Journal of Abnormal Child Psychology, Vol. 21, No.

6, 1993

Relation of Parental Support and Control to

Adolescents' Externalizing Symptomatology and
Substance Use: A Longitudinal Examination of
Curvilinear Effects
Eric Stice, 1,2 Manuel Barrera, Jr., 1 and Laurie Chassin 1

Past research has generated inconsistent findings regarding the relation of

parental control and support to adolescent problem behaviors. Using two waves
of data collected 1 year apart, the current study examined the influence of
parental control and support on adolescents' externalizing symptoms, alcohol
use, and illicit substance use. A sample of adolescents and their parents (N
= 454) was studied, within which approximately half of the adolescents were
at high risk because of parental alcoholism. Multiple-regression analyses of
cross-sectional data showed a negative quadratic relation between parental
control and adolescent externalizing symptomatology, and between parental
control and adolescent illicit substance use. Parental control had a negative
linear relation to adolescent alcohol use. Parental support showed a negative
quadratic relation to adolescent illicit substance use, and negative linear
relations to adolescent alcohol use and externalizing symptoms. Although
longitudinally adjusted contemporaneous results were consistent with
cross-sectional findings, parental support and control were prospectively related
only to adolescent alcohol use. The quadratic relations suggest that adolescents
who receive either extreme of parental support or control are at risk for problem
behaviors.

Manuscript received in final form April 26, 1993.

This research was supported by grant DA05227 from the National Institute on Drug Abuse
to Laurie Chassin (principal investigator) and Manuel Barrera, Jr. (coprincipal investigator).
Parts of the cross-sectional results were presented at the biennial meeting of the Southwestern
Society for Research in Child Development, Tempe, Arizona, March 1992. A version of the
longitudinal results was presented at the annual convention of the Western Psychological
Association, Portland, April-May 1992. The authors thank Heather Shaw, Patrick Curren,
and Steven West for their helpful comments regarding this manuscript.
1Arizona State University, Tempe, Arizona 85287.
2Address all correspondence to Eric Stice, Department of Psychology, Arizona State
University, Tempe, Arizona 85287-1104.

609
0o91-0627/93/1200-o609507,00/0@ 1993 PlenumPublishingCorporation
610 Stice, Barrera, and Chassin

Parenting styles have been linked to the development of childhood and

adolescent problem behaviors such as externalizing disorders, alcohol use,
and illicit substance use (Baumrind, 1991; Patterson, DeBaryshe, & Ram-
sey, 1989). In many conceptualizations, parenting is thought to be com-
posed of two relatively orthogonal dimensions: parental control and support
(Maccoby & Martin, 1983; Rollins & Thomas, 1979). Disturbances in either
or both of these parenting dimensions can impair children's social, emo-
tional, and cognitive functioning. The major aim of this study was to ex-
amine the relation of these two parenting dimensions to adolescents'
externalizing symptoms, alcohol use, and illicit substance use.

The Relation of Parental Control to Adolescent Problem Behaviors

Social learning models suggest that parents promote children's anti-

social behavior through inconsistent monitoring and noncontingent re-
sponding to children's behavior (Patterson et al., 1989). Additionally, lax
or inconsistent parental discipline may result in a disruption in children's
identification with their parents (Hirschi, 1969), which in turn may interfere
with the internalization of parental and societal values and norms. These
omissions leave children lacking in self-control which is thought to result
in externalizing behaviors and substance use. Although these theories ex-
plain the relation between low levels of parental control and problem be-
haviors, they do not account for the curvilinear relations between parental
control and problem behaviors that have been found in some studies (e.g.,
Pandina & Schuele, 1983; Rollins & Thomas, 1979). Extreme parental con-
trol is thought to preclude a successful child-parent identification. Exces-
sive parental discipline and monitoring may cause rebelliousness in the
adolescent that takes the form of externalizing behaviors and illicit sub-
stance use. Moreover, children who experience highly restrictive parenting
may seek social support outside of the family from delinquent peer groups
(Klein, Jorgensen, & Miller, 1978), which has been linked to deviance and
substance use initiation (Jacob & Leonard, 1991).
There are data to support both theories that predict linear and quad-
ratic relations between parental control and externalizing symptoms. Several
studies found negative relations between parental control and children's
problem behavior (Baumrind, 1991; Lamborn, Mounts, Steinberg, & Dorn-
busch, 1991; Patterson & Reid, 1984; Patterson & Stouthamer-Loeber,
1984). However, researchers have also reported curvilinear relations between
parental control and childhood aggression (Becker, Peterson, Luria, Shoe-
maker, & Hellmer, 1962; McCord, McCord, & Howard, 1961; Rollins &
Thomas, 1979), and childhood delinquency (Glueck & Glueck, 1968; West
Parenting Styles and Adolescent Problem Behavior 611

& Farrington, 1973). Findings concerning the relation of parental control

to adolescent substance use are more complex. As with externalizing behav-
iors, research has found both negative (Baumrind, 1991; Dishion & Loeber,
1985; Lamborn et al., 1991; Mercer & Kohn, 1980) and quadratic (Akers,
Krohn, Lanza-Kaduce, & Radosevich, 1979; Braucht, Brakarsh, Follingstad,
& Berry, 1973; Pandina & Schuele, 1983; Rollins & Thomas, 1979) relations
between parental control and adolescent alcohol and illicit substance use.
Surprisingly, researchers have also reported positive relations between pa-
rental control and adolescent drug use (Brook, Whiteman, & Gordon, 1983;
Kandel, Kessler, & Margulies, 1978).
These mixed findings are consistent with the supposition that there
is a negative quadratic relation between parental control and problem be-
haviors. First, if there is a negative quadratic relation between control and
problem behaviors, and only a linear solution was tested, the best fit solu-
tion would be a negative linear one. Indeed, most researchers apparently
did not check for quadratic effects. For example, Foxcroft and Lowe (1991)
noted that curvilinear effects are apparent in the results of some studies
but were not reported.
Second, researchers often dichotomized the parenting variables, re-
sulting in three or four types of parenting styles (e.g., Lamborn et al., 1991).
Although focusing on parenting styles (such as authoritarian and indulgent
parenting styles) has heuristic value, this approach renders it impossible to
find anything other than a linear relation between these parenting variables
and problem behaviors. Hence, there may be quadratic effects between
control and problem behaviors that are obscured by categorizing continuous
variables.
Another explanation for the inconsistent findings may be sampling
variation. If the relation between parental control and problem behavior is
truly quadratic, but the sampling methods used in some studies produced a
truncated sample of problem behaviors, then linear findings would be ex-
pected. In fact, many of the studies reporting negative linear effects between
parental control and substance use employed high school samples that were
likely biased toward better adjusted adolescents (e.g., Mercer & Kohn,
1980). However, in a study that included adolescents referred for substance
abuse treatment, quadratic effects were obtained (Pandina & Schuele, 1983).
Interestingly, the impact of parental control may also differ depending
on whether the criterion is alcohol use or illicit substance use. Regarding
the relation of parental control to adolescent alcohol use, all of the re-
searchers reported negative linear findings. However, for parental control's
relation to adolescent illicit substance use, researchers reported negative,
positive, and quadratic findings. The first aim of the present study was to
examine the form of the relations between parental control and adolescent
612 Stice, Barrera, and Chassin

