Parenting and Childhood Anxiety: Theory, Empirical Ndings, and Future Directions

Journal of Child Psychology and Psychiatry 44:1 (2003), pp 134–151
Parenting and childhood anxiety: theory, empirical

¢ndings, and future directions
Jeffrey J. Wood, Bryce D. McLeod, Marian Sigman, Wei-Chin Hwang,
and Brian C. Chu
University of California, Los Angeles, USA
Theories of anxiety development suggest that parental acceptance, control, and modeling of anxious
behaviors are associated with children’s manifestations of anxiety. This paper reviews research pub-
lished in the past decade on the relation between parenting and childhood anxiety. Observed parental
control during parent–child interactions was consistently linked with shyness and child anxiety dis-
orders across studies. Mixed support for the role of parental acceptance and modeling of anxious
behaviors was found in observational studies. However, there was little evidence supporting the con-
tention that self-reported parenting style was related to children’s trait anxiety. Because of limitations
associated with past research, inferences about the direction of effects linking parenting and child
anxiety cannot be made. A conceptual framework based on recent models of anxiety development (e.g.,
Vasey & Dadds, 2001) is presented to aid in the interpretation of extant research findings and to provide
suggestions for future research and theory development. Improved methodological designs are pro-
posed, including the use of repeated-measure and experimental designs for examining the direction of
effects.
Are particular parenting styles or behaviors linked but also expand upon such theory and provide
with childhood anxiety? Clinical anxiety is one of the recommendations to advance the field.
most common psychiatric problems experienced by The definition of ‘anxiety’ varies across subdis-
school-aged children (Bell-Dolan & Brazeal, 1993; ciplines and studies. For this paper, we adopted
Bowen, Offord, & Boyle, 1990; see also Schniering, Craske’s (1999) definitions of and distinctions be-
Hudson, & Rapee, 2000). Although the problem is tween trait anxiety (also referred to as negative affect
widespread and affects children throughout child- or neuroticism, comprising nonspecific symptoms of
hood and adolescence, the etiology and sequelae of fear, worry, and other negative mood states not
childhood anxiety remain complex and elusive. One unique to a single disorder) and anxiety disorders
important area that has been emphasized as con- (e.g., social phobia, separation anxiety disorder).
tributing to the development of childhood anxiety is Trait anxiety is viewed as a continuous characteristic
parenting (e.g., Chorpita & Barlow, 1998; Craske, that, when elevated, represents a generalized vul-
1999; Vasey & Dadds, 2001). In support, behavioral nerability to mood disorders, but may not cause
genetic studies conducted in the past decade have clinically significant functional impairment by itself,
suggested that the ‘shared environment,’ possibly whereas anxiety disorders represent specific anxiety
including childrearing experiences, accounts for a symptom clusters that cause distress or impairment.
significant amount of the variance in childhood an- We review the relation between parenting and both
xiety symptoms and disorders (see Eley, 2001). Some trait anxiety and anxiety disorders in children, and
naturalistic studies have also found a linkage be- use the term ‘childhood anxiety’ when referring to
tween parenting behavior and childhood anxiety both. Distinctions between the two are noted when
(e.g., Whaley, Pinto, & Sigman, 1999). But the con- relevant.
sistency of this association across studies and the
direction of effects remain in question, primarily
Traditional perspectives on parenting
because a current synthesis of the literature is
and childhood anxiety
lacking. A previous meta-analysis (Gerlsma, Em-
melkamp, & Arrindell, 1990) and two review articles From the 1990s to the present, most studies exam-
(Masia & Morris, 1998; Rapee, 1997) on this topic ining the relation between parenting style or behav-
focused primarily on retrospective studies based on ior and childhood anxiety have focused on three
adult reports. Rapee’s review also included some parenting dimensions: acceptance, control, and
studies of children, but these were primarily from the modeling of anxious behaviors. These categories –
1970s and 1980s. To address this gap, this paper particularly the first two – are traditional groupings
reviews and discusses the most recent empirical within the literature. The first parenting category,
evidence regarding the relation between parenting acceptance, refers to a general parenting approach
and childhood anxiety. In doing so, we not only characterized by interactional warmth and respon-
review current theories of child anxiety development, siveness (including acceptance of children’s feelings
Ó Association for Child Psychology and Psychiatry, 2003.
Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
Parenting and childhood anxiety 135
and behaviors, active listening, praise, use of re- conceptualized as a moderator rather than a direct
flection, etc.) as well as emotional and behavioral predictor of children’s psychosocial outcomes (Dar-
involvement in children’s lives and activities (e.g., ling & Steinberg, 1993). Parenting style is a general
Maccoby, 1992). Some studies investigate the op- pattern of caregiving that provides a context for
posite of acceptance, namely parental criticism and specific episodes of parental childrearing behaviors;
rejection. It is hypothesized that parents who dem- but it does not refer to a specific act or set of acts of
onstrate acceptance of children’s expressions of parenting. In contrast, parenting ‘practices’ or ‘be-
negative affect – rather than criticizing and minim- haviors’ are conceptualized as specific kinds of par-
izing children’s feelings – help promote children’s ental interactions with children in specific
emotion regulation by allowing children to learn, situations. Thus, a self-report item that assesses an
through trial and error, to tolerate negative affect accepting parenting style might be, ‘My parent un-
(Gottman, Katz, & Hooven, 1997), thus reducing derstands how I feel,’ whereas an item assessing
children’s sensitivity to anxiety. specific parenting behaviors might be, ‘Today before
The second parenting category, control, is defined school, my parent let me know s/he understood how
as a pattern of excessive regulation of children’s ac- I was feeling.’ Parenting practices or behaviors are
tivities and routines, autocratic parental