RESEARCH REPORT doi:10.1111/j.1360-0443.2007.01855.

Alcohol-specific rules, personality and adolescents’

alcohol use: a longitudinal person–environment study

Haske Van Der Vorst1, Rutger C. M. E. Engels1, Maja Deković2, Wim Meeus2 & Ad A. Vermulst1
Behavioural Science Institute, Radboud University Nijmegen, the Netherlands1 and Department of Child and Adolescent Studies, Utrecht University, the
Netherlands2

ABSTRACT

Aims To examine the bi-directional associations between providing alcohol-specific rules and adolescents’ alcohol
use. Further, to explore person–environment interactions, we tested whether Big Five personality traits moderate the
assumed association between providing alcohol-specific rules and adolescents’ alcohol use. Design Longitudinal data
(three waves in 2 years) from 428 families, consisting of both parents and two adolescents (aged 13–16 years) were
used for the analyses. Analyses were conducted on four samples: a group of older adolescents and a group of younger
adolescents who already consumed alcohol, and a group of older and younger adolescents who were not drinking at
baseline measurement. Findings In general, results of structural equation modelling showed that providing clear
alcohol-specific rules lowers the likelihood of drinking initiation, regardless of the age of the youngsters. Once adoles-
cents have established a drinking pattern, the impact of parental alcohol-specific rules declined or even disappeared.
Finally, the Big Five personality traits did not moderate the association between providing alcohol-specific rules and
adolescents’ alcohol involvement. Conclusions In sum, in particular during the initiation phase of drinking, parents
could prevent the drinking of their offspring, regardless of the age or personality of their youngsters, by providing clear
alcohol-specific rules.

Keywords Adolescents, alcohol-specific rules, alcohol, personality, person–environment interactions.

Correspondence to: Haske van der Vorst, Behavioural Science Institute, Radboud University Nijmegen, PO Box 9104, 6500 HE Nijmegen, the Netherlands.
E-mail: h.vandervorst@bsi.ru.nl
Submitted 9 May 2006; initial review completed 19 September 2006; final version accepted 13 March 2007

INTRODUCTION older and younger adolescent children within a family

Studies focusing on the influence of parents in the devel-
opment of adolescents’ drinking have examined prima-
rily the influence of parental alcohol use and general
parenting practices (e.g. [1,2]). Recent research has paid
attention to the role of alcohol-specific socialization,
which refers to activities parents undertake specifically to
manage or prevent their children’s drinking behaviour
(e.g. providing alcohol-specific rules, showing disap-
proval of adolescents’ drinking or talking at home about
alcohol use) [3–5].
Providing alcohol-specific rules seems to be the most
influential form of alcohol-specific socialization on ado-
lescents’ drinking [3,4,6]. Van der Vorst et al. [4] showed
strong negative associations between imposing alcohol-
specific rules and adolescents’ alcohol use (bs ranging
between -0.26 and -0.52). This finding was observed for
and was based on the separate perceptions of family
members (father, mother, adolescents) on alcohol-specific
rules underscoring the consistency of this finding.
Further, Van Zundert et al. [7] showed that the direct
association between monitoring and adolescents’ alcohol
use disappears when taking alcohol-specific rule-setting
into account. This finding suggests that alcohol-specific
rules are embedded in the level of behavioural control in
a family, but also that the rules about alcohol in particu-
lar are related to drinking in adolescence. In addition, Yu
[5] demonstrated that parents who prohibit their adoles-
cents to drink reduce the likelihood of current adoles-
cents’ alcohol involvement. A longitudinal study by
Jackson et al. [3] showed that parents who permitted their
children to drink at home were more likely to have
alcohol-consuming children 2 years later (see also Wood
et al. [6]).

© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction Addiction, 102, 1064–1075