externalizing behavior, alcohol use, and illicit substance use, using a con-
tinuous measure of parental control. Both linear and quadratic relations
were examined.

The Relation of Parental Social Support to Adolescent Problem Behaviors

A supportive parental relationship and strong parent-child bond are

thought to promote adolescents' internalization of their parent's conven-
tional attitudes and behaviors (Baumrind, 1991; Jacob & Leonard, 1991).
Parental support is also thought to decrease the likelihood that adolescents
will affiliate with a deviant peer group which has been linked to the de-
velopment of problem behaviors (Kandel & Andrews, 1987; Patterson &
Stouthamer-Loeber, 1984). Finally, supportive parents are thought to con-
tribute to children's self-acceptance and self-efficacy, which are important
for healthy emotional and psychosocial development.
Investigators have reported negative relations between parental support
and conduct problems (Wolchik, Beals, and Sandier, 1989), delinquency (Pat-
terson & Stouthamer-Loeber, 1984; West & Farrington, 1973), problem be-
havior (Lamborn et al., 1991), adolescent alcohol use (Barnes, 1984; Barnes,
Farrell, & Cairns, 1986; Margulies, Kessler, & Kandel, 1977), and adolescent
illicit substance use (Brook, Whiteman, & Gordon, 1983; Kandel et al., 1978;
Mercer & Kohn, 1980; Pandina & Schuele, 1983). However, two studies
found a quadratic effect of parental support on adolescents' substance use
(Baumrind, 1991; Smart, Chibucos, & Didier, 1990). Foxcroft and Lowe
(1991) suggested that extremes of parental support may reflect enmeshment
and disengagement, which are both dysfunctional for child development. The
inconsistent pattern of findings suggest that it would be important to check
for quadratic relations between support and problem behaviors.
Previous studies have differed in the operationalization of support,
which is a multifaceted construct including domains such as provisions of
companionship, intimacy, affection, instrumental aid, and expression of ad-
miration (Barrera, 1986; Furman & Buhrmester, 1985). Relatively few inves-
tigators have used measures of support with adequate content validity.
Moreover, as described above, studies that dichotomize families as high and
low support cannot detect curvilinear effects. Consequently, a second aim of
the current study was to examine the form of the relation between parental
support and adolescent problem behaviors, using a continuous measure of
support with adequate content validity, and testing for quadratic effects.
In addition to the points detailed above, there are two general meth-
odological limitations of this body of literature. First, many studies rely
completely on adolescent self-reports, which raises questions about the va-
Parenting Styles and Adolescent Problem Behavior 613

lidity of these findings. Accordingly, the current study used multiple re-
porters. Second, most previous research has been cross-sectional, so that
the directionality of effects cannot be determined. Thus, the current study
used a longitudinal design.
Two analytic approaches were used with the current longitudinal data.
Because prior symptoms are often the best predictor of future symptoms,
it is important to control for initial symptoms in longitudinal research
(Monroe, 1983). Thus, we examined if parenting at wave 1 predicted wave
2 problem behavior, after controlling for wave 1 problem behavior. How-
ever, criticisms have been leveled at traditional prospective designs (Gollob
& Reichardt, 1987; Rogosa, 1988; Rogosa & Willett, 1985). One central
problem is selecting the appropriate time lag to capture the effect. If the
time lag chosen is too long for the variables under study, the prospective
analyses will yield nonsignificant results (Gollob & Reichardt, 1987). Ac-
cordingly, a longitudinally adjusted contemporaneous analysis was em-
ployed that is not subject to the difficulty in specifying the time lag. In this
analysis the effects of wave 2 parenting on wave 2 problem behavior were
assessed, while adjusting for initial levels of both wave 1 parenting and
problem behavior. This technique was suggested by Cronbach and Furby
(1970) as a method of measuring concomitant change across different do-
mains, and is a simple extension of analysis of partial variance as proposed
by Cohen and Cohen (1983). Although this approach cannot demonstrate
temporal ordering like prospective analyses, it represents an improvement
over cross-sectional analyses because the effects of the initial levels of the
predictors as well as the criteria are partialed o u t : A s noted above, this
analysis technique also does not present the time lag specification difficul-
ties inherent in traditional prospective analyses.

Aims of the Present Study

The purpose of this study was to examine the relations of parental

control and support to adolescent externalizing symptomatology, alcohol
use, and illicit substance use. The main focus was to determine the form
of the relation of parental control and support to adolescent problem be-
haviors. Because of the focus on problem behaviors, this study oversampled
adolescent children of alcoholics who are at high risk for these problems
(Chassin, Rogosch, & Barrera, 1991; Sher, Walitzer, Wood, & Brent, 1991).
Accordingly, approximately half of the current sample consisted of children
of alcoholics and the remainder were children of nonalcoholic parents. 3
3A published report, using wave 1 data from this project, addressed the effects of parental
alcoholism on adolescent psychopathology (Chassin et al., /991).
614 Stice, Barrera, and Chassin

We attempted to address limitations of past literature by using multiple

reporters, a longitudinal design, a broadly operationalized parental support
measure, and continuous measures of control and support. The two-wave
design allowed us to test the hypotheses in the initial wave of data, replicate
the findings in the second wave of data, conduct prospective analyses, and
perform longitudinally adjusted contemporaneous analyses.

