a cross-sectional research design, data are
gathered once. It is the design used most often to
study adolescent drug use, but it is inadequate
for studying peer infiuence because selection effects cannot be controlled. Studies with a panel
research design, in which data are collected from
the same subjects more than once and the data
are linked by subject, may potentially remove
selection effects because prior adolescent behavior can be analytically controlled when examining the association between friend behavior and
subsequent adolescent behavior. Moreover,
when a panel design is used and prior adolescent
behavior is controlled, the effects of projection of
adolescent behavior to fiiend behavior is reduced
or eliminated because adolescent behavior is the
source of projection.
It follows that when a panel design is used and
prior adolescent behavior is controlled, the correlation between perceived reports of friend behavior and adolescent behavior should be
substantially smaller than the correlation obtained with a cross-sectional design. That is the
case. In the North Carolina studies referred to
above, the cross-sectional bivariate correlations
between adolescent reports of their own and
their friends' drinking and smoking were reduced to a range of 0.07-0.12 from 0.38-0.56
when panel data were used and initial drug behavior was introduced as a covariate in the association between friend behavior at the first
Karl E. Bauman & Susan T. Ennett
data collection point and the subsequent behavior of the adolescent (p<0.01, n=11431203)." Although the reduced strength of
relationships in longitudinal data might be in
part attributable to other factors, such as the lag
between data collection points, it is likely due in
large part to the removal of selection and projection effects.
Despite the reduction in the relationship, it is
noteworthy in the North Carolina studies and in
other panel research that the statistically
significant associations between friend behavior
and subject use onset do not disappear completely with control for prior subject use. This
supports the existence of at least a moderate peer
influence effect.'^- '2-34, 36, 4o. 43. ss-eo
Social networks
The assumption that drug use begins and is
nourished in peer groups is summarized by Oetting & Beauvais:*'
...adolescent drug use is very strongly linked
to membership in small groups of people, including pairs such as best friends and
boyfriend-girlfriend. These small groups are
peer clusters, in which (a) drugs are made
available; (b) the youth learns to use them; (c)
there is a sharing of beliefs, attitudes, values,
and rationales for drug use; and (d) drug use
plays an important role in group membership
and identification (p. 19).
friendship circles that typify adolescent peer
Clearly, friendship groups must be identified
before drug use within them, outside them and
between them can be considered. The most direct approach to identifying friendship groups is
by application of social network analysis. Network analysis uses aggregated data on the relationships among individuals in a social system
(e.g. the friendship links among adolescents who
go to the same school) to identify groups. Network analysis identifies friendship groups by
analysing present and absent fnendship links between adolescents. Friendship links are identified
by asking adolescents to name their fi^ends;
overlapping links are then used to detect groups
and other friendship patterns. Note that this
approach to group identification is completely
different from inferring group membership from
the behavior that adolescents attribute to friends.
The use of social network analysis requires data
from both adolescents and friends, and it must
be possible to link their data. Few studies have
obtained the necessary data.
The identification of peer groups and assessment of drug use by peer group members and
non-members have been done in only two samples."*'' *' As described above, Cohen examined
determinants of peer group homogeneity in high
school friendship groups.*' The results showed
that homogeneity was higher among peer groups
than among non-peer group members for a variety of individual characteristics, including
cigarette smoking, liquor use and beer drinking.
The results did not show, however, whether peer
This image of why adolescents begin to use group homogeneity resulted from members bedrugs is based on the premise that peer groups ing users or non-users, nor did the study comcause adolescents to use drugs through peer pare use and non-use among peer group
group infiuence. It has been suggested that ado- members and non-members.
lescents isolated from peer groups may be proMore recently, Ennett & Bauman conducted
tected from using drugs because that potent social network analyses of adolescents in five
junior high schools and investigated whether
source of influence is eliminated.'
However, few studies of adolescent drug use adolescents who were group members or nonhave either measured peer groups as groups or members differed in smoking.*' Ninety per cent
compared the drug use of peer group members of adolescents who were peer group members
and non-members. As such, there is little evi- were non-smokers; in fact, 68% of the peer
dence that adolescent drug use is strongly linked groups were composed entirely of non-smokto membership in 'peer clusters',*' 'friendship ers.** In contrast, only 28% of smokers were
circles','" or 'social groups'.'^ The closest that group members. This finding, which was consistresearchers have typically come to directly exam- ent across the school samples, is in contrast to
ining drug use patterns in peer social units is in the belief that smoking is a peer group phenomthe context of friendship pairs'**' "** or crowds of enon, or that social isolation is protective.
adolescents;*^' *^ however, these are not the
The implication of the findings is that to the
Peers and adolescent drug use
extent that influence occurs in peer groups, for
nearly all adolescents it favors non-smoking
rather than smoking. Because most groups were
composed of non-smokers, the influence for
most adolescents could not have been smokers
influencing non-smokers to smoke. Furthermore, because smokers were most likely to be
isolates, they were immune from the influence
associated with being a member of a peer group.
