PROCESS

Parental Separation and Children’s Behavioral/

Emotional Problems: The Impact of Parental
Representations and Family Conflict
STEPHANIE STADELMANN, PH.D. n
SONJA PERREN, PH.D.w
MAUREEN GROEBEN, PH.D. n
KAI VON KLITZING, M.D.z

To read this article’s abstract in both Spanish and Mandarin Chinese, please visit the article’s
full-text page on Wiley InterScience (http://interscience.wiley.com/journal/famp).

In this longitudinal study, we examine whether the effect of parental separation on

kindergarten children’s behavioral/emotional problems varies according to the level of
family conflict, and children’s parental representations. One hundred and eighty seven
children were assessed at ages 5 and 6. Family conflict was assessed using parents’
ratings. Children’s parental representations were assessed using a story-stem task. A
multiinformant approach (parent, teacher, child) was employed to assess children’s be-
havioral/emotional problems. Bivariate results showed that separation, family conflict,
and negative parental representations were associated with children’s behavioral/emo-
tional problems. However, in multivariate analyses, when controlling for gender and
symptoms at age 5, we found that children of separated parents who showed negative
parental representations had a significantly greater increase in conduct problems between
5 and 6 than all other children. In terms of emotional symptoms and hyperactivity,
symptoms at 5 and (for hyperactivity only) gender were the only predictors for symptoms
1 year later. Our results suggest that kindergarten children’s representations of parent-
child relationships moderate the impact of parental separation on the development of
conduct problems, and underline play and narration as a possible route to access the
thoughts and feelings of young children faced with parental separation.

Keywords: Kindergarten Age; Parental Separation; Parental Representations; Family

Conflict; Behavioral/Emotional Problems

Fam Proc 49:92–108, 2010

n
Department of Child & Adolescent Psychiatry, University of Basel, Switzerland.
w
Jacobs Center for Productive Youth Development, University of Zurich, Switzerland
z
Department of Child and Adolescent Psychiatry, Psychotherapy and Psychosomatics, University of
Leipzig, Germany
Correspondence concerning this article should be addressed to Stephanie Stadelmann, University
of Basel, Child & Adolescent Psychiatry, Scaffhauserrheinweg 55, CH-4058 Basel, Switzerland.
E-mail: stephanie.stadelmann@medizin.uni-leipzig.de

92
Family Process, Vol. 49, No. 1, 2010 r FPI, Inc.
STADELMANN, PERREN, GROEBEN, & VON KLITZING / 93
INTRODUCTION

P arental separation is rated as one of their most stressful life events by both parents
and children (Davies & Cummings, 1994) and was found to strongly influence the
quality of family relationships even two decades later (Ahrons, 2007). As family roles,
relationships, and circumstances change children often tend to be depressed, anxious,
angry, demanding and noncompliant, and experience a drop in school performance
(Amato, 2001; Hetherington & Stanley Hagan, 1999). In this study, we examine whether
parental separation, family conflict, and children’s parental representations are pre-
dictors of individual changes in children’s behavioral/emotional problems at kindergarten
age and whether the adjustment of children who experienced parental separation differs
according to the level of family conflict and to children’s representations of caregivers.
For kindergarten children, the family is a highly salient context for understanding
psychopathological development. A wealth of evidence from cross-sectional and long-
itudinal studies indicates that parental separation as an important life event is asso-
ciated with an increased rate of behavioral/emotional problems in children (Amato, 2001;
Kelly, 2000). The greatest effects of divorce on child adjustment relate to behavioral
problems. Effects on emotional symptoms (depression/anxiety) are far more inconsistent
(Amato & Keith, 1991; Hetherington & Stanley Hagan, 1999). Just as important as the
evidence of mean differences is the marked individual variation in children’s socio-
emotional development among those experiencing a parental separation. Some children
are severely affected by separation, while others seem to be relatively unaffected.
Moreover, the effects of separation are not necessarily adverse. Children who move from
a conflictual, abusive, or negligent family situation to a more harmonious one may show
diminished problems following separation (Amato, Loomis, & Booth, 1995). This sheds
light on new directions in research which seek to identify mediating and moderating risk
and protective factors in the association between separation and child outcome (Amato
et al., 1995; Cummings & Davies, 2002). As a potential moderator, marital conflict re-
ceives much empirical attention, and research has indicated that it is marital conflict
rather than the break-up of the family that is primarily responsible for many of the
problems seen in children whose parents separate (Kelly, 2000). However, current re-
search on marital conflict also proposes that the effects on children are a function of
children’s perceptions of conflict in relation to themselves and their families rather than
simply reflecting the frequency or characteristics of conflict (Cummings & Davies, 2002;
Fincham, 1998). Thus, to understand the effects of parental separation and of marital
conflict on children, exploring psychological response processes underlying children’s
development and adjustment over time is of great importance. Here, children’s
representations of family relationships play an important role.
Children’s representations have been conceptualized in terms of working models
in attachment theory, and schemas or scripts in social cognitive approaches
(e.g., Bretherton, 1990; Lemerise & Arsenio, 2000). Common to these theoretical
perspectives is the notion that children internalize important aspects of care-giving
experiences, and that these mental structures influence children’s emotional and
behavioral responses to new social encounters. The evaluation of children’s play
narratives, elicited by story stems that reflect socioemotional dilemmas, has proved
to be one way of gaining access to young children’s thoughts and feelings about
emotionally significant relationships (Bretherton & Oppenheim, 2003). Evidence has
emerged to suggest that children’s representations are affected by histories of