METHOD

Subjects

At wave 1 the sample consisted of 454 adolescents (214 females and

240 males), and their parents. The adolescents ranged in age from 10.5 to
15.5 (average age = 12.7) at wave 1 data collection. Families were catego-
rized as either Hispanic (22.9%) or non-Hispanic Caucasians (77.1%) ac-
cording to self-reports. Parental education ranged from grade school to
graduate school with a modal response of some college attendance. Because
the attrition rate was only 1.2% from wave 1 to wave 2, attrition analyses
were not performed. The sample was comprised of children of alcoholics
(COALs) and matched controls, with 54% of the adolescents in the sample
residing with a biological alcoholic parent. There were 211 alcoholic fathers
and 59 alcoholic mothers. Twenty-four families had two alcoholic parents.
Parental alcoholism was characterized by an early onset of drinking prob-
lems, with 74.6% of the alcoholic fathers, and 58.5% of the alcoholic moth-
ers reporting drinking problems before age 25. Of the alcoholic parents,
31% of the fathers and 33% of the mothers reported receiving alcoholism
treatment. Regarding recency, at the initial interview 32% of the alcoholic
fathers and 46.4% of the alcoholic mothers reported the occurrence of al-
cohol abuse or dependency symptoms within the last year.

Recruitment Procedures

COA families were recruited using community telephone surveys,

questionnaires from a health maintenance organization, and court driving
while intoxicated (DWI) records. COAs had to meet the following criteria:
Anglo or Hispanic ethnicity, Arizona residency, ages 10 to 16, English-
speaking, and no cognitive limitations that would preclude interview (e.g.,
psychosis). Direct interview data had to confirm that a biological and cus-
todial parent met DSM-III criteria for lifetime diagnosis of alcohol abuse
or dependence by the Diagnostic Interview Schedule (DIS) or by spouse-
Parenting Styles and Adolescent Problem Behavior 615

report on Family History Research Diagnostic Criteria (FH-RDC) if a par-

ent could not be interviewed. After COA families were selected, they were
matched with control families. Potential controls were identified through
reverse directories that listed households in the same neighborhoods as
COA families. Telephone screening interviews were used to find partici-
pants who matched COA families on (a) child's age within 1 year, (b) family
structure, and (c) ethnicity of alcoholic parent. Families were matched for
socioeconomic status using property value codes in reverse directories or
parents' report of income. Direct interviews verified that neither parent
met DSM-I11 or F H - R D C criteria for lifetime diagnosis of alcohol abuse
or dependence. The recruitment procedures are presented in greater detail
elsewhere (Chassin, Barrera, Bech, & Kossak-Fuller, 1992).

Procedure

When subjects were recruited, the study was presented as an attempt

to understand why some children have an easy and successful time adjusting
to adolescence, and others develop problems. Data were collected in annual
computer-assisted interviews with the adolescents and their parents. Inter-
view items required close-ended responses which were entered directly into
laptop computers. To minimize contamination, family members were inter-
viewed separately during the same time interval by different interviewers
(in all but 11 cases). Confidentiality was assured with a Certificate of Con-
fidentiality from the Department of Health and Human Services. Inter-
viewers were blind to the hypotheses of the study. Interviews lasted 1 to 2
hr and families received $50 for their participation.

Measures

Covariates. Adolescent age, gender, and ethnicity, 4 as well as parental

education and lifetime alcoholism diagnosis (by R D C or DIS criteria) were
used as covariates. 5

4For the regression analyses, a dichotomous variable was created that coded families as
Hispanic if either the adolescent, the mother, or the father classified themselves as Hispanic,
otherwise they were coded as non-Hispanic Caucasians.
5Diagnosis of parent antisocial personality disorder (APD) was also available. Although
research has found that parental APD is related to adolescent externalizing behaviors,
preliminary analyses indicated that the precision of the analyses was not improved by the
inclusion of this covariate. Covariates are used for the purpose of controlling extraneous
variables statistically, thereby increasing the precision of the analyses (Pedhazur, 1982). There
were no differences in the pattern of significant findings with and without parental APD
diagnosis in the equation, indicating that the precision of the analyses was not improved by
616 Stice, Barrera, and Chassin

Parenting Variables. Perceived parental support was measured with an

adapted version of the Network of Relationships Inventory (Furman &
Buhrmester, 1985). Adolescents answered six questions about the types of
social support they received over the preceding 3 months from each of their
parents. This scale assessed the domains of companionship, guidance, inti-
macy, affection, admiration, and reliable alliance. Because the adolescents'
reports of mother and father support were correlated (wave 1 r = .60, wave
2 r = .63), they were combined by averaging. Cronbach's alpha for the pa-
rental support scale was .88 for wave 1 and .89 for wave 2. In families with
only one parent, that parent's social support score was used. Perceived pa-
rental control was measured with 10 items from the nonenforcement and
inconsistent discipline subscales of the Child Report of Parental Behavior
Inventory originally constructed by Schaefer (1965). Adolescents rated the
items separately for mother and father parenting over the past 3 months.
Because the adolescents' reports of mother and father control were corre-
lated (.72 for both wave 1 and wave 2), the two were combined. The internal
consistency for the scale was .89 for wave 1 and .90 for wave 2. In families
with only one parent, that parent's control score was used. Because both
parenting variables were assessed exclusively with adolescent reports, they
should be regarded as measures of perceived social support and control.
Externalizing Symptoms. The adolescents' level of externalizing symp-
tomatology was assessed using mother, father, and adolescent reports of
items from the Child Behavior Checklist (Achenbach & Edelbrock, 1983).
All items reflected a 3-month time window. This scale consisted of 22 items
that loaded on the externalizing factor for both girls and boys ages 12 to
16. In adolescent self-report, the response scale was expanded to a 5-point
scale to increase the variance. Because mother and father reports of ex-
ternalizing were correlated (wave 1 r = .65, wave 2 r = .66), the two reports
were combined. Cronbach's alpha for the parent report was .92 for wave
1 and .93 for wave 2. Cronbach's alpha for the adolescent report was .88
for both waves 1 and 2.
Substance Use. Adolescents self-reported their frequency of consump-
tion during the past 3 months of beer/wine/winecoolers, and hard liquor,
as well as their frequency of heavy drinking (five or more drinks on one
occasion). The alcohol use measure was formed by summing across these
three items. This variable was raised to the .25th power to normalize the
distribution. Cronbach's alpha for the alcohol use measure was .85 for wave
1 and .86 for wave 2. Adolescents also self-reported their frequency of use

the inclusion of this variable. Omitting APD as a covariate permitted the inclusion of 99
families that would have been dropped because of missing APD diagnoses. Similarly, family
structure was omitted as a covariate because preliminary analyses indicated that the pattern
of significant effects was the same with and without the inclusion of this covariate.
Parenting Styles and Adolescent Problem Behavior 617

