Social Competence in Children of Alcoholic Parents Over Time

Abstract

In the current study, the authors tested the hypothesis that children of alcoholic parents (COAs) show deficits in social competence that begin in early childhood and escalate through middle adolescence. Teachers, parents, and children reported on the social competence of COAs and matched controls in a community sample assessed from ages 6 to 15. Hierarchical linear growth models revealed different patterns of change in social competence across development as a function of the reporter of various indicators of competence. Moreover, female COAs showed deficits in social competence in early childhood that receded in adolescence and that varied across subtypes of parent alcoholism. Implications of these findings for understanding the development of social competence in children, and at-risk children in particular, are discussed.

Full text

Social Competence in Children of Alcoholic Parents Over Time
Andrea M. Hussong
University of North Carolina at Chapel Hill
Robert A. Zucker and Maria M. Wong
University of Michigan
Hiram E. Fitzgerald
Michigan State University
Leon I. Puttler
University of Michigan
In the current study, the authors tested the hypothesis that children of alcoholic parents (COAs) show
deficits in social competence that begin in early childhood and escalate through middle adolescence.
Teachers, parents, and children reported on the social competence of COAs and matched controls in a
community sample assessed from ages 6 to 15. Hierarchical linear growth models revealed different
patterns of change in social competence across development as a function of the reporter of various
indicators of competence. Moreover, female COAs showed deficits in social competence in early
childhood that receded in adolescence and that varied across subtypes of parent alcoholism. Implications
of these findings for understanding the development of social competence in children, and at-risk children
in particular, are discussed.
Keywords: parent alcoholism, social competence, social development, peer relationship, at-risk youth
The development of social competence is a fundamental aspect
of children’s adjustment. Children with friends have greater aca-
demic success and are less likely to be aggressive, lonely, and
depressed compared with children without friends (Parker &
Asher, 1987, 1993; Rockhill, 1995). Moreover, both lower peer
acceptance and greater peer rejection in preadolescents have been
linked to lower feelings of self-worth and psychopathology in
adulthood (Bagwell, Newcomb, & Bukowski, 1998), demonstrat-
ing the potential long-term consequences of social competence.
Less understood than the impact of social competence, however,
is the development of social competence itself. Theories of social
competence posit that individual children change over time in their
social functioning, suggesting a pattern of intraindividual variabil-
ity, and that children vary from one another in the extent to which
their social competence changes over time, suggesting a pattern of
interindividual variability (e.g., Berndt & Burgy, 1996; Eccles et
al., 1989). Unfortunately, previous studies primarily use cross-
sectional or short-term (i.e., two time point) longitudinal designs
that do not permit distinctions between these two types of change
over time. Greater use of high-risk, prospective research designs
with multiple assessment periods overcomes this limitation (Cur-
ran, Bollen, Paxton, Kirby, & Chen, 2002; Kalverboer, 1988;
Ledingham, 1990). High-risk designs allow us to better model and
understand low-base rate behaviors, such as child maladjustment
and social deficits, and multiple assessment designs can disentan-
gle intraindividual and interindividual change by using recent
statistical advances (Raudenbush, 2001).
One group of children who show an elevated risk for a broad
range of problems is children of alcoholic parents (COAs; West &
Prinz, 1987). Research concerning COAs has largely focused on
risk for maladaptive behaviors in this population, indicating that
COAs are more likely to show internalizing symptoms, aggression,
and substance abuse than their peers (Chassin, Rogosch, & Bar-
rera, 1991; Puttler, Zucker, Fitzgerald, & Bingham, 1998; Sher,
1991). Little attention has been given to whether COAs show
deficits in adaptive behavior in addition to such excesses in mal-
adaptive behavior. To address these issues, the current study
examined individual trajectories of social competence spanning
the ages of 6 to 15 in a high-risk sample of COAs and matched
controls.
The Development of Social Competence
Competence is typically defined as a global construct in which
“the competent individual is one who is able to make use of
environmental and personal resources to achieve a good develop-
mental outcome” (Waters & Sroufe, 1983, p. 81). Although studies
vary widely in operationalizations of social competence, making
and maintaining friendships, fostering popularity and social accep-
tance, and developing skills in relating to peers are consistent
markers of the construct (Berndt & Burgy, 1996; Rose-Krasnor,
1997). Moreover, these three dimensions are often intertwined in
Andrea M. Hussong, Department of Psychology, University of North
Carolina at Chapel Hill; Robert A. Zucker, Maria M. Wong, and Leon I.
Puttler, Department of Psychiatry and Addictions Research Center, Uni-
versity of Michigan; Hiram E. Fitzgerald, Department of Psychology,
Michigan State University.
Andrea M. Hussong received support from National Institute on Drug
Abuse Grants DA12912 and DA15398 in the writing of this article. The
work was also supported by National Institute on Alcohol Abuse and
Alcoholism Grant R37 AA 07065 to Robert A. Zucker and Hiram E.
Fitzgerald. We thank Susan Refior, study director of field operations, for
her sustained effort and skill in sustaining this study’s viability over such
a long period of time. We also thank the participating families for their
willingness to engage with us in what is an essentially altruistic activity,
over so many years.
Correspondence concerning this article should be addressed to Andrea
M. Hussong, Department of Psychology, University of North Carolina at
Chapel Hill, College of Arts and Sciences, Campus Box 3270, Davie Hall,
Chapel Hill, NC 27599-3270. E-mail: hussong@unc.edu
Developmental Psychology Copyright 2005 by the American Psychological Association
2005, Vol. 41, No. 5, 747–759 0012-1649/05/$12.00 DOI: 10.1037/0012-1649.41.5.747
747

assessments of overall social competence (Bracken, 1992; Harter,
1982; Marsh & O’Neill, 1984).
Such measures have been used to assess social competence across
a variety of reporters, including children, parents, teachers, and peers,
though few studies actually test patterns of intraindividual variability,
or trajectories, of social competence over time. Cole, Maxwell, et al.
(2001) examined change over time in self-reported social competence
among 3rd to 11th graders as assessed by Harter’s (1982, 1985) social
concept measure. In this study, normative trajectories of self-reported
social competence increased until 7th grade, when they began to
plateau. Gender differences emerged in social self-concept in high
school, with boys showing greater social competence. Chan, Ramey,
Ramey, and Schmitt (2000) also examined change over time in social
competence through parent and teacher reports in a sample of kin-
dergarteners who completed four annual assessments. Different pat-
terns of intraindividual change emerged as a function of observer,
such that parent reports increased more from kindergarten to 2nd
grade than from 2nd grade to 3rd grade, evidencing a curvilinear
increase over time. However, teacher reports showed a linear decrease
in social competence over time. Gender differences in this study
indicated that girls showed greater social competence than did boys in
teacher reports but not in parent reports. Together, these findings
suggest that both intraindividual patterns of change in social compe-
tence over time as well as gender differences in social competence are
likely to vary by reporter, which presumably in part reflects the social
presentation differences found in these various contexts.
Previous studies provide support for interindividual stability in
social competence, such that those children who are most compe-
tent at one age are likely to be the most competent at another
(Eccles et al., 1989). For example, Masten et al. (1995) showed
prospective consistency (with a stability of .45) in a latent factor of
social competence composed of child self-reports and peer nomi-
nations in late childhood and of child and parent interviews in late
adolescence. Parent and teacher reports, however, may show a
different trend. On the basis of these findings, we used a sample of
children aged 6 to 15 to test gender differences in both intraindi-
vidual and interindividual change in various indicators of social
competence assessed over time through multiple reporters (Cole,
Jacquez, & Maschman, 2001; Eccles et al., 1989; Wigfield, Eccles,
Maciver, Reuman, & Midgley, 1991).
COAs
Little research has directly examined the social competence of
COAs, although peer relationships are thought to play a key role in
many theories of psychopathology to which this population is
prone (Coie, Belding, & Underwood, 1988; Dishion, Capaldi,
Spracklen, & Li, 1995; Oetting & Beauvais, 1987). Early deficits
in social skills, such as having few friends and less group partic-
ipation, are associated with later problems in acceptance by non-
deviant peers and forming friendships with them (Parker, Rubin,
Price, & DeRosier, 1995). As such, social deficits at school entry
may be an early warning sign that intervention is needed (Ladd,
1990).
Little is known about social competence in COAs during child-
hood and adolescence. On the basis of the retrospective reports of
college students, Segrin and Menees (1996) found no differences
in the social skills (i.e., communication– expressivity and sensitiv-
ity) of COAs and their peers in early childhood, although problems
with retrospective reports and child reports of parent alcoholism
weakened this study. More broadly, social skill deficits have been
found in children of parents with psychiatric problems. For exam-
ple, Larsson, Knutsson-Medin, Sundelin, and Trost von Werder
(2000) reported greater problems in peer and sibling relationships
among children of parents admitted to a Swedish psychiatric unit
compared with similar-aged schoolchildren. Chan et al. (2000)
also found social skills deficits, as reported by teachers, in children
of parents with lower education and income, which are often
associated with parent alcoholism, although no differences in
social skills were noted in parent reports.
Although a coherent picture has not emerged from the impov-
erished literature to date, we postulate that differences in social
competence may vary within COAs given the substantial hetero-
geneity of stress and conflict, as well as genetic liability, within
these families (Puttler et al., 1998; Wong, Zucker, Fitzgerald, &
Puttler, 1999). We expected those COAs with two, rather than one,
alcoholic parent may not benefit from the potential compensating
protection offered by a nonalcoholic parent and thus would show
greater social competence deficits (as suggested by Werner, 1986,
though also see Curran & Chassin, 1996). In addition, children
whose parents have experienced alcoholism more recently, as
opposed to those whose parents were recovered, may also experi-
ence more acute chaos, stress, and unpredictability in the home
environment, also increasing risk for social competence deficits
(Pillow, Barrera, & Chassin, 1998). Finally, we examined social
competence deficits in children with alcoholic mothers versus
fathers. Given alcoholic parents’ frequent role as primary caretak-
ers and the impact of assortative mating, in which a family with an
alcoholic mother often also has an alcoholic father, children of
alcoholic mothers may be expected to show greater social compe-
tence deficits. However, paternal alcoholism has also been asso-
ciated with greater risk for behavioral dysregulation and difficult
temperament, perhaps related to a genetic liability, each of which
may increase risk for social competence deficits as well (Tarter et
al., 1999). As such, we explored the relation between maternal and
paternal alcoholism and children’s social competence over time.
In sum, we examined interindividual and intraindividual
changes in children’s trajectories of social competence over time
as well as how these trajectories relate to parent alcoholism in a
heterogeneous sample of COAs and matched controls. Given the
well-documented heterogeneity among alcoholic adults, to better
understand whether such early social skill deficits are equally
evident across this risk group, we tested whether evidence for
social deficits varied as a function of the number of alcoholic
parents in the family, the recency of the parent’s alcoholism, and
the gender of the alcoholic parent. Moreover, we considered the
potential for gender differences, evident in normative samples,
within each of these hypotheses. Although multiple reporters of
social competence were examined, reporters varied in the indica-
tors of social competence assessed. As such, cross-reporter com-
parisons are interpreted with respect to potential differences as a
function either of reporter or of the variation in subcomponents of
social competence across measures assessed.
Method
Participants
Families in the current study are participants in an ongoing, prospective,
multiwave study that is tracking a community sample of families with high
748
HUSSONG, ZUCKER, WONG, FITZGERALD, AND PUTTLER

