Review Article
Examining the association between parenting and childhood anxiety: A meta-analysis

https://doi.org/10.1016/j.cpr.2006.09.002Get rights and content

Abstract

Theoretical models emphasize the role of parenting in the development and maintenance of child anxiety, but reviews of the empirical literature have provided mixed support for existing theories. To help clarify the role parenting plays in childhood anxiety, we conducted a meta-analysis of 47 studies testing the association between parenting and child anxiety. Across these studies, parenting accounted for only 4% of the variance in child anxiety. Moderator tests indicated that methodological factors (i.e., how child anxiety and parenting were conceptualized and assessed) may be a source of inconsistent findings within the literature. In addition, our analyses revealed that parental control was more strongly associated with child anxiety than was parental rejection. Specific subdimensions within parental rejection and control differed in their association with child anxiety (e.g., autonomy-granting accounted for 18% of the variance, but warmth < 1%), indicating that efforts to disaggregate parenting dimensions may inform theory development and future research. Overall, however, the modest association between parenting and child anxiety suggests that understanding the origins of children's anxiety will require identifying factors other than parenting that account for the bulk of the variance.

Introduction

Anxiety disorders are among the most common psychiatric problems experienced by children and they appear to remain stable and problematic for many youths throughout childhood and adolescence (Benjamin et al., 1990, Boyd et al., 2000, Last et al., 1996). Theoretical models of anxiety disorders have emphasized the influence of parenting on the development, maintenance, and amelioration of childhood anxiety (e.g., Chorpita and Barlow, 1998, Craske, 1999, Dadds and Roth, 2001, ⁎Fox et al., 2005, Krohne, 1990, Manassis and Bradley, 1994, Rapee, 2001, Vasey and Dadds, 2001). However, four reviews of the empirical evidence linking parenting to childhood anxiety have provided mixed support for existing theories (see Gerlsma et al., 1990, Masia and Morris, 1998, Rapee, 1997) and have failed to resolve a fundamental question: Are parenting practices substantially associated with childhood anxiety?

Traditional models of childhood anxiety sought to explain the development of anxiety in terms of single main effects and focused primarily on the broad parenting dimensions of acceptance versus rejection and psychological granting of autonomy versus psychological control (see Gerlsma et al., 1990, Masia and Morris, 1998, Rapee, 1997, Wood et al., 2003). Both represent bipolar parenting dimensions, with positive parenting practices (e.g., acceptance) at one end of the continuum and negative parenting practices (e.g., rejection) at the other end of the continuum. For simplicity, we will use the terms rejection and control throughout the rest of this paper. Parental rejection connotes low levels of parental warmth, approval, and responsiveness (i.e., coldness, disapproval, and unresponsiveness) (e.g., Clark and Ladd, 2000, Maccoby, 1992). Parental rejection is hypothesized to undermine children's emotion regulation by increasing sensitivity to anxiety (e.g., Gottman, Katz, & Hooven, 1997). Hence, parental rejection is hypothesized to put children at an increased risk for developing anxiety problems. Parental control involves excessive parental regulation of children's activities and routines, encouragement of children's dependence on parents, and instruction to children on how to think or feel (e.g., Barber, 1996, Steinberg et al., 1989). Some theoretical models (see, e.g., Chorpita and Barlow, 1998, Krohne, 1990, Manassis and Bradley, 1994, Rapee, 2001, Rubin and Mills, 1991) hypothesize that when parents are highly controlling in contexts when it is developmentally appropriate for children to act independently (e.g., attending elementary school), children may experience decreased self-efficacy, and thus, increased anxiety (Wood, 2006) — e.g., about their ability to function on their own within their environments. Conversely, some models (see, e.g., Chorpita and Barlow, 1998, Wood et al., 2003) have hypothesized that parental encouragement of children's autonomy and independence (e.g., in novel contexts) may augment children's perceptions of mastery over the environment, leading to anxiety reduction.

