Child and Adolescent Mental Health Volume **, No. *, 2018, pp. **–** doi:10.1111/camh.

12269

The direct and indirect effect of attachment

insecurity and negative parental behavior on anxiety
in clinically anxious children: it’s down to dad
Sonja Breinholst, Marie Tolstrup & Barbara Hoff Esbjørn
Center for Anxiety, Department of Psychology, University of Copenhagen, Oester Farimagsgade 2A, 1353 Copenhagen,
Denmark. E-mail: Sonja.breinholst@psy.ku.dk

Background: Theoretically, insecure attachment and negative parental behaviors are risk factors for childhood
anxiety. However, few empirical studies have examined their relative contribution including differences
between mothers and fathers. To date, only one study has examined a mediational model including these fac-
tors, albeit in a nonclinical sample. Method: This study ameliorates this limitation by investigating direct and
indirect relations between maternal and paternal attachment and behaviors, and clinical anxiety in children
(mean age 9.6 years). The study recruited 54 families. Anxiety symptoms were measured by the Spielberger
State-trait Inventory for Children, attachment relationships by the Security Scale, and parental behaviors using
the Rearing Behavior Questionnaire. Results: Neither insecure attachment relationship with mother nor
maternal negative behavior was a significant predictor of anxiety in children. However, insecure attachment
to father accounted for a significant proportion of variance in anxiety, and also fully mediated the relationship
between paternal rejection and anxiety. Conclusions: Our findings highlight the importance of assessing the
relative contribution of risk factors and the importance of including fathers when investigating the develop-
ment and maintenance of childhood anxiety.

Key Practitioner Message

• Fathers and mothers might play different roles regarding childhood anxiety.
• Paternal variables might need to be included in assessment of childhood anxiety and anxiety case formulation.
• Fathers might need to be included in the treatment of anxiety.

Keywords: Anxiety; parent-child interaction; attachment; mediation

approval, and responsiveness. It’s hypothesized that

Introduction
Anxiety is a highly prevalent and persistent psychiatric
disorder in childhood and adolescence (Kessler & Wang,
2008). Approximately 10% of all children meet the diag-
nostic criteria for an anxiety disorder during childhood
(Copeland, Angold, Shanahan, & Costello, 2014). Conse-
quences of anxiety include negative influences on fami-
lies (Lebowitz et al., 2013) and societal financial burdens
(Bodden, Dirksen, & B€ ogels, 2008), making identifying
risk factors leading to anxiety pivotal. This study investi-
gates the unique influence and mutual relation between
two risk factors well-established in the literature: inse-
cure attachment and negative parental behaviors. It also
assesses these in relation to mothers and fathers,
respectively.
Negative parental behaviors and anxiety
Risk factors such as negative parental behaviors have
received theoretical and empirical attention. Rejection,
overprotection, and psychological control are theoreti-
cally associated with childhood anxiety disorders
(McLeod, Wood, & Avny, 2011). Parental rejection is
characterized by reduced levels of parental warmth,
parental rejection may cause children to develop nega-
tive assumptions about themselves and their competen-
cies as well as negative/threatening expectations of the
world, increasing the risk of developing anxiety (Cres-
well, Murray, Stacey, & Cooper, 2011). Opposite to par-
ental rejection, parental warmth and acceptance
behaviors signal that assistance is available, reducing
fear and anxiety (B€ogels & Brechman-Toussaint, 2006).
Theoretically, parental overprotection (rather than
autonomy granting) minimizes experiences with unex-
pected or difficult situations, reducing children’s devel-
opment of self-efficacy and feelings of mastery and
control. This is hypothesized to increase children’s
perceptions of threat and danger (B€ ogels & Brechman-
Toussaint, 2006). Finally, psychological control is
characterized by excessive parental vigilance and intru-
sion. Its interference with children’s experiences of
control is also considered to induce fear and anxiety
(Thirlwall & Creswell, 2010).
Theoretical associations between negative parental
behaviors and anxiety in children and adolescents have
been supported by three empirical reviews (McLeod,
Wood, & Weisz, 2007; Van der Bruggen, Stams, &

© 2018 Association for Child and Adolescent Mental Health.

Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA
2 Sonja Breinholst, Marie Tolstrup & Barbara Hoff Esbjørn
B€ogels, 2008; Waite, Whittington, & Creswell, 2014).
Van der Bruggen et al. (2008) found moderate effect
sizes for negative parental behaviors, with the strongest
for overprotection (d = .93), psychological control
(d = .52), rejection (d = .41). This study corroborated
previous findings of parental control as more strongly
associated with anxiety than parental rejection (McLeod
et al., 2007). Waite et al. (2014) found consistent evi-
dence for an association between anxiety and parental
control in primarily nonclinical adolescent samples,
whereas evidence for a relationship between anxiety and
parental rejection was less consistent. Despite theoreti-
cal emphasis on negative parental behaviors as an
important contributing factor in childhood anxiety, the
overall association between childhood anxiety and nega-
tive parental behaviors is modest, explaining only 4% of
variance in anxiety (McLeod et al., 2007). Consequently,
attention has turned to other factors within the
parent-child relationship, for example, child-parent
attachment.