levels of alcohol use disorder, along with a community contrast sample of
families drawn from the same neighborhoods who do not show the high-
substance abuse profile (Zucker & Fitzgerald, 1991; Zucker et al., 2000).
Recruitment followed a high-risk design in which children of alcoholic and
nonalcoholic parents were targeted through a rolling recruitment proce-
dure. The sample in the current study can be characterized by three cohort
assessments (see Figure 1). In the first cohort, alcoholic fathers were
identified through district courts and invited to participate if they had a
male biological child aged 3–5, were Caucasian, lived with the child, and
were coupled with the child’s biological mother (whose substance abuse
was not constrained). Exclusion criteria included the presence of fetal
alcohol syndrome in the target child. (Fetal alcohol syndrome was assessed
by master-level clinical psychologists trained in early child assessment
who had experience using the Fetal Alcohol Study Group guidelines [Sokol
& Claren, 1989] to evaluate facial dysmorphology, growth retardation, etc.
to make the classification. For more details see Bingham, Fitzgerald,
Fitzgerald, & Zucker, 1996, and Noll, Zucker, Fitzgerald, & Curtis, 1992).
A second subset of alcoholic fathers was uncovered during neighborhood
canvassing for contrast families (see below). This group also needed to
have at least 1 son between 3 and 5 years of age and have a father coupled
with the boy’s biological mother, whose drug involvement was also free to
vary. Parents in the contrast group were recruited through canvassing in the
same neighborhoods where the alcoholic families lived; parents were
required to show an absence of lifetime history of substance abuse or
dependence and were matched to alcoholic families on the basis of neigh-
borhood residence and child age. A total of 335 boys (from 335 families)
participated in this first cohort assessment when the male target child was
age 3–5 years (Wave 1). At the time these data were analyzed, families had
been assessed over four waves of data collection, separated by 3-year
intervals. A total of 231, 249, and 248 boys had reports on their functioning
available at Waves 2– 4, respectively, yielding an overall participation rate
of 88% for boys with at least two waves of data in the sample (see Zucker
et al., 2000, for a detailed description of sample criteria and recruitment
procedures).
A second cohort involved girls in the Cohort 1 families who were
recruited when Cohort 1 boys were at Wave 2. Because Cohort 1 inclusion
criteria involved having families with at least 1 male child with no
restriction on the presence of other children, fewer families had female
children. To compensate for these target families having fewer girls than
boys, we used a broader age range to recruit girls, and girls were enrolled
if they were aged 3–11, with those aged 3–5 receiving the Wave 1 battery,
those aged 6 –8 receiving the Wave 2 battery, and those aged 9 –11
receiving the Wave 3 battery. Similarly, the third cohort contained all
additional siblings of the male target child in Cohort 1 who were aged 3–11
at the time of data collection, with assessment batteries structured by age
as for Cohort 1. The siblings in Cohorts 2 and 3 were reassessed in all
subsequent waves of data collection and received measures that paralleled
the male target children in Cohort 1 on the basis of age of assessment.
Because children in Cohorts 2 and 3 were recruited later in time and could
enter the study at older ages, fewer waves of data collection are available
for these participants by design. A total of 149 girls (from 149 families)
comprised Cohort 2 and an additional 69 siblings (from 62 families)
comprised Cohort 3.
Across all three cohorts, 553 children from 335 families were eligible for
inclusion in the current analyses. However, to ensure similar samples for
analyses across the multiple reporters of social competence, we constrained
Figure 1. Sample description. COAs ⫽ children of alcoholic parents.
749
SOCIAL COMPETENCE IN CHILDREN