In keeping with the dominant theoretical models, most empirical studies have focused on the linkage between childhood anxiety and the broad parenting dimensions of rejection and control. Early empirical efforts primarily relied on adult retrospective reports. Papers reviewing this group of studies concluded that anxious adults generally remember their parents as being rejecting and controlling (see Gerlsma et al., 1990, Masia and Morris, 1998, Rapee, 1997). These findings are consistent with existing theories, but may not generalize to the association between parenting and anxiety during childhood. Retrospective reports have strengths, but they have been criticized because they may not provide a reliable measure of actual parenting (see Holden & Edwards, 1989). For instance, evidence suggests that adults with mood disorders give biased accounts of parenting that do not reflect their actual childhood experiences (Lewinsohn & Rosenbaum, 1987). A recent review excluded retrospective reports and focused exclusively on studies that employed concurrent measures of parenting and anxiety in childhood or adolescence (Wood et al., 2003). But this review drew no clear conclusions regarding the parenting and childhood anxiety association, partially because a narrative review does not allow for an examination of the magnitude of effects across studies or potential moderating variables.

None of these four reviews used quantitative methods for synthesizing findings across studies. As a result, these reviews are limited in their ability to characterize the parenting–childhood anxiety association and cannot accurately assess the extent to which variations in parenting are actually associated with child anxiety. Methodologists (e.g., Rosenthal, 1991, Schmidt, 1992) caution that statistical tests from individual studies can be misleading, given the dependence of such metrics on sample size and other sample characteristics, and that a better way to assess associations is to rely on effect size (ES) values examined within meta-analyses that synthesize those values across multiple studies. Like any statistical tool, meta-analyses are most informative when applied to the most representative data possible (e.g., the complete collection of studies carried out) and when used with safeguards against distortion (e.g., small sample bias). In the present paper, we present such a meta-analysis, relying on a highly representative set of studies and using stringent procedures to assess the strength of association.

Such an approach is useful in assessing the extent to which variations in parenting are actually associated with child psychological problems. Recent controversy over the long-accepted premise that parental childrearing practices are associated with child psychological problems raises questions about the extent to which parental behavior impacts children's psychological well-being (Harris, 2002, Kagan, 2003, Maccoby, 2002, Rutter, 2002). The need to examine this broad question is highlighted by findings of Rothbaum and Weisz's (1994) meta-analysis testing the association between parenting and child externalizing problems. These authors found that, in contrast to the theoretical literature which strongly emphasized parenting as a determinant, parenting actually accounted for less than 6% of the variance in child externalizing problems. This rather surprising finding raises the question of whether or not internalizing problems such as child anxiety are strongly associated with parenting.

The theoretical literature cited earlier certainly asserts that a strong association between parenting and childhood anxiety exists, but other lines of evidence point to non-parenting factors that might be considerably more powerful in their impact. For example, twin studies (see e.g., ⁎Eley et al., 2003, Rice et al., 2002, van Beijsterveldt et al., 2004) have suggested that additive genetic effects may account for 30–80% of the variance in children's trait anxiety, and that the non-shared environment (i.e., biological and social environmental influences that affect one sibling but not another) might account for a substantial proportion of additional variance. Nonetheless, most behavioral genetic studies suggest that there is at least some role of the “shared environment” in children's anxiety, which can include parenting influences. Meta-analytic methods permitted us to address the extent to which parenting behavior may be linked with child anxiety, to further evaluate this possibility.

Meta-analytic methods also allow the examination of variables that may moderate the association between parenting and childhood anxiety. Moderator variables can be the source of inconsistent findings across studies and can serve to mask the effects of other variables. Narrative reviews cannot identify modest but reliable effects associated with moderator variables, and thus are limited in their ability to identify potential sources of inconsistent findings. Elucidating possible sources of inconsistent findings can help inform theory development and provide direction for future research. Thus, a major goal of the present study was to examine whether a set of methodological and theoretical variables moderated the association between parenting and childhood anxiety.