Alcohol-Specific
Rules T1
Figure 1 Conceptual model of the
drinking sample on the associations
Adolescents'
between alcohol-specific rules and ado- Alcohol Use T1
lescents’ alcohol consumption at three
waves
Almost all the studies concentrating on alcohol-
specific rules are cross-sectional, and therefore they do
not provide insight into the direction of the effect of
imposing alcohol-specific rules: does imposing alcohol-
specific rules really prevent youngsters from drinking
alcohol as most scholars indicate, based on theoretical
assumptions concerning parent–child effects (e.g. [8]), or
are parents adjusting their rules to the drinking of their
offspring? That the relationship between parenting and
adolescents’ problem behaviour is not unidirectional—
the direction estimated in most empirical studies—but
that adolescents are also influencing their parents’
behaviour, so-called bi-directionality, has been demon-
strated in several studies (e.g. [9,10]). With respect to
alcohol, parental strict control prevents early adoles-
cents’ drinking, but early adolescents’ drinking also leads
to a decline in the degree to which parents provide strict
control [11]. In other words, parents are responding to
their children’s alcohol involvement by loosening their
control. Thus, a longitudinal research design measuring
the bi-directional association between providing alcohol-
specific rules and adolescents’ drinking will reveal insight
into the direction of effects. To develop effective alcohol
prevention programmes focusing on parents, it is essen-
tial to establish the direction of the association between
alcohol-specific rules and adolescents’ alcohol use over
time.
To understand the development of alcohol consump-
tion one should consider the complexity of factors affect-
ing adolescents’ behaviours. These factors reflect the
interactions between individual characteristics, such as
personality, and environmental factors, such as alcohol-
specific socialization [12,13]. Behaviour, in this case
drinking, is always an outcome of ongoing processes
between the person and the environment [14], and there-
fore person–environment interactions could be defined as
a set of processes that are based on a similar set of expe-
riences having different consequences, depending on the
characteristics of an individual [15].
© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction
Alcohol-specific rules and personality 1065
Alcohol-Specific Alcohol-Specific
Rules T2 Rules T3
Adolescents' Adolescents'
Alcohol Use T2 Alcohol Use T3
With regard to drinking, we assume that adolescents
could react differently when parents impose alcohol-
specific rules simply because they differ in their personal-
ity. For instance, impulsive adolescents might find it more
difficult to obey the alcohol-specific rules of their parents
than other adolescents, probably due also to their vulner-
ability for alcohol consumption [16]. On the contrary,
imposing strict rules about drinking might even result in
abstaining from alcohol for conscientious youngsters, as
they are more likely to obey their parents and are already
less prone to drink alcohol [17,18].
That personal characteristics and the social environ-
ment of an individual interact is often argued in theories
of the development of problem behaviour [13,19].
Empirical studies, however, exploring personality–
environment interactions are rare with regard to problem
behaviour, and in particular with regard to the develop-
ment of alcohol use (see for exceptions [13,20]). Engels
et al. [12] demonstrated that aggressive men with less
self-control were most at risk for developing problem
drinking in young adulthood if they were raised in a non-
structured environment. Further, Goodwin et al. [21]
demonstrated that high-risk Danish boys (e.g. those with
a family history of alcoholism) who experienced prob-
lems with attention and impulsiveness at the age of
19–20 years were most likely to be diagnosed with alco-
holism at the age of 29–30 years. In sum, there is some
empirical evidence that personality has an influence on
the relation between family factors and alcohol use.
The aim of the present study was to examine the
bi-directional relationships between providing alcohol-
specific rules and adolescents’ alcohol consumption, con-
trolling for previous behaviour (Figs 1 and 2). We used
structural equation modelling techniques to assess these
hypotheses. Our main interests were the so-called cross-
lagged paths which refer to the longitudinal associations
between alcohol-specific rules and adolescents’ drinking.
In addition, the influence of parents might depend on the
drinking experience of adolescents (e.g. [22]). Therefore,
Addiction, 102, 1064–1075
1066 Haske van der Vorst et al.
Alcohol-Specific Alcohol-Specific
Rules T1 Rules T2
Adolescents'
Alcohol Use T2
we tested our hypotheses for a sample of adolescents who
were already engaged in drinking at baseline measure-
ment and for a sample of adolescents who were not drink-
ing at baseline. Further, we examined whether each
personality trait of the Big Five factor model (extraver-
sion, conscientiousness, agreeableness, emotional stabil-
ity, resourcefulness) [23,24] moderated the association
between providing alcohol-specific rules and adolescents’
drinking. We used three-wave full-family data (father,
mother and two adolescents of the same family), which
enabled us to compare the perceptions of each family
member on alcohol-specific rules and to explore possible
differences between older and younger adolescents
within a family.
METHOD
Participants and procedure
Data for this study were collected as part of a broader
longitudinal survey (‘Family and Health’), which exam-
ined different family processes in relation to various
health behaviours in adolescence (see the cross-sectional
study by van der Vorst et al. [4]). A sample of Dutch
families with at least two adolescent children aged
13–16 years were asked (by mail) to participate in the
study. The addresses of these families were derived from
the registers of 22 municipalities in the Netherlands; 885
of the families approached agreed to participate by
returning the included response form. These families
were then contacted by telephone to establish whether
they fulfilled all the inclusion criteria; i.e. the parents had
to be married or living together, and the adolescents and
their parents should be biologically related. Families with
twins or with adolescents who had mental or physical
disabilities were excluded from the project. Because an
equal division of education and an equal amount of
sibling dyads was aimed for (i.e. boy–boy, boy–girl, girl–
girl, girl–boy), a further selection was made. Finally, a
total of 428 Dutch families participated in this longitudi-
© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction
Alcohol-Specific
Rules T3
Figure 2 Conceptual model of the
non-drinking adolescents at T1 on the
Adolescents' associations between alcohol-specific
Alcohol Use T3 rules and adolescents’ alcohol consump-
tion at three waves
nal research. At the second and third waves, respectively,
416 and 404 families participated.
The families were interviewed three times with an
interval of 1 year at home in the presence of a trained
interviewer. An extensive questionnaire had to be com-
pleted by each family member individually. This took
about 2 hours. The family members were not permitted to
discuss the questions or answers with each other. Each
family received €30 after the four family members had
completed the questionnaire. At the end of the third wave
five holiday cheques of €1000 were raffled between the
families who participated in all three waves of the study.
Each family consisted of both parents and two adoles-
cent children; 95% of the participants were of Dutch
origin, and most of the remainder was born in another
West European country or in Indonesia. The mean
age of the older siblings was 15.22 years at time 1
(T1; SD = 0.60; range 14–17 years), and that of the
younger siblings was 13.36 years at T1 (SD = 0.50;
range 13–15 years). Fathers’ mean age was 46 years
(SD = 4.00) and mothers’ was 44 years (SD = 3.57).
About one-third of both siblings followed special or low
education, one-third followed an intermediate general
education and the remainder followed the highest level of
secondary school in the Netherlands (i.e. preparatory
college and university education). A total of 52.8% of the
older adolescents and 47.7% of the younger adolescents
were boys.
Measures
Alcohol consumption
The adolescents were asked about the intensity and fre-
quency of their drinking. Intensity of drinking was mea-
sured by asking the number of glasses of alcohol the
adolescents had been drinking during the previous week
during weekdays and at weekends into contexts at home
and outside the home [25]. These divisions were made to
allow the adolescents to recall more accurately what they
Addiction, 102, 1064–1075