the sample to those children with parent alcoholism data and at least one
observation (i.e., wave of assessment) on each of the three social compe-
tence outcome variables. The resulting analysis sample included 373 chil-
dren from 265 families (32 female controls, 78 female COAs, 63 male
controls, and 200 male COAs; aged 2–15 across waves; participants were
predominantly [98%] Caucasian; parent education ranged from 8% of
families in which neither parent had a high school education to 24% of
families in which at least one parent was a college graduate; no information
on disability status or sexual orientation was available). Although all
children were included in analyses for the three social competence out-
comes, the pattern of missing data within the sample varied by outcome,
such that a different number of repeated observations were available for
each outcome. Specifically, 203 children had all three assessments (at ages
6 – 8, 9 –11, and 12–15) of parent-reported social competence available,
113 had two waves, and 57 had a single wave of data. For teacher-reported
social competence, 78, 168, and 129 had three, two, and one wave of data
available, respectively, and 116, 189, and 68 had three, two, and one wave
of self-reported social competence data, respectively. Because of this data
structure, we used missing data techniques in the statistical analyses that
are capable of incorporating diverse patterns of missingness without pro-
ducing biased findings (see Results).
Two sets of attrition analyses were conducted. We tested whether the
180 children excluded from all analyses because of missing reports of all
social competence variables differed on child gender and parent alcoholism
from the 373 retained for analyses. No differences were found on parent
alcoholism status,
␹
2
(1, N ⫽ 553) ⫽ 0.12, p ⫽ .73, nor on participant
gender,
␹
2
(1, N ⫽ 553) ⫽ 1.98, p ⫽ .16. We also estimated the effects of
missing data, whether because of attrition or design, by calculating the
number of missing variables out of all central variables (the three social
competence variables at Waves 2– 4, COA, and child’s gender) for each of
the 373 participants included in analyses. Correlations between the number
of missing variables and the nine continuous outcome measures (where
present) revealed no significant associations. T tests showed that COAs
were more likely to have missing data than their peers, t(371) ⫽⫺2.18,
p ⬍ .05, and, as expected by design, girls had more missing data than boys,
t(371) ⫽ 4.24, p ⬍ .001. These missing data effects suggest some potential
bias if listwise deletion procedures were used in analyses; thus, missing
data were addressed through the use of appropriate analytic techniques that
retained all 373 participants in statistical procedures (see Results).
Procedure
Each family completed a primarily in-home assessment conducted by
trained staff who were blind to family diagnostic status (Wong et al., 1999).
Although protocol length varied by wave of assessment, parent assess-
ments typically involved 9 –10 hr of data collection, and child assessments
were typically 7 hr, each spread over seven testing sessions. Families were
compensated $300 for their involvement if the assessment was carried out
on a 1-child family and $375 if 2 children were involved. A variety of
age-appropriate tasks (e.g., questionnaires, semistructured interviews, and
interactive tasks) were administered. As special arrangements were made
to assess families who had relocated, no families were lost for this reason.
During the age 6 – 8 assessment, parents were asked to provide the name of
the child’s school, the child’s grade, and the name of the child’s teacher.
Parents were also asked to sign a release form. This form was mailed to the
child’s teacher along with a letter of introduction, three surveys about the
child, and a $10 check as compensation. Similar teacher data were obtained
in assessments for ages 9–11 and 12–15. At the point when the multiple-
teacher class schedule began, the child was asked to identify the teacher
with whom he or she spent the most time (ruling out home room and
physical education class).
Measures
Parent alcoholism (alcohol use disorder). Parental alcohol use disor-
der at Wave 1 was assessed by the Diagnostic Interview Schedule—III
(Robins, Helzer, Croughan, & Ratcliff, 1980), the Short Michigan Alcohol
Screening Test (Selzer, Vinokur, & van Rooijan, 1975), and the Drinking
and Drug History Questionnaire (Zucker, Fitzgerald, & Noll, 1990). On the
basis of information collected by all three instruments, a diagnosis of
alcohol use disorder (lifetime as well as past 3 years) was made by a trained
clinician using Diagnostic and Statistical Manual of Mental Disorders (4th
ed.; American Psychiatric Association, 1994) criteria. The availability of
three sources of information collected over three different sessions, sepa-
rated in some instances by as much as several months, served as an
across-method validity check on respondent replies. Given the volume of
material collected as well as the spacing between sessions, it is unlikely
respondents would recall their specific replies. In cases of discrepant
information, the data represented by the majority of information sources
were used in establishing the diagnosis. Interrater reliability for the diag-
nosis was excellent (
␬
⫽ .81).
In the current analyses, children were coded as having an alcoholic
parent if either parent met lifetime criteria for alcohol abuse or dependence
at the age 6 –8 assessment period, which is the first assessment period for
social competence outcomes. (Lifetime parent alcoholism diagnoses were
combined with subsequently assessed diagnoses in the past 3 years to
determine whether either parent had ever met criteria for an alcohol abuse
or dependence disorder when a given child was assessed at ages 6– 8.)
Because lifetime parent alcoholism diagnoses are a function of time and
may differ for siblings within the same family depending on their ages, all
indicators of parent alcoholism were treated as individual-, rather than
family-level, variables in the current analyses.
Three indices of parent alcoholism coded for heterogeneity within this
risk indicator. On the basis of the lifetime indicator of parent alcoholism,
one set of indicators coded separately for maternal (n ⫽ 128 of 373
participants) and paternal (n ⫽ 267 of 373 participants) alcoholism. A
second indicator was formed by a set of dummy variables that distin-
guished between families without an alcoholic parent (n ⫽ 95), with a
recently alcoholic parent (within the past 3 years prior to the age 6 –8
assessment; n ⫽ 206), and with a recovered alcoholic parent (n ⫽ 72). A
third indicator summed the number of alcoholic parents present in a family
(95 with zero alcoholic parents, 161 with one alcoholic parent, and 117
with two alcoholic parents).
Social competence. Reports of social competence were assessed by
self-, parent, and teacher report. See Table 1 for descriptive statistics of
outcome measures by child age.
Children completed Harter’s Self-Perception Profile (Harter, 1985,
1988) measure in Waves 2– 4. In Waves 2–3, the children were aged 6 –11
and the child version of this questionnaire was administered, whereas in
Wave 4 children were 12–15 and the adolescent version was used. To
create a consistent scale across waves for use in longitudinal analyses, we
retained only similar items across the two versions to avoid measurement
differences over time (similar to Cole, Jacquez, & Maschman, 2001;
sample items assess finding it hard to make friends and having lots of
friends). The response scale for these items followed the original format,
ranging from 1– 4. Retained items resulted in a five-item scale assessing
self-reported social competence at each of the three waves (
␣
⫽ .65, .72,
and .81 for boys and .61, .74, and .79 for girls at Waves 1–3, respectively).
The teacher report measure was created using two items from the School
Performance Questionnaire (Piejak, Fitzgerald, Zucker, & von Eye, 2004)
and nine items concerning classroom activity from the Revised Class Play
Questionnaire (RCPQ; Fitzgerald et al., 1993; Noll et al., 1992). Items from
the School Performance Questionnaire were chosen to capture teacher
report of social competence (i.e., two items rating social development and
child likability, respectively), whereas those from the RCPQ were chosen
on the basis of previous analyses by Piejak et al. (2004) to construct a
social competence measure from this instrument (sample items assess ease
of making friends and having many friends). Both instruments shared a
5-point response scale. Exploratory factor analyses of these 11 items most
strongly supported a two-factor solution as indicated by incremental vari-
ance indicators and the principle of simple structure (Loehlin, 1992). Using
750
HUSSONG, ZUCKER, WONG, FITZGERALD, AND PUTTLER

promax rotation, we found similar solutions for each of three waves of data
analyzed. The first factor contained seven items from the School Perfor-
mance Questionnaire and RCPQ that reflected having lots of friends and
being popular, and the second factor contained items from the RCPQ that
reflected the development of leadership skills. For purposes of the current
study, a single scale based on the average of seven items on the first factor
was retained as the teacher report variable for social competence (
␣
⫽ .88,
.89, and .88 for boys and .83, .90, and .90 for girls at Waves 1–3,
respectively).
Parent report items of social competence were derived from the com-
petency subscales of the Child Behavior Checklist (Achenbach, 1991).
Four items assessed through mother and father report of the number of
close friends, time spent with friends, and getting along with siblings and
other kids formed this scale. Available parent reports were combined into
a single measure (across-reporter correlations ranged from .41 to .56, p ⬍
.001, over waves). All items were rated on a 3-point scale, except for the
number of friends that was collapsed to three categories (0 –1, 2–3, 4 or
more) as guided by the observed distribution. These eight items were
averaged to form a parent report scale for social competence (
␣
⫽ .62, .54,
and .77 for boys and .55, .67, and .64 for girls at Waves 1–3, respectively).
Results
We used a series of hierarchical linear models (HLM; Rauden-
bush & Bryk, 2002) to test whether heterogeneity in parent alco-
holism predicted three indicators of social competence. These
models allowed us to account for the three levels of structure in
these data (i.e., repeated time points nested within individuals and
siblings nested within families) while examining longitudinal re-
lations between parent alcoholism and trajectories of social com-
petence over time. Another benefit of the HLM approach is that the
empirical Bayes estimation procedures permit inclusion of partic-
ipants with incomplete data over time. The approximately normal
distribution of our outcome measures permitted linear estimation
methods, although robust test statistics are reported here. Outlier
analyses revealed no significant influential cases in the data.
1
Trajectories of Social Competence
Prior to testing relations between various predictors and trajec-
tories of social competence over time, we conducted separate
unconditional growth models to examine the trajectories underly-
ing each of the three social competence outcomes. Because of both
the differences in the assessment schedule as a function of gender
and the various patterns of missingness in the data, we elected to
analyze the data as a cohort sequential design, with age of the child
rather than wave of assessment as a marker of time, such that
measures of social competence began at age 6 and ended at age 15.
Means for each social competence index as a function of age are
presented in Table 1. As these means indicate, we would expect
our unconditional HLM analyses to show an increase over time in
self-reported social competence, a slight decrease over time in
teacher reports, and little change over time in parent reports.
The unconditional linear growth model comprises four key
parameters that define the average or group trajectory for a given
outcome as well as individual variability in the set of trajectories
constituting the sample. The average trajectory is defined in terms
of a mean intercept and slope, modeled as fixed effects. Interin-
dividual variability across participants’ trajectories is then re-
flected in variation among participants’ trajectories from the mean
intercept and mean slope, modeled as random effects. For the
current analyses, we coded the fixed intercept to reflect the average
level of social competence for 6-year-olds in the sample and the
fixed slope to reflect the amount of change in social competence
per year over observed assessments from age 6 to 15. As such, the
random intercept reflects the extent to which participants differ
from one another in their social competence at age 6, and the
random slope reflects the extent to which participants differ in the
rate of change characterizing their social competence over time.
Because the current study involves a three-level design, in which
siblings are nested within families, we also modeled intrafamilial
variability in the repeated observations of social competence
1
To examine the influence of potential outliers, we performed individ
-
ual regression analyses for each participant to derive an estimated intercept
and slope for his or her trajectory. On the basis of these trajectory
parameters, we estimated time-specific measures for each social compe-
tence indicator. Potential outliers were identified as (a) participants with
large discrepancies between observed and estimated time-specific mea-
sures for a given social competence indicator and (b) large estimated
intercept or slope factors with respect to the group mean. Unconditional
HLM analyses were reestimated dropping these potential outliers and
compared with results when these cases were included. No substantive
changes in trajectory parameters resulted from dropping these cases, indi-
cating that they were not influential. As such, all cases were retained in
analyses presented here.
Table 1
Descriptive Statistics for Social Competence Indicators
Age
Self-report (range ⫽ 1–4) Parent report (range ⫽ 1–3) Teacher report (range ⫽ 1–5)
NMSD
␣
NMSD
␣
NMSD
␣
6 46 3.04 0.63 .40 49 2.21 0.26 .48 35 3.48 0.72 .84
7 87 3.00 0.76 .65 87 2.32 0.32 .65 65 3.56 0.75 .81
8 88 2.94 0.77 .66 89 2.32 0.26 .47 62 3.29 0.76 .80
9 118 3.06 0.77 .74 116 2.32 0.33 .69 95 3.53 0.79 .90
10 111 3.12 0.67 .70 111 2.30 0.29 .57 80 3.49 0.76 .89
11 100 3.16 0.65 .73 98 2.31 0.32 .59 84 3.59 0.80 .90
12 86 3.33 0.67 .79 107 2.36 0.34 .72 92 3.40 0.92 .93
13 79 3.33 0.55 .75 116 2.32 0.35 .68 89 3.44 0.67 .85
14 55 3.40 0.52 .79 78 2.33 0.29 .55 60 3.38 0.87 .91
15 24 3.27 0.66 .88 41 2.32 0.38 .81 33 3.22 0.85 .88
751
SOCIAL COMPETENCE IN CHILDREN