during the past 3 months of marijuana/hashish, amphetamines, quaaludes/

barbiturates, tranquilizers, hallucinogens, cocaine/crack, opiates, and inha-
lant drugs. Items were summed to form a composite measure of illicit sub-
stance use. This variable was also raised to the .25th power to normalize
the distribution. Cronbach's alpha for the illicit substance use measure was
.77 and .75 for waves 1 and 2, respectively.
Because of the young age of the sample, the prevalence of adolescent
substance use was generally low. Accordingly, the current study is best
viewed as an examination of substance use initiation. At the wave 2 data
collection, 26% of the adolescents reported alcohol use over the past 3
months. Within the past 3 months, 15% reported drinking beer/winecoolers
one to two times, 3% reported drinking beer/winecoolers three times, 4%
reported drinking beer/winecoolers six to nine times, and 3% reported
drinking beer/winecoolers over 10 times. Regarding illicit substance use,
7% of the adolescents reported illicit substance use over the past 3 months
at wave 2. These rates are comparable to national data for older adoles-
cents (ages 12 to 17), as indicated by the National Household Survey (Na-
tional Institute on Drug Abuse, 1989).

RESULTS

Preliminary multiple-regression analyses revealed no substantively

meaningful Covariate x Predictor interactions, a condition that must be
satisfied when using covariates (Pedhazur, 1982). Specifically, tests were
performed for two-way interactions between each covariate (parental alco-
holism status, ethnicity, age, gender, and parental education) and each pre-
dictor (support and control). Multiple-regression analyses also indicated
that there were no substantively meaningful interactions between parental
alcoholism status and any of the other covariates which would have indi-
cated that the effects of parental alcoholism were moderated by that co-
variate. Further, analyses found no parental support by control interaction
that would have necessitated the inclusion of an interaction vector in the
analyses. The zero-order correlation matrix among the predictor and cri-
terion variables for waves 1 and 2 are provided in Table I.
To investigate the joint and unique effects of parental control and
support on the criteria, hierarchical multiple-regression analyses were per-
formed with covariates entered on step 1 and parental support and control
entered on step 2. For each wave of data, multiple-regression analyses were
performed for the following criteria: adolescent and parent reports of ex-
ternalizing symptoms, adolescent reports of alcohol use, and adolescent re-
ports of illicit substance use. To test the prediction that each parenting
618 Stice, Barrera, and Chassin

Table I. Zero-Order Correlations Between Parental Support, Parental Control, and Criteria a
1 2 3 4 5 6
Wave 1
1. Parental support .45 b -.42 b -.23 b -.283 -.27 b
2. Parental control -.42 b -.15 b -.27 b -.22 b
3. Adolescent reports of
externalizing symptoms .420 .41 b .290
4. Parent reports of externalizing
symptoms .20 b .22 b
5. Adolescent reports of alcohol
use .51 b
6. Adolescent reports of substance
use
Wave 2
1. Parental support ,47 b -.39 b -.23 b -.32 b -.27 b
2. Parental control -.48 b -.20 b -.350 -.26 b
3. Adolescent reports of
externalizing symptoms .45 b .43 b .29 b
4. Parent reports of externalizing
symptoms .22 t' .28 b
5, Adolescent reports of alcohol
use .49 b
6. Adolescent reports of substance
use
aNote: All significance tests were two tailed.
bp < .001.

variable would have a unique effect, the partial regression coefficients (B)
associated with parental control and support were tested for significance.
Table II provides the F-changes, Bs, and the amount of variance accounted
for by the joint and unique effects of parental support and control on all
criteria for waves 1 and 2. The joint effects of parental control and social
support were statistically significant for all criteria, and nearly all of the
unique effects of parental control and support were significant.
Quadratic effects were tested with hierarchical multiple-regression
analyses where covariates were entered on step 1, linear terms for parental
control and support were entered on step 2, and a power vector repre-
senting the quadratic trend for either control or support was entered on
step 3. For each wave of data, multiple-regression analyses were performed
for all criteria. Table III presents the F-changes and the amount of variance
accounted for by the quadratic effects of control and support on all criteria
for waves 1 and 2. Parental control showed a quadratic relation to adoles-
cent and parent reports of externalizing symptoms at wave 1, and to parent
reports of externalizing symptoms and adolescent reports of illicit substance
use at wave 2. Support was related quadratically to adolescent reports of
Parenting Styles and Adolescent Problem Behavior 619

Table II. Joint and Unique Effects of Parental Support and Control on Externalizing
Symptomatology and Substance Use: Cross-Sectional Results from Waves 1 and 2a
Joint effects of
support and control, Support Control
F change unique, unique,
Dependent variables (% of variance) B (% of variance) B (% of variance)
Wave 1
Adolescent reports of
externalizing symptoms 57.6 (18.9) a -.178 (5.7) a -.208 (5.1) a
Parent reports of
externalizing symptoms 12.3 (4.9) a -.065 (3.5) `/ -.010 (0.I), n.s.
Adolescent reports of
alcohol use 15.6 (5.7) `/ -.117 (2.2) a -.100 (1A) c
Adolescent reports of illicit
substance use 15.9 (6.3) a -.066 (3.5) a -.029 (0.5), n.s.
Wave 2
Adolescent reports of
externalizing symptoms 58.0 (18.6) a -.121 (2.9) a -.253 (8.0) a
Parent reports of
externalizing symptoms 14.6 (5.8) a -.062 (2.9) a -.036 (0.6) b
Adolescent reports of
alcohol use 23.8 (8.5) u -.129 (2.4) a -.161 (2.3) d
Adolescent reports of illicit
substance use 13.9 (5.8) a -.057 (2.6) a -.039 (0.8) b
aNote: Joint tests, df =2/434; unique tests, df = 1/434. Column 1 presents the combined effects
of parental support and control on the criteria when entered after the covariates were in
the equations. Columns 2 and 3 present the unique effects of parental support and control
for all criteria as assessed by the partial regression coefficient (B) from the full regression
model.
bp < .10.
~p < .05.
< .001.