had consumed that week. The scores on these four ques-
tions were summed to obtain an indication of the total
amount of alcohol consumption the adolescents drank
during a week. Frequency of drinking was assessed by
asking the adolescents how many times they drank
alcohol during the past 4 weeks. They had to respond on
a six-point scale: (1) ‘have not been drinking’; (2)
‘1–3 days a month’; (3) ‘1–2 days a week’; (4), ‘3–4 days
a week’; (5) ‘5–6 days a week’; and (6) ‘every day’ [26].
Rules about alcohol
We developed a 10-item scale to measure the degree to
which parents permit their children to consume alcohol
in various situations, such as ‘in the absence of parents at
home’ or ‘at a friends’ party’ [4]. Thus, we asked each
family member what rules the parents had or what they
prohibited concerning alcohol. Respondents had to
answer to what degree these rules were applicable at their
home. Response categories ranged from (1) ‘completely
applicable’ to (5) ‘not applicable at all’. Higher scores indi-
cate having stricter rules about alcohol consumption. We
have some preliminary psychometric features of the rules
concerning alcohol scale. The scale seems to have a good
content validity, as explorative factor analysis revealed
one factor (items loaded between 0.56 and 0.93), but also
the reliability was high with Cronbach’s as between 0.80
and 0.93 across reporters and waves. Further, divergent
validity seem to be satisfactory. Alcohol-specific rules was
namely moderately correlated (varying between 0.34
and 0.50 depending on the perspective of the respon-
dents) with alcohol-specific monitoring (parents knowing
when their adolescent child is drinking, where and with
whom). Finally, predictive validity is shown by the results
of this study.
Personality
We used the short version of the Big Five questionnaire
[24]. Five personality dimensions were assessed on the
basis of the five-factor personality model [23]. For 30
traits, the adolescents were asked to what degree they
possessed each trait. We used the items of the short version
of the Big Five questionnaire, assessed at T1. Response
categories ranged from (1) ‘absolutely disagree’ to (7)
‘absolutely agree’. The dimension ‘extraversion’ was mea-
sured with (recodes of) items such as quiet, withdrawn
and shy (older adolescents (O): a = 0.86; younger adoles-
cents (Y): a = 0.79). The dimension ‘agreeableness’ was
assessed with items such as kind, likeable and cooperative
(O: a = 0.77; Y: a = 0.78). The dimension ‘conscientious-
ness’ was measured with items such as organized, orderly
and efficient (O: a = 0.85; Y: a = 0.84). The dimension
‘emotional stability’ was assessed with (recodes of) items
such as nervous, fearful and sensitive (O: a = 0.73; Y:
© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction
Alcohol-specific rules and personality 1067
a = 0.74), and ‘resourcefulness’ with items such as cre-
ative, artistic and versatile (O: a = 0.68; Y: a = 0.63) (see
[23] for more information about the psychometric proper-
ties of the short version of the Big Five questionnaire).
Analyses strategy
We selected a sample including adolescents who partici-
pated at all three waves (n = 401) for both adolescents.
Thus, all missing data on the basis of non-participation
were omitted from the analyses. Next, we divided both
samples of the younger and older adolescents into two
groups: adolescents who were not drinking alcohol at
baseline measurement (nolder = 111; nyounger = 229) and
adolescents who reported that they had been drinking at
baseline (nolder = 290; nyounger = 172). We conducted all
analyses on these four samples.
First, descriptive analyses were conducted to calculate
the means and standard deviations of adolescents’ inten-
sity and frequency of drinking in each group. Secondly,
paired t-tests were used to compare the responses of the
four family members on the scale on alcohol-specific rules
at all three waves in each group. This will gain insight
into possible differences in alcohol-specific rules towards
older and younger adolescents in a family, into possible
differences in experiences of parents and adolescents, but
also into the stability of alcohol-specific rules over time.
Thirdly, to test our longitudinal models, we applied struc-
tural equation modelling using version 4.2 of the Mplus
statistical package [27]. We tested three models for the
older adolescents (mother, father and adolescent reports
on alcohol-specific rules) as well as for the younger ado-
lescents in each group (thus a total of 12 models). [The
correlations between alcohol-specific rules based on the
reports of the mothers and alcohol-specific rules based on
the reports of the fathers were not high, varying between
0.33 and 0.55. This indicates that they are separate con-
structs. In line with this, we have tried to make a latent
‘parental rules’ construct of the fathers’ and mothers’
reports. However, the variables do not seem to load
together. The factor loadings of the fathers are too low
(around 0.40), which is another indication that fathers’
and mothers’ perspectives should be measured sepa-
rately.] In each model, assessments of alcohol use were
based on self-reports of the particular adolescents (older
or younger adolescents). The distributions of the inten-
sity of drinking variables were very skewed. Therefore,
the number of glasses of alcohol was divided into six
classes for the structural equation models: 1 (no glasses),
2 (one to five glasses), 3 (six to 10 glasses), 4 (11–20
glasses), 5 (21–30 glasses) and 6 (> 30 glasses a week).
In the first model parental rules about alcohol use
were based on the reports of the adolescents, in the
second model parental rules about alcohol use were based
Addiction, 102, 1064–1075
1068 Haske van der Vorst et al.
on the reports of the mothers and in the third model on
the reports of the fathers (see Fig. 1 for the conceptual
model for adolescents who reported to consume alcohol at
T1). The models of the non-drinking groups were the
same as the models of the drinking groups, except that the
alcohol use variable of the adolescents at T1 was omitted
due to lack of variance (all adolescents were non-drinkers
at T1; Fig. 2). [Correlation tables of the model variables
can be obtained from the first author of this paper.]
The variables in the models as depicted in Figs 1 and
2 are treated as latent variables. The alcohol-specific
rules were measured by 10 items. To overcome the
problem of estimating too many parameters in relation
to the sample size, the 10 items were split up into two
equivalent parts (parcels) according to the recommen-
dations of Bandalos & Finney [28] and Nasser & Wisen-
baker [29]. The indicators for adolescents’ alcohol use
were the two alcohol variables intensity and frequency,
both measured with six ordered categories. The factor
loadings of the 12 models were mainly above 0.80. The
lowest loading was 0.65. The conclusion is that the
parcels/indicators represent adequately the underlying
latent variables.
Because the alcohol indicators are non-normal and
more ordered categorical (ordinal) than metric, the most
suited estimator available in Mplus is the weighted least-
square estimator with mean- and variance-adjusted c2
statistic (WLSMV-estimator). However, if the data are
ordinal with at least five categories and moderately non-
normal (skewness < 2 and kurtosis < 7) an adequate
alternative is the maximum likelihood (ML) estimator
[30]. The skewness and kurtosis was < 2 for all alcohol
variables, while there were six categories. For these
reasons and because multiple group testing (see below)
is somewhat complicated using the WLSMV-estimator,
we decided to use the ML-estimator for all our analyses.
Moreover, we tested the 12 models with both estimators
resulting in identical results. The low percentage of
missing data (< 2%) was treated with the help of the full-
information maximum likelihood (FIML) approach: esti-
mation of parameters using all the available information
in the data case by case, maximizing the likelihood of the
observed data [31] with the help of the expectation
maximization (EM) algorithm via maximum likelihood
[32].
To assess the moderating effects of the Big Five per-
sonality traits on the associations between providing
alcohol-specific rules and adolescents’ alcohol consump-
tion, we tested all models in relation to each personality
trait under two conditions: low and high on the particu-
lar personality trait [33]. The median split method was
used to divide the sample, for instance, into a high extra-
version group and a low extraversion group (for addi-
tional details on the procedures to test moderating
© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction
effects, see van der Vorst et al. [11]). Because the number
of respondents was reduced as a consequence of the
median split method, our aim was to be very economical
with the free parameters of the baseline model. The
structural parameters were kept free (10 bs and three
disturbance correlations for the model shown in Fig. 1
and seven bs with two disturbance correlations for the
model shown in Fig. 2). All the other parameters were
fixed at the values resulting from the 12 model estimates.
A baseline c2 was computed with no equality constraints
between the structural parameters of the two groups
(unconstrained model). Next, all the bs were constrained
to be equal for both groups. The c2 of this constrained
model was compared with the c2 of the unconstrained
model. If c2 increases significantly one or more bs would
be significantly different across groups. To observe which
bs were different between two groups, additional differ-
ence tests for each individual b were conducted. These
analyses were performed for each of the Big Five person-
ality traits for all 12 groups. The fit of the models was
assessed by the following global fit indexes: c2, compara-
tive fit index (CFI), root mean square error of approxima-
tion (RMSEA) and standardized root mean square
residual (SRMR) [34].
RESULTS
Descriptives
The older adolescents within the subsample of drinkers
consumed on average 5.4 glasses of alcohol a week at T1
(Table 1). The amount of alcohol consumed increased
significantly to 8.1 glasses a week per year later, and to
10.1 glasses a week 2 years later. The older adolescents in
the non-drinking sample at T1 consistently drank less
than the older adolescents of the drinking sample,
namely an average 2.2 glasses of alcohol a week at T2
and 4.9 glasses a week at T3. The increase of intensity of
drinking for the non-drinking older adolescents was also
significant. The younger adolescents in the drinking
sample consumed 2.0 glasses of alcohol a week at T1, 3.8
at T2 and 6.9 at T3. The younger adolescents in the non-
drinking sample also drank less than the adolescents of
the younger drinking sample. They drank on average 1.3
glasses of alcohol a week at T2 and 3.2 glasses 1 year
later. The increase in intensity of alcohol use for both
groups of the younger adolescents was significant The
older adolescents in both groups increased their fre-
quency of drinking significantly from T2 to T3. We also
found an increase in frequency of drinking for the
younger adolescents in the non-drinking group,
although they generally drank less often than the older
adolescents. However, the younger adolescents for the
drinking group did not increase significantly the fre-
quency of their drinking.
Addiction, 102, 1064–1075
 Alcohol-specific rules and personality 1069