Fam. Proc., Vol. 49, March, 2010

94 / FAMILY PROCESS

exposure to marital conflict (Cummings & Davies, 2002; Grych & Fincham, 1990).
Shamir, Du Rocher Schudlich, and Cummings (2001) found that negative marital
conflict strategies were predictive of more negative representations of family systems
in 5–8-year-old children’s story-stem narratives. Grych, Wachsmuth Schlaefer, and
Klockow (2002) showed that children who experienced extensive domestic violence
showed significantly more negative representations of self and mothers, within more
incoherent stories, than children without such negative experiences. Research also
documents an association between children’s representations and their psycho-
pathological development. Children who display a high level of negative representa-
tions and a low level of positive/coherent representations of caregivers in their play
narratives after socioemotional dilemmas were found to show higher levels of beha-
vioral/emotional problems and increased social impairment compared with other
children (e.g., Oppenheim, Nir, Warren, & Emde, 1997b; Shields, Ryan, & Cicchetti,
2001). In a recent study on kindergarten children, Stadelmann, Perren, von Wyl, and
von Klitzing (2007) showed that negative parental representations were associated
with the development of conduct problems, whereas positive parental representations
were associated with the development of prosocial behavior. Grych, Fincham, Jouriles,
and McDonald (2000) were able to show that children’s evaluations of self-blame and
threat mediated the association between interparental conflict and their own ad-
justment problems, specifically internalizing problems. Although studies have shown
significant associations between family background, children’s cognitive-emotional
processes and adjustment, and support the notion that children’s perceptions of the
family background shape the impact of family events on aspects of self-functioning,
much more research is needed on the interplay between these aspects if develop-
mental pathways in children are to be better understood. Specifically, there is a need
for studies to investigate young children’s cognitive-emotional processes as mediators
or moderators of the impact of parental separation on their adjustment.
In the present study, we examine whether parental separation, family conflict, and
children’s parental representations are predictors of individual changes in children’s
behavioral/emotional problems at kindergarten age. We hypothesized that children from
families where the parents had separated would show a greater increase in behavioral/
emotional problems at kindergarten age than children of nonseparated families. Fur-
thermore, we hypothesized that children from families with high levels of conflict and
children who display many negative parental representations in their play narratives
would show a greater increase in behavioral/emotional problems at kindergarten age.
Specifically, we were interested in whether family conflict and children’s negative par-
ental representations moderate the impact of parental separation on behavioral/emo-
tional problems in kindergarten age. In the main analyses, we control for gender as
various studies found gender differences in children’s representations (e.g., Clyman,
2003; von Klitzing, Kelsay, Emde, Robinson, & Schmitz, 2000) as well as in behavioral
and emotional problems (Zahn-Waxler, Klimes Dougan, & Slattery, 2000).

METHOD
Sample
One hundred and eighty-seven children (76 girls, 111 boys) participated in the
study and were assessed at the ages of 5 and 6 years (age at first assessment:
mean ¼ 5.27, SD ¼ 0.39; second assessment: mean ¼ 6.23, SD ¼ 0.35). Ninety-three

www.FamilyProcess.org
STADELMANN, PERREN, GROEBEN, & VON KLITZING / 95
children were recruited through their kindergarten classes in the city of Basel. Kin-
dergarten classes were selected from different city districts representing various so-
cioeconomic and ethnic backgrounds in the city of Basel. After the kindergarten
teacher had agreed to participate, parents were informed at a parents’ meeting and
given written information on the study. Ninety-eight families agreed to participate
(participation rate 74.2%). Another 64 children were part of our longitudinal study on
family relationships that began when the mothers of these children were pregnant
(Perren et al., 2003; Perren, von Wyl, Stadelmann, Bürgin, & von Klitzing, 2006). To
increase the number of children with behavioral/emotional problems in the clinical
range, we included a group of 30 children who had been referred to our child psy-
chiatric outpatient unit for clinical evaluation and treatment. The referrals were
generally initiated by the children’s teachers and/or parents, and followed behavioral
problems (mostly dysregulated aggressive behavior and/or hyperactivity) or emotional
symptoms (anxiety and/or depressive symptoms). Parents and children were asked
whether they would be willing to participate in our study, and if they agreed they were
assessed at intake into the clinic. For bivariate analyses all available data were used
(N ¼ 155–187). For longitudinal analyses, we only used complete datasets (N ¼ 154).
The participating families were mainly Swiss-German, while 41 families (22%) had
an immigrant background. In 94 families (51%) both parents had a higher education,
in 43 families (24%) at least one of the parents had a higher education; in 46 families
(25%) both parents had only received a basic education (9 years or less of schooling) or
a professional qualification (vocational training). In 4 cases, we only had data on the
mother (N ¼ 2 had higher education and N ¼ 2 were low educated).
Parents reported on parental separation (yes/no) and children’s age at separation.
Thirty-three (18%) children (16 girls, 17 boys) experienced parental separation before
the age of 5. Nineteen of these children (58%) lived with their mother, nine children
(27%) lived with both parents (different households, shared custody), one child (3%)
lived with its father, and four children (12%) lived in alternative settings such as with
mother and stepfather, with the grandmother, or in a children’s home.
Families were only included into this group if parents remained separated from one
another during the study interval. Children’s age at separation ranged from 5 months
to 5 years of age. Behavioral/emotional problems, family conflict, and children’s
parental representations were not significantly associated with children’s age at
separation. Therefore, we did not control for this variable in our main analyses.
Comparing the community and clinical sample, we found that parents of children
from the clinical sample had a significantly lower educational level than those of the
‘‘community’’ children ( p  .05). The two samples did not differ from each other in
terms of migration background or parental separation. However, children in the
clinical sample showed significantly higher levels of conduct problems, hyperactivity,
and emotional symptoms at both ages, and experienced significantly more conflict in
their families than children of the community sample ( p  .05). No sample differences
were found in terms of children’s parental representations.

Measures
Parents, children, and teachers reported on children’s behavioral/emotional pro-
blems twice (age 5/6). Parents also reported on family conflict at the same ages.
Children’s parental representations were assessed using a story-stem task (age 5/6).

Fam. Proc., Vol. 49, March, 2010

96 / FAMILY PROCESS

All children were interviewed individually by a trained psychologist. Teachers and

parents completed questionnaires within 4 weeks of the child interview.
Behavioral/emotional problems (age 5/6)
Parents and teachers completed the symptom scales of the Strengths and
Difficulties Questionnaire (SDQ, Goodman, 1997). Children’s behavioral/emotional
problems were assessed according to three scales: conduct problems (e.g., ‘‘often has
temper tantrums or a hot temper’’), hyperactivity/inattention (e.g., ‘‘restless,
overactive, cannot stay still for long’’), and emotional symptoms (e.g., ‘‘many worries,
often seems worried’’). Each scale consists of 5 items that are rated on a 3-point scale.
Foreign-language speaking parents completed validated versions of the SDQ in their
first language. Internal consistency was moderate to high (parents: Cronbach’s
a ¼ .57  .79; teachers: Cronbach’s a ¼ .73  .88).
Children’s self-report of behavioral/emotional problems was assessed by means of a
standardized puppet interview. The Berkeley Puppet Interview (BPI, Measelle, Ab-
low, Cowan, & Cowan, 1998) is carried out by means of two identical hand puppets
that make two opposing statements on a topic, after which the child can give his/her
own statement. The interview is videotaped and subsequently scored by independent
raters. Each item is rated on a 7-point scale. Our interviewers were trained by the
authors of the instrument. Interrater reliability was established first with the authors
of the instrument, and then with the raters of the research group (average ICC ¼ .97).
To link child reports of behavioral/emotional problems to parent and teacher re-
ports, we aggregated the original BPI subscales. ‘‘Emotional symptoms’’ encompass
depression, separation anxiety, and over-anxiety (Cronbach’s a ¼ .69  .74, 20 items).
‘‘Conduct problems’’ encompass oppositionality/defiance and overt aggression to
peers (Cronbach’s a ¼ .64  .74, 13 items). ‘‘Hyperactivity/inattention’’ includes im-
pulsivity plus a single item on inattention (Cronbach’s a ¼ .58  .59, 7 items). The
interviews were conducted in Swiss German.
As recommended by Kraemer et al. (2003) we aggregated child, parent, and teacher
reports. This allows the integration of information on children’s characteristics from
different perspectives and contexts. Previous analyses of the cross-sectional data
showed that the aggregation of the three informants yielded reliable and valid in-
formation on children’s behavioral/emotional problems (Perren et al., 2006). In this
study, cross-informant agreements (child, parents, and teachers) were moderate
to high (intraclass correlations between .39 and .53). To combine child, parent, and
teacher reports on behavioral/emotional problems, the average scores of each in-
formant were first z-standardized and then averaged across informants (mean scores).

Family conflict (age 5/6)

Family conflict was assessed by parent reports using the subscale ‘‘conflict’’ of the
German version of the Family Environment Scale (Moos & Moos, 1981; Schneewind,
Beckmann, & Hecht-Jackl, 1985). The subscale conflict describes the amount of anger
and aggression expressed in the family (8 items, a ¼ .76  .79). Items are rated on a
4-point scale.
We used the linear variables to analyze bivariate associations. For multivariate
analyses, we combined longitudinal data on family conflict (age 5/6). We first calcu-
lated the mean of family conflict between the two time points. We then dichotomized
this variable using the 75th percentile as cut-off criterion: low level of family

www.FamilyProcess.org
STADELMANN, PERREN, GROEBEN, & VON KLITZING / 97
conflicto75th percentile of the mean of age 5/6Ffamily conflict (N ¼ 117); high level
of family conflict  75th percentile of the mean of age 5/6Ffamily conflict (N ¼ 38).
Children’s parental representations in play narratives (age 5/6)
The MacArthur Story-Stem Battery (MSSB, Bretherton & Oppenheim, 2003),
designed for children from 3 to 7 years of age, uses standardized story stems to elicit
children’s responses across a range of socioemotional dilemmas. All stems consist of a
brief narrative introduced by the experimenter, and end with a dilemma. In this
study, after the presentation of a warm-up story, each child was given eight story
stems. Bretherton and Oppenheim (2003) outline detailed procedures for adminis-
tering seven of the eight socio-emotional dilemmas (‘‘hot soup,’’ ‘‘lost Barney,’’
‘‘mother’s headache,’’ ‘‘three is a crowd,’’ ‘‘lost keys,’’ ‘‘bathroom shelf,’’ ‘‘exclu-
sion’’). The last stem was devised by the project group (‘‘sandcastle’’: a friend wants
the child to join in aggravating a younger child; parents standing aside). Children’s
responses to the story stems were videotaped.
Children’s narratives were coded for negative parental representations by an in-
dependent coder, using the MacArthur Narrative Coding Manual (Oppenheim, Emde, &
Warren, 1997a; Robinson & Mantz-Simmons, 2003). Negative representations en-
compass harsh, punitive, rejecting, and ineffectual behavior of parental figures toward
child figures. Parental representations were coded as present or absent in each story.
Means were taken of the individual scores for each theme across all eight stories for each
subject. Interrater reliability was established between the coder and the last author of
the present study, who has been trained by the authors of the instrument (k ¼ .86).
For bivariate analyses, we used the linear variables. For multivariate analyses we
combined longitudinal data on parental representations (age 5/6) and placed the children
into two categories: None ¼ child did not show negative parental representations in play
narratives (N ¼ 101). At least once ¼ child showed negative parental representations at
age 5 and/or at age 6 (N ¼ 53). We decided to use this procedure because of distribution
properties (highly skewed distribution of the variable at both ages).

RESULTS
First, we report descriptive results and bivariate associations between study vari-
ables. Second, we present results on multivariate analyses.

Descriptive Results and Bivariate Associations Between Study Variables

Bivariate cross-sectional and longitudinal associations
To analyze associations between family conflict, parental representations, and be-
havioral/emotional problems, we computed Pearson correlations coefficients (Table 1).
Family conflict and children’s parental representations in play narratives were not
significantly associated with each other, but both aspects showed significant correla-
tions with behavioral/emotional problems. Family conflict was positively associated
with conduct problems and emotional symptoms (age 5, 6, 5 with 6, 6 with 5) and with
hyperactivity (age 5, 5 with 6). Negative parental representations were positively
associated with conduct problems (age 5, 6, 5 with 6) and with hyperactivity (age 6).
No significant associations were found between parental representations and
emotional symptoms. Finally, behavioral/emotional problems showed significantly
positive correlations with each other (age 5, 6, 5 with 6, 6 with 5).

Fam. Proc., Vol. 49, March, 2010

 98
/
TABLE 1
Bivariate Cross-sectional and Longitudinal Associations between Symptoms, Family Conflict, and Negative Parental Representations

1 2 3 4 5 6 7 8 9 10

Symptoms age 5
1. Conduct p. F
2. Hyperactivity .674 n n F
3. Emotional s. .421 n n .491 n n F
Symptoms age 6
4. Conduct p. .642 n n .411 n n .227 n n F
5. Hyperactivity .489 n n .744 n n .328 n n .504 n n F
6. Emotional s. .283 n n .362 n n .622 n n .261 n n .332 n n F
Family conflict
7. Age 5 .277 n n .238 n n .270 n n .196 n n .173 n .189 n F
8. Age 6 .198 n .077 .250 n n .282 n n .118 .241 n n .643 n n F
Negative parental
representation
9. Age 5 .196 n .124  .017 .298 n n .050  .015 .078 .069 F
10. Age 6 .026 .058  .071 .153 n .153 n n  .084 .055  .028 .353 n n F

Note. po.01. po.05 (Pearson’s correlation coefficient, two-tailed).

nn n
www.FamilyProcess.org

FAMILY PROCESS
STADELMANN, PERREN, GROEBEN, & VON KLITZING / 99
Stability
To analyze stability of family conflict, parental representations, and behavioral/
emotional problems over 1 year, we computed Pearson correlations coefficients
(Table 1). Family conflict showed high stability between age 5 and 6. Negative
parental representations showed moderate stability between 5 and 6. Behavioral/
emotional problems were also highly stable between 5 and 6.

Frequency of behavioral/emotional problems

Based on the SDQ norms (http://www.sdqinfo.com; Goodman, 1997), at the age of 5,
36% of children (N ¼ 57) of the community sample (77% [N ¼ 23] of the clinical sample)
were in the clinical range on at least one of the symptom scales, according to parent and/
or teacher reports. In detail, 21% of children (N ¼ 33) of the community sample (20%
[N ¼ 6] of the clinical sample) had behavioral problems only (i.e., conduct problems and/
or hyperactivity/inattention); 9% (N ¼ 14) of the community sample (17% [N ¼ 5] of the
clinical sample) had emotional symptoms only; and 6% (N ¼ 10) of the community sample
(40% [N ¼ 12] of the clinical sample) had both behavioral and emotional problems.

Group differences of separated and nonseparated families

To analyze group differences between separated and nonseparated families, we
computed univariate variance analyses (see Table 2). Children who experienced
parental separation showed significantly more emotional symptoms and, with a trend
toward significance, more conduct problems at age 6 compared with children
of nonseparated families. No group differences were found concerning behavioral/
emotional problems at the age of 5. Furthermore, children of separated families
showed a significantly higher level of negative parental representations in their play
narratives at age 5, but no group differences were found concerning parental re-
presentations at the age of 6. In addition, parents of separated families reported sig-
nificantly more family conflict compared with parents of nonseparated families when

TABLE 2
Group Differences Nonseparation vs. Separation

Nonseparated Separated

M SD M SD F p
Symptoms age 5
Conduct problems  0.04 .69 .14 .80 1.829 .178
Hyperactivity  0.01 .72 .13 .95 0.889 .347
Emotional symptoms  0.01 .70 .16 .93 1.325 .251
Symptoms age 6
Conduct problems  0.04 .66 .23 .82 3.705 .056
Hyperactivity  0.04 .68 .14 .79 1.621 .205
Emotional symptoms  0.04 .68 .24 .78 4.128 .044
Family conflict
Age 5 2.15 .47 2.36 .55 5.337 .022
Age 6 2.14 .46 2.17 .45 0.112 .738
Negative parental representations
Age 5 0.03 .08 .07 .18 5.032 .026
Age 6 0.04 .08 .05 .11 0.463 .497

Fam. Proc., Vol. 49, March, 2010

100 / FAMILY PROCESS

children were 5 years old. Again, no group differences were found concerning family
conflict at the age of 6.

Gender differences
To analyze gender differences, we calculated univariate variance analyses. At both
ages boys showed higher levels of conduct problems (age 5: girls: M ¼  .23, SD ¼ .59,
boys: M ¼ .14, SD ¼ .75; F ¼ 12.37, p ¼ .001; age 6: girls: M ¼  .24, SD ¼ .57, boys:
M ¼ .18, SD ¼ .75; F ¼ 16.58, p ¼ .000) and of hyperactivity than girls (age 5: girls:
M ¼  .16, SD ¼ .73, boys: M ¼ .14, SD ¼ .77; F ¼ 6.68, p ¼ .011; age 6: girls: M ¼  .23,
SD ¼ .58, boys: M ¼ .17, SD ¼ .74; F ¼ 15.14, p ¼ .000). Girls and boys did not differ
from each other in terms of emotional symptoms. In addition, boys presented
significantly more negative parental representations in their play narratives than
girls at the age of 6 (girls: M ¼ .02, SD ¼ .05, boys: M ¼ .06, SD ¼ .10; F ¼ 10.47,
p ¼ .001). At the age of 5 this difference only showed a trend toward significance (girls:
M ¼ .02, SD ¼ .06, boys: M ¼ .05, SD ¼ .12; F ¼ 3.03, p ¼ .083). No gender differences
were found regarding family conflict.

Multivariate Analyses with Interaction Effects

To analyze the predictive value of parental separation, family conflict, and negative
parental representations on changes of behavioral/emotional problems between
the age of 5 and 6, and to test whether family conflict and negative parental
representations have an impact on the strength of the effect of separation on beha-
vioral/emotional problems, we computed General Linear Models with interaction
effects (see Table 3). As explained in the methods, we categorized family conflict and
negative parental representations (low vs. high levels of family conflict; no negative
parental representations vs. child shows negative representations at least at one time
point). Parental separation, family conflict, negative parental representations as main
effects, and the interaction effects between parental separation and family conflict and
between parental separation and negative parental representations were all analyzed as
predictors of changes in behavioral/emotional problems between the ages of 5 and 6.
Gender and behavioral/emotional problems at age 5 were included as control variables.

TABLE 3
Multivariate Associations with Interaction Effects (N ¼ 154)

Symptoms age 6

Conduct Emotional
problems Hyperactivity symptoms

Predictors F p F p F p

Symptoms age 5 74.013 .000 166.171 .000 70.059 .000

Gender 4.087 .045 10.772 .001 0.109 .742
Family conflict (F) 5.302 .023 0.422 .517 0.933 .336
Negative parental representation. (N) 4.668 .032 0.084 .772 0.284 .595
Parental separation (PS) 6.589 .011 1.967 .163 2.107 .149
PS  F 0.030 .864 0.344 .559 0.067 .796
PS  N 3.927 .049 0.007 .934 0.542 .463

Note: Conduct problems: R2 ¼ .49; hyperactivity: R2 ¼ .59; emotional symptoms: R2 ¼ .41.

www.FamilyProcess.org
STADELMANN, PERREN, GROEBEN, & VON KLITZING / 101
Concerning conduct problems, we found that the main effects of symptoms at age 5,
gender (girls: M ¼  .20, SD ¼ .58, boys: M ¼ .17, SD ¼ .75), parental separation
(nonseparated: M ¼  .04, SD ¼ .65, separated: M ¼ .24, SD ¼ .84), family conflict (low
level: M ¼  .12, SD ¼ .60, high level: M ¼ .39, SD ¼ .86), and negative parental re-
presentations (none: M ¼  .06, SD ¼ .66; at least once: M ¼ .14, SD ¼ .76), were sig-
nificant predictors of symptoms at the age of 6. In addition, we found a significant
interaction effect between separation and negative parental representations in pre-
dicting conduct problems. As can be seen in Figure 1, children from families where the
parents had separated who showed negative parental representations in their nar-
ratives also showed a significantly higher increase in conduct problems at the age of 6
compared with children of separated families without negative representations, and
compared with children of nonseparated families.
Concerning hyperactivity, corresponding symptoms at the age of 5 and gender were
significant predictors of symptoms at the age of 6. None of the other variables were
found to be significant predictors.
Concerning emotional symptoms at the age of 6, symptoms at the age of 5 was the
only predictor of symptoms 1 year later.

DISCUSSION
Our hypothesis that parental separation, family conflict, and children’s negative
parental representations would predict a greater increase in behavioral/emotional
problems at kindergarten age was partially supported. Our longitudinal study showed
that all three of these aspects were risk factors for an increase in conduct problems
during the kindergarten years. The effect of parental separation was particularly
strong if children also had negative parental representations.

Conduct Problems
In support of our hypothesis, multivariate analyses showed that parental separation
predicted higher levels of conduct problems of children at the age of 6, and explained an
increase in conduct problems between 5 and 6. We did not find that this effect varied

0.8
non-separated /
0.7 no negative represent.
(N=81)
z-scores: conduct problems

0.6
non-separated /
0.5 negative represent.
(N=43)
0.4 separated /
no negative represent.
0.3 (N= 20)
0.2 separated /
negative represent.
0.1 (N=10)

0
–0.1
–0.2
age 5 age 6
FIGURE 1. Conduct problems (age 5 and age 6) predicted by parental separation and negative
parental representations (age 5/6).

Fam. Proc., Vol. 49, March, 2010

102 / FAMILY PROCESS

according to the level of family conflict. This is partly in line with Morrison and Coiro
(1999) and Cheng, Dunn, O’Connor, Golding, and Kingdom (2006), who showed that
parental separation predicted an increase in behavioral/emotional problems in 4–9-year-
old children over and above the effect of preseparation parental quarrels. Nonetheless,
like Morrison and Coiro (1999), we found that family conflict is an additional significant
predictor of an increase in conduct problems at kindergarten age.
In support of our hypothesis, negative representations (at age 5 and/or 6) turned out to
be a further additional predictor of an increase in conduct problems between 5 and 6.
Moreover, children’s representations also had an impact on the strength of the association
between separation and conduct problems. Those children of separated parents who de-
scribed the behavior of parental figures toward child figures in their play narratives as
harsh, punitive, rejecting, or ineffectual had a significantly higher increase in conduct
problems between the ages of 5 and 6 than did other children, while the absence of ne-
gative parental representations seems to be protective for children’s development after the
experience of their parents’ separation. This result suggests that the quality of internal
representations of relationships in the face of parental separation has a moderational
impact on the development of conduct problems at an early state of development. Ac-
cording to social-cognitive models, children internalize important aspects of care-giving
experiences, which guide their behavior in new social situations (Bretherton, 1990;
Lemerise & Arsenio, 2000). Thus, children’s accounts of parental behavior might be ex-
pected to correspond to their experiences within their own families (e.g., Grych et al., 2002;
Shamir et al., 2001). Generally, children need parents who are warm, supportive, and
responsive to their needs, who exert structured positive discipline (Eisenberg & Fabes,
1998), and who manage conflict well, with the ability to build an interparental consensus
to coparent their children effectively, not allowing children to get caught up in the par-
ental acrimony (Ahrons, 2007; Grych & Fincham, 1990; Hetherington & Stanley Hagan,
1999). Such positive parenting might also have a positive impact on children’s internal
representations of social conflict and on their behavioral adjustment following separation.
In our study, children’s parental representations were significantly associated with par-
ental separation but not with parents’ ratings of the amount or quality of family conflict.
This result is not consistent with the findings of the studies mentioned above. On the one
hand, family conflict, assessed by a questionnaire, may be a relatively imperfect measure
of actual family relationships and parents’ reports on family conflict are subject to a
possible ‘‘social desirability’’ effect and might not reflect actual family life. On the other
hand, however, the result suggests that children’s representations are not simply a mirror
of their actual family life. They also reflect the children’s own way of comprehending and
coping with social situations. As such, they represent the at least partially independent
emotional and cognitive processes of the child, and are markedly affected by the child’s
gender (Oppenheim et al., 1997b). Representations might also be an expression of chil-
dren’s competence in regulating their emotions when confronted with a stressful stimulus
(Clyman, 2003). Furthermore, it is important to keep in mind that the as-if nature of play
offers the child an opportunity to test behaviors and to act out aggressive impulses and
wishes in a safe space (Marans, Dahl, Marans, & Cohen, 1993).

Hyperactivity/Inattention
None of our analyses pointed to separation as a risk factor for hyperactivity, and only
in bivariate analyses did we find significant positive associations with family conflict and

www.FamilyProcess.org
STADELMANN, PERREN, GROEBEN, & VON KLITZING / 103
negative parental representations. In multivariate analyses, gender and hyperactivity at
age 5 were the only predictors for hyperactivity 1 year later. This might indicate on the
importance of neurobiological explanations for hyperactivity disorder (Nigg & Casey,
2005). On the other hand, the results are not in line with studies of family context and
hyperactivity, which show that parents of children with attention deficit/hyperactivity
disorder experience higher levels of stress, are more likely to display negative behavior
toward their children (DuPaul, McGoey, Eckert, & VanBrakle, 2001), and more often
separate (Barkley, 1990) than parents of control children. Despite the different
associations of conduct problems and hyperactivity in our study, it is important to
recognize that there is a significant overlap between conduct problems and hyperactivity
symptoms. It might therefore be that the associations of separation, negative parental
representations, and family conflict with behavioral problems such as conduct problems
and hyperactivity are not symptom specific.

Emotional Symptoms
In bivariate analyses, we found that children of separated parents and with high
levels of conflict in their families showed more emotional symptoms than other chil-
dren. But again, in multivariate analyses emotional symptoms at the age of 5 were the
only predictor of symptoms 1 year later. Developmental research has indeed identified
that cognitive-emotional and family environment risk factors have an impact on the
origin and development of emotional symptoms in children (Zahn-Waxler et al., 2000;
Zalsman, Brent, & Weersing, 2006). Nevertheless, the results are far more incon-
sistent compared with those of behavioral problems (Amato & Keith, 1991; Hether-
ington & Stanley Hagan, 1999; Oppenheim et al., 1997b; von Klitzing et al., 2000). One
explanation for the observation that none of the family environment factors were
relevant for emotional symptoms in our study is that emotional symptoms were less
frequent and less variable compared with conduct problems. This might also reflect
the problem that emotional symptoms are harder to detect than behavioral problems
in young children. Moreover, despite ‘‘the hope that narrative methods will go beyond
the external manifestation of children’s problems and provide an entrée to the
internal experience of young children’’ (Oppenheim et al., 1997b, p. 291), it turned out
to be difficult to detect narrative characteristics of emotional symptoms. In our study,
children with emotional symptoms might have been less likely to use parental figures
in their play narratives. Instead, performance aspects in narratives could be of special
importance in research on emotional symptoms. In order to regulate negative feelings
when confronted with conflict themes in story stems, children with depressive or
anxiety symptoms might modify or avoid talking about certain things (Clyman, 2003;
Stadelmann, 2006).

Methodological Limitations and Strengths

In this paper, we are interested in the impact of parental separation as an im-
portant life event on children’s socioemotional development. Parental separation is
just one step in a series of family transitions that affect family relationships and
children’s adjustment. Children’s family and life situations change during and after
separation. The interactions between various individual, family, and extra-familial
risk and protective factors that undermine or promote the well-being of children,
produce diverse developmental trajectories (Hetherington & Stanley Hagan, 1999).

Fam. Proc., Vol. 49, March, 2010

104 / FAMILY PROCESS

We focused on two possible risk factorsFfamily conflict and children’s parental

representations in the context of socioemotional dilemmasFand analyzed whether
they have an impact on the strength of the association between separation and chil-
dren’s adjustment. Because of our sample size, we did not take into account further
details of the family or living situation (e.g., new partner, quantity/quality of contact
of both parents after separation). Another limitation of this study concerns the as-
sessment of family conflict. We did not control which family system parents focused on
when answering the questionnaire. In separated families, this could have been the
family system with a new partner.
We also found higher levels of negative representations and family conflict in se-
parated families (compared with nonseparated families) only at the age of 5 but not at
the age of 6. Even though this was an unexpected finding the result indicates that after
a parental separation family relationships do not necessarily remain highly conflictual
but can also ease and move toward a more harmonious situation (Amato et al., 1995).
On the other hand, our results showed significant differences in behavioral/emo-
tional problems between separated and nonseparated families at the age of 6 but not
at the age of 5 (even though there was stability in behavioral/emotional problems
across ages). We might speculate whether this is an indication of a delayed effect,
which has been frequently demonstrated in the divorce literature (e.g., Kelly, 2000).
So far, the majority of studies on children’s narratives have been conducted in
community samples (Robinson, 2007). In this study, we included a subsample of 30
children who had been referred to our outpatient department by their parents and/or
teachers. Children of this sample were assessed in our research department whereas
children of the community sample were assessed in their kindergarten or at home.
There were substantially more children with clinically relevant behavioral/emotional
problems in the clinical subsample and families in this sample experienced a higher
level of conflict. In contrast, negative parental representations did not differ in the
referred and the nonreferred sample. We speculate that in the clinical setting, chil-
dren might be more inhibited. To assess play narratives accurately in clinical set-
tingsFin comparison with research settings such as the kindergartenFchildren
might need more emotional support in order to feel comfortable and free to narrate.
On the other hand, the result might also be due to the large number of children in the
nonreferred sample (36%) who also exhibited clinically relevant behavioral/emotional
problems. Because of the different recruitment schemes of the samples we saw the
need to control for study affiliation, but found that results remained the same.
Separation, family conflict, and children’s parental representations are not the only
influences on children’s behavioral/emotional problems in kindergarten age. It is clear
that behavioral/emotional problems are multidetermined and that a comprehensive
account will include numerous biological as well as other socializing factors, such as
parental psychopathology and parents’ educational level. In our study, we also found
that behavioral/emotional problems at the age of 5 were always the strongest
predictors for behavioral/emotional problems in the following year. One of these
determinating factors, for which we also controlled, was the children’s gender. At both
ages, boys showed more conduct problems and hyperactivity. Male gender was even
able to account for an increase in hyperactivity and conduct problems between
the ages of 5 and 6. Metaanalytic research has found modest evidence that boys are
more strongly affected by their parents’ separation than girls (Amato, 2001).
However, in longitudinal studies girls were found to be affected at an older age than

www.FamilyProcess.org
STADELMANN, PERREN, GROEBEN, & VON KLITZING / 105
boys (e.g., Oldehinkel, Ormel, Veenstra, De Winter, & Verhulst, 2008). It would have
been important to investigate gender as a moderator in this study. However, our
sample size was too small to include another interaction term.

Clinical Implications
Systematic insights into children’s ways of dealing with relevant themes and con-
flict in their play narratives, and the finding that the quality of caregiver re-
presentations contributed to a differential outcome in children of separated parents,
can inform our psychotherapeutic techniques. Most families who require assistance in
negotiating the divorce process respond to low-intensity interventions. The programs
generally offer psychoeducation for parents and focus on specific communication,
conflict-resolution, and parenting skills. Some of these programs have been shown to
be effective on children’s adjustment as well as on the quality of parent-child re-
lationships (e.g., Parent Management Training [PMT], Patterson, DeGarmo, & For-
gatch, 2004; New Beginnings Program [NBP], Wolchik et al., 2000). However,
evidence provided in this study underlines to practitioners the relevance of an early
therapeutic engagement based on the complete family system, and the need to con-
sider young children in their family therapeutic enactments, and to support the ca-
pacity of these young children to deal with family conflict in a constructive way. Grych
(2005) notes that even though parents bear primary responsibility for managing
postdivorce conflict, programs for children may be useful particularly if they help
children to develop skills for coping with situations where they may be caught up in
parental patterns of negative attribution and interparental conflict. Children’s pro-
grams are not as widespread as parent-focused programs, and focus mostly on school-
age children. Sometimes they form part of a program that includes a group for parents
(e.g., Wolchik et al., 2000), or may be school-based groups (e.g., Pedro-Carroll, 1997).
Surprisingly, the dual-component programs have demonstrated little positive effect
beyond that gained when only parents participate in the group (Wolchik et al., 2000).
In contrast, the school-based Children of Divorce Intervention Program demonstrated
both short- and long-term effects on children’s adjustment (Pedro-Carroll, 1997).
Thus, a definite conclusion about the efficacy of child-oriented prevention programs
cannot yet be drawn. Another example of a treatment which focuses on parents and
children is the Integrative Family Therapy for Disputes involving Child Custody and
Visitation (IFT-DCCV). This family-based approach provides intense therapy for high-
conflict families who do not respond to low-intensity treatments, incorporating mul-
tiple therapy session formats, with a solution-oriented focus and drawing upon a wide
range of intervention techniques such as psychoeducation, mediation, cognitive, and
narrative techniques (Lebow & Rekart, 2007). The specific intervention strategies for
children in this program are tailored to the children’s age. For young children the
authors recommend the use of stories that enable feelings to be processed in fantasy as
launching points for dialogue. Overall, our study underlines the relevance of ap-
proaches that integrate play and narration to promote mental health in young chil-
dren who have experienced a parental separation.

REFERENCES
Ahrons, C.R. (2007). Family ties after divorce: Long-term implications for children. Family
Process, 46(1): 53–65.

Fam. Proc., Vol. 49, March, 2010

106 / FAMILY PROCESS

Amato, P.R. (2001). Children of divorce in the 1990s: An update of the Amato and Keith (1991)
meta-analysis. Journal of Family Psychology, 15(3): 355–370.
Amato, P.R., & Keith, B. (1991). Parental divorce and the well-being of children: A meta-ana-
lysis. Psychological Bulletin, 110(1): 26–46.
Amato, P.R., Loomis, L.S., & Booth, A. (1995). Parental divorce, marital conflict, and offspring
well-being during early adulthood. Social Forces, 73(3): 895–915.
Barkley, R.A. (1990). Attention-deficit hyperactivity disorder: A handbook for diagnosis and
treatment. New York: Guilford Press.
Bretherton, I. (1990). Communication patterns, internal working models, and the inter-
generational transmission of attachment relationships. Infant Mental Health Journal, 11(3):
237–252.
Bretherton, I., & Oppenheim, D. (2003). The MacArthur Story Stem Battery: Development,
administration, reliability, validity, and reflections about meaning. In R.N. Emde, D.P. Wolf,
& D. Oppenheim (Eds.), Revealing the inner worlds of young children. The MacArthur Story
Stem Battery and parent-child-narratives (pp. 55–80). New York: Oxford University Press.
Cheng, H., Dunn, J., O’Connor, T.G., Golding, J., & Kingdom, A.S.T.B.U. (2006). Factors
moderating children’s adjustment to parental separation: Findings from a community study
in England. Journal of Abnormal Child Psychology, 34(2): 239–250.
Clyman, R.B. (2003). Portrayals in maltreated children’s play narratives: Representations or
emotion regulation? In R.N. Emde, D.P. Wolf, & D. Oppenheim (Eds.), Revealing the inner
worlds of young children. The MacArthur Story Stem Battery and parent-child narratives
(pp. 201–221). New York: Oxford University Press.
Cummings, E.M., & Davies, P.T. (2002). Effects of marital conflict on children: Recent advances
and emerging themes in process-oriented research. Journal of Child Psychology and Psy-
chiatry, 43(1): 31–63.
Davies, P.T., & Cummings, E.M. (1994). Marital conflict and child adjustment: An emotional
security hypothesis. Psychological Bulletin, 116(3): 387–411.
DuPaul, G.J., McGoey, K.E., Eckert, T.L., & VanBrakle, J. (2001). Preschool children with
attention-deficit/hyperactivity disorder: Impairments in behavioral, social, and school
functioning. Journal of the American Academy of Child and Adolescent Psychiatry, 40(5):
508–515.
Eisenberg, N., & Fabes, R.A. (1998). Prosocial development. In N. Eisenberg (Ed.), Social,
emotional and personality development (5th ed., Vol. 3, pp. 701–778). New York: John Wiley
& Sons.
Fincham, F.D. (1998). Child development and marital relations. Child Development, 69(2): 543–
574.
Goodman, R. (1997). The strengths and difficulties questionnaire: A research note. Journal of
Child Psychology and Psychiatry, 38(5): 581–586.
Grych, J.H. (2005). Interparental conflict as a risk factor for child maladjustment: Implications
for the development of prevention programs. Family Court Review, 43, 97–108.
Grych, J.H., & Fincham, F.D. (1990). Marital conflict and children’s adjustment: A cognitive-
contextual framework. Psychological Bulletin, 108(2): 267–290.
Grych, J.H., Fincham, F.D., Jouriles, E.N., & McDonald, R. (2000). Interparental conflict and
child adjustment: Testing the mediational role of appraisals in the cognitive-contextual
framework. Child Development, 71(6): 1648–1661.
Grych, J.H., Wachsmuth Schlaefer, T., & Klockow, L.L. (2002). Interparental aggression and
young children’s representations of family relationships. Journal of Family Psychology,
16(3): 259–272.
Hetherington, E.H., & Stanley Hagan, M. (1999). The adjustment of children with divorced
parents: A risk and resiliency perspective. Journal of Child Psychology and Psychiatry, 40(1):
129–140.

www.FamilyProcess.org
STADELMANN, PERREN, GROEBEN, & VON KLITZING / 107
Kelly, J.B. (2000). Children’s adjustment in conflicted marriage and divorce: A decade review of
research. Journal of the American Academy of Child and Adolescent Psychiatry, 39(8): 963–
973.
Kraemer, H.C., Measelle, J.R., Ablow, J.C., Essex, M.J., Boyce, W.T., & Kupfer, D.J. (2003). A
new approach to integrating data from multiple informants in psychiatric assessment and
research: Mixing and matching contexts and perspectives. American Journal of Psychiatry,
160(9): 1566–1577.
Lebow, J., & Rekart, K.N. (2007). Integrative family therapy for high-conflict divorce with
disputes over child custody and visitation. Family Process, 46(1): 79–91.
Lemerise, E.A., & Arsenio, W.F. (2000). An integrated model of emotion processes and cognition
in social information processing. Child Development, 71(1): 107–118.
Marans, S., Dahl, E.K., Marans, W., & Cohen, D.J. (1993). Aggressivity in play: Discussions with
oedipal children. In A.J. Solnit, D.J. Cohen, & P.B. Neubauer (Eds.), The many meanings of
play: A psychoanalytic perspective (pp. 275–296). New Haven: Yale University Press.
Measelle, J.R., Ablow, J.C., Cowan, P.A., & Cowan, C.P. (1998). Assessing young children’s
views of their academic, social, and emotional lives: An evaluation of the self-perception
scales of the Berkeley puppet interview. Child Development, 69(6): 1556–1576.
Moos, R.H., & Moos, B.S. (1981). Family Environment Scale. Manual. Palo Alto: Consulting
Psychologists Press.
Morrison, D.R., & Coiro, M.J. (1999). Parental conflict and marital disruption: Do children
benefit when high-conflict marriages are dissolved? Journal of Marriage and the Family,
61(3): 626–637.
Nigg, J.T., & Casey, B.J. (2005). An integrated theory of attention deficit/hyperactivity disorder
based on the cognitive and affective neurosciences. Development & Psychopathology, 17, 785–
806.
Oldehinkel, A.J., Ormel, J., Veenstra, R., De Winter, A.F., & Verhulst, F.C. (2008). Parental
divorce and offspring depressive symptoms: Dutch developmental trends during early ado-
lescence. Journal of Marriage and Family, 70(2): 284–293.
Oppenheim, D., Emde, R.N., & Warren, S.L. (1997a). Children’s narrative representations of
mothers: Their development and associations with child and mother adaptation. Child De-
velopment, 68(1): 127–138.
Oppenheim, D., Nir, A., Warren, S.L., & Emde, R.N. (1997b). Emotion regulation in mother-
child narrative co-construction: Associations with children’s narratives and adaptation.
Developmental Psychology, 33(2): 284–294.
Patterson, G.R., DeGarmo, D., & Forgatch, M.S. (2004). Systematic changes in families fol-
lowing prevention trials. Journal of Abnormal Child Psychology, 32(6): 621–633.
Pedro-Carroll, J.L. (1997). The children of divorce intervention program: Fostering resilient
outcomes for school-aged children. In G.W. Albee & T. Gullotta (Eds.), Primary prevention
works (pp. 213–238). Thousand Oaks, CA: Sage.
Perren, S., von Wyl, A., Simoni, H., Stadlmayr, W., Bürgin, D., & von Klitzing, K. (2003).
Parental psychopathology, marital quality, and the transition to parenthood. American
Journal of Orthopsychiatry, 73(1): 55–64.
Perren, S., von Wyl, A., Stadelmann, S., Bürgin, D., & von Klitzing, K. (2006). Associations
between behavioral/emotional difficulties in kindergarten children and the quality of their
peer relationships. Journal of the American Academy of Child and Adolescent Psychiatry,
45(7): 867–876.
Robinson, J. (2007). Story stem narratives with young children: Moving to clinical research and
practice. Attachment and Human Development, 9(3): 179–185.
Robinson, J., & Mantz-Simmons, L. (2003). The MacArthur Narrative Coding System: One
approach to highlighting affective meaning making in the MacArthur Story Stem Battery. In
R.N. Emde, D.P. Wolf, & D. Oppenheim (Eds.), Revealing the inner worlds of young children.

Fam. Proc., Vol. 49, March, 2010

108 / FAMILY PROCESS

The MacArthur Story Stem Battery and parent-child-narratives (pp. 81–105). New York:
Oxford University Press.
Schneewind, K.A., Beckmann, M., & Hecht-Jackl, A. (1985). Familiendiagnostisches Testsystem
(Vol. Bericht 8.2): Forschungsberichte aus dem Institutsbereich Persönlichkeitspsychologie
und Psychodiagnostik. Universität München, Institut für Psychologie.
Shamir, H., Du Rocher Schudlich, T., & Cummings, E.M. (2001). Marital conflict, parenting
styles, and children’s representations of family relationships. Parenting: Science and Prac-
tice, 1(12): 123–151.
Shields, A., Ryan, R.M., & Cicchetti, D. (2001). Narrative representations of caregivers and
emotion dysregulation as predictors of maltreated children’s rejection by peers. Develop-
mental Psychology, 37(3): 321–337.
Stadelmann, S. (2006). Konfliktbasierte Spielnarrative von Kindergartenkindern: Zu-
sammenhänge mit der Verhaltensregulation und dem familiären Umfeld. Unpublished Dis-
sertation, from http://www.freidoc.uni-freiburg.de/volltexte/2819; accessed December 2009.
Stadelmann, S., Perren, S., von Wyl, A., & von Klitzing, K. (2007). Associations between family
relationships and symptoms/strengths at kindergarten age: What is the role of children’s
parental representations? Journal of Child Psychology and Psychiatry, 48(10): 996–1004.
von Klitzing, K., Kelsay, K., Emde, R.N., Robinson, J., & Schmitz, S. (2000). Gender-specific
characteristics of 5-year-olds’ play narratives and associations with behavior ratings. Journal
of the American Academy of Child and Adolescent Psychiatry, 39(8): 1017–1023.
Wolchik, S.A., West, S.G., Sandler, I.N., Tein, J., Coatsworth, D., Lengua, L., et al. (2000). An
experimental evaluation of theory-based mother and mother-child programs for children of
divorce. Journal of Consulting and Clinical Psychology, 68, 843–856.
Zahn-Waxler, C., Klimes Dougan, B., & Slattery, M.J. (2000). Internalizing problems of child-
hood and adolescence: Prospects, pitfalls, and progress in understanding the development of
anxiety and depression. Development and Psychopathology, 12(3): 443–466.
Zalsman, G., Brent, D.A., & Weersing, V.R. (2006). Depressive disorders in childhood and
adolescence: An overview epidemiology, clinical manifestation and risk factors. Child and
Adolescent Psychiatric Clinics of North America, 15(4): 827–841.

www.FamilyProcess.org