Insecure attachment and anxiety disorder
According to attachment theory, the quality of the par-
ent-child attachment relationship more likely impacts
the psychological development of the child (Bowlby,
1973; ¸ Cassidy & Berlin, 1994). Children’s susceptibility
to fear and anxiety is hypothesized to depend on the
physical/emotional availability and sensitivity of their
caregiver (Bowlby, 1973). A secure attachment relation-
ship is considered pivotal for a healthy development. It is
formed by caregivers’ consistent sensitivity to children’s
attachment behaviors, serving as a secure base from
which children can explore their surroundings, and a
safe haven to which children can return when dis-
tressed. However, in insecure attachment relationships,
the caregiver is unable or unwilling to consistently and
sensitively react to these behaviors. The absence of the
caregiver’s availability may cause chronic anxiety
(Bowlby, 1973; Cassidy & Berlin, 1994). Thus, insecure
attachment is supposedly associated with fear and anxi-
ety (Bosquet & Egeland, 2006; Bowlby, 1958).
Recently, empirical evidence for the theoretical link
between insecure attachment and anxiety has been eval-
uated in four reviews (Brumariu & Kerns, 2010; Colon-
nesi et al., 2011; Groh, Roisman, van IJzendoorn,
Bakermans-Kranenburg, & Fearon, 2012; Madigan,
Atkinson, Laurin, & Benoit, 2013). An association
between insecure attachment and higher anxiety levels
in children and adolescents was found, but results were
mixed and inconsistent. These inconsistencies might be
due to methodological differences such as which type of
insecure attachment was included, and how attachment
was measured. Furthermore, the reviews differed with
regard to whether anxiety symptoms, anxiety disorders,
or internalizing disorders were investigated. Last, differ-
ent age groups were investigated in the different reviews.
Colonnesi et al. (2011) assessed only secure and inse-
cure ambivalent attachment and anxiety symptoms in
children, finding effect sizes to be moderate (r = .30).
Brumariu and Kerns (2010), Groh et al. (2012), and
Madigan et al. (2013) investigated ambivalent, avoidant,
and disorganized attachment and internalizing symp-
toms, not anxiety symptoms specifically. Groh et al.
(2012), as well as Madigan et al. (2013) measured
attachment in infancy through observational studies
Child Adolesc Ment Health 2018; *(*): **–**
only, whereas Colonnesi et al. (2011), and Brumariu
and Kerns (2010) also included questionnaires, projec-
tive task and interviews in children up to 18 years of age.
Groh et al. (2012), and Madigan et al. (2013) only
included mother-child attachment studies, while Colon-
nesi et al. (2011), and Brumariu and Kerns (2010)
included studies on attachment to both parents, peers
and attachment figures in general. The association
between internalizing symptoms and attachment in the
meta-analysis by Groh et al. (2012) was smaller than
previously reported (d = .15) and was explained by inse-
cure avoidant attachment (d = .17). Madigan et al.
(2013) corroborated this finding. Both studies contradict
the findings by Brumariu and Kerns (2010), who
reported that the effect of avoidant attachment disap-
peared when analyses included disorganized attach-
ment. A final limitation in the reviews is that most
participants included in the analyses were nonclinically
anxious, for example, four of 61 studies in the analysis
by Madigan et al. (2013), and six studies out of 55 in the
review by Brumariu and Kerns (2010) were clinical,
thus, potentially reducing the validity of findings for clin-
ically anxious children. Inconsistent findings and
diverse methods therefore warrant further investigation.
The interaction between negative parental
behaviors and attachment
Parental behaviors in the child’s first year will lead to the
child displaying a specific pattern in the attachment
behavior (Cassidy & Berlin, 1994). Hence, parental dis-
play of neglectful behavior may cause children to either
withdraw or increase attempts at proximity to the par-
ent. Both behaviors indicate insecure attachment (Van
IJzendoorn, 1995). Similarly, parental over-control may
reduce openness to children’s needs and emotional
expressions, triggering dependency from the child due to
low self-efficacy or withdrawal in order to gain control.
Theoretically associated with insecure attachment, this
may be an antecedent to anxiety (Van IJzendoorn,
1995). Empirical testing of these suggestions requires a
model testing the indirect relationship between attach-
ment relationship and parental behavior.