Table 1 Means and standard deviations of frequency and intensity of drinking.

Drinking adolescents at T1 Non-drinking adolescents at T1

Older Younger Older Younger

M SD M SD M SD M SD

Frequency T1 2.6a 0.70 2.3a 0.59

Frequency T2 2.6a 0.84 2.3a 0.80 1.8a 0.83 1.5a 0.66
Frequency T3 2.7b 0.85 2.4a 0.83 2.1b 0.89 1.8b 0.81
Intensity T1 5.4a 5.43 2.0a 2.08
Intensity T2 8.1b 7.93 3.8b 4.35 2.2a 4.35 1.3a 2.60
Intensity T3 10.1c 10.19 6.9c 7.05 4.9b 8.59 3.2b 5.41

Comparisons are made over time for both adolescents for both groups. The results should be read vertically in a column. a,b,cMeans that do not share
superscripts are significantly different (P < 0.05).

Table 2 Means and standard deviations of alcohol-specific rules.

O MO FO Y MY FY

M SD M SD M SD M SD M SD M SD

Drinking adolescents at T1
Rules T1 2.97a 0.80 3.98a 0.65 3.99b 0.68 3.69a 0.76 4.51b 0.50 4.50b 0.60
Rules T2 2.52a 0.72 3.52b 0.70 3.50b 0.75 3.23a 0.76 4.12b 0.69 4.13b 0.68
Rules T3 2.28a 0.67 3.13b 0.72 3.11b 0.76 2.82a 0.76 3.71b 0.73 3.73b 0.77
Non-drinking adolescents at T1
Rules T1 4.03a 0.85 4.62b 0.51 4.50b 0.66 4.33a 0.71 4.79b 0.40 4.75b 0.40
Rules T2 3.28a 1.00 4.19b 0.68 4.05b 0.87 3.80a 0.85 4.45b 0.63 4.40b 0.68
Rules T3 2.88a 1.00 3.60b 0.89 3.60b 0.86 3.34a 0.93 4.04b 0.78 4.03b 0.82

O = Older adolescents, MO = mothers’ perspectives about the older adolescents, FO = fathers’ perspectives about the older adolescents, Y = younger
adolescents, MY = mothers’ perspectives about the younger adolescents, FY = fathers’ perspective about the younger adolescents. Comparisons are made
between reports of older adolescents, fathers and mothers at each wave. The same holds for the younger adolescents. The results should be read
horizontally. a,bMeans that do not share superscripts are significantly different (P < 0.05).

Perceptions of alcohol-specific rules drinking as well as of non-drinking adolescents

appeared to be less strict towards the 15-year-old
Comparisons of responses of the adolescents and
younger adolescents than towards the older adolescents
parents of the drinking samples as well as the non-
at that age.
drinking samples showed a consistent pattern. We
found for each sample strong significant differences in
Structural equation modelling
perceptions of providing alcohol-specific rules at all
three measurements (Table 2). Both parents reported Drinking samples
that they imposed stricter rules than both youngsters
The fit of the three models of the older adolescents and of
reported them to do at all three measurements.
the three models of the younger adolescents was satisfac-
However, the rules on alcohol imposed by parents
tory (Table 3).
declined significantly during the 2 years according to
the four family members. Further, although the parents
Models of the older adolescents
treated their offspring differently (they were stricter with
the younger adolescents than with the older ones at all All cross-sectional associations between alcohol-specific
three waves) fathers and mothers were, on average, rules and older adolescents’ alcohol use were significant
similar in their rule-setting to their older adolescent and for the three models (Table 3), with the exception of cor-
also to their younger adolescent. Finally, when the relation of the model of the mothers at T2 and of the
younger adolescents reached the age of the older ado- model of the adolescents at T3. It is noteworthy that cor-
lescents at the first measurement (15 years), parents of relations between cross-sectional latent variables are

© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction Addiction, 102, 1064–1075
1070 Haske van der Vorst et al.

Table 3 Standardized estimates and fit measures for models of drinking adolescents at T1 tested for reports by adolescents, fathers and
mothers.