illicit substance use at both waves and to parent reports of externalizing

symptoms at wave 2.
To depict the nature of the quadratic effects, we adopted the graphing
procedure recommended by Hayduk (1987). Because all of the quadratic ef-
fects had a negative sloping curvilinear shape, two illustrative examples were
selected. Examples of quadratic effects for both support and control, on ado-
lescent as well as parent-reported outcomes, were chosen. Observed values
of parental control were entered into the full regression equation predicting
parental reports of externalizing symptoms, and observed values of parental
support were entered into the full regression equation predicting child reports
of illicit substance use. The predicted values are plotted in Fig. 1.
Longitudinal relations between parenting and adolescent problem be-
haviors over the 1-year period were investigated using hierarchical multi-
ple-regression analyses. First, prospective analyses were performed to
620 Stice, Barrera, and Chassin

Table III. F-Change and the Percent of Variance Explained by the Quadratic Effects of
Parental Support and Control: Cross-Sectional Results from Waves 1 and 2a
Quadratic effects of Quadratic e f f e c t s o f
parental support, parental control,
F change F change
Dependent variables (% of variance) (% of variance)
Wave
Adolescent reports of externalizing
symptoms 3.58 (0.6) o 4.92 (0.8) c
Parent reports of externalizing symptoms 1.07 (0.2), n.s. 7.01 (1.4) a
Adolescent reports of alcohol use 0.94 (0.2), n.s. 0.50 (0.1), n.s.
Adolescent reports of illicit substance
use 16.26 (3.1) e 3.77 (0.7) b
Wave 2
Adolescent reports of externalizing
symptoms 2.65 (0.4), n.s. 0.19 (0.0), n.s.
Parent reports of externalizing symptoms 21.67 (4.0) e 12.16 (2.3) e
Adolescent reports of alcohol use 0.73 (0.1), n.s. 0.27 (0.1), n.s.
Adolescent reports of illicit substance
use 6.25 (1.3) c 5.43 (1.1) c
aNote: For wave 1 tests, df = 1/433; for wave 2 tests, df = 1/428. Quadratic terms were entered
after the linear terms for parental support and control and covariates were in the equations.
bp < .10.
~p < .05.
< .01.
ep < .001.

ascertain if wave 1 parenting was predictive of wave 2 problem behaviors

while controlling for wave 1 problem behaviors. Separate hierarchical mul-
tiple-regression analyses were performed for all wave 2 criteria where co-
variates and the wave 1 measure of the criterion was entered on step 1,
and wave 1 support and control were entered on step 2. Table IV presents
the F-changes, Bs, and the percentages of variance accounted for in the
prospective analyses. The only significant prospective finding was the joint
effect of parental support and control on adolescent alcohol use. There
were no statistically significant quadratic effects in the prospective analyses.
Second, longitudinally adjusted contemporaneous analyses were per-
formed. These hierarchical multiple-regression analyses examined the rela-
tions between wave 2 parenting and wave 2 problem behavior after adjusting
for wave 1 levels of parenting and problem behaviors. These analyses were
accomplished by entering wave 2 parental control and support after the co-
variates, wave 1 parental control and support, and the wave 1 version of
the criteria were already entered into the equation. Significance at this step
indicated that parenting at wave 2 predicted problem behavior at wave 2,
above and beyond the effects predictable from initial levels of these vari-
ables. Table IV reports the F-changes, Bs, and the percentages of variance
Parenting Styles and Adolescent Problem Behavior 621

J 'E
~m .20

.1o

[
lI,
2~N~. 5
I
-l.0
I
-0.5
I
0.0
I
0.5
I
!.0

Parental Control

o
8
8

N 4

,s

I ~ J
~ J J I I I j I
~2.0 -1.0 O.O 1.0

Parental Support

Fig. 1. Predicted values of externalizing symptoms

from the regression of parental reports of external-
izing symptoms on support and the linear and quad-
ratic terms of control at wave 1, and predicted values
of adolescent illicit substance use from the regres-
sion o f adolescent reports of illicit substance use on
control and the linear and quadratic terms of sup-
port at wave 1.

accounted for by the joint and unique effects of parental control and support
in the longitudinally adjusted contemporaneous analyses. All of the joint
effects and many of the unique effects were statistically significant.
622 Stice, Barrera, and Chassin

Table IV. Percent of Variance Accounted for by the Joint and Unique Effects of Parental
Support and Control in the Prospective and Longitudinally Adjusted Contemporaneous
Analysed

Joint effects of
support and control, Support Control
Wave 2 dependent F change unique, unique,
variables (% of variance) B (% of variance) B (% of variance)
Prospective analyses using wave 1 predictors
Adolescent reports of
externalizing symptoms 0.03 (0,0), n.s. -.01 (0.0), n,s. -.01 (0.0), n.s.
Parent reports of
externalizing symptoms 0.48 (0.1), n.s. -,01 (0.0), n.s. -.00 (0.0), n.s.
Adolescent reports of
alcohol use 5.9 (1.9)a -.07 (0.6)b -.08 (0.6)t'
Adolescent reports of illicit
substance use 2.1 (0.7), n.s. -.02 (0.3), n.s. -.01 (0.1), n.s.
Longitudinally adjusted contemporaneous analyses using wave 2 predictors
Adolescent reports of
externalizing symptoms 21.4 (4.8)e -.08 (0.8) d - . 1 8 (3.0) e
Parent reports of
externalizing symptoms 6.2 (1.0)d -.04 (0.9)e -.00 (0.0), n.s.
Adolescent reports of
alcohol use 7.1 (2.2)e - . 0 8 (0.7) b - . 1 3 (1.1) ~
Adolescent reports of illicit
substance use 4.6 (1.5)c -.03 (0.5)b -.04 (0.6)b
aNote: Prospective analyses: joint tests, df =2/434; unique tests, df = 1/434. Longitudinally
adjusted contemporaneous analyses: joint tests, df =2/432; unique tests, df = 1/432. For
prospective analyses column 1 presents the joint effects of parental support and control when
entered after the covariates. For longitudinally adjusted contemporaneous analyses column
1 presents the joint effects of wave 2 parental support and control when entered after the
covariates and wave 1 variables. Columns 2 and 3 present the unique effects of parental
support and control as assessed by the partial regression coefficient (B) from the full
regression models.
bp < .10.
~ < .05.
< .01.
ep < .001.