Our examination of potential moderator variables begins with methodological variables. It is important to examine whether the way in which the target constructs (i.e., parenting and childhood anxiety) are measured moderates the strength of association, because variation in measurement is often the source of inconsistent findings. For example, Rothbaum and Weisz (1994) found that the association between parenting and childhood externalizing behavior was significantly reduced when parenting was measured using questionnaire measures as opposed to observational measures. We therefore examined whether methodological factors moderate the parenting–childhood anxiety association.

In addition, we assessed whether findings might be moderated by the manner in which parenting is construed and operationalized. The parenting dimensions of “rejection” and “control” emphasized in traditional theoretical models may lack specificity (Wood, 2006). Discontent with these models has led some to hypothesize that subdimensions of parental rejection (e.g., aversiveness, withdrawal, and a lack of warmth) and control (e.g., overinvolvement and a lack of autonomy-granting) may be differentially associated with childhood anxiety. For instance, Rapee (2001) emphasized the importance of parental overinvolvement, whereas Chorpita and Barlow (1998) stressed the importance of parental autonomy-granting. If specific parenting subdimensions are differentially associated with childhood anxiety, then the use of the broad parenting categories may produce uneven effects due to heterogeneity among the subdimensions. At present, though, limited empirical evidence can be brought to bear on this issue since past reviews have not examined these parenting subdimensions. Thus, an aim of this paper was to assess whether subdimensions of the traditional parenting categories were differentially associated with childhood anxiety.

We also assessed whether the strength of the parenting-anxiety association might depend in part on child age or gender, or parent gender. One hypothesis we addressed was that the association between parenting and childhood anxiety is strongest in early childhood. Some models posit that parenting is particularly influential early in a child's life when the parents represent the main socializing agent — for example, early experiences with limited control are hypothesized to contribute to the development of anxiety (Chorpita, 2001). But it is also plausible that associations are stronger in later childhood as a result of the cumulative influences that child–parent interactions have on the development of anxiety over time (cf. Maccoby, 1992, Patterson, 1982, Rothbaum and Weisz, 1994). With regard to child gender, more girls than boys suffer from anxiety disorders and symptoms (see Weiss & Last, 2001) suggesting that higher levels of anxiety exhibited by some girls may ensnare parents in an overprotective or aversive cycle of behavior. To our knowledge, the role of parent's gender has not been the focus of theory nor has it been addressed in any reviews. On the one hand, women are more likely to suffer from anxiety symptoms than are men (see Craske, 1999) and may thus be more likely to model anxious behavior to their children (Whaley, Pinto, & Sigman, 1999). On the other hand, studies have suggested that fathers’ affect may have a greater influence than mothers’ on the mood of family members, including children (e.g., Larson & Richards, 1994). In sum, an aim of the present study was to examine whether child age and gender, and parent gender, moderate the parenting–childhood anxiety association.

Finally, we examined whether the magnitude of the parenting–childhood anxiety association varied as a function of several study design factors that are often examined in meta-analyses related to child and adult psychopathology. These factors included: (a) diagnostic status (i.e., whether the participants were diagnosed with an anxiety disorder), (b) type of anxiety (i.e., social anxiety symptoms versus other anxiety symptoms), (c) anxiety assessment strategy (continuous variable approach vs. categorical group comparison), (d) measurement technology of anxiety measure (i.e., the type of measure used to characterize anxiety), (e) child anxiety informant (i.e., who provided the report on children's anxiety), (f) measurement technology of parenting measure (i.e., the type of measure used to characterize parenting), (g) parenting informant (i.e., who provided the information on parenting), and (h) the subdimensions of parental rejection and control (i.e., withdrawal, aversiveness, warmth, overinvolvement, autonomy-granting).