Although theoretical models of childhood anxiety
often include multiple risk factors, few studies have
empirically investigated relative contributions of nega-
tive parental behaviors and insecure attachment. One
study investigated the influence of different risk factors
like maternal behavior, mother-child attachment and
maternal anxiety on the development of anxiety disor-
ders in 202 (102 behaviorally inhibited, 100 behaviorally
noninhibited) children aged 4–9 years (Hudson & Dodd,
2012). Mother-child attachment was measured with the
preschool version of the Strange Situation Procedure
(Cassidy & Marvin, 1992) and maternal behavior
through a speech preparation task, a 5-Min Speech
Sample (Hudson & Rapee, 2001), and the Parent Protec-
tion Scale (Thomasgard, Metz, Edelbrock, & Shonkoff,
1995). The authors reported that maternal over-involve-
ment – but not negativity or insecure attachment – at
4 years of age, significantly predicted anxiety diagnoses
at 9 years of age. No moderation effects were found
between these variables (Hudson & Dodd, 2012). A limi-
tation was that fathers were not included. This is very
common in those types of studies. Since fathers are as
pivotal attachment figures as mothers (Lieberman,
© 2018 Association for Child and Adolescent Mental Health.
doi:10.1111/camh.12269
Doyle, & Markiewicz, 1999), this study cannot exclude
attachment with fathers as a potential mechanism influ-
encing anxiety development.
In general, few studies on attachment as well as on
anxiety include fathers. The limited number of studies
on father-child attachment, and their rather broad cate-
gorization of potential associated problems in children
(e.g., depression, externalizing disorders), have resulted
in these studies being excluded from two out of four
reviews on the topic (Groh et al., 2012; Madigan et al.,
2013). Results of the few studies on attachment and the
broader child problem behavior, which have included
fathers alongside mothers, are mixed. For example, a
study by Fagot (1995) found nothing particularly regard-
ing fathers. Rothbaum, Rosen, Pott, and Beatty (1995)
found that only maternal acceptance and maternal
attachment security was significant for child wellbeing,
and that only maternal attachment predicted later prob-
lems in children. Most of the studies including fathers
alongside mothers, report on a combined parental
result, or fail to find anything special regarding the dif-
ference between the parents (please see the reviews by
Groh et al., 2012; Madigan et al., 2013 for further
details).
Overall, the paucity of literature calls for empirical
studies on interactions between parental behaviors and
attachment including fathers. To our knowledge, the
only study examining this with both parents is a study
using a sample of nonclinical children (Breinholst,
Esbjørn, & Reinholdt-Dunne, 2015). The study tested
whether insecure attachment mediated the association
between negative parental behaviors and anxiety.
Results indicated that insecure attachment and negative
parental behaviors accounted for a small and unique
proportion of variance in anxiety symptoms. In contrast
to the findings by Hudson and Dodd (2012), examining
the distinctive influence of mothers’ and fathers’ behav-
iors revealed significance only in maternal behavior. The
relation between anxiety symptoms and both maternal
rejection and maternal psychological control was par-
tially mediated by the child’s report of insecure attach-
ment to the mother (Breinholst et al., 2015). However,
methodological differences between studies may help
explain this difference.
This study aimed to expand existing empirical knowl-
edge regarding the unique contribution of insecure paren-
tal attachment and negative behavior on childhood
anxiety disorders and the interaction between these risk
factors regarding mothers and fathers, given the field is
relatively unexplored but still important in understand-
ing/preventing anxiety in children and understanding
their social world, which includes influences from both
parents. The study attempted to reproduce findings by
Breinholst et al. (2015) in a clinical sample, so that
insight may be found concerning anxiety disorders; prior
studies have only been conducted using normative sam-
ples, and consequently aren’t applicable to anxiety dis-
orders. Based on the literature, these hypotheses were
assessed: (a) attachment insecurity and negative paren-
tal behaviors each account for a significant and unique
proportion of variance in childhood anxiety, (b) attach-
ment insecurity mediates the relation between negative
parental behaviors and anxiety disorders in children.
Due to contradictory findings in previous studies, no
hypotheses were made regarding differences between
© 2018 Association for Child and Adolescent Mental Health.
Parental behavior, attachment and childhood anxiety 3
maternal and paternal variables and their influence on
childhood anxiety disorders.
Method
Participants, procedure, and ethical considerations
Fifty-nine families were eligible for the present cross-sectional
study. Five declined; they were offered nonresearch-related
treatment elsewhere. Ultimately, 54 families participated. Of
these, 28 (51.9%) were boys. Mean age was 9 years, 6 months
(SD = 1 year, 7 months). All the families were self-referred but
some of them were informed about the clinic and encouraged to
attend by their general practitioners, school psychologists, tea-
cher etc. All the children met diagnostic criteria for either gener-
alized anxiety disorder, separation anxiety disorder, specific
phobia or social phobia as the primary diagnosis using Anxiety
Disorders Interview Schedule for Children – Child and Parent
versions ADIS C/P (Silverman, Albano, & Barlow, 1996). Chil-
dren’s full-scale IQ needed to be higher than or equal to 70
(WISC-III; Wechsler, 1991), and at least one parent was a native
Danish speaker. Written consents from the parents were
received.