through a random family-level intercept.
2
Results for the three
models are reported in Table 2.
For self-reported social competence, significant fixed-effect
components of the trajectories indicated that average social com-
petence at age 6 was 2.90 (i.e., the intercept), with a significantly
increasing linear trend of .06 per year (i.e., the slope), such that
average levels of self-reported social competence increased from
2.90 at age 6 to 3.44 over the course of the 9-year interval. A
significant random effect in the intercept was also found, indicat-
ing that participants differed from one another in their self-
reported social competence at age 6. Moreover, a marginally
significant random effect in the slope indicated some confidence
that individuals varied from one another in the extent to which
their social self-competence scores change over time.
For the parent report model, the significant fixed intercept
indicated an average of 2.30 for social competence at age 6, but a
nonsignificant fixed slope indicated no change or a flat average
trajectory of social competence over time. Significant random
effects, however, showed reliable individual differences in both
the levels of social competence at age 6 (as reflected in the random
intercept term) and in the rate of change in social competence over
time (as reflected in the random slope term). Thus, although the
average group trajectory for parent-reported social competence is
a flat line reflecting a consistent level of 2.30, individual partici-
pants had trajectories that reliably varied from one another in both
their intercepts and slopes.
Finally, in the teacher report model, a significant fixed effect of
the intercept and a nonsignificant fixed effect of the slope were
found. This average trajectory was characterized by a 3.49 level of
social competence at age 6 and by no change in social competence
over time. Significant individual differences were found in the
random intercept for this model, such that participants reliably
differed from one another in their age 6 levels of teacher-reported
social competence but not in the random effect for the slope
parameters.
Parent Alcoholism and Social Competence
Three models for each social competence outcome, or a total of
nine models, were conducted for the effects of heterogeneity in
parent alcoholism on trajectories of social competence. For each
model, a series of individual (or Level 2) predictors were added to
the model that included indicators of parent alcoholism, child
gender, and the interaction between the two. These variables were
included in predictions of the random intercept and the random
slope of the social competence trajectories.
3
This model allowed us
to examine whether boys and COAs show early decrements in
social competence (in the prediction of the random intercept term)
and escalating patterns of social competence problems over time
(through their interaction with age in predicting social compe-
tence). These models were tested hierarchically, such that an initial
2
Intrafamilial variation is modeled in the social competence out
-
comes to correct standard errors for nesting of siblings within families
in the study design. However, intrafamilial variation is not central to the
questions of the current study and thus will not be discussed further.
Small numbers of children within families (ranging from 1 to 4) limit
our ability to interpret the intrafamilial variation effects meaningfully.
In HLM reduced-model notation, the final models were estimated as follows:
Y
tik
⫽ (
␥
000
) ⫹ (
␤
10k
* AGE
tik
) ⫹ (U
00k
⫹ r
0ik
⫹ r
1ik
⫹ e
tik
), where Y
represents the outcome, t represents time, i represents individual children,
k represents family,
␥
000
represents the fixed intercept,
␤
10k
represents the
fixed slope, U
00k
represents the random effect for intrafamilial variation in
social competence, r
0ik
represents the random effect for the intercept, r
1ik
represents the random effect for the slope, and e represents residual variance.
3
Although some researchers suggest that nonsignificant variance in the
random slopes precludes the inclusion of predictors of the random slopes,
more recent suggestions have indicated that such predictors are indeed
appropriate and simply reflect the interaction of a given predictor with time
that might be modeled as either a fixed or random effect. Following the
more recent suggestion, we thus modeled all predictors in our models as
both main effects and as interactions with time (or predictors of the random
slope) across all three social competence models. In HLM reduced-model
notation, the final models were estimated as follows:
Y
tik
⫽ (
␥
000
) ⫹ (
␤
10k
* AGE
tik
) ⫹ (
␤
o1k
* ALC ⫹
␤
02k
* SEX ⫹
␤
03k
* ALC * SEX ⫹
␤
11k
* ALC * AGE ⫹
␤
12k
* SEX * AGE
⫹
␤
13k
* ALC * SEX * AGE) ⫹ (U
00k
⫹ r
0ik
⫹ r
1 ik
⫹ e
tik
)
where Y represents the outcome, t represents time, i represents individual
children, k represents family,
␥
000
represents the fixed intercept,
␤
10k
represents the fixed slope, all other
␤
pqk
represent the main and interactive
effects of individual predictors on social competence, ALC represents
parent alcoholism, U
00k
represents the random effect for intrafamilial vari-
ation in social competence, r
0ik
represents the random effect for the inter-
cept, r
1ik
represents the random effect for the slope, and e represents
residual variance.
Table 2
Social Competence Unconditional Trajectories
Effect
Self-report
(range ⫽ 1–4)
Parent report
(range ⫽ 1–3)
Teacher report
(range ⫽ 1–5)
Coefficient Test statistic Coefficient Test statistic Coefficient Test statistic
Fixed
Intercept 2.90 57.90*** 2.30 114.17*** 3.49 59.31***
Slope 0.06 6.47*** 0.00 0.71 ⫺0.01 ⫺1.55
Random
Intercept 0.39 384.53*** 0.14 377.21*** 0.33 259.81***
Slope 0.02 337.21
†
0.02 372.75* 0.02 244.91
Note. Coefficients and test statistics are betas and z tests for fixed effects and standard errors and Wald’s
chi-square for random effects. All fixed effects are reported with robust estimation.
†
p ⬍ .10. * p ⬍ .05. *** p ⬍ .001.
752
HUSSONG, ZUCKER, WONG, FITZGERALD, AND PUTTLER