Adolescents Mothers Fathers

Older Younger Older Younger Older Younger

n 290 172 290 172 290 172

Cross-sectional
Rules T1–Alcohol T1 -0.41*** -0.43*** -0.34*** -0.34** -0.26*** -0.17
Rules T2–Alcohol T2 -0.17** -0.17** -0.09 -0.09 -0.17** -0.06
Rules T3–Alcohol T3 -0.07 -0.10 -0.10** -0.05 -0.04 -0.01
Stability paths
Rules T1–Rules T2 0.69*** 0.64*** 0.70*** 0.68*** 0.65*** 0.67***
Rules T2–Rules T3 0.53*** 0.57*** 0.57*** 0.47*** 0.59*** 0.52***
Rules T1–Rules T3 0.16* 0.12 0.18* 0.33*** 0.16* 0.22***
Alcohol T1–Alcohol T2 0.65*** 0.76*** 0.65*** 0.76*** 0.62*** 0.79***
Alcohol T2–Alcohol T3 0.63*** 0.78*** 0.60*** 0.87*** 0.59*** 0.66***
Alcohol T1–Alcohol T3 0.16 -0.19 0.16 -0.30 0.17 -0.11
Cross-lagged paths
Rules T1–Alcohol T2 0.08 -0.11 0.10 -0.18* -0.01 -0.00
Rules T2–Alcohol T3 0.07 0.06 -0.09 0.05 -0.09 0.09
Alcohol T1–Rules T2 0.07 -0.10 -0.02 0.01 0.09 -0.04
Alcohol T2–Rules T3 0.09 -0.03 0.04 -0.06 0.01 -0.05
Fit measures
d.f. 38 38 38 38 38 38
c2 60.248 41.305 55.967 48.241 53.412 44.995
P 0.016 0.328 0.030 0.123 0.050 0.202
CFI 0.987 0.997 0.990 0.991 0.992 0.994
RMSEA 0.043 0.022 0.040 0.040 0.037 0.033
SRMR 0.029 0.031 0.035 0.032 0.024 0.043

*P < 0.05; **P < 0.01; ***P < 0.001. CFI: comparative fit index, RMSEA: root mean square error of approximation, SRMR: standardised root mean square
residual.

decreasing over time, as these correlations are controlled the cross-lagged paths from adolescents’ drinking to pro-
for previous influences and must be interpreted as partial viding alcohol-specific rules.
correlations. Further, rules about alcohol and adoles-
cents’ alcohol use showed strong stability over time in all Non-drinking sample
models. More importantly, all cross-lagged associations
The fit of the six models was satisfactory (Table 4).
were not significant, which implies that providing
alcohol-specific rules does not prevent older adolescents’ Models of the older adolescents
drinking and that parents do not become more permissive
as their offspring consume more alcohol. According to the older adolescents themselves, as well as
the mothers and fathers, alcohol-specific rules (T1) were
Models of the younger adolescents strongly negatively related to older adolescents’ drinking
(T2; Table 4). This implies that imposing strict rules about
Alcohol-specific rules was cross-sectionally significantly
alcohol prevents 15-year-olds who do not yet drink
related only to younger adolescents’ alcohol use in the
alcohol to start consuming alcohol intensively and fre-
model of the younger adolescents themselves (T1, T2)
quently. This longitudinal effect was also found a year
and in the model of the mothers at T1. Alcohol-specific
later (from T2 to T3) in the model with adolescent reports.
rules predict rules over time, and previous drinking pre-
Alcohol use of the older adolescents (T2) was not associ-
dicts drinking later on. According to mothers, rules about
ated significantly with alcohol-specific rules over time
alcohol at T1 were related negatively to adolescents’
(T3).
alcohol use at T2. This indicates that rules about alcohol
prevent younger adolescents’ alcohol involvement later
Models of the younger adolescents
on. The other cross-lagged paths from alcohol specific
rules to alcohol use of younger adolescents in the three The preventive effect of providing alcohol-specific rules
models we tested were not significant. The same holds for (T1) on adolescents’ drinking behaviour (T2) appeared to

© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction Addiction, 102, 1064–1075
 Alcohol-specific rules and personality 1071

Table 4 Standardized estimates and fit measures for models of non-drinking adolescents at T1 tested for reports by adolescents,
fathers and mothers.

Adolescents Mothers Fathers

Older Younger Older Younger Older Younger

n 111 229 111 229 111 229

Cross-sectional
Rules T2–Alcohol T2 -0.24** -0.41*** -0.21*** -0.26*** -0.29*** -0.15**
Rules T3–Alcohol T3 -0.10 -0.16*** -0.17** -0.11** -0.09 -0.10*
Stability paths
Rules T1–Rules T2 0.61*** 0.62*** 0.80*** 0.61*** 0.51*** 0.59***
Rules T2–Rules T3 0.57*** 0.64*** 0.74*** 0.60*** 0.79*** 0.46***
Rules T1–Rules T3 0.22** 0.13 -0.01 0.27*** 0.16* 0.10
Alcohol T2–Alcohol T3 0.64*** 0.58*** 0.68*** 0.58*** 0.73*** 0.61***
Cross-lagged paths
Rules T1–AlcoholT2 -0.34*** -0.13 -0.54*** -0.02 -0.28*** -0.21*
Rules T2–Alcohol T3 -0.30*** -0.10 -0.11 -0.21** -0.07 -0.08
Alcohol T2–Rules T3 -0.04 0.00 -0.02 -0.05 0.14 -0.00
Fit measures
d.f. 23 23 23 23 23 23
c2 39.031 31.583 50.275 37.235 35.904 23.732
P 0.020 0.109 0.001 0.031 0.042 0.419
CFI 0.979 0.995 0.965 0.991 0.984 1.000
RMSEA 0.079 0.040 0.103 0.052 0.071 0.012
SRMR 0.042 0.024 0.048 0.031 0.043 0.022

*P < 0.05; **P < 0.01; ***P < 0.001. CFI: comparative fit index, RMSEA: root mean square error of approximation, SRMR: standardised root mean square
residual.