To summarize, in the present study, we conducted a meta-analysis of studies examining the linkage between parenting and childhood anxiety to provide a more definitive statement about the overall strength of this association and to assess the impact of potential moderators. This is a particularly appropriate time for a meta-analysis because new theoretical frameworks for understanding the role that parenting plays in the development, maintenance, and amelioration of childhood anxiety are emerging, yet the strength of the parenting–childhood anxiety association has not yet been established (see Chorpita and Barlow, 1998, Craske, 1999, Dadds and Roth, 2001, Rapee, 2001, Wood, 2006). Our goal was to contribute to the development of these new frameworks by examining the overall degree of covariation of parenting and childhood anxiety. We sought to avoid some of the limitations of previous reviews by excluding adult retrospective reports and by using the quantitative approach of meta-analysis to synthesize the entire body of research in this area. We considered including longitudinal and experimental studies in the study set, but our search found no experimental studies and too few longitudinal studies to warrant a meta-analysis. Thus, only studies examining the concurrent association between parenting and childhood anxiety were included in this meta-analysis, but we did include a brief overview of the longitudinal studies to date.

Section snippets

Selection of studies

We conducted a literature search for studies presenting quantitative data on the association between parenting and childhood anxiety involving mothers and/or fathers, and children from infancy through adolescence. A computer based information search was conducted on the PsychInfo computer database, which indexes (with key terms) and abstracts articles. The search covered from January 1887 up to April 2004, and we used the following 12 anxiety-related key terms: internaliz-, anxi-, fear-,

Study-level analysis

Our study-level analysis focused upon the association between parenting and childhood anxiety. For the parenting–childhood anxiety association the weighted mean ES was .21, reflecting a relation in which more negative parenting was associated with more child anxiety, and the 95% confidence interval did not include zero. This ES meets criteria for a small effect and suggests that parenting was associated with approximately 4% of the variance in childhood anxiety. The homogeneity analysis was

Discussion

The primary goal of the present study was to identify the nature and strength of associations between parenting and childhood anxiety. Influential theories in the field have stressed the role of parenting, but reviews of the empirical literature have suggested that the association is inconsistent from study to study (see e.g., Masia and Morris, 1998, Rapee, 1997). The meta-analytic data synthesized here indicate that the connection between parenting and child anxiety was small in magnitude,

Acknowledgement

This paper was supported in part by a training grant from the National Institute of Mental Health (NIMH) awarded to Bryce D. McLeod (National Research Service Award; F31-MH64993) and grants awarded to John R. Weisz from NIMH (R01-MH49522 and R01-MH068806) and the John D. and Catherine T. MacArthur Foundation (Initiative on Youth Mental Health, and Research Network on Youth Mental Health).

References (100)

  • C.G. Last et al.

    A prospective study of childhood anxiety disorders

    Journal of the American Academy of Child & Adolescent Psychiatry

    (1996)
  • K. Manassis et al.

    The development of childhood anxiety disorders: Toward an integrated model

    Journal of Applied Developmental Psychology

    (1994)
  • P. ⁎Muris et al.

    Parental rearing practices, fearfulness, and problem behavior in clinically referred children

    Personality and Individual Differences

    (1996)
  • P. ⁎Muris et al.

    Worry in children is related to perceived parental rearing and attachment

    Behavior Research and Therapy

    (2000)
  • R.M. Rapee

    Potential role of childrearing practices in the development of anxiety and depression

    Clinical Psychology Review

    (1997)
  • K. ⁎Travillion et al.

    The role of maternal discipline and involvement in peer rejection and neglect

    Journal of Applied Developmental Psychology

    (1993)
  • A. Bandura

    Self-efficacy conception of anxiety

    Anxiety Research

    (1988)
  • B. Barber

    Parental psychological control: Revisiting a neglected construct

    Child Development

    (1996)
  • M. ⁎Barrera et al.

    Family and peer social support as specific correlates of adolescent depressive symptoms

    Journal of Abnormal Child Psychology

    (1992)
  • P. ⁎Barrett et al.

    Do parent and child behaviours differentiate families whose children have obsessive-compulsive disorder from other clinic and non-clinic families?

    Journal of Child Psychology and Psychiatry

    (2002)
  • I. ⁎Berg et al.

    Are school phobic adolescents overdependent?

    British Journal of Psychiatry

    (1971)
  • H.B. ⁎Biller et al.