The families participated in intake assessment procedures
from March 2009 to September 2010 at the University Clinic,
where all the questionnaires were administered by trained clini-
cal psychologists or clinically trained students receiving ongo-
ing supervision. Families received cognitive behavioral therapy
free of charge.
Children met criteria for these primary diagnoses: Separation
Disorder 46.3% (25), Generalized Anxiety Disorder 22.2% (12),
Specific Phobia 25.9% (14), and Social Phobia 5.3% (3). Of the
54 children, 41 (75.9%) met criteria for a secondary diagnosis,
and 20 (37%) met criteria for a tertiary diagnosis, most of which
were comorbid anxiety disorders. The participating families
were rather homogenous, primarily Caucasian, well-educated
with 13% of the sample having an household income of
300,000–5000,000 Danish Kroner, 22.2% of the sample having
a household income of 500,000–750,000 Danish Kroner, and
38.9 earning more than 750,000 Danish Kroner, hence, a gen-
erally low risk sample.
The study complied with all ethical standards regarding
research conducted using children and was approved by the
Institutional Ethical Review Board, Department of Psychology,
University of Copenhagen. Written informed consent from all
participating parents – and assent from all participating chil-
dren – was obtained.
Questionnaires
Questionnaires were translated to Danish using a translation/
back-translation procedure (Brislin, 1970). Original question-
naires were translated to Danish by a staff member, after which
a second staff member, who hadn’t read the original question-
naire, performed the back-translation. The back-translations
were approved by the original questionnaire authors.
State-trait anxiety inventory for children. The trait scale of
Spielberger State-Trait Inventory for Children, STAICtrait (Spiel-
berger, Edwards, Lushene, Montuori, & Platzek, 1973) is a
widely used self-report measure of anxiety. Children rated 20
items on a 3-point scale. Higher scores indicated higher anxiety
levels. An example item is “I feel like crying.” The child reports
whether this happens “almost never,” “sometimes,” or “often.”
The STAIC displays sound psychometric properties with regard
to both reliability and validity (Spielberger et al., 1973). In our
study, Cronbach’s alpha was .91.
Security scale. The Security Scale (SEC) measures children’s
perception of the attachment relationship with their mother or
father in middle childhood (Kerns, Aspelmeier, Gentzler, & Gra-
bill, 2001; Kerns, Klepac, & Cole, 1996). It is a self-report mea-
sure consisting of 15 items scored on a 4-point scale. Higher
scores indicate greater security. The child is presented with
4 Sonja Breinholst, Marie Tolstrup & Barbara Hoff Esbjørn
information about two types of children: “Some kids find it easy
to trust their mom” BUT “Other kids find it difficult to trust their
mom.” The child decides which of the statements is most like
them. Second, the child indicates whether the statement is “very
much true” or “somewhat true.” The scale holds good psycho-
metric properties with good internal consistency and test-retest
reliability (Kerns et al., 2001). In this study, children’s percep-
tions of their attachment to both their mother and father were
assessed separately, which has been done in prior studies of
self-competence (Diener, Russell, Behunin, & Wong, 2007). In
our study, Cronbach’s alpha was .76 and .92 for mothers and
fathers, respectively.
Rearing behavior questionnaire. The Rearing Behavior
Questionnaire (RBQ) measures perceptions of parental rearing
behaviors. In this study, we assessed mothers’ and fathers’ per-
ceptions of their rearing behaviors independently (B€ ogels &
Melick, 2004). It consists of 33 items loading on three subscales:
psychological control (“I am always telling my child how it should
behave”), granting of autonomy versus overprotection (“I encour-
age my child to make its own decisions” vs. “I am overprotective of
my child”), and acceptance versus rejection (“I enjoy being with
my child” vs. “I do not like to have my child around the house”).
Each item is rated on a 4-point scale ranging from “not true at all”
to “very true.” Psychometric properties are reported as satisfac-
tory (B€
ogels & Melick, 2004; Verhoeven, B€ ogels, & van der Brug-
gen, 2012). In our study, Cronbach’s alpha for the subscales
(psychological control, granting of autonomy vs. overprotection
and acceptance vs. rejection) were .81, .54, and .69, respectively,
for mothers and .85, .75, and .39 for fathers.
Results
SPSS version 22 software was used for analysis. We exam-
ined age and gender effects on our target variables (STAIC,
SEC and all RBQ). A median split between younger chil-
dren (7–9 years) and older children (10–12 years) was per-
formed prior to independent sample t-tests. We found no
age difference for anxiety, though we found older children
reported more secure attachment to mothers than younger
children (M (SD) = 3.4 (0.3) vs. 3.1 (0.5); t (49) = 3.4,
p = .002). Fathers of older children also reported being
more overprotective than fathers of younger children (M
(SD) = 39 (4.6) vs. 36 (3.2); t (37) = 2.4; p = .021). Age
was therefore included in the main analyses. No significant
gender differences were found.