be significant in the model of the fathers. In the model
based on information of the mothers, this significant
association was observed a year later (T2 to T3). Thus,
parents who provided strict rules on alcohol consumption
are less likely to have a drinking younger adolescent later
on than permissive parents.
Multi-group analyses
To establish whether adolescents scoring low or high on
each of the five personality traits (extraversion, conscien-
tiousness, agreeableness, emotional stability, resourceful-
ness) differ in their reaction on alcohol-specific rules
setting, we conducted multi-group analyses in every
model based on drinking adolescents and non-drinking
adolescents (five multi-group analyses per model for a
total of 12 models). Here, only differences between cross-
lagged effects are discussed, because these effects are our
main interest.
Drinking adolescents
We found one significant different cross-lagged path
for the drinking groups, namely in the model of older
adolescents (paternal reports) concerning the personal-
ity trait ‘agreeableness’. For older adolescents scoring
high on agreeableness, alcohol-specific rules were
related negatively significantly to adolescents’ drinking,
and for adolescents scoring low on agreeableness,
alcohol-specific rules were not related to alcohol use
(T2–T3; Dc2 (10) = 21.129, P < 0.01; bhigh = -0.23,
P < 0.01; blow = -0.06, P > 0.05). In all other models
no differences between the group scoring low on a trait
and the group scoring high on a specific personality
trait were observed. This indicates that personality does
not moderate the association between providing
alcohol-specific rules and adolescents’ alcohol use
in the case that adolescents already drink alcohol at
T1.
Non-drinking adolescents
All cross-lagged associations were not significantly differ-
ent between the group scoring low on a trait and the
group scoring high on a trait for the non-drinking
samples.
Further, it should be stressed that despite the large
number of multi-group analyses (60 models were tested
in total), we found a difference in paths for only one multi-
group analysis in the drinking group of older adolescents.
DISCUSSION
The aim of the present study was to determine the
bi-directional associations between providing alcohol-

© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction Addiction, 102, 1064–1075
1072 Haske van der Vorst et al.
specific rules and adolescents’ drinking behaviour using
longitudinal data from parents and their children. In
addition, we assessed whether each trait of the Big
Five personality model moderates the association
between alcohol-specific rules and adolescents’ alcohol
consumption.
First of all, providing alcohol-specific rules was related
cross-sectionally strongly negatively to the alcohol con-
sumption of both older and younger adolescents who
were already drinking and who still were in the initiation
phase of drinking. These findings correspond with those
of Jackson et al. [3] and Yu [5]. More importantly, provid-
ing alcohol-specific rules prevented older and younger
adolescents’ alcohol use from starting to consume
alcohol. The validity of this finding is illustrated by the
consistent results across the different family members,
and across the two children in a family, with the excep-
tion of the younger adolescents with their perspective. On
the other hand, imposing alcohol-specific rules did not
prevent the alcohol use of adolescents who had already
started to drink. Thus, if an adolescent has established a
drinking pattern, the impact of parents declines or even
disappears, at least with respect to alcohol-specific rules.
However, in the case that an adolescent has not started to
drink regularly, no matter whether he/she is 13 or 15
years old, parents seem to be able to control their off-
spring drinking by providing alcohol-specific rules. This
might be an interesting implication for health organiza-
tions that develop alcohol prevention programmes
focusing on parental involvement. Alcohol prevention
programmes should make parents aware that they play a
role in preventing youth drinking by setting rules before
their children have established a drinking pattern.
Further, all family members reported that parents
were stricter towards the younger adolescents than
towards the older ones at all three measurements,
proving clearly that parents are treating their youngsters
differently with concern to alcohol use. Parents seem to
become more permissive towards adolescents’ alcohol use
over the years. On the basis of the mean scores on
alcohol-specific rules, it is quite clear that parents become
less strict over time. Relatively, however, the results show
that parents are fairly stable in imposing alcohol-specific
rules. Moreover, it is interesting that when the younger
adolescents reached the age of the older adolescents at
the first measurement (15 years), parents appeared to be
less strict towards the 15-year-old younger adolescents
than towards the older adolescents at that age. This
finding on birth order indicates that parents are influ-
enced, at least partly, by the behaviour of their offspring.
Nevertheless, this pattern was not supported by the
bi-directional results of our longitudinal models. For both
drinking and abstaining adolescents at the first wave,
alcohol use did not affect alcohol-specific rule-setting.
© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction
Thus, parents do not seem to become more permissive
because they notice that their youngsters are drinking.
It is noteworthy that we do not know whether parents
are consistent in providing alcohol-specific rules in the
short term. Do they impose the same rules each day, or
are they more permissive in some occasions than others?
When parents fluctuate strongly in their alcohol-specific
rule-setting in daily life, it is perhaps not a clear signal
towards adolescents, leading to a higher likelihood of
adolescent drinking [35]. Thus, it might be that it is not
the mean level of alcohol-specific rules that is important,
but the flexibility (or rigidity) of the rule-setting. On the
other hand, rigidity of parents might be perceived as an
indication of inflexibility to deal with youngsters’ needs
(see [36]), and lead subsequently to more adolescents’
alcohol involvement. Nevertheless, it would be important
to investigate the flexibility of parents in their alcohol-
specific rule-setting by, for instance, observing how
parents and adolescents communicate about alcohol-
specific socialization.
Person–environment interactions
We argued that the complexity of factors affecting ado-
lescents should be considered fully in order to understand
the development of drinking, for instance by testing
person–environment interactions [12,13,20]. Person–
environment issues have been a topic for debate for years
in the field of developmental psychopathology, but have
generally been neglected in the research on alcohol use
[13,19,20]. We hypothesized that the impact of pro-
viding alcohol-specific rules on adolescents’ drinking
depends on adolescents’ characteristics, e.g. the Big Five
personality traits of adolescents. However, no consistent
significant findings were observed that could confirm our
hypothesis. Again, the advantages of using multi-
informant data and data on younger as well as older ado-
lescents, in relation to the large numbers of analyses
conducted, underline the robustness of our findings.
Therefore, we think this finding is a meaningful contribu-
tion to the scientific debate, showing that personality and
the environment are not necessarily related (as scholars
have been suggesting), at least not with regard to paren-
tal socialization and adolescents’ alcohol use. Neverthe-
less, this might seem striking considering that the older
non-drinking group of adolescents is a somewhat special
group in the Netherlands; the majority of Dutch youth
have already been drinking alcohol at the age of 15 [37].
In this light one might assume that personality charac-
teristics would be a relevant explanation for the atypical
behaviours of these adolescents. However, in the current
study we showed that personality is not playing a role
here. On the basis of this we might conclude that im-
posing alcohol-specific rules affect only the drinking of
Addiction, 102, 1064–1075