However, as other researchers have suggested,
adolescents may be influenced by a peer group
they aspire to join and model the typical behavior of that group." The prototypical behavior of
most peer groups in this study was non-smoking,
however, suggesting that even for isolates, peer
group influence must have been in the direction
of non-smoking rather than smoking.
We conclude that regardless of the importance
to adolescent drug use that has been ascribed to
peer groups, few studies have examined peer
groups directly. Social network analysis is a direct method to identify peer groups and to examine drug use in the context of peer groups.
Additional considerations
Evaluation of other ways to assess peer influence
At the beginning of this paper, we noted that the
evidence usually offered for a strong peer
influence effect is the strong cross-sectional correlation that has been found between adolescent
drug behavior and the drug behavior that adolescents attribute to their friends. We briefly evaluate here other approaches that have been used to
assess peer influence.
Adolescents have been asked why they began
to use drugs, and references to friends are taken
as indicative of friend influence. This retrospective method relies on adolescents' ability to know
why they behave the way they do; many psychological theories and studies suggest that direct
questioning of individuals frequently fails to
identify the correct etiology. ^'"'^ Moreover, because teachers and the media inform young people that they use drugs because of their friends,
perhaps that is why adolescents give that as a
reason for using drugs. Another approach is to
ask adolescents whether they were with friends
when they flrst used drugs, or whether their
friends gave them drugs, or if their friends pressured them to use drugs.''"" This method is
limited by its failure to provide comparable in-
formation from subjects who did not use drugs,
thereby precluding the ability to identify an association, the most elementary information required for inferring causality.'*
Another approach to the study of peer
influence is to have adolescents describe the
different types of groups at their schools. The
four or flve groups identifled in this way then are
given labels, such as 'dirts', 'hot-shots', 'jocks'
and 'regulars'. The variation in smoking behavior across these groups is attributed in part to
peer group influence.'^' *'• "• '* However, this
approach does not measure peer groups. The
aggregates formed might most appropriately be
considered 'crowds' because they consist of collectives of stereotyped individuals who may not
have intimate ties rather than the smaller and
closer knit primary groups typically considered in
the adolescent drug use literature." Comparing
the four or flve aggregates identifled in studies of
crowds with the much larger number of peer
groups identifled in studies of dyads and peer
groups makes it clear that adolescents have many
more friendship pairs and small groups that are
commonly considered to be most influential on
behavior.^' Moreover, subjects may use the drug
behavior of crowds to define and differentiate
crowds in a school, yielding a measurement redundancy when correlating crowd membership
and drug behavior." Finally, studies of crowds
have yet to separate influence and selection effects, and projection may misleadingly inflate
associations between friend and subject drug behavior. The net effect is that studies of crowds
may substantially overestimate the impact of
peer influence on adolescent drug behavior.
Other studies have considered normative expectations to be evidence of peer influence. Normative expectations are the estimates of drug use
prevalence that adolescents assign to general reference groups, such as to 'your peers' and to
'others your own age'. Researchers have noted
that normative expectations substantially exceed
drug prevalence in populations, and that adolescents with high normative expectations are more
likely to use drugs than adolescents with low
normative expectations.^^' ^''' ^' The association
between normative expectations and drug use
has been offered as evidence of peer influence
and used to justify focusing on reducing normative expectation in drug prevention programs.^^
However, if adolescents use their beliefs about
the behavior of their friends to gauge the behav-
Karl E. Bauman & Susan T. Ennett
ior of their more generalized reference groups,
and if the behavior they ascribe to their friends is
largely projected from their own behavior, then
peer influence is overestimated by the correlation
between normative expectations and individual
drug behavior. Moreover, if adolescents select
friends with similar drug behavior, and if drug
behavior is a cause of normative expectations,
then the correlation between normative expectations and drug behavior could be explained as
readily by selection as by influence.'*
Studies of the assumed mediators of the association between friend and adolescent drug
behavior also offer evidence that peer influence
may be overestimated. Brown & Clasen found
that when directly measuring pressure for drug
use and other misconduct, there was little peer
pressure and it explained only a small portion of
variance in misconduct.*^' *^ These flndings are
consistent with those of Urberg and her colleagues who found that when peer pressure is
examined empirically, it does not account for the
relationship between friend and adolescent
smoking." Findings such as these reduce our
confidence that peer influence is such a strong
determinant of adolescent drug behavior.
Many theories specific to the use of drugs or
related behaviors^' ^^^^ and more general behavioral and social theories*'"" include social
influence as a component. Although these theories recognize peer influence as a determinant of
behavior, some do not assign it the prominence
that now exists in the research literature on
adolescent drug use. Peers sometimes are included among a constellation of significant others, including parents, who influence adolescent
behavior. Our review is consistent with these
theoretical perspectives in suggesting that the
centrality ofren awarded to peer influence on
adolescent drug behavior should be tempered by
integrating it with other causal factors and recognizing that the importance frequently ascribed to
peer influence may have been in part due to the
other factors emphasized in this review. Aseldne's consideration of selection effects and the
use of perceived reports of fiiend behavior in the
context of the relative worth of socialization and
social control theories to understanding marijuana use is particularly instructive in this regard.'"