Contribution of attachment and parental
behaviors to child anxiety
Two multiple regression analyses were run to test our
hypothesis that attachment and parental behaviors
would account for a unique proportion of variance in
child anxiety. One model included maternal variables,
the other paternal variables. Age, SEC, and RBQ sub-
scales were entered as independent variables with STAIC
as the dependent variable. The hypothesis was partially
supported; the maternal variables model did not provide
a significant model fit, but, the paternal variables model
was significant and explained 28.9% of the variance in
child anxiety. Child-reported attachment to fathers was
the only significant contributor (B = 5.6; SE B = 2.2;
b = .39; p = .02).
Insecure attachment as a mediator between
negative parental behaviors and anxiety
Our second hypothesis was that attachment insecurity
would mediate the relation between negative parental
behaviors and anxiety disorders in children. Mediational
Child Adolesc Ment Health 2018; *(*): **–**
analyses were only conducted on paternal variables as
these yielded a significant model fit in the previous anal-
ysis. Analyses were conducted separately for each pater-
nal RBQ scale (rejection, overprotection, psychological
control). OLS regressions were used to assess the extent
to which the criteria for mediation in the causal step
strategy were met (Baron & Kenny, 1986).
The hypothesis was partially supported; the causal
step analysis revealed no mediation regarding paternal
overprotection and psychological control. However, a full
mediational effect of insecure attachment was found
between paternal rejection and anxiety, as the signifi-
cant association between anxiety disorder and paternal
rejection disappeared when accounting for insecure
attachment to father (Figure 1).
Discussion
This study set out to replicate findings from a previous
study investigating the direct and indirect relationship
between parent-child attachment and negative parental
behaviors separately for mothers and fathers in a clini-
cally anxious sample.
We hypothesized that attachment insecurity and neg-
ative parental behaviors would directly and indirectly
influence clinically anxious children’s self-reported anxi-
ety levels. These hypotheses were partially supported.
The maternal variables model did not yield a significant
model fit. Child-reported attachment to mother and
mothers’ report of their behavior did not significantly
explain the variance in child-reported anxiety. The lack
of significant influence of maternal insecure attachment
and negative behavior on child anxiety matches findings
by Hudson and Dodd (2012), although the lack of influ-
ence of overprotection contradicts their findings. In our
study, only the paternal variables model was significant.
This contrasts with findings by Breinholst et al. (2015).
They found insecure attachment to mothers but not
fathers, as well as perceptions of negative maternal but
not paternal behavior, contributed significantly to child
anxiety. Our study also found that child-reported inse-
cure attachment to father fully mediated between
fathers’ report on rejection and child-reported anxiety
levels.
We offer two explanations for the difference in results
between the current and earlier study: first, it may be
due strictly to methodological differences regarding the
informant level. Our study applied multiple informants
with parental self-report for parental behaviors and child
report on anxiety and attachment to minimize the risk of
single-informant bias. The previous study applied only
Insecure
a = .36* Attachment b = –.42*
Rejection Anxiety
c = .38* (c ´ = –.16)
Figure 1. Standardized regression coefficients for the relation-
ship between rejection and anxiety as mediated by attachment
insecurity
© 2018 Association for Child and Adolescent Mental Health.
doi:10.1111/camh.12269
child report on all variables (Breinholst et al., 2015). It is
possible that the magnitude of associations was inflated
because of shared method variance. Findings may also
have been influenced by the level of anxiety. Clinically
anxious children endorse cognitive distortions that may
contribute to perceiving, for instance, their parent’s
behavior overly negatively (Van Brakel, Muris, B€ ogels, &
Thomassen, 2006). Furthermore, many studies have
reported low agreement between parental and child
report on parenting behavior (B€ ogels & Melick, 2004;
B€ogels, van Oosten, Muris, & Smulders, 2001), with chil-
dren reporting lower levels of positive parental behav-
iors. Therefore, maternal rearing behavior might not
have been as negative and crucial for clinical anxiety
levels, as suggested by the previous study. Future
research could apply partner reports instead of child
and parental reports on parental behaviors to solve this
issue. B€ ogels and Melick (2004) have argued for this
solution as spouses’ reports on each other were more
highly correlated with children’s reports than parental
self-report.