model tested the main effects of parent alcoholism and child
gender on the social competence trajectories and a second model
tested the additional effect of the interaction term. Three sets of
predictors were used to test for parent alcoholism effects where
heterogeneity in this risk factor was modeled first as a function of
gender of the alcoholic parent (i.e., the unique effects of having an
alcoholic father and/or an alcoholic mother), the recency of parent
alcoholism, and the number of alcoholic parents in the family.
Results from these models are reported in Table 3.
In the self-reported social competence model, the interaction
between having an alcoholic father and the child’s gender was a
significant predictor of initial levels of social competence (i.e.,
intercepts). No other effects of paternal or maternal alcoholism
were found. To better understand this interaction, we probed mean
trajectories of self-reported social competence as a function of
child’s gender and parent alcoholism (techniques for testing such
interactions are described in detail in Curran, Bauer, & Wil-
loughby, in press). Girls with alcoholic fathers self-reported less
social competence at age 6 compared with girls without alcoholic
fathers (
␤
⫽⫺.48), t(367) ⫽⫺2.27, p ⬍ .05, whereas boys with
and without alcoholic fathers did not differ from one another in
their social competence at age 6 (
␤
⫽ .08), t(367) ⫽ 0.56, ns.
An interaction between child’s gender and the recency of parent
alcoholism predicting the intercept of self-reported social compe-
tence trajectories was also found. Though no differences in self-
reported social competence were found for those without an alco-
Table 3
Parental Alcoholism and Trajectories of Social Competence
Model
Self-report Parent report Teacher report
Coefficient t Coefficient t Coefficient t
Predicting random intercept
1
Sex 0.15 1.34 ⫺0.06 ⫺1.57 0.19 1.73†
Mom alcoholic ⫺0.01 ⫺0.12 ⫺0.01 ⫺0.13 ⫺0.21 ⫺1.52
Dad alcoholic ⫺0.07 ⫺0.57 ⫺0.02 ⫺0.45 ⫺0.09 ⫺0.69
Sex ⫻ Mom Alcoholic ⫺0.02 ⫺0.07 ⫺0.12 ⫺1.28 0.15 0.61
Sex ⫻ Dad Alcoholic ⫺0.57 ⫺2.36* 0.04 0.47 ⫺0.70 ⫺3.34***
2
Sex 0.15 1.35 ⫺0.06 ⫺1.48 0.18 1.55
Recovered alcoholic ⫺0.15 ⫺0.97 ⫺0.02 ⫺0.29 ⫺0.14 ⫺0.88
Recent alcoholic ⫺0.13 ⫺1.23 0.00 0.04 ⫺0.27 ⫺2.05*
Sex ⫻ Recovered Alcoholic ⫺0.40 ⫺1.34 0.08 0.68 ⫺0.13 ⫺0.51
Sex ⫻ Recent Alcoholic ⫺0.54 ⫺2.15* ⫺0.01 ⫺0.08 ⫺0.69 ⫺2.79**
3
Sex 0.15 1.34 ⫺0.06 ⫺1.52 0.19 1.71†
Number alcoholic parents 0.04 ⫺0.73 ⫺0.01 ⫺0.57 ⫺0.15 ⫺2.05*
Sex ⫻ Number Alcoholic Parents ⫺0.29 ⫺2.20* ⫺0.04 ⫺0.77 ⫺0.28 ⫺2.02*
Predicting random slope
1
Sex ⫺0.03 ⫺1.32 0.02 2.45* ⫺0.01 ⫺0.40
Mom alcoholic ⫺0.01 ⫺0.48 ⫺0.01 ⫺1.11 0.01 0.31
Dad alcoholic 0.00 0.22 0.01 1.13 0.00 ⫺0.06
Sex ⫻ Mom Alcoholic 0.01 0.19 0.01 0.85 ⫺0.03 ⫺0.80
Sex ⫻ Dad Alcoholic 0.07 1.40 ⫺0.02 ⫺0.91 0.13 3.20**
2
Sex ⫺0.03 ⫺1.47 0.02 2.09* ⫺0.01 ⫺0.56
Recovered alcoholic 0.03 1.10 0.02 1.54 0.02 0.84
Recent alcoholic 0.01 0.53 0.00 ⫺0.15 0.00 ⫺0.12
Sex ⫻ Recovered Alcoholic 0.07 1.19 ⫺0.01 ⫺0.24 0.06 1.08
Sex ⫻ Recent Alcoholic 0.06 1.17 ⫺0.02 ⫺1.27 0.10 2.06*
3
Sex ⫺0.03 ⫺1.31 0.02 2.35* ⫺0.01 ⫺0.39
Number alcoholic parents 0.00 0.27 0.00 ⫺0.07 0.00 0.23
Sex ⫻ Number Alcoholic Parents 0.04 1.40 0.00 ⫺0.13 0.05 1.85
†
Variance accounted for
1 .0133 .2544 .0042
2 .0256 .0612 .0488
3 .0277 .2476 .0786
Note. All fixed effects are reported with robust estimation. Main effects and interactions are tested hierarchically such that reported main effects above
are from models that do not include interaction effects. Variance accounted for indicates the proportion of variance in the outcome (intercept) accounted
for by the set of predictors with respect to the baseline (unconditional) models as reported in Table 1. The degree of freedom for the t tests is 367 for Models
1 and 2 and 369 for Model 3.
†
p ⬍ .10. * p ⬍ .05. ** p ⬍ .01. *** p ⬍ .001.
753
SOCIAL COMPETENCE IN CHILDREN

holic parent and those whose parents had recovered from
alcoholism, children with a recently alcoholic parent differed from
those without an alcoholic parent as a function of child gender.
Specifically, girls with a recently alcoholic parent reported less
social competence at age 6 than did girls with nonalcoholic parents
(
␤
⫽⫺.52), t(367) ⫽⫺2.47, p ⬍ .05. No effect of parent
alcoholism was found on boys’ self-reported social competence
(
␤
⫽ .01), t(367) ⫽ 0.12, ns.
The interaction between child’s gender and the number of
alcoholic parents also predicted the intercept of trajectories of
self-reported social competence. Probing of this interaction
showed that having more alcoholic parents predicted lower self-
reported social competence in girls (
␤
⫽⫺.25), t(369) ⫽⫺2.26,
p ⬍ .05, but not in boys (
␤
⫽ .04), t(369) ⫽ 0.62, ns. No other
effects of the number of alcoholic parents were found.
In the parent-reported social competence model, no main or
interactive effects involving maternal and paternal alcoholism
were found.
4
Child’s gender did predict the trajectory intercepts
and significantly predicted the trajectory slopes, such that parents
reported somewhat greater social competence at age 6 but decreas-
ing rates of change in boys compared with girls. No effects of the
number of alcoholic parents or recency of parent alcoholism were
found for parent-reported social competence trajectories.
In the teacher-reported models of social competence, paternal
alcoholism interacted with child’s gender to predict both trajectory
intercepts and slopes. No effects of maternal alcoholism were
found. Probing of the finding for paternal alcoholism showed that
teachers reported lower social competence at age 6 (
␤
⫽⫺.62),
t(367) ⫽⫺3.58, p ⬍ .001, and less decreasing slopes over time
(
␤
⫽ .10), t(367) ⫽ 2.95, p ⬍ .01, for girls with alcoholic fathers
than for girls without alcoholic fathers. For boys, teacher reports of
social competence at age 6 (
␤
⫽⫺.08), t(367) ⫽⫺0.55, ns, and
of changes in social competence over time (
␤
⫽⫺.03), t(367) ⫽
⫺1.40, ns, did not differ for those with and without an alcoholic
father. (See Figure 2 for estimated trajectories depicting these
results.)
Recency of parent alcoholism also interacted with child’s gen-
der to predict teacher-reported social competence. Girls with re-
cent alcoholic parents showed less social competence at age 6
(
␤
⫽⫺.79), t(367) ⫽⫺3.64, p ⬍ .001, and less steep decreases
in social competence over time (
␤
⫽ .07), t(367) ⫽ 1.99, p ⬍ .05,
compared with girls whose parents were never diagnosed with
alcoholism. Boys with a recent alcoholic parent versus nonalco-
4
We tested the impact of scale reliability by replicating our models
within a latent growth curve modeling framework using Mplus (Muthe´n &
Muthe´n, 1998) for the parent report model, where we corrected for unre-
liability in the measurement structure. No substantive changes in key
findings emerged.
Figure 2. Trajectories for teacher-reported social competence by father’s alcoholism. Alc ⫽ alcoholic;
Non-Alc ⫽ nonalcoholic.
754
HUSSONG, ZUCKER, WONG, FITZGERALD, AND PUTTLER