Our conclusion that peer influence for drug
use initiation might be overestimated has substantial additional implications. One is that the
contribution of other predictive variables may be
underestimated. Peer influence is considered a
major determinant of adolescent drug behavior
in large part because variables that describe
fiiends usually are more strongly related to adolescent drug use than other variables. In studies
that do not adjust for projection and selection
effects, the contribution of other variables to the
explanation of drug use is automatically underestimated relative to peer influence. Another implication of the overestimation of peer influence is
that when the major determinant of a behavior is
believed to be known then there can be a hesitancy to examine critically the favored determinant and a bias against seeking other causes.
This can, of course, delay progress toward
understanding behavior. The influence of families, for example, may be substantially underestimated because the importance of peer influence
is exaggerated.'^
Prevention programs
If peer influence is not as important a determinant of drug use as is commonly assumed, then
the promise of prevention programs and policies
focused on peer influence might not be realized.
Hawkins and his colleagues, in an extensive review of etiological research on adolescent drug
use and drug prevention programs, provide a
detailed portrayal of the common wisdom that
adolescent drug use will be substantially reduced
by peer resistance interventions.' The US Surgeon General's report on cigarette smoking by
youth provides another detailed consideration of
programs based on the assumption that peer
influence is a major factor in smoking onset.'^
Descriptions of programs based on other prevention strategies also assume that peer influence is
a major factor in adolescent drug use.'^' '^
The review we present in this paper suggests
that such programs will be less effective than
hoped because peer influence on drug use may
be substantially less than commonly assumed.
Perhaps this explains why drug abuse prevention
curricula that emphasize peer influences have
only moderate^'-«'-'» and short-lived"'"-"'^ behavioral effects. Whereas the 'Just Say No' campaign to reduce drug use in the United States
discourages friend influence, the findings of re-
Peers and adolescent drug use
search on social networks suggests that most
adolescents are in peer groups without smokers.
Therefore, 'Just Say Yes' to friends may yield the
more favourable public health response.
Applicability across drugs and other variables
Much of the research on peer influence has been
conducted with alcohol, tobacco or marijuana as
the drug of interest, but our flndings may generalize to other drugs such as to cocaine and
heroin. On the other hand, the principles discussed in this paper also may vary by many
factors, including the type of drug used. Similarly, we recognize that the relative influence of
peers and other variables on drug use may vary
by stage of use, such as whether progressing to
experimentation or to problem use.^''"'' '"* We
do not elaborate on these possibilities, and in
this sense our analysis is incomplete.
In addition, peer effects may vary across categories of additional variables. For example, no
peer effect was found in the predominantly
African-American sample studied longitudinally
by Farrell & Danish,'"' and other studies suggest
that associations between drug behaviors of
friends are weakest for African Americans.'"*
Moreover, we conflned our review to research on
adolescents and recognize that it might not apply
to younger children and adults. Influence and
selection effects, projection and social networks
may vary by gender, and we have ignored that
possibility here. More recognition of the evidence that peer influence might be overestimated
is a prerequisite to the more reflned analyses
suggested by these caveats.
Other sources of spuriousness
There are ways in addition to those emphasized
in this paper in which peer influence might be
spuriously increased. For example, the association between friend and adolescent behavior
might be spurious because of a common association with environmental and background variables.^"' '"' We have not considered that
possibility in detail here because many environmental and background variables have been controlled in longitudinal studies and a correlation
between friend and adolescent drug behavior has
remained.'^' ^^' '*• '"* Future research needs to
continue to determine whether the association is
produced spuriously by such variables.
Selection and projection effects for other friend characteristics
Other friend-related variables associated with
drug use, such as friend attitudes toward drugs
and peer group norms pertaining to drug use,
have been studied as correlates of adolescent
drug use with cross-sectional designs and with
reliance on perceived rather than actual measures of these predictor variables. Findings from
such studies may be limited in ways similar to
those in which perceived friend behavior is the
primary independent variable. For example,
friendships may be formed on the basis of common attitudes, and the extent to which this selection is not separated from influence when
studying the impact of friend attitudes on behavior could yield an overestimation of influence
effects. Moreover, adolescents' perceptions of
the attitudes of their friends and the norms of
their peer group may result at least in part from
the adolescents' projection of attitudes and
norms to friends; therefore, the association of
perceived friend attitudes and norms with adolescent behavior may reflect things other than
influence. Investigators are advised either to control for such factors or to provide the caveat that
what they consider to be influence may be partially or entirely spurious because of the effects of
selection and projection.
Behaviors other than drugs
Peer influence is considered to be fundamental
to many more behaviors than those involving
drugs.*'" **• ""-"' It is beyond our scope to consider in detail here how the arguments in the
present review might generalize to behaviors
other than those with drugs. We do note, however, that friend homogeneity in adolescent sexual behavior has been attributed to peer
influence, whereas research suggests that it may
be due to friend selection*" and that perceived
reports of friend sexual behavior could reflect
projection."^ Moreover, that influence can be
confused for selection has been suggested for
political orientation, delinquency, educational
aspiration, value similarity and a host of other
variables.*'- "'• '"' We leave to others the challenge of extending more fully this consideration
of the magnitude of peer influence to other adolescent behaviors.