adolescents who have not already established a drinking
pattern, despite the adolescents’ personality.
Nevertheless, regardless of our conclusion concern-
ing personality and providing alcohol-specific rules, it is
meaningful to speculate why we found such a lack of
support for our hypothesis. The general assumption is
that the personality of an individual is a composite of
scores on a continuum of each of the five personality
traits [20]. Perhaps a certain composite makes an adoles-
cent more sensitive to the impact of alcohol-specific rules
instead of just one trait (see [23]). In the current study we
tested exclusively the moderating effect of each single
trait, while perhaps specific constellations of personality
factors are, in particular, putting youth at risk for a lack of
effect in providing alcohol-specific rules on their drinking
behaviour. Secondly, it could be that other individual
characteristics are of influence in the association
between alcohol-specific rules and adolescents’ drinking,
such as self-control or aggression. For example, the risk
for problem drinking increases for young adult men who
are aggressive and experience low levels of family func-
tioning [12]. It might be that providing alcohol-specific
rules has less effect on especially the drinking of aggres-
sive male adolescents. Thirdly, the explanation is method-
ological. We conducted a median-split method to divide
our sample. Although this is a commonly used method, it
has a limitation, namely adolescents scoring in the
middle of the range of each trait are also included in
the analyses. Analyses conducted on adolescents with
more extreme scores might show significant person–
environment interactions. However, we were not able to
perform this type of analysis as we would have encoun-
tered problems with statistical power. Finally, our lack of
findings might be due to small samples. [We also con-
ducted multi-group analyses on the whole sample of
older and younger adolescents, so combining drinkers
and non-drinkers at T1. Both these samples contained
428 adolescents. We did not find significant moderating
effects of each personality trait in these samples. In this
case a lack of power could not be the reason for not
finding significant results.] For smaller samples it is more
difficult to obtain significant effects, especially when one
is testing a complex model such as ours.
Strengths, limitations and implications for future
research
Although the current study has several strengths, such
as its longitudinal design and the multi-informant data, it
is limited by some factors. Although we selected families
carefully on the basis of, for instance, educational level or
the proportion of sibling dyads, the findings should not be
generalized to single-parent families, to families with no
biological relation between parent and child or to families
© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction
Alcohol-specific rules and personality 1073
with twins. Secondly, it might be that our sample does not
reflect the Dutch situation accurately, as we do not have
any information about the families who did not respond
on our recruitment letter. Thirdly, one should be careful
in generalizing our results to other countries with other
drinking cultures. Dutch adolescents drink more fre-
quently than adolescents from other European countries
[38]. It is difficult to predict whether providing alcohol-
specific rules would be more effective—in terms of con-
tinuation of drinking—in cultures where children start
drinking at a later age, drink generally less often and
intensively, or where the societal norms on youth alcohol
are less permissive. Therefore, it would be important to
investigate the role of alcohol-specific rules in other
(drinking) cultures. Further, providing alcohol-specific
rules might prevent heavy drinking or problem drinking
in adolescence but also later in life. It would be interesting
to find out what the long-term effects are of providing
alcohol-specific rules instead of the rather short-term
effects as we did. In addition, it would be important to
explore how drinking peers affect the impact of alcohol-
specific socialization, as the effect of a peer-drinking
culture might interfere with the effect parents have on
their offspring [39]. Finally, it is still unclear how parents
sanction their child after he or she break the rules about
alcohol. Not perceiving consequences after not following
a rule might give an adolescent the impression that the
alcohol-specific rules are not as strict as the adolescent
thought in the first place. This might, in turn, affect
future drinking. Unfortunately, we have no data on
parental sanctions.
In sum, the current study is one of the first which
provides insight into the impact of alcohol-specific rules
on the development of adolescents’ alcohol consumption,
as well as the role of the adolescents’ personality in this.
The findings yield substantial evidence that imposing
strict rules prevents the uptake of drinking, regardless of
their personality, but in the case that adolescents are
already involved in drinking, the impact of parents
declines or even disappears.
Acknowledgements
Rutger Engels was supported by a fellowship of the Dutch
Organization of Scientific Research (NWO) during the
preparation of this manuscript.
References
1. Barnes G. M., Farrell M. P. Parental support and control as
predictors of adolescent drinking, delinquency and related
problem behaviours. J Marr Fam 1992; 54: 763–76.
2. Reifman A., Barnes G., Dintcheff B. A., Farrell M. P., Uhteg
L. Parental and peer influences on the onset of heavier
drinking among adolescents. J Stud Alcohol 1998; 59: 311–
17.
Addiction, 102, 1064–1075
1074 Haske van der Vorst et al.
3. Jackson C., Henriksen L., Dickinson D. Alcohol-specific
socialization, parenting behaviors and alcohol use by chil-
dren. J Stud Alcohol 1999; 60: 362–7.
4. van der Vorst H., Engels R. C. M. E., Meeus W., Deković M.,
Van Leeuwe J. The role of alcohol-specific socialization in
adolescents’ drinking behaviour. Addiction 2005; 100:
1464–76.
5. Yu J. The association between parental alcohol-related
behaviors and children’s drinking. Drug Alcohol Depend
2003; 69: 253–62.
6. Wood M. D., Read J. P., Mitchell R. E., Brand N. H. Do
parents still matter? Parent and peer influences on alcohol
involvement among recent high school graduates. Psychol
Addict Behav 2004; 18: 19–30.
7. van Zundert R. M. P., van der Vorst H., Vermulst A. A.,
Engels R. M. C. E. Pathways to alcohol use among Dutch
students in regular education and education for adolescents
with behavioral problems: the role of parental alcohol use,
general parenting practices, and alcohol-specific parenting