The second explanation is that our study recruited a
sample of clinically anxious children, whereas the
previous study recruited typically developing youth. Sev-
eral studies have found that maternal behavior and
mother-child attachment predicts anxiety symptoms in
non-clinical samples (Brown & Whiteside, 2008; Muris,
Meesters, Merckelbach, & H€ ulsenbeck, 2000), in line
with Breinholst et al. (2015). In relation to clinical anxi-
ety levels, Hudson and Dodd (2012) found that maternal
overprotection but neither negativity nor mother-child
attachment predicted anxiety disorders. This finding is
partly corroborated by our study, where child attach-
ment to mother did not predict clinical anxiety. However,
as we included fathers, significant results emerged with
father-child attachment predicting clinical levels of anxi-
ety. The different results may indicate that parents offer
different contributions depending on their child’s level of
difficulties; a suggestion the literature supports (B€ ogels
& Phares, 2008). It is possible that biology and society
render parents to treat their children differently, result-
ing in mothers and fathers promoting different risk and
protective factors for children (B€ ogels & Phares, 2008).
However, it’s critical to note that very few studies exam-
ine the role of paternal behaviors, and even fewer exam-
ine children’s attachment to fathers. In a meta-analysis
of observational studies of parental behaviors in child
anxiety, only one of 22 studies included fathers, and only
eight included mixed father/mother samples, with an
overrepresentation of mothers (Van der Bruggen et al.,
2008). Some of the few studies including fathers show
that paternal behavior – especially control – may influ-
ence confidence of socially anxious children, moreso
than maternal behavior. However this is reversed in
samples of normal or low socially anxious children, with
maternal behavior being more influential than paternal
(e.g., B€
ogels, Stevens, & Majdandzic, 2011; Greco & Mor-
ris, 2002). Findings that there are differences between
maternal and paternal influence depending on the clini-
cal anxiety level is similar to findings in this study. How-
ever, a further problem is that most studies on fathers
don’t distinguish between children with clinical anxiety,
children with anxiety symptoms, and children display-
ing internalizing disorders (B€ ogels & Phares, 2008; Van
© 2018 Association for Child and Adolescent Mental Health.
Parental behavior, attachment and childhood anxiety 5
der Bruggen et al., 2008). This prohibits direct compar-
ison of effects of maternal and paternal variables.
Overall, these limitations and the findings from this
study highlight the need for further studies assessing
the unique roles of mothers and fathers concerning both
attachment and parental behaviors, as well as their
reciprocal relationship and potential interactions.
Strengths and limitations
Our study provides new insight on the relationship
between two well-established risk factors for child-
hood anxiety disorders. Strengths of the study are
the use of a clinical sample of anxious children and
the inclusion/comparison of the role of mothers and
fathers, seeking to bridge the gap in the present
body of research. However, our study is not without
limitations, the most critical being the cross-sectional
design. This prohibits us from offering conclusions
regarding cause and effect in relationships of attach-
ment, parental behaviors, and child anxiety. Future
studies should include longitudinal designs or experi-
mental manipulations before our model may be con-
firmed. An additional limitation is our relatively small
sample; it limits our statistical power and examina-
tion of complicated models (Fritz & MacKinnon,
2007), as well as reliance on self-report measures.
Future studies should incorporate multiple infor-
mants on the same variables. The measures used
were also chosen specifically to accommodate the age
and developmental phase of the children, which
might make it difficult to generalize into other ages
and phases. Furthermore, it is a typical two-parent,
middle-class Caucasian sample; caution must be
taken before generalizing results to other samples.
Low alphas on some of the RBQ constructs are
another concern. A minor limitation concerning mea-
sures is that attachment was assessed using a
dimensional rather than categorical approach. It was
therefore impossible to distinguish between different
classifications of insecure attachment regarding anxi-
ety. However, some attachment categories are rela-
tively rare; very large samples are required to prevent
reducing power. Therefore, the majority of studies
within this field report results on secure versus inse-
cure attachment (Lubiewska & van de Vijver, 2013).
Overall, limitations and findings from this study high-
light the need for further studies assessing the unique
roles of mothers and fathers concerning both attach-
ment and parental behaviors, as well as their reciprocal
relationship and potential interactions. The implications
for doing this are that fathers and mothers might play
different roles regarding childhood anxiety and, based
on this knowledge, paternal variables might need to be
included in the assessment of childhood anxiety and
case formulation, as well as that fathers might need to
be included in the treatment of anxiety.
Acknowledgement
This study was made possible by funding from Helse Founda-
tion (Grant number 2008A438) and Egmont Foundation (Grant
number 831‐2205). The authors would like to thank participat-
ing children and parents, as well as all students and staff at
Center for Anxiety who helped to collect data. They are also
grateful for statistical guidance from Associate Professor
6 Sonja Breinholst, Marie Tolstrup & Barbara Hoff Esbjørn Child Adolesc Ment Health 2018; *(*): **–**

Mathias Gondam. The authors have declared that they have no Diener, M. L., Russell, I. A., Behunin, M. G., & Wong, M. S.
competing or potential conflicts of interest. (2007). Attachment to mother and fathers during middle
childhood. Associations with child gender, grade, and compe-
tence. Social Development, 17, 84–101.