Karl E. Bauman & Susan T. Ennett
2. FLAY, B . R . , D'AVBRNAS, J. R., BEST, J. A.,
International considerations
The studies we identified that separated
influence and selection effects, measured perceived and actual reports of friend behavior or
used social network analysis to form peer
groups were based on adolescents living in the
United States. The extent to which our review
can be generalized to other countries must await
research in those countries. Only a few studies
of adolescent drug use have compared adolescents across countries, and they found both
similarities and differences in prevalence and
correlates.'"""* Given the importance generally
ascribed to peer influence and the qualifications
we describe, such research appears to be needed.
The evidence identified here suggests that peer
influence may be a less important determinant of
the initiation of drug behavior than commonly
assumed because selection and projection may
increase the association between friend and adolescent behavior. This qualification is rarely
mentioned in the literature. The evidence also
suggests that peer influence plays a limited role
in adolescent drug use, however, and in this
regard we recommend a more critical look at the
power of peer influence rather than declare that
further consideration of peer influence is unnecessary. We also suggest that social network
analysis be considered for resolving the problem
that peer groups rarely have been measured in
research on adolescent drug use.
KERSELL, M . W . & RYAN, K . B . (1983) Cigarette
smoking: why young people do it and ways of
preventing it, in: M C G R A T H , P. J. & FIRESTONE,
P. (Eds) Pediatric and Adolescent Behavioral Medicine: Issues in Treatment, pp. 132-183 (New York,
3. GLYNN, T . J. (1981) From family to peer: a
review of transitions of influence among drugusing youth. Journal of Youth and Adolescence, 10,
(1985) Childhood predictors and prevention of
adolescent substance abuse, in: JONES, C . & BATTJES, R. (Eds) Etiology of Drug Abuse: Implications
for Prevention, pp. 75-126 (Rockville, M D , US
Y. (1992) Risk and protective factors for alcohol
and other drug problems in adolescence and early
adulthood: implications for substance abuse prevention. Psychological Bulletin, 112, 64-105.
6. LEVENTHAL, H . & CLEARY, P. D . (1980) T h e
smoking problem: a review of research and theory in behavioral risk modification. Psychological
Bulletin, 88, 370-405.
7. NEWCOMB, M . D . & BENTLER, P. M. (1989)
Substance use and abuse among children and
teenagers, American Psychologist, 44, 242-248.
8. PERRY, C . L . & MURRAY, D . M . (1985) The
prevention of adolescent drug abuse: implications
from etiological, developmental, behavioral, and
environmental models, Journal of Primary Prevention, 6, 31-52.
9. WERNER, M . J. (1991) Adolescent substance
abuse: risk factors and prevention strategies.
Maternal and Child Health Technical Information
Bulletin, 1-16.
10. HUBA, G. & BENTLER, P. (1980) The role of peer
and adult models for drug taking at different
stages of adolescence. Journal of Youth and
Adolescence, 9, 449-465.
This work was supported in part by grant
DA02480 from the National Institute on Drug
Abuse of the US Department of Health and
Human Services. Our special thanks to Dr J.
Richard Udry, Dr Vangie Foshee and anonymous reviewers for their helpful comments on
earlier drafts of this paper.
(1989) Peer group influence as a factor in smoking behavior of adolescents. Adolescence, XXIV,
ent's Guide to Prevention (Washington, D C , US
Government Printing Office).
VICES (1994) Preventing Tobacco Use Among
Young People: A Report of the Surgeon General
(Washington, D C , US Government Printing
14. ENNETT, S . T . & BAUMAN, K . E . (1991) Media-
tors in the relationship between parental and peer
characteristics and beer drinking by early adolescents. Journal of Applied Social Psychology, 2 1 ,
1. BAUMRIND, D . & MOSELLE, K . A. (1985) A
15. GRAHAM, J. W., MARKS, G . & HANSEN, W . B .
developmental perspective on adolescent drug
(1991) Social influence processes affecting
adolescent substance use. Journal of Applied Psychology, 76, 291-298.
16. KANDEL, D . B . (1985) On processes of peer
influence in adolescent drug use: a developmen-
abuse, in: BROOK, J. S., LETTIERI,
D . J. &
BROOK, D . W . (Eds) Alcohol and Substance Abuse
in Adolescence, pp. 41-67 (New York, T h e
Haworth Press).
Peers and adolescent drug use
tal perspective, in: BROOK, J. S., LETTIERI, D . J . ,
BROOK, D . W . & STIMMEL, B . (Eds) Alcohol and
Substance Abuse in Adolescence, pp. 139-163 (New
York, The Haworth Press).
17. KANDEL, D . B . & ANDREWS, K . (1987) Processes
of adolescent socialization by parents and by
peers. International Journal of the Addictions, 22,
TTie Milbank Memorial Fund Quarterly, 66, 137171.