holic parents did not differ from one another in their social com-
petence at age 6 (
␤
⫽⫺.10), t(367) ⫽⫺0.65, ns, or in their rates
of change in social competence over time (
␤
⫽⫺.03), t(367) ⫽
⫺1.10, ns. (Estimated trajectories for groups defined by child
gender and recency of parent alcoholism are plotted in Figure 3.)
No differences were found between children whose parents had
recovered from alcoholism prior to the past 3 years and children
whose parents were never diagnosed as alcoholic.
The number of alcoholic parents also interacted with child
gender in predicting the intercepts and (marginally) the slopes of
teacher-reported social competence trajectories. Teachers reported
less social competence at age 6 (
␤
⫽⫺.36), t(369) ⫽⫺2.84, p ⬍
.01, and marginally less steep slopes (
␤
⫽ .04), t(369) ⫽ 1.75, p ⬍
.10, in girls with more alcoholic parents. Among boys, the number
of alcoholic parents was unrelated to both teacher reports of age 6
social competence (
␤
⫽⫺.08), t(369) ⫽⫺0.93, ns, and to change
in social competence over time (
␤
⫽⫺.01), t(369) ⫽⫺0.69, ns.
Discussion
In contrast to a literature focused on the development of deficits
and psychopathology in at-risk youth, the current study focused on
resilience as evident through the development of social compe-
tence in children at risk for psychopathology due to having an
alcoholic parent. HLMs revealed different patterns of change in
social competence across development as a function of the reporter
of various indicators of social competence. Moreover, female
COAs compared with female non-COAs showed lower social
competence in early childhood that receded in adolescence and
that varied across subtypes of parent alcoholism. We discuss each
of these findings in turn before focusing on implications for
understanding the development of social competence more
generally.
Trajectories of Social Competence
As just noted, a key finding was that the form of normative
trajectories of social competence varied as a function of the ob-
server and specific indicator of the child’s social competence.
Given differences in the subcomponents of social competence
tapped by measures given to different reporters, however, we
could not disentangle these two sources of variance in our findings.
We thus discuss the findings as they relate to previous studies on
reporter effects and then with respect to other measurement issues
across reporters.
Largely consistent with the limited literature, social competence
increased over time when assessed by child’s self-report and
showed no change over time when assessed by teachers’ report
Figure 3. Estimated trajectories for teacher-reported social competence by recency of parent alcoholism. Rec
Alc ⫽ recent alcoholic parent; Non-Alc ⫽ nonalcoholic parent.
755
SOCIAL COMPETENCE IN CHILDREN

(Chan et al., 2000; Cole, Jacquez, & Maschman, 2001; Cole,
Maxwell, et al., 2001). Although evidence of interindividual dif-
ferences in social competence upon school entry was found re-
gardless of reporter, only children’s and parents’ reports of social
competence showed interindividual differences in intraindividual
change over time. As such, teacher reports of social competence at
age 6 were highly predictive of how socially competent teachers
perceived children to be into adolescence.
Reporter differences in trajectories of social competence may be
a function of the items given to each reporter, the changing
expectations of each reporter for social competence with the
child’s age, the changing experience of teachers and parents as
children succeed or fail at different, age-graded role demands, and,
for teacher-reports in particular, changes in the social context in
which such behaviors are observed. Items assessing teacher report
of social competence focused on making friends, popularity, and
possessing socially desirable traits. Given stability in average
levels of social competence over time (i.e., average intraindividual
stability), these items may be reliable indicators of social compe-
tence at any age, although the skills needed to have friends and the
experience of these friendships themselves are likely to change
significantly over time. Moreover, self-reported social competence
items focused on making friends and popularity but also on satis-
faction with social relationships. Perhaps as adolescents become
more active participants in selecting and shaping their social
experiences, it is their satisfaction with their social competence
that increases over time, creating an average increasing pattern of
intraindividual change in self-reported social competence over
time.
Changing expectations and social settings may also impact
trajectories of social competence over time. This is perhaps most
easily illustrated with teacher reports. As children enter new school
environments, such as the transition to middle school, teachers
change as do teacher expectations for behavior and institutionally
imposed consequences and restraints on behavior (Eccles et al.,
1989; Wigfield et al., 1991). Disruption of peer networks and of
how teachers interact with students may impact teacher ratings of
social competence over time. Although our teacher reports of
social competence reflect changes in reporters as well as changes
in the child’s development, teachers are not able to rate social
development in children with respect to their past accomplish-
ments but must rely on the social behavior of other children for
comparison. This perhaps makes teachers better reporters of inter-
individual than of intraindividual change in social competence.
Nonetheless, the degree of average intraindividual stability was
striking in the current study.
COAs and Social Competence
Consistent with our hypothesis, COAs showed deficits in both
self-reported and teacher-reported social competence. However,
this effect was limited as a function of child gender and hetero-
geneity among COAs. These results led us to three key findings.
First, gender differences in social competence deficits associated
with parent alcoholism indicated that this risk is specific to girls.
Given that many of the early negative outcomes in COAs have
been documented primarily in boys or found to be more likely to
occur in boys (Fitzgerald, Zucker, Puttler, Caplan, & Mun, 2000;
Puttler et al., 1998), the current finding urges researchers to cast a
broader net in assessing adjustment indicators when evaluating
gender differences in risk associated with parent alcoholism. Gen-
der socialization theory would suggest that a greater emphasis on
social skills development in girls may create different expectations
for social competence in boys and girls, leading others to judge
deficits in girls more harshly. This possibility is supported not only
by the pattern of gender differences in teacher reports at age 6 but
also by greater reported social competence in girls than in boys
from nonalcoholic families, reflecting the expected difference. To
the extent that girls also internalize these gender-related expecta-
tions for social interaction, this same pattern of gender differences
should emerge in the self-reports of social competence, a predic-
tion consistent with the current findings.
Second, early deficits in social competence at age 6 disappeared
over time, an effect driven by relatively greater decreases in
low-risk, rather than high-risk, girls’ social competence. The de-
crease in self-reported social competence over time in girls may
reflect a lessening of the need for normative ego-protective mech-
anisms with increasing age. As suggested in work regarding social
skills in children with attention-deficit/hyperactivity disorder,
young children have a tendency to overestimate their competence
(Gresham, Lane, MacMillan, Bocian, & Ward, 2000). Given the
emphasis on social skills associated with the female gender role,
this tendency may be more evident in girls than in boys for the
particular domain of social competence. This overestimation is
posited to protect young children’s self-esteem, permitting chil-
dren the opportunity to enact new skills, to suffer setbacks with
fewer repercussions, and to explore their social interactions in a
manner that maintains self-confidence and ego resilience. As chil-
dren develop, their more advanced metacognitive skills (e.g., per-
spective taking, social comparison) and experiences yield to more
realistic estimations of social competence, resulting in decreasing
social competence over time as is evident in low-risk girls in the
current study. Absence of this protective mechanism in young
female COAs may thus reflect a point of vulnerability in early
childhood.
In a second, potentially co-occurring, mechanism, female COAs
may not show the same levels of social competence as their female
peers because of early risk for externalizing and internalizing
symptomatology or temperament-related difficulties. For COA
boys, the normative style of rough and tumble play characterizing
peer interactions as well as different thresholds over time for
friendship interactions may result in fewer problems in developing
normative levels of social competence. As low-risk girls develop a
greater emphasis on dyadic friendships and patterns of relational
aggression and exclusion arise in early adolescence as a normative
style of peer interaction, the relative advantage in social compe-
tence of these girls over female COAs may diminish.
Third, not all COAs manifested risk for early social competence
deficits; rather, this effect was most evident in those with paternal
rather than maternal alcoholism, recent as opposed to recovered
alcoholic parents, and having two alcoholic parents rather than
one. As suggested by Zucker, Wong, Puttler, and Fitzgerald
(2003), such findings call for a more differentiated definition of
risk in the study of resilience. Having an active alcoholic parent, or
having two alcoholic parents, may increase risk for social compe-
tence deficits due to the greater stress and chaos in the home
environment and the more direct observation of parental problems,
including social interaction deficits characterizing the parent– child
relationship (Loukas, Zucker, Fitzgerald, & Krull, 2003). Notably,
children of parents who were recovered alcoholics (i.e., had expe-
756
HUSSONG, ZUCKER, WONG, FITZGERALD, AND PUTTLER