Q u a d r a t i c r e l a t i o n s w e r e t e s t e d in t h e l o n g i t u d i n a l l y a d j u s t e d c o n t e m -
p o r a n e o u s a n a l y s e s by e n t r y o f p o w e r v e c t o r s for w a v e 2 p a r e n t a l s u p p o r t
o r c o n t r o l a f t e r t h e p r e v i o u s l y specific w a v e 2 l i n e a r effects w e r e in t h e
equation. Quadratic relations were detected between parental support and
p a r e n t r e p o r t s o f e x t e r n a l i z i n g s y m p t o m s [ F c h a n g e (1, 429) = 6.87, p <
.01], b e t w e e n p a r e n t a l c o n t r o l a n d p a r e n t r e p o r t s o f e x t e r n a l i z i n g s y m p t o m s
[ F c h a n g e (1, 429) = 3.82, p < .05], a n d b e t w e e n p a r e n t a l c o n t r o l a n d
child r e p o r t s o f illicit s u b s t a n c e u s e I F c h a n g e (1, 431) = 2.84, p < .10].
Parenting Styles and Adolescent Problem Behavior 623

As with the cross-sectional quadratic results described earlier, negative

sloping quadratic effects were obtained.

DISCUSSION

Summary of Findings

In cross-sectional analyses the joint and unique effects of parental

support and control were generally negatively related to adolescent and
parent reports of externalizing symptomatology, and adolescent reports of
alcohol and illicit substance use. Parental control showed quadratic rela-
tions to adolescent externalizing symptoms and illicit substance use, but
not to adolescent alcohol use. Quadratic effects were also found for pa-
rental support's relation to adolescent externalizing symptoms and illicit
substance use. The only prospective finding when controlling for initial
problem behaviors was the joint effect of support and control on adolescent
alcohol use. However, after adjusting for initial levels of parental support,
control, and adolescent problem behaviors, wave 2 parental support and
control were predictive of all wave 2 problem behaviors. Generally the mag-
nitude of the effects was stronger for adolescent-reported criteria than for
parent-reported criteria.

Effects of Parental Control

The quadratic relation between parental control and adolescent ex-

ternalizing symptomatology was a novel finding. Although this relation has
been found with substance use, it has not been reported in relation to gen-
eral adolescent externalizing symptomatology. The results are consistent
with findings reported for specific types of externalizing problems (aggres-
sion and delinquency) in younger age groups (Glueck & Glueck, 1968; Rol-
lins & Thomas, 1979; West & Farrington, 1973). Parental control also
showed quadratic relations to adolescent illicit substance use, but not to
adolescent alcohol use. Other studies have found quadratic relations be-
tween parental control and adolescent illicit substance use (e.g., Pandina
& Schuele, 1983), and negative relations between parental control and ado-
lescent alcohol use (e.g. Baumrind, 1991). However, previous research has
also reported both negative relations between parental control and illicit
substance use (Dishion & Loeber, 1985) and positive relations (Brook et
al., 1983; Kandel et al., 1978).
624 Stice, Barrera, and Chassin

It is interesting to speculate why quadratic effects would be found

for illicit substance use but not for alcohol use. These findings might reflect
the fact that alcohol use is a more normative parental behavior in our so-
ciety than is illicit substance use. Extreme parental control is thought to
disrupt parent-child relationships, resulting in rebellion against parental
norms, rather than internalization of them. Illicit substance use, because it
is a less normative behavior for parents, may be perceived by adolescents
as a more rebellious behavior than alcohol use. Adolescents with overly
controlling parents might be less likely to emulate their parents' use of
alcohol, but rather might use illicit substances in acts of rebellion. This
explanation is consistent with the work of Braucht and associates (1973),
who noted that illicit substance use is more likely to serve as an act of
rebellion for adolescents than alcohol use.
Alternatively, a reverse direction of influence might explain the find-
ings. High levels of adolescent illicit substance use might be more likely
than high levels of adolescent alcohol use to elicit increased control at-
tempts from parents. This explanation would be consistent with the lack
of prospective effects of control on adolescent illicit substance use and the
presence of prospective effects of control on adolescent alcohol use.
When interpreting the differential effects of parenting on adolescent
illicit substance use and alcohol use, it is important to acknowledge that
the curvilinear relation between parental control and adolescent illicit sub-
stance use for wave 1 was not large nor completely replicated in wave 2.
However, by necessity the power vectors representing quadratic effects
were highly collinear with the linear effect vectors, resulting in a stringent
test for quadratic effects.
These findings are generally supportive of compliance training theory
(Patterson et al., 1989) and adolescent identification theory (Hirschi, 1969),
but also suggest that these linear theories might break down at high levels
of parental control. Extremely high parental control, as well as extremely
low parental control, were associated with adolescent externalizing symp-
toms and illicit substance use. These findings suggest that extreme parental
control is not a deterrent of adolescent externalizing behaviors and illicit
substance use as implied by previous reports of negative linear relations.
Parental overcontrol may disrupt the parent-child bond so that, instead of
adolescents internalizing parental norms, they are more likely to rebel
against them and identify with deviant peers. However, regarding adoles-
cent alcohol use, high levels of parental control appear to be more bene-
ficial. Additional research on the differential effects of parental control on
various adolescent problem behaviors is needed before firm conclusions can
be drawn concerning these findings. Independent replication would be an
important first step in addressing this issue. Future research in this arena
Parenting Styles and Adolescent Problem Behavior 625

should utilize continuous measures of parental control and support in order

to detect nonlinear relations and to increase statistical power.

Effects of Parental Support

The quadratic relation between parental support and adolescent ex-

ternalizing symptomatology has not been previously reported. However, the
quadratic relation between parental support and adolescent illicit substance
use was consistent with previous reports (Baumrind, 1991; Smart et al.,
1990). Other studies have found negative linear relations between parental
support and adolescent substance use, but these studies have not tested for
quadratic effects (e.g., Brook et al., 1983; Kandel & Andrews, 1987). Finally,
the negative relation between parental support and alcohol use was in accord
with past findings (e.g., Barnes et al., 1986; Margulies et al., 1977).
Family systems theory offers an explanation for the quadratic effects
of parental support by asserting that moderate cohesion is optimal for family
functioning but that extremes of disengagement and enmeshment are prob-
lematic (Foxcroft & Lowe, 1991; Minuchin, 1974). A warm and supportive
parent-child bond with a balanced amount of cohesiveness encourages the
child's identification and internalization of the parent's conventional atti-
tudes and behaviors (Jacob & Leonard, 1991). This is thought to lead to
social competence and low levels of externalizing behavior (Baumrind,
1991). The findings indicate that a moderate level of parental support may
be ideal, suggesting that excessive cohesiveness inhibits optimal social de-
velopment. Although focusing primarily on marital relations, Coyne, Wort-
man, and Lehman (1988) discussed how emotional overinvolvement could
produce negative consequences for support recipients.