31. BAUMAN, K . E . & FISHER, L . A. (1986) On the
measurement of friend behavior in research on
friend influence and selection: findings from
longitudinal studies of adolescent smoking and
drinking. Journal of Youth and Adolescence, 15,
S . J., SANTI, S . M . ,
32. BAILEY, S . L . & HUBBARD, R . L . (1990) Devel-
SMITH, E . A. & BROWN, K . S . (1988) Preventing
opmental variation in the context of marijuana
initiation among adolescents. Journal of Health
and Social Behavior, 31, 58-70.
33. BAUMAN, K . E . (1980) Predicting Adolescent Drug
Use: The Utility Structure and Marijuana (New
York, Praeger).
cigarette smoking among school children. Annual
Review of Public Health, 9, 161-201.
19. BOTVIN, G. & McAusTER, A. (1982) Cigarette
smoking among children and adolescents: causes
and prevention, in: ARNOLD, C . B . (Ed.) Ad-
vances in Disease Prevention, pp. 222-249 (New
York, Springer).
20. GLYNN, T . J. (1989) Essential elements of
school-based smoking prevention programs.
Journal of School Health, 59, 181-188.
21. HANSEN, W . B . (1992) School-based substance
abuse prevention: a review of the state of the art
in curriculum, 1980-1990, Health Education Research, 7, 403-430.
22. HANSEN, W . B . & GRAHAM, J. W. (1991) Pre-
venting alcohol, marijuana, and cigarette use
among adolescents: peer pressure resistance
training versus establishing conservative norms.
Preventive Medicine, 20, 414-430.
23. McAusTER, A. L., PERRY, C . & MACCOBY, N .
(1979) Adolescent smoking: onset and prevention. Pediatrics, 63, 650-658.
24. SussMAN, S. (1989) Two social influence perspectives of tobacco use development and prevention. Health Education Research, 4, 213-223.
VICES (1987) Drug Abuse and Drug Abuse Research: the second triennial report to congress
from the Secretary, Department of Health and
Human Services (Rockville, M D , US Department of Health and Human Services).
26. KANDEL, D . B . (1986) Processes of peer
influences in adolescence, in: SILBEREISEN, R . ,
K., EYFERTH, K . & RUDINGER, G. (Eds) Development as Action in Context, pp. 203-227 (Berlin,
27. ROWE, D. C , WOULBROUN, E. J. & GULLEY, B. .
L. (1994) Peers and friends as nonshared environmental influences, in: HETHERINGTON, E .
34. BAUMAN, K . E . , FISHER, L . A., BRYAN, E . S . &
CHENOWETH, R . L . (1984) Antecedents, subjective expected utility, and behavior: a panel study
of adolescent cigarette smoking. Addictive Behaviors, 9, 121-136.
D. & BENTLER, P . M . (1987) A multivariate
model of the determinants of cigarette smoking
among adolescents. Journal of Health and Social
Behavior, 28, 273-289.
36. CHASSIN, L . , PRESSON, C . C , SHERMAN, S . J.,
MONTELLO, D. & MCGREW, J. (1986) Changes
in peer and parental influence during adolescence: longitudinal versus cross-sectional perspectives on smoking initiation. Developmental
Psychology, 22, 327-334.
37. CLAYTON, R . R . & RITTER, C . (1985) The epi-
demiology of alcohol and drug abuse among adolescents, in: BROOK, J. S., LETTIERI, D . J. &
BROOK, D . W . (Eds) Alcohol and Substance Abuse
in Adolescence, pp. 69-97 (New York, T h e
Haworth Press).
38. FosHEE, V. & BAUMAN, K . (1992) Parental and
peer characteristics as modifiers of the bondbehavior relationship: an elaboration of control
theory. Journal of Health and Social Behavior, 33,
39. HUBA, G. J., WiNGARD, J. A. & BENTLER, P . M .
(1979) Beginning adolescent drug use and peer
and adult interaction patterns. Journal of Consulting and Clinical Psychology, 47, 265-276.
40. JESSOR, R . & JESSOR, S . L . (1975) Adolescent
development and the onset of drinking: a longitudinal study. Journal of Studies on Alcohol, 36,
M., REISS, D . & PLOMIN, R . (Eds) Separate Social
41. KANDEL, D . B . , TREIMAN, D . , FAUST, R . & SIN-
Worlds of Siblings: The Impact of Nonshared Environment and Development (Hillsdale, NJ,
Lawrence Erlbaum Associates, Publishers).
GLE, E. (1976) Adolescent involvement in legal
and illegal drug use: a multiple classification
analysis. Social Forces, 55, 438-458.
42. KROSNICK, J. A . & JUDD, C. M. (1982) Transitions in social influences at adolescence: who
induces cigarette smoking? Developmental Psychology, 18, 359-368.
28. BAUMAN, K . E . & ENNETT, S . T . (1994) Peer
influence on adolescent drug use, American Psychologist, 49, 820-822.