rienced alcoholism but not in the past 3 years) showed similar
levels of social competence as children with nonalcoholic parents,
suggesting that the risks associated with deficits in this domain
may be more short term than those evident in other domains.
Finally, paternal alcoholism has also been associated with greater
risk for behavioral dysregulation and difficult temperament, per-
haps related to a genetic liability, each of which may increase risk
for social competence deficits as well (Tarter et al., 1999). Our
results suggest that not all COAs manifest risk or resilience sim-
ilarly and that part of the difference in outcomes within this risk
group may be due to heterogeneity in the type of alcoholism in the
family and in the gender of the child (Zucker, Fitzgerald, & Moses,
1995).
Conclusions
The current findings point to the need for finer-grained analyses
that may better differentiate the development of various dimen-
sions of social competence over time as a function of social
context and societal expectation. This study offers evidence that
some COAs lag behind their peers in the development of social
competence. This evidence is found in the self- and teacher reports
of girls. Strengths of the current study that lend support to these
findings include the use of a community-based, high-risk sample,
inclusion of multiple reporters of social competence, and the
longitudinal assessment of these youth. However, limitations of
the study include the use of nonparallel measures of social com-
petence across reporters and low reliability of the parent report
measure of social competence. Although the small size of certain
subsamples of children (e.g., girls in nonalcoholic families) may
raise issues concerning generalizability, the attention to represen-
tativeness in sample recruitment and retention counter this poten-
tial limitation.
Given the modest effects of parent alcoholism and child gender
on these indicators of social competence, other influences on and
periods for change in the development of social competence over
time should be considered. For example, future studies may want
to include more fine-grained assessment around periods of greater
fluctuations in social self-competence such as the transition to
middle school (Eccles et al., 1989; Wigfield et al., 1991). Associ-
ated predictors may also provide greater insight into the factors
that more clearly define resilient outcomes in the face of risk
associated with parent alcoholism. Lessons for the study of risk
and resilience offered by the current study include an emphasis on
heterogeneity in the outcomes of social competence among COAs.
Although some indicators of parent alcoholism were associated
with greater deficits in social competence (e.g., paternal alcohol-
ism, having a greater number of alcoholic parents, and having
more recent parent alcoholism), others were not. Moreover, parent
alcoholism appears to have a time-limited influence on social
competence, perhaps suggesting the presence of more proximal
influences that distinguish among the resilient and distressed
COAs in adolescence. Consistent with other areas of study within
the field of resilience, research to identify developmentally linked
mediational mechanisms that account for variability in adolescent
social competence among this risk group is needed (Zucker et al.,
2003).
This study contributes to our understanding of social compe-
tence in general and to the complexities that are inherent in
studying the development of competence in a high-risk population.
Although the predictors of parent alcoholism and children’s gender
account for only small to moderate variance in the social devel-
opment outcomes, the results contribute to a very limited literature
on prosocial outcomes among high-risk youth and thus add to their
practical significance. In this case, risk is manifested as a delay in
the development of normative social competence for female
COAs. The extent to which such delays may be redressed through
the use of social skills training curricula for these children is not
yet known. Our findings suggest that such training may be bene-
ficial if considered in conjunction with the specificity in risk for
such targeted deficits in the daughters of a subset of alcoholic
parents. However, what is not clear is whether such changes in
social competence might be useful as well in addressing other
areas of social risk for COAs, such as the tendency to associate
with deviant peers. The intersection of social competence and this
well-known risk factor for multiple negative outcomes associated
with parent alcoholism is an important area of further inquiry.
References
Achenbach, T. M. (1991). “Comorbidity” in child and adolescent psychi-
atry: Categorical and quantitative perspectives. Journal of Child and
Adolescent Psychopharmacology, 1, 271–278.
American Psychiatric Association. (1994). Diagnostic and statistical man-
ual of mental disorders (4th ed.). Washington, DC: Author.
Bagwell, C., Newcomb, A., & Bukowski, W. (1998). Preadolescent friend-
ship and peer rejection as predictors of adult adjustment. Child Devel-
opment, 69, 140 –153.
Berndt, T., & Burgy, L. (1996). Social self-concept. In B. A. Bracken (Ed.),
Handbook of self-concept: Developmental, social, and clinical consid-
erations (pp. 171–209). Oxford, England: Wiley.
Bingham, C. R., Fitzgerald, W. W., Fitzgerald, H. E., & Zucker, R. A.
(1996). Prenatal history, parental alcoholism, and the development of
preschool aged sons of alcoholics [Abstract]. Alcoholism: Clinical and
Experimental Research, 20, 469.
Bracken, B. A. (1992). The Multidimensional Self Concept Scale. Austin,
TX: Pro-Ed.
Chan, D., Ramey, S., Ramey, C., & Schmitt, N. (2000). Modeling intra-
individual changes in children’s social skills at home and at school: A
multivariate latent growth approach to understanding between-settings
differences in children’s social skill development. Multivariate Behav-
ioral Research, 35, 365–396.
Chassin, L., Rogosch, F., & Barrera, M. (1991). Substance use and symp-
tomatology among adolescent children of alcoholics. Journal of Abnor-
mal Psychology, 100, 449 –463.
Coie, J. D., Belding, M., & Underwood, M. (1988). Aggression and peer
rejection in childhood. In B. B. Lahey & A. E. Kazdin (Eds.), Advances
in clinical child psychology (Vol. 11, pp. 125–158). New York: Plenum
Press.
Cole, D. A., Jacquez, F. M., & Maschman, T. L. (2001). Social origins of
depressive cognitions: A longitudinal study of self-perceived compe-
tence in children. Cognitive Therapy and Research, 25, 377–395.
Cole, D. A., Maxwell, S. E., Martin, J. M., Peeke, L. G., Seroczynski,
A. D., Tram, J. M., et al. (2001). The development of multiple domains
of child and adolescent self-concept: A cohort sequential longitudinal
design. Child Development, 72, 1723–1746.
Curran, P. J., Bauer, D. J., & Willoughby, M. T. (in press). Testing and
probing interactions in hierarchical linear growth models. In C. S.
Bergeman & S. M. Boker (Eds.), The Notre Dame Series on Quantitative
Methodology: Vol. 1. Methodological issues in aging research. Mah-
wah, NJ: Erlbaum.
Curran, P. J., Bollen, K. A., Paxton, P., Kirby, J., & Chen, F. N. (2002).
The noncentral chi-square distribution in misspecified structural equa-
757
SOCIAL COMPETENCE IN CHILDREN