Joint Effects of Parental Support and Control

The joint effects of parental support and control were demonstrated

repeatedly in the present study. The variance accounted for by the joint
effects of these variables on adolescent externalizing symptomatology, al-
cohol use, and illicit substance use ranged from 5 to 19%. This pattern of
findings was stable and was replicated across reporters, as well as across
waves of measurement. Although the longitudinally adjusted contempora-
neous analyses were consistent with the cross-sectional results, there were
few significant prospective effects. It is important to note that the longitu-
dinally adjusted contemporaneous findings do not imply unidirectional ef-
fects. These analyses are an improvement over cross-sectional analyses
626 Stice, Barrera, and Chassin

because they adjust for prior levels of the predictors and criteria, but they
cannot identify the direction of influence.
The prospective findings provide little evidence that parenting exerts
a temporal effect on problem behaviors during adolescence. Two interpre-
tations may be given regarding the minimal prospective effects. First, the
lack of effects may be due to a misspecification of the causal time lag that
is operating between the variables. As Gollob and Reichardt (1987) dis-
cussed, if the time lag chosen is too long for the variables under study, the
prospective analyses will yield nonsignificant results. Thus, parenting might
exert its effects over a shorter time period than a year. The fact that the
longitudinal adjusted contemporaneous results converge with the cross-sec-
tional results suggests that perhaps the causal time lag between parenting
and problem behavior is shorter than 1 year. Second, reverse directionality
may be responsible for the lack of prospective effects, with an adolescent's
problem behaviors eliciting extremes in parental control and support. In-
deed, research has found that child behavior influences parent behavior
(Lytton, 1990).

Limitations of the Present Research

It is also important to consider some of the study's limitations. For

externalizing symptomatology, parent reports did not perfectly replicate
child reports. There are two factors that may account for this fact. First,
adolescents might be considered to be more informed reporters than par-
ents for many symptoms of psychopathology. Parents only observe a small
segment of the adolescent's behavior, and are not exposed to many of the
other environments such as the school setting or the peer setting in which
the adolescent interacts. Research has documented that parents underre-
port child externalizing symptoms compared to child self-reports (Weiss-
man et al., 1987). This fact may attenuate the correlations for the parent
reports. Second, the imperfect replication may be due to the fact that non-
independent data sources were used. That is, the within-reporter correla-
tions might be biased high because adolescents reported both the predictors
and the criteria. However, because most of the cross-reporter results were
significant, and several of the effects were actually stronger for parent re-
ports than adolescent reports, this explanation is unlikely.
Another limitation of this study was the reliance on adolescent per-
ceptions of parental support and control. More confidence could have been
placed in the findings if multiple reporters or behavioral observations of
parenting were employed. Finally, it is important to acknowledge that the
present study focuses only on parenting influences. Many other important
Parenting Styles and Adolescent Problem Behavior 627

variables that were not included in the present model are related to the
development of adolescent problem behaviors, such as peer influences, pa-
rental modeling, and temperament.

Directions for Future Research

The results of the present study provide additional evidence of the

relation between parenting and adolescent problem behaviors. Future re-
search should explore possible mediators of the relations of parental sup-
port and control to adolescent problem behaviors. The quality of the
parent-child bond could be tested as one potential mediator. Regarding
parental control, parent's use of appropriate reinforcement and punishment
strategies could be examined as possible mediators. Future studies might
examine the relative importance of the different dimensions of the social
support construct, such as provisions of companionship, affection, instru-
mental aid, and expression of admiration, to adolescent externalizing prob-
lems. Finally, investigators also need to examine the probable bidirectional
effects between parenting and problem behaviors. A fuller understanding
of the bidirectional effects would significantly advance our understanding
of the relations between parenting and adolescent problem behaviors.

REFERENCES

Achenbach, T., & Edelbrock, C. (1983). Manual for the Child Behavior Checklist and Revised
ChiM Behavior Profile. Burlington: University of Vermont, Department of Psychiatry.
Akers, R. L, Krohn, M. D., Lanza-Kaduce, L., & Radosevich, M. (1979). Social learning and
deviant behavior: A specific test of a general theory. American Sociological Review, 44,
636-655.
Barnes, G. M. (1984). Adolescent alcohol abuse and other problem behaviors: Their
relationships and common parental influences. Journal of Youth and Adolescence, 13,
329-348.
Barnes, G. M., Farrell, M. P., & Cairns, A. L. (1986). Parental socialization factors and
adolescent drinking behaviors. Journal of Marriage and the Family, 48, 27-36.
Barrera, M., Jr. (1986). Distinctions between social support concepts, measures, and models,
American Journal of Community Psychology, 14, 413-445.
Baumrind, D. (1991). The influence of parenting style on adolescent competence and
substance use. Journal of Early Adolescence, 11, 56-95.
Becker, W. C., Peterson, D. R., Luria, Z., Shoemaker, D. J., & Hellmer (1962). Relations of
factors derived from parent-interview ratings to behavior problems of five-year-olds. Child
Development, 33, 509-535.
Braucht, G. N., Brakarsh, D., Follingstad, D., & Berry, K. L. (1973). Deviant drug use in
adolescence: A review of psycbosocial correlates. Psychological Bulletin, 79, 92-106.
Brook, J. S., Whiteman, M., & Gordon, A. (1983). Stages of drug use in adolescence:
Personality, peer, and family correlates. Developmental Psychology, 19, 269-277.
628 Stice, Barrera, and Chassin