practices. J Fam Psychol 2006; 20: 456–67.
8. Darling N., Steinberg L. Parenting style as context: an inte-
grative model. Psychol Bull 1993; 113: 487–96.
9. Kerr M., Stattin H. Parenting of adolescents: action or reac-
tion? In: Booth A., Crouter A., editors. Children’s influences
on family dynamics: the neglected side of family relationships.
Mahwah, NJ: Lawrence Erlbaum; 2003: 121–51.
10. Stice E., Barrera M. A longitudinal examination of the recip-
rocal relations between perceived parenting and adoles-
cents’ substance use and externalizing behaviors. Dev
Psychol 1995; 31: 322–34.
11. van der Vorst H., Engels R. C. M. E., Meeus W., Deković M.,
Vermulst A. Parental attachment, parental control and
early development of alcohol use: a longitudinal study.
Psychol Addict Behav 2006; 20: 107–16.
12. Engels R. M. C. E., Vermulst A. A., Dubas J. S., Bot S. M.,
Gerris J. Long-term effects of family functioning and child
characteristics on problem drinking in young adulthood.
Eur Addict Res 2005; 11: 32–7.
13. Rose R. J. A developmental behavioural–genetic perspective
on alcoholism risk. Alcohol Health Res World 1998; 22: 131–
43.
14. Sroufe L. A., Egeland B. Illustrations of person–
environment interaction from a longitudinal study. In:
Wachs T. D., Plomin R., editors. Conceptualization and mea-
surement of organism–environment interaction. Washington,
DC: American Psychological Association; 1991: 64–84.
15. Rutter M., Pickles A. Person–environment interactions: con-
cepts, mechanisms, and implications for data analysis. In:
Wachs T. D., Plomin R., editors. Conceptualization and mea-
surement of organism–environment interaction. Washington,
DC: American Psychological Association; 1991: 105–36.
16. Walton K. E., Roberts B. W. On the relationship between
substance use and personality traits: abstainers are not mal-
adjusted. J Res Pers 2004; 38: 515–35.
17. Cook M., Young A., Taylor D., Bedford A. P. Personality
correlates of alcohol consumption. Pers Individ Diff 1998;
24: 641–7.
18. Tucker J. S., Friedman H. S., Tomlinson-Keasey C., Schwartz
J. E., Wingard D. L., Criqui M. H. et al. Childhood psychoso-
cial predictors of adulthood smoking, alcohol consumption
and physical activity. J Appl Soc Psychol 1995; 25: 1884–
99.
19. Rutter M., Dunn J., Plomin R., Simonoff E., Pickles A.,
Maughan B. et al. Integrating nature and nurture: implica-
© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction
tions of person–environment correlations and interactions
for developmental psychopathology. Dev Psychopathol
1997; 9: 335–64.
20. O’Connor B. P., Dvrok T. Conditional associations between
parental behavior and adolescent problems: a search for
personality–environment interactions. J Res Pers 2001; 35:
1–26.
21. Goodwin D. W., Knop J., Jensen P., Gabrielle W. F., Pennick
E. C. Thirty-year follow-up of men at high risk for alcohol-
ism. Ann NY Acad Sci 1994; 708: 1210–13.
22. Spijkerman R., Van Den Eijnden R. J. J. M., Overbeek G. J.,
Engels R. C. M. E. The impact of peer and parental norms
and behaviour on adolescent drinking: the mediating role of
drinker prototypes. Psychol Health 2007; 22: 7–29.
23. Dubas J. S., Gerris J. R. M., Janssens J. M. A. M., Vermulst A.
A. Personality types of adolescents: concurrent correlates,
antecedents, and type–parenting interactions. J Adolesc
2002; 25: 79–92.
24. Vermulst A. A., Gerris J. R. M. QBF: Quick Big Five Persoon-
lijkheidstest Handleiding [Quick Big Five Personality Test
Manual]. Leeuwarden, the Netherlands: LDC Publications;
2005.
25. Engels R. C. M. E., Knibbe R. A., Drop M. J. Why do late
adolescents drink at home? A study on psychological well-
being, social integration and drinking context. Addict Res
1999; 7: 31–46.
26. Engels R. C. M. E., Knibbe R. A. Alcohol use and intimate
relationships in adolescence: when love comes to town.
Addict Behav 2000; 25: 435–9.
27. Muthén L. K., Muthén B. O. Mplus User’s Guide, 4th edn. Los
Angeles, CA: Muthén & Muthén: 1998–2006.
28. Bandalos D. L., Finney S. J. Item parceling issues in struc-
tural equation modeling. In: Marcoulides G. A., Schu-
macker R. E., editors. New developments and techniques in
structural equation modeling. Mahwah, NJ: Lawrence
Erlbaum Associates; 2001, p. 269–96.
29. Nasser F., Wisenbaker J. A Monte Carlo study investigating
the impact of item parceling strategies on parameter esti-
mates and their standard errors in CFA. Struct Equat Model
2006; 13: 204–28.
30. Finney S. J., DiStefano C. Non-normal and categorical data
in structural equation modeling. In: Hancock G. R., Mueller
R. O., editors. Structural equation modeling. A second course.
Greenwich: Information Age Publishing (IAP); 2006,
p. 269–314.
31. Wothke W. Longitudinal and multigroup modeling with
missing data. In: Little T. D., Schnabel K. U., Baumert J.,
editors. Modeling longitudinal and multilevel data. Mahwah,
NJ: Lawrence Erlbaum Associates; 2000, p. 219–40.
32. Kaplan D. Structural equation modeling. Foundations and
extensions. Thousand Oaks, CA: Sage Publications; 2000.
33. Jöreskog K., Sörbom D. PRELIS 2: User’s Reference Guide.
Chicago, IL: Scientific Software International, Inc.; 1996.
34. Marsh H. W., Balla J. R., McDonald R. P. Goodness-of-fit
indexes in confirmatory factor analysis: the effect of sample
size. Psychol Bull 1996; 103: 391–410.
35. Engels R. C. M. E., Bot S. M. Social influences on adolescent
substance use. In: De Ridder D., De Wit J., editors. New per-
spectives on health behaviour: the role of self-regulation. Chich-
ester: John Wiley; 2006, p. 19–20.
36. Hollenstein T., Granic I., Stoolmiller M., Snyder J. Rigidity in
parent–child interactions and the development of external-
izing and internalizing behavior in early childhood.
J Abnorm Child Psychol 2004; 32: 595–607.
Addiction, 102, 1064–1075

37. Poelen E. A. P., Scholte R. H. J., Engels R. C. M. E., Boomsma
D. I., Willemsen G. Prevalence and trends of alcohol use and
misuse among adolescents and young adults in the Nether-
lands from 1993 to 2000. Drug Alcohol Depend 2005; 79:
413–21.
38. Hibell B., Andersson B., Bjarnasson T., Ahlström S., Bala-
kireva O., Kokkevi A. et al. Alcohol and Other Drug Use Among
Students in 35 European Countries. The ESPAD Report 2003.
© 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction
Alcohol-specific rules and personality 1075
The Swedish Council for Information on Alcohol and
Other Drugs (CAN), Council of Europe, Co-operation Group
to Combat Drug Abuse and Illicit Trafficking in Drugs
(Pompidou Group); 2004.
39. Mounts N. S. Adolescents’ perceptions of parental manage-
ment of peer relationships in an ethnically divers sample.
J Adolesc Res 2004; 19: 446–67.
Addiction, 102, 1064–1075