29. CHASSIN, L . , PRESSON, C . C . & SHERMAN, S . J.
(1990) Social psychological contributions to the
understanding and prevention of adolescent
cigarette smoking. Personality and Social Psychology BuUetin, 16, 133-151.
30. CLEARV, P . D . , HITCHCOCK, J. L., SEMMER, N . ,
FUNCHBAUGH, L. J. & PiNNEY, J. M. (1988)
Adolescent smoking: research and health policy.
43. CONRAD, K . M . , FLAY, B . R . & HILL, D . (1992)
Why children start smoking cigarettes: predictors
of onset, British Journal of Addiction, 87, 17111724.
44. HuNDLEBY, J. D. & MERCER, W . G . (1987)
Family and friends as social environments and
their relationship to young adolescents' use of
Karl E. Bauman & Susan T. Ennett
alcohol, tobacco, and marijuana. Journal of Marriage and the Family, 49, 151—164.
45. COHEN, J. M . (1977) Sources of peer group
homogeneity. Sociology of Education, 50, 227—
46. KANDHL, D . B . (1978) Homophily, selection, and
socialization in adolescent friendships, American
Journal of Sociology, 84, 427-436.
47. BniY, J. O. G. & UDRY, J. R. (1985) Patterns of
adolescent friendship and effects on sexual behavior. Social Psychology Quarterly, 48, 27-41.
48. FISHER, L . A. & BAUMAN, K . E . (1988) Influence
and selection in the friend-adolescent relationship: findings from studies of adolescent smoking
and drinking. Journal of Applied Social Psychology,
18, 289-314.
49. ENNETT, S. T . & BAUMAN, K . E . (1994) The
contribution of influence and selection to adolescent peer group homogeneity: the case of adolescent cigarette smoking. Journal of Personality
and Social Psychology, 67, 653-666.
50. ASELTINE, R. H. (1995) A reconsideration of
parental and peer influences on adolescent deviance. Journal of Health and Social Behavior, 36,
51. URBERG, K . A . , SHYU, S.-J. & LIANG, J. (1990)
Peer influence in adolescent cigarette smoking,
Addictive Behaviors, 15, 247-255.
52. IANNOTTI, R. J. & BUSH, P. J. (1992) Perceived
vs. actual friends' use of alcohol, cigarettes, marijuana, and cocaine: which has the most
influence? Journal of Youth and Adolescence, 21,
53. HOLMES, D . S. (1968) Dimensions of projection.
Psychological Bulletin, 69, 248-268.
54. MARKS, G . & MILLER, N . (1987) Ten years of
research on the false-consensus effect: an empirical and theoretical review. Psychological Bulletin,
102, 72-90.
(1991) Grade changes in peer influence on adolescent cigarette smoking: a comparison of two
measures. Addictive Behaviors, 16, 21-28.
61. OETTING, E . R . & BEALTVAIS, F . (1986) Peer
cluster theory: drugs and the adolescent. Journal
of Counselling and Development, 65, 17—22.
62. MosBACH, P. & LEVENTHAL, H . (1988) Peer
group identification and smoking: implications
for intervention. Journal of Abnormal Psychology,
97, 238-245.
63. SussMAN, S., DENT, C . W . , STACY, A. W. et al.
(1990) Peer-group association and adolescent tobacco use. Journal of Abnormal Psychology, 99,
64. COLHIWAN, J. S. (1961) The Adolescent Society
(New York, The Free Press of Glencoe).
65. DouvAN, E. & ADELSON, J. (1966) The Adolescent
Experience (New York, Wiley).
66. SHERIF, M . & SHERIF, C . W . (1964) Exploration
into Conformity and Deviation of Adolescents (New
York, Harper and Row).
67. ENNETT, S . T . & BAUMAN, K . E . (1993) Peer
group structure and adolescent cigarette smoking: a social network analysis. Journal of Health
and Social Behavior, 34, 226-236.
68. ENNETT, S. T . , BAUMAN, K . E . & KOCH, G . G .
(1994) Variability in cigarette smoking within
and between adolescent friendship cliques. Addictive Behaviors, 19, 295-305.
69. HARVEY, J. H. & WEARY, G . (1985) Attribution:
Basic Issues and Applications (San Diego, CA,
Academic Press, Inc.).
(1982) Judgment Under Uncertainty: Heuristics and
Biases (New York, NY, Cambridge University
71. NiSBETT, R. E. & Ross, L. (1980) Human Inference: Strategies and Shortcomings of Social Judgment
(Englewood Cliffs, NH, Prentice-Hall).
55. MULLEN, B . , ATKINS, J. L., CHAMPION, D . S. et
72. WEARY, G . , STANLEY, M . A. & HARVEY, J. H.
al. (1985) The false consensus effect: a metaanalysis of 115 hypothesis tests. Journal of Experimental Social Psychology, 21, 262-283.
(1989) Attribution (New York, NY, Springer-Verlag).
73. BiGLAN, A., SEVERSON, H . , BAVRY, J. & Mc-
CONNELL, S. (1983) Social influence and adolescent smoking: a first look behind the bam.