tion models: Finite sample results from a Monte Carlo simulation.
Multivariate Behavioral Research, 37, 1–36.
Curran, P. J., & Chassin, L. (1996). A longitudinal study of parenting as a
protective factor for children of alcoholics. Journal of Studies on Alco-
hol, 57, 305–313.
Dishion, T. J., Capaldi, D., Spracklen, K. M., & Li, F. Z. (1995). Peer
ecology of male adolescent drug use. Development and Psychopathol-
ogy, 7, 803– 824.
Eccles, J. S., Wigfield, A., Flanagan, C. A., Miller, C., Reuman, D. A., &
Yee, D. (1989). Self-concepts, domain values, and self-esteem—Rela-
tions and changes at early adolescence. Journal of Personality, 57,
283–310.
Fitzgerald, H. E., Sullivan, L. A., Ham, H. P., Zucker, R. A., Bruckel, S.,
Schneider, A. M., et al. (1993). Predictors of behavior problems in
three-year-old sons of alcoholics: Early evidence for the onset of risk.
Child Development, 64, 110 –123.
Fitzgerald, H. E., Zucker, R. A., Puttler, L. I., Caplan, H. M., & Mun, E.-Y.
(2000). Alcohol abuse/dependence in women and girls: Etiology, course,
and subtype variations. Alcoscope: International Review of Alcoholism
Management, 3, 6 –10.
Gresham, F. M., Lane, K. L., MacMillan, D. L., Bocian, K. M., & Ward,
S. L. (2000). Effects of positive and negative illusory biases: Compar-
isons across social and academic self-concept domains. Journal of
School Psychology, 38, 151–175.
Harter, S. (1982). The Perceived Competence Scale for Children. Child
Development, 53, 87–97.
Harter, S. (1985). Manual for the Self-Perception Profile for Children.
Denver, CO: University of Denver.
Harter, S. (1988). Manual for the Self-Perception Profile for Adolescents.
Unpublished manuscript, University of Denver.
Kalverboer, A. (1988). Follow-up of biological high-risk groups. In M.
Rutter (Ed.), Studies of psychosocial risk: The power of longitudinal
data (pp. 114 –137). New York: Cambridge University Press.
Ladd, G. (1990). Having friends, keeping friends, making friends, and
being liked by peers in the classroom: Predictors of children’s early
school adjustment? Child Development, 61, 1081–1100.
Larsson, B., Knutsson-Medin, L., Sundelin, C., & Trost von Werder, A. C.
(2000). Social competence and emotional/behavioural problems in chil-
dren of psychiatric inpatients. European Child & Adolescent Psychiatry,
9, 122–128.
Ledingham, J. E. (1990). Recent developments in high-risk research.
Advances in Clinical Child Psychology, 13, 91–137.
Loehlin, J. C. (1992). Latent variable models: An introduction to factor,
path, and structural analysis (2nd ed.). Hillsdale, NJ: Erlbaum.
Loukas, A., Zucker, R. A., Fitzgerald, H. E., & Krull, J. L. (2003).
Developmental trajectories of disruptive behavior problems in sons of
alcoholics: Effects of parent psychopathology, family conflict, and child
undercontrol. Journal of Abnormal Psychology, 112, 119–131.
Marsh, H. W., & O’Neill, R. (1984). Self Description Questionnaire III.
The construct validity of multidimensional self-concept ratings by late
adolescents. Journal of Educational Measurement, 21, 153–174.
Masten, A. S., Coatsworth, J. D., Neemann, J., Gest, S. D., Tellegen, A., &
Garmezy, N. (1995). The structure and coherence of competence from
childhood through adolescence. Child Development, 66, 1635–1659.
Muthe´n, L. K., & Muthe´n, B. O. (1998). Mplus user’s guide [Computer
software and manual]. Los Angeles: Authors.
Noll, R. B., Zucker, R. A., Fitzgerald, H. E., & Curtis, W. J. (1992).
Cognitive and motoric functioning of sons of alcoholic fathers and
controls: The early childhood years. Developmental Psychology, 28,
665– 675.
Oetting, E. R., & Beauvais, F. (1987). Common elements in youth drug
abuse: Peer clusters and other psychosocial factors. Journal of Drug
Issues, 17(1–2), 133–151.
Parker, J. G., & Asher, S. R. (1987). Peer relations and later personal
adjustment—Are low-accepted children at risk? Psychological Bulletin,
102, 357–389.
Parker, J. G., & Asher, S. R. (1993). Friendship and friendship quality in
middle childhood: Links with peer group acceptance and feelings of
loneliness and social dissatisfaction. Developmental Psychology, 29,
611– 621.
Parker, J. G., Rubin, K. H., Price, J. M., & DeRosier, M. E. (1995). Peer
relationships, child development, and adjustment: A developmental psy-
chopathology perspective. In D. Cicchetti & D. J. Cohen (Eds.), Devel-
opmental psychopathology: Vol. 2. Risk, disorder, and adaptation (pp.
96 –161). Oxford, England: Wiley.
Piejak, L. A., Fitzgerald, H. M., Zucker, R. A., & von Eye, A. (2004).
Predictors of academic functioning of sons of male antisocial and
nonantisocial alcoholics during the elementary school years. Unpub-
lished manuscript, Michigan State University and University of
Michigan.
Pillow, D. R., Barrera, M., & Chassin, L. A. (1998). Using cluster analysis
to assess the effects of stressful life events: Probing the impact of
parental alcoholism on child stress and substance use. Journal of Com-
munity Psychology, 26, 361–380.
Puttler, L. I., Zucker, R. A., Fitzgerald, H. E., & Bingham, C. R. (1998).
Behavioral outcomes among children of alcoholics during the early and
middle childhood years: Familial subtype variations. Alcoholism: Clin-
ical and Experimental Research, 22, 1962–1972.
Raudenbush, S. W. (2001). Comparing personal trajectories and drawing
causal inferences from longitudinal data. Annual Review of Psychology,
52, 501–525.
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models:
Applications and data analysis methods (2nd ed.). Thousand Oaks, CA:
Sage.
Robins, L. N., Helzer, J. E., Croughan, J., & Ratcliff, K. S. (1980). The
NIMH Diagnostic Interview Schedule: Its history, characteristics and
validity. St. Louis, MO: Washington University School of Medicine.
Rockhill, C. M. (1995, March). Depressed children’s friendships. Poster
session presented at the biennial meeting of the Society for Research in
Child Development, Indianapolis, IN.
Rose-Krasnor, L. (1997). The nature of social competence: A theoretical
review. Social Development, 6, 111–135.
Segrin, C., & Menees, M. M. (1996). The impact of coping styles and
family communication on the social skills of children of alcoholics.
Journal of Studies on Alcohol, 57, 29–33.
Selzer, M. L., Vinokur, A., & van Rooijan, L. (1975). A self-administered
short Michigan Alcoholism Screening Test (SMAST). Journal of Stud-
ies on Alcohol, 36, 117–126.
Sher, K. J. (1991). Children of alcoholics: A critical appraisal of theory
and research. Chicago: University of Chicago Press.
Sokol, R. J., & Claren, S. K. (1989). Guidelines for use of terminology
describing the impact of prenatal alcohol on the offspring. Alcoholism:
Clinical and Experimental Research, 13, 597–598.
Tarter, R. E., Vanyukov, M., Giancola, P., Dawes, M., Blackson, T.,
Mezzich, A., & Clark, D. B. (1999). Etiology of early age onset sub-
stance use disorder: A maturational perspective. Development and Psy-
chopathology, 38, 115–142.
Waters, E., & Sroufe, L. A. (1983). Social competence as a developmental
construct. Developmental Review, 3, 79 –97.
Werner, E. E. (1986). Resilient offspring of alcoholics: A longitudinal
study from birth to age 18. Journal of Studies on Alcohol, 47, 34 –40.
West, M. O., & Prinz, R. J. (1987). Parental alcoholism and childhood
psychopathology. Psychological Bulletin, 102, 204 –218.
Wigfield, A., Eccles, J. S., Maciver, D., Reuman, D. A., & Midgley, C.
(1991). Transitions during early adolescence—Changes in children’s
domain-specific self-perceptions and general self-esteem across the tran-
sition to junior high school. Developmental Psychology, 27, 552–565.
Wong, M. M., Zucker, R. A., Fitzgerald, H. E., & Puttler, L. I. (1999).
Heterogeneity of risk aggregation for alcohol problems between early
758
HUSSONG, ZUCKER, WONG, FITZGERALD, AND PUTTLER

and middle childhood: Nesting structure variations. Development and
Psychopathology, 11, 727–744.
Zucker, R. A., & Fitzgerald, H. E. (1991). Early developmental factors and
risk for alcohol problems. Alcohol Health and Research World, 15,
18 –24.
Zucker, R. A., Fitzgerald, H. E., & Moses, H. D. (1995). Emergence of
alcohol problems and the several alcoholisms: A developmental per-
spective on etiologic theory and life course trajectory. In D. Cicchetti &
D. J. Cohen (Eds.), Developmental psychopathology: Vol. 2. Risk, dis-
order, and adaptation (pp. 677–711). New York: Wiley.
Zucker, R. A., Fitzgerald, H. E., & Noll, R. B. (1990). Drinking and drug
history. Unpublished manuscript, Michigan State University.
Zucker, R. A., Fitzgerald, H. E., Refior, S. K., Puttler, L. I., Pallas, D. M.,
& Ellis, D. A. (2000). The clinical and social ecology of childhood for
children of alcoholics: Description of a study and implications for a
differentiated social policy. In H. E. Fitzgerald, B. M. Lester, & B. S.
Zuckerman (Eds.), Children of addiction: Research, health and policy
issues (pp. 109 –141). New York: Routledge Falmer.
Zucker, R. A., Wong, M. M., Puttler, L. I., & Fitzgerald, H. E. (2003).
Resilience and vulnerability among sons of alcoholics: Relationship to
developmental outcomes between early childhood and adolescence. In S.
Luthar (Ed.), Resilience and vulnerability: Adaptation in the context of
childhood adversities (pp. 76–103). New York: Cambridge University
Press.
Received March 15, 2004
Revision received December 6, 2004
Accepted January 17, 2005 䡲
New Editors Appointed, 2007–2012
The Publications and Communications (P&C) Board of the American Psychological Association
announces the appointment of three new editors for 6-year terms beginning in 2007. As of January
1, 2006, manuscripts should be directed as follows:
• Journal of Experimental Psychology: Learning, Memory, and Cognition (www.apa.org/journals/
xlm.html), Randi C. Martin, PhD, Department of Psychology, MS-25, Rice University, P.O. Box
1892, Houston, TX 77251.
• Professional Psychology: Research and Practice (www.apa.org/journals/pro.html), Michael C.
Roberts, PhD, 2009 Dole Human Development Center, Clinical Child Psychology Program,
Department of Applied Behavioral Science, Department of Psychology, 1000 Sunnyside Avenue,
The University of Kansas, Lawrence, KS 66045.
• Psychology, Public Policy, and Law (www.apa.org/journals/law.html), Steven Penrod, PhD,
John Jay College of Criminal Justice, 445 West 59th Street N2131, New York, NY 10019-1199.
Electronic manuscript submission. As of January 1, 2006, manuscripts should be submitted
electronically through the journal’s Manuscript Submission Portal (see the Web site listed above
with each journal title).
Manuscript submission patterns make the precise date of completion of the 2006 volumes uncertain.
Current editors, Michael E. J. Masson, PhD, Mary Beth Kenkel, PhD, and Jane Goodman-
Delahunty, PhD, JD, respectively, will receive and consider manuscripts through December 31,
2005. Should 2006 volumes be completed before that date, manuscripts will be redirected to the new
editors for consideration in 2007 volumes.
In addition, the P&C Board announces the appointment of Thomas E. Joiner, PhD (Department
of Psychology, Florida State University, One University Way, Tallahassee, FL 32306-1270), as
editor of the Clinician’s Research Digest newsletter for 2007–2012.
759
SOCIAL COMPETENCE IN CHILDREN