Chassin, L., Barrera, M., Bech K., & Kossak-Fuller, J. (1992). Recruiting a community sample
of adolescent children of alcoholics: A comparison of three subject sources. Journal of
Studies on Alcohol, 53, 316-319.
Chassin, L., Rogosch, F., & Barrera, M. (1991). Substance use and symptomatology among
adolescent children of alcoholics. Journal of Abnormal Psychology, 100, 449-463.
Cohen, J., & Cohen, P. (1983). Applied multiple regression~correlation analysis for the behavioral
sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Coyne, J. C., Wortman, C. B., & Lehman, D. R. (1988). The other side of support: Emotional
overinvolvement and miscarried helping. In B. Gottlieb (Ed.), Marshaling social support:
Formats, processes, and effects (pp. 305-330). Newbury Park, CA: Sage.
Cronbach, L. J., & Furby, L. (1970). How we should measure "change"-----or should we?
Psychological Bulletin, 74, 68-80.
Dishion, T. J., & Loeber, R. (1985). Adolescent marijuana and alcohol use: The role of parents
and peers revisited. American Journal of Drug and Alcohol Abuse, 11, 11-25.
Foxcroft, D. R., & Lowe, G. (1991). Adolescent drinking behaviour and family socialization
factors: A meta-analysis. Journal of Adolescence, 14, 255-273.
Furman, W., & Buhrmester, D. (1985). Children's perceptions of the personal relations in
their social networks. Developmental Psychology, 21, 1016-1024.
Glueck, S., & Glueck, E. T. (1968). Delinquents and non-delinquents in perspective. Cambridge,
MA: Harvard University Press.
Gollob, H. F., & Reichardt, C. S. (1987). Taking account of time lags in causal models. Child
Development, 58, 80-92.
Hayduk, L. A. (1987). Structural equation modeling with LISREL. Baltimore: Johns Hopkins.
Hirschi, T. (1969). Causes of delinquency. Berkeley: University of California Press.
Jacob, T., & Leonard, K. (1991, November) Family and peer influences in the development of
adolescent alcohol abuse. Paper presented at the NIAAA conference, Working Group on
the Development of Alcohol-Related Problems in High-Risk Youth: Establishing
Linkages Across Biogenetic and Psyehosocial Domains, Washington DC.
Kandel, D. B., & Andrews, K. (1987). Processes of adolescent socialization by parents and
peers. The International Journal of the Addictions, 22, 319-342.
Kandel~ D. B., Kessler, R. C., & Margulies, R. Z. (1978). Antecedents of adolescent initiation
into stages of drug use: A developmental analysis. In D. B. Kandel (Ed.), Longitudinal
research on drug use (pp. 73-99). Washington, DC: Hemisphere.
Klein, D. M., Jorgensen, S. R., & Miller, B. C. (1978). Research methods and developmental
reciprocity in families. In R. M. Lerner & G. B. Spanier (Ed.), Child influences on marital
and family interaction: A life-span perspective (pp. 107-135). New York: Academic Press.
Lamborn, S. D., Mounts, N. S., Steinberg, L., & Dornbusch, S. M. (1991). Patterns of
competence and adjustment from authoritative, authoritarian, indulgent and neglectful
families. Child Development, 62, 1049-1065.
Lytton, H. (1990). Child and parental effects in boys' conduct disorder: A reinterpretation.
Developmental Psychology, 26, 683-697.
Maccoby, E. E., & Martin, J. A. (1983). Socialization in the context of the family: Parent-child
interaction. In P. H. Mussen & E. M. Hetherington (Eds.), Handbook of child psychology
(4th ed.): Vol. iv. Socialization, personality and social development. New York: Wiley.
Margulies, R. Z., Kessler, R. C., & Kandel, D. B. (1977). A longitudinal study of onset of
drinking among high-school students. Journal of Studies on Alcohol, 38, 897-912.
McCord, W., McCord, J., & Howard, A. (1961). Familial correlates of aggression in
non-delinquent male children. Journal of Abnormal and Social Psychology, 62, 79-93.
Mercer, G., & Kohn, P. (1980). Child-rearing factors, authoritarianism, drug use attitudes,
and adolescent drug use: A model. Journal of Genetic Psychology, 136, 159-171.
Minuchin, S. (1974). Family and family therapy. New York: Tavistock.
Monroe, S. M. (1983). Social support and disorder: Toward an untangling of cause and effect.
American Journal of Community Psycholog),, 11, 81-97.
National Institute of Drug Abuse. (1989). Highlights of the 1988 National Household Survey
on Drug Abuse, Rockville, MD: National Institute on Drug Abuse.
Parenting Styles and Adolescent Problem Behavior 629

Pandina, R. J., & Schuele, J. A. (1983). Psychosocial correlates of alcohol and drug use of
adolescent students and adolescents in treatment. Journal of Studies on Alcohol, 44,
950-973.
Patterson, G. R., DeBaryshe, B. D., & Ramsey, E. (1989). A developmental perspective on
antisocial behavior. American Psychologist, 44, 329-335.
Patterson, G. R., & Reid, J. B. (1984). Social interactional processes within the family: The
study of the moment-by-moment family transactions in which human social development
is imbedded. Journal of Applied Developmental Psychology, 5, 237-262.
Patterson, G. R., & Stouthamer-Loeber, M. (1984). The correlation of family management
practices and delinquency. Child Development, 55, 1299-1307.
Pedhazur, E. J. (1982). Multiple regression in behavioral research: Explanation and prediction.
San Francisco: Holt, Rinehart and Winston.
Rogosa, D. (1988). Myths about longitudinal research. In K. W. Schaie, R. T. Campbell, W.
Meredith, & S. C. Rawlings (Eds.), Methodological issues in aging research (pp. 171-209).
New York: Springer.
Rogosa, D. R., & Willett, J. B. (1985). Understanding correlates of change by modeling
individual differences in growth. Psychometrika, 50, 203-228.
Rollins, B. C., & Thomas, D. L. (1979). Parental support, power, and control techniques in
the socialization of children. In W. R. Burr, R. Hill, F. I. Nye, & I. L. Reiss (Eds.),
Contemporary theories about the family (Vol. 1, pp. 317-364). New York: The Free Press.
Schaefer, E. S. (1965). Children's reports of parental behavior: An inventory. Child
Development, 36, 413-424.
Sher, K. J., Walitzer, K. S., Wood, P., & Brent, E. E. (1991). Characteristics of children of
alcoholics: Putative risk factors, substance use and abuse, and psychopathology. Journal
of Abnormal Psychology, 100, 427-448.
Smart, L. S., Chibucos, T. R., and Didier, L. A. (1990). Adolescent substance use and
perceived family functioning. Journal of Family Issues, 11, 208-227.
Weissman, M. M., Wickramaratne, P., Warner, V., John, K., Pursoff, B. A., Merikangas, K.
R., & Gammon, G. D. (1987). Assessing psychiatric disorders in children. Arch&es of
General Psychiatry, 44, 747-753.
West, D. J., & Farrington, D. P. (1973). Who becomes delinquent? Nw York: Crane, Russak
& Co.
Wolchik, S. A., Beals, J., & Sandier, I. N. (1989). Mapping children's support networks:
Conceptual and methodological issues. In D. Belle (Ed.), Children's social networks and
social supports. New York: Wiley.