Health Education, September/October, 14-18.
56. BAUMAN, K . E . , BOTVIN, G . J., BOTVIN, E . M . &
BAKER, E . (1992) Normative expectations and
the behavior of significant others: an integration
of traditions in research on adolescents' cigarette
smoking. Psychological Reports, 71, 568-570.
A. (1985) Smoking onset among teens: an
empirical analysis of social situations. Addictive
Behaviors, 10, 1-13.
57. BAUMAN, K . E . & KOCH, G . G . (1983) Validity
of self-reports and descriptive and analytical conclusions: the case of cigarette smoking by adolescents and their mothers, American Journal of
Epidemiology, 118, 90-98.
58. BROOK, J. S., WHITEMAN, M . , GORDON, A. S.,
NOMURA, C . & BROOK, D . W . (1986) Onset of
adolescent drinking: a longitudinal study of intrapersonal and interpersonal antecedents, in:
Alcohol and Substance Abuse in Women and Children, pp. 91-110 (New York, The Haworth
59. KANDEL, D . B . (1978) Longitudinal Research on
Drug Use: Empirical Findings and Methodological
Issues (Washington, DC, Hemisphere-Wiley).
60. URBERG, K . A., CHENG, C . - H . & SHYU, S.-J.
M. S. (1985) Peer pressure and drug use—
exploding the myth. Adolescence, XX, 949-958.
STOUFFER, S . A . (1950) Some observations on
study design, American Journal of Sociology, 55,
ECKERT, P. (1983) Beyond the statistics of adolescent smoking, American Journal of Public
Health, 73, 439-441.
DOLCINI, M. M. & ADLER, N . E . (1994) Perceived competencies, peer group affiliation, and
risk behavior among early adolescents. Health
Psychology, 13, 496-506.
BROWN, B . B . (1989) The role of peer groups in
Peers and adolescent drug use
adolescents' adjustment to secondary school, in:
BERNDT, T . J. & LADD, G . W . (Eds) Peer Rela-
BANNING, J. (1991) Prevention, peer clusters,
and the paths to drug abuse, in: DONOHEW, L . ,
tionships in Child Development, pp. 188-215 (New
York, Wiley).
G . J.,
SYPHER, H . E . & BUKOSKI, W . J. (Eds) Persuasive
E . M . , BAKER, E . ,
DusENBURY, L. & GOLDBERG, C . J. (1992) The
false consensus effect: predicting adolescents' tobacco use from normative expectations, Psychological Reports, 70, 171-178.
8 1 . SussMAN, S., D E N T , C . W . , MBSTEL-RAUCH, J. et
al. (1988) Adolescent nonsmokers, triers, and
regular smokers' estimates of cigarette smoking
prevalence: when do overestimations occur and
by whom? Journal of Applied Social Psychology, 18,
82. BROWN, B . B . , CLASEN, D . R . & EICHER, S . A.
(1986) Perceptions of peer pressure, peer conformity dispositions, and self-reponed behavior
among adolescents. Developmental Psychology, 22,
Communication and Drug Abuse Prevention (Hillsdale, NJ, Lawrence Erlbaum Associates).
BANGERT-DROWNS, R . L . (1988) T h e effects of
school-based substance abuse education—a
meta-analysis. Journal of Drug Education, 18,
HALEBSKY, M . A . (1987) Adolescent alcohol and
substance abuse: parent and peer effects. Adolescence, 22, 961-967.
ToBLER, N . S. (1992) Drug prevention programs
can work: research lindings. Journal of Addictive
Diseases, 11, 1-28.
BRUVOLD, W . H . (1993) A meta-analysis of adolescent smoking prevention programs, American
Journal of Public Health, 83, 872-880.
99. ENNETT, S . T . , TOBLER, N . S., RINGWALT, C . L .
multidimensionality of peer pressure in adolescence. Journal of Youth and Adolescence, 14,
& FLEWELUNG, R . L . (1994) How effective is
drug abuse resistance education? A meta-analysis
of Project DARE outcome evaluations, American
Journal of Public Health, 84, pp. 1394-1401.
84. FLAY, B . R . & PETRAITIS, J. (1991) Methodolog-
100. ELUCKSON, P . L . , BELL, R . M . & M C G U I G A N , K .
83. CLASEN, D . R . & BROWN, B . B . (1985) The
ical issues in drug use prevention research: theoretical foundations, in: LEUKEFELD, C . G . &
BUKOSKI, W. J. (Eds) Drug Abuse Prevention Intervention Research: Methodological Issues, pp. 81—
109 (Washington, D C , US DHSS).
85. HuBA, G. J. & BENTLER, P. M. (1982) A developmental theory of drug use: derivation and assessment of a causal modeling approach, in:
BALTES, B . P . & BRIM, O . G . Jr (Eds) Life-span
Development and Behavior, pp. 147—203 (New
York, NY, Academic Press).
jESSOR, R. & JESSOR, S. L . (1977) Problem Behavior and Psychosodal Development (New York, NY,
Academic Press).