References Fagot, B. I. (1995). Classification of problem behaviors in young
children: A comparison of four systems. Journal of Applied
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator
developmental Psychology, 16, 95–106.
variable distinction in social psychological research: Concep-
Fritz, M. S., & MacKinnon, D. P. (2007). Required sample size to
tual, strategic, and statistical considerations. Journal of Per-
detect the mediated effect. Psychological science, 18, 233–
sonality and Social Psychology, 51, 1173–1182.
239.
Bodden, D. H., Dirksen, C. D., & B€ ogels, S. M. (2008). Societal
Greco, L. A., & Morris, T. L. (2002). Paternal child-rearing style
burden of clinically anxious youth referred for treatment: A
and child social anxiety: Investigation of child perceptions
cost-of-illness study. Journal of Abnormal Child Psychology,
and actual father behavior. Journal of Psychopathology and
36, 487–497.
Behavioral Assessment, 24, 259–267.
B€
ogels, S. M., & Brechman-Toussaint, M. L. (2006). Family
Groh, A. M., Roisman, G. I., van IJzendoorn, M. H., Bakermans-
issues in child anxiety: Attachment, family functioning, par-
Kranenburg, M. J., & Fearon, R. (2012). The significance of
ental rearing and beliefs. Clinical Psychology Review, 26,
insecure and disorganized attachment for children’s internal-
834–856.
izing symptoms: A meta-analytic study. Child Development,
B€
ogels, S. M., & Melick, M. V. (2004). The relationship between
83, 591–610.
child-report, parent self-report, and partner report of per-
Hudson, J. L., & Dodd, H. F. (2012). Informing early interven-
ceived parental rearing behaviors and anxiety in children and
tion: Preschool predictors of anxiety disorders in middle
parents. Personality and Individual Differences, 37, 1583–
childhood. PLoS ONE, 7, e42359.
1596.
Hudson, J. L., & Rapee, R. M. (2001). Parent–child interactions
B€
ogels, S., & Phares, V. (2008). Fathers’ role in the etiology, pre-
and anxiety disorders: An observational study. Behaviour
vention and treatment of child anxiety: A review and new
Research Theraphy, 39, 1411–1427.
model. Clinical Psychology Review, 28, 539–558.
Kerns, K. A., Aspelmeier, J. E., Gentzler, A. L., & Grabill, C. M.
B€
ogels, S., Stevens, J., & Majdandzi c, M. (2011). Parenting and
(2001). Parent-child attachment and monitoring in middle
social anxiety: Fathers’ versus mothers’ influence on their
childhood. Journal of Family Psychology, 15, 69–81.
children’s anxiety in ambiguous social situations. Journal of
Kerns, K. A., Klepac, L., & Cole, A. (1996). Peer relationships
Child Psychology and Psychiatry, 52, 599–606.
and preadolescents’ perceptions of security in the child-
B€
ogels, S. M., van Oosten, A., Muris, P., & Smulders, D. (2001).
mother relationship. Developmental Psychology, 32, 457–
Familial correlates of social anxiety in children and
466.
adolescents. Behaviour Research and Therapy, 39, 273–
Kessler, R. C., & Wang, P. S. (2008). The descriptive epidemi-
287.
ologv of commonly occurring mental disorders in the United
Bosquet, M., & Egeland, B. (2006). The development and main-
States.
tenance of anxiety symptoms from infancy through adoles-
Lebowitz, E. R., Woolston, J., Bar-Haim, Y., Calvocoressi, L.,
cence in a longitudinal sample. Development and
Dauser, C., Warnick, E., . . . & Vitulano, L. A. (2013). Family
Psychopathology, 18, 517–550.
accommodation in pediatric anxiety disorders. Depression
Bowlby, J. (1958). The nature of the child’s tie to his mother.
and Anxiety, 30, 47–54.
International Journal of Psycho-Analysis, 39, 350–373.
Lieberman, M., Doyle, A. B., & Markiewicz, D. (1999). Develop-
Bowlby, J. (1973). Attachment and Loss. Vol. 1: Attachmen. Lon-
mental patterns in security of attachment to mother and
don: Basic Books.
father in late childhood and early adolescence: Associations
Breinholst, S., Esbjørn, B. H., & Reinholdt-Dunne, M. L. (2015).
with peer relations. Child development, 70, 202–213.
Effects of attachment and rearing behavior on anxiety in nor-
Lubiewska, K., & van de Vijver, F. J. R. (2013). Attachment types
mal developing youth: A mediational study. Personality and
or dimensions: Evidence from adult attachment scale across
Individual Differences, 81, 155–161.
three generations. Symposium at 6th International Attach-
Brislin, R. W. (1970). Back-translation for cross-cultural
ment Conference, 2013, Pavia, Italy.
research. Journal of cross-cultural psychology, 1, 185–216.
Madigan, S., Atkinson, L., Laurin, K., & Benoit, D. (2013).