    Perceived maternal control, anxiety, and opposite sex role preference among elementary school girls

    The Journal of Psychology

    (1972)
  • C. ⁎Booth et al.

    Predicting social adjustment in middle childhood: The role of preschool attachment security and maternal style

    Social Development

    (1994)
  • C.P. Boyd et al.

    Prevalence of anxiety and depression in Australian adolescents: Comparisons with worldwide data

    Journal of Genetic Psychology

    (2000)
  • J.H. ⁎Bray

    Family relationships and children's adjustment in clinical and nonclinical stepfather families

    Journal of Family Psychology

    (1992)
  • D.T. Campbell et al.

    Convergent and discriminant validation by the multitrait–multimethod matrix

    Psychological Bulletin

    (1959)
  • B.F. Chorpita

    Control and the development of negative emotion

  • B.F. Chorpita et al.

    The development of anxiety: The role of control in the early environment

    Psychological Bulletin

    (1998)
  • K.E. Clark et al.

    Connectedness and autonomy support in parent–child relationships: Links to children's socioemotional orientation and peer relationships

    Developmental Psychology

    (2000)
  • J. Cohen

    Statistical power analysis for the behavioral sciences

    (1988)
  • A.M. Connell et al.

    The association between psychopathology in fathers versus mothers and children's internalizing and externalizing behavior problems: A meta-analysis

    Psychological Bulletin

    (2002)
  • P.A. Cowan et al.

    What an intervention design reveals about how parents affect their children's academic achievement and behavior problems

  • M.G. Craske

    Anxiety disorders: Psychological approaches to theory and treatment

    (1999)
  • M.R. Dadds et al.

    Family processes in the development of anxiety problems

  • J.E. ⁎Dumas et al.

    “Balance of power”: A transactional analysis of control in mother–child dyads involving socially competent, aggressive, and anxious children

    Journal of Abnormal Psychology

    (1995)
  • M. ⁎Eastburg et al.

    Shyness and perceptions of parental behavior

    Psychological Reports

    (1990)
  • T.C. ⁎Eley et al.

    A twin study of anxiety-related behaviours in pre-school children

    Journal of Child Psychology and Psychiatry

    (2003)
  • J.R. Fleiss

    Statistical methods for rates and proportions

    (1981)
  • A.C. ⁎Fletcher et al.

    Adolescents’ well-being as a function of perceived interparental consistency

    Journal of Marriage and the Family

    (1999)
  • R. Forehand et al.

    Maternal and paternal parenting: Critical dimensions in adolescent functioning

    Journal of Family Psychology

    (1993)
  • N.A. ⁎Fox et al.

    Behavioral inhibition: Linking biology and behavior within a developmental framework

    Annual Review of Psychology

    (2005)
  • J.M. Gottman et al.

    Meta-emotion: How families communicate emotionally

    (1997)
  • S. Graham et al.

    Ethnicity, peer harassment, and adjustment in middle school: An exploratory study

    The Journal of Early Adolescence

    (2002)
  • L.A. ⁎Greco et al.

    Paternal child-rearing style and child social anxiety: Investigation of child perceptions and actual father behavior

    Journal of Psychopathology and Behavioral Assessment

    (2002)
  • G.T. ⁎Harold et al.

    Mom and dad are at it again: Adolescent perceptions of marital conflict and adolescent psychological distress

    Developmental Psychology

    (1997)
  • J.R. Harris

    Beyond the nurture assumption: Testing hypotheses about the child's environment

  • L.V. Hedges

    Fixed effects models

  • L. ⁎Hernandez-Guzman et al.

    Parent–child interactions predict anxiety in Mexican adolescents

    Adolescence

    (1996)
  • L.A. ⁎Hersov

    Persistent non-attendance at school

    Journal of Child Psychology & Psychiatry

    (1960)
  • E.D. ⁎Hibbs et al.

    Factors affecting expressed emotion in parents of ill and normal children

    American Journal of Orthopsychiatry

    (1993)
  • Cited by (809)

    View all citing articles on Scopus
    View full text