AKERS, R . L . (1977) Deviant Behavior: A Social
Learning Approach (Belmont, CA, Wadsworth
BANDURA, A . (1977) Social Learning Theory (Englewood Cliffs, NJ, Prentice-Hall, Inc.).
FiSHBEiN, M. & AjZEN, I. (1975) Belief, Attitude,
Intention and Behavior: An Introduction to Theory
and Research (Reading, MA, Addison-Wesley
Publishing Company).
HiRSCHi, T. (1969) Causes of Delinquency (Berkeley, CA, tJniversity of California Press).
SUTHERI^VND, B. H. & CRESSEY, D . R . (1966)
Principles of Criminology (Philadelphia, Lippincott).
92. BAUMAN, K . E . , FOSHEE, V. A., LINZER, M . A. &
KOCH, G . G . (1990) The effect of parental smoking classification on the association between
parental and adolescent smoking. Addictive Behaviors, 15, 413-422.
93. BEISECKER, A . E . (1991) Interpersonal approaches to drug abuse prevention, in: DoNOHEw, L., SYPHER, H . E . & BuKOSKi, W. J. (Eds)
Persuasive Communication and Drug Abuse Prevention, pp. 229-238 (Hillsdaie, NJ, Lawrence Erlbaum Associates).
(1993) Preventing adolescent drug use: longterm results of a junior high program, American
Journal of Public Health, 83, 856-861.
101. FLAY, B . R . , KOEPKE, D . , THOMSON, S . J. et al.
(1989) Six-year follow-up of the first Waterloo
school smoking prevention trial, American Journal
of Public Health, 79, 1371-1376.
A. I., PiRiE, P. & LuEPKER, R. V. (1988) Fourand five-year follow-up results from four seventhgrade smoking prevention strategies. Journal of
Behavioral Medicine, 11, 395-405.
103. RoWE, D. C. (1992) Family background and the
horizontal transmission of smoking. Paper presented at the Society for Research in Adolescence,
Washington, D C , 19-22 March.
104. SHEDLER, J. & BLOCK, J. (1990) Adolescent drug
use and psychological health: a longitudinal inquiry, American Psychologist, 45, 612-630.
105. FARRELL, A . D . & DANISH, S. J. (1993) Peer drug
associations and emotional restraint: Causes or
consequences of adolescents' drug use? Journal of
Consulting and Clinical Psychology, 6 1 , 327—334.
106. HEADEN, S . W . , BAUMAN, K . E . , DEANE, G . D .
& KOCH, G . G . (1991) Are the correlates of
cigarette smoking initiation different for black
and white adolescents? American Journal of Public
Health, 81, 854-858.
107. JussiM, L. & OsGOOD, D . W. (1989) Influence
and similarity among friends: an integrative
model applied to incarcerated adolescents. Social
Psychology Quarterly, 52, 98-112.
108. KANDEL, D . B . , KBSSLER, R . C . & MARGUUES,
R. Z. (1978) Antecedents of adolescent initiation
into stages of drug use: a development analysis.
Journal of Youth and Adolescence, 7, 13-40.
109. BROWN, B . B . (1990) Peer groups and peer cultures, in: FELDMAN, S . S . & E L U O T T , G . R .
(Eds) At the Threshold: The Developing Adolescent,
Karl E. Bauman & Susan T. Ennett
cohol use among adolescents in Israel, France,
pp. 171-196 (Cambridge, MA, Harvard Univerand the United States, Journal of Youth and Adosity Press).
lescence, 11, 89-113.
110. CLAUSEN, J. A. (1968) Perspectives on childhood
socialization, in: CLAUSEN, J. A. (Ed.) Socializati- 114. JESSOR, R., YOUNG, H . B., YOUNG, E . B. & TESI,
on and Society, pp. 130-181 (Boston, Little,
G. (1970) Perceived opportunity, alienation, and
Brown and Company).
drinking behavior among Italian and American
111. SAVIN-WILLIAMS, R. C . & BERNDT, T . J. (1990)
Friendship and peer relations, in: FELDMAN, S. S.
youth. Journal of Personality and Social Psychology,
15, 189-201.
& ELUOTT, G . R. (Eds) At the Threshold: The 115. KANDEL, D . B. & ADLER, I. (1982) Socialization
Developing Adolescent, pp. 277-307 (Cambridge,
MA, Harvard University Press).
112. WiLCOX, S. & UDRY, J. R. (1986) Autism and
accuracy in adolescent perceptions of friends'
sexual attitudes and behavior. Journal of Applied 116.
Social Psychology, 16, 361-374.
113. ADLER, I. & KANDEL, D . (1982) A cross-cultural
comparison of sociopsychological factors in al-
into marijuana use among French adolescents: a
cross-cultural comparison with the United States,
Journal of Health and Social Behavior, 23, 295309.
KANDBL, D . B., ADLER, I. & SUDIT, M . (1981)
The epidemiology of adolescent drug use in
France and Israel, American Journal of Public
Health, 71, 256-264.