Brown, A. M., & Whiteside, S. P. (2008). Relations among per-
Attachment and internalizing behavior in early childhood: A
ceived parental rearing behaviors, attachment style, and
meta-analysis. Developmental Psychology, 49, 672.
worry in anxious children. Journal of Anxiety Disorders, 22,
McLeod, B. D., Wood, J. J., & Avny, S. B. (2011). Parenting and
263–272.
child anxiety disorders. In D. McKay & E. A. Storch (Eds.),
Brumariu, L. E., & Kerns, K. A. (2010). Parent-child attachment
Handbook of child and adolescent anxiety disorders (pp. 213–
and internalizing symptoms in childhood and adolescence: A
228). New York: Springer.
review of empirical findings and future directions. Develop-
McLeod, B. D., Wood, J. J., & Weisz, J. R. (2007). Examin-
ment and Psychopathology, 22, 177–203.
ing the association between parenting and childhood anx-
Cassidy, J., & Berlin, L. J. (1994). The insecure/ambivalent pat-
iety: A meta-analysis. Clinical Psychology Review, 27,
tern of attachment: Theory and research. Child development,
155–172.
65, 971–991.
Muris, P., Meesters, C., Merckelbach, H., & H€ ulsenbeck, P.
Cassidy, J., & Marvin, R. S. (1992). Attachment organization in
(2000). Worry in children is related to perceived parental rear-
preschool children: Procedures and coding manual. University
ing and attachment. Behaviour Research and Therapy, 38,
of Virginia: Unpulished manuscript.
487–497.
Colonnesi, C., Draijer, E. M., Stams, G. J., Van der Bruggen, C.
Rothbaum, F., Rosen, K. S., Pott, M., & Beatty, M. (1995). Early
O., Bogels, S. M., & Noom, M. J. (2011). The relation between
parent-child relationships and later problem behavior: A lon-
insecure attachment and child anxiety: A meta-analytic
gitudinal study. Merill-palmer Quarterly, 41, No. 2, 133–151.
review. Journal of Clinical Child and Adolescent Psychology,
Silverman, W. K., Albano, A. M., & Barlow, D. H. (1996). Manual
40, 630–645.
for the ADIS-IV-C/P. New York: Psychological Corporation.
Copeland, W. E., Angold, A., Shanahan, L., & Costello, E. J.
Spielberger, C. D., Edwards, C. D., Lushene, R. E., Montuori, J.,
(2014). Longitudinal patterns of anxiety from childhood to
& Platzek, D. (1973). State-Trait Anxiety Inventory for Children:
adulthood: The great smoky mountains study. Journal of the
Professional manual. Redwood City, CA: Mind Garden, Inc.
American Academy of Child & Adolescent Psychiatry, 53, 21–
Thirlwall, K., & Creswell, C. (2010). The impact of maternal con-
33.
trol on children’s anxious cognitions, behaviours and affect:
Creswell, C., Murray, L., Stacey, J., & Cooper, P. (2011). Parent-
An experimental study. Behaviour Research and Therapy, 48,
ing and child anxiety. Anxiety disorders in children and
1041–1046.
adolescents, 2, 299–322.

© 2018 Association for Child and Adolescent Mental Health.

doi:10.1111/camh.12269
Thomasgard, M., Metz, W., Edelbrock, C., & Shonkoff, J. P.
(1995). Parent-child relationship disorders: I. Parental over-
protection and the development of the Parent Protection
Scale. Developmental Behavior Pediatrics, 16, 244–250.
Van Brakel, A. M., Muris, P., B€ogels, S. M., & Thomassen, C.
(2006). A multifactorial model for the etiology of anxiety in
non-clinical adolescents: Main and interactive effects of
behavioral inhibition, attachment and parental rearing. Jour-
nal of Child and Family Studies, 15, 568–578.
Van der Bruggen, C. O., Stams, G. J. J., & B€
ogels, S. M. (2008).
Research review: The relation between child and parent anxi-
ety and parental control: A meta-analytic review. Journal of
Child Psychology and Psychiatry, 49, 1257–1269.
Van IJzendoorn, M. (1995). Adult attachment representations,
parental responsiveness, and infant attachment: A meta-
© 2018 Association for Child and Adolescent Mental Health.
Parental behavior, attachment and childhood anxiety 7
analysis on the predictive validity of the Adult Attachment
Interview. Psychological bulletin, 117, 387–403.
Verhoeven, M., B€ogels, S. M., & van der Bruggen, C. C. (2012).
Unique roles of mothering and fathering in child anxiety;
moderation by childs age and gender. Journal of Child and
Family Studies, 21, 331–343.
Waite, P., Whittington, L., & Creswell, C. (2014). Parent-child
interactions and adolescent anxiety: a systematic review. Psy-
chopathology Review, 1, 51–76.
Wechsler, D. (1991). WISC-III: Wechsler intelligence scale for
children. San Antonio, TX: Psychological Corporation.
Accepted for publication: 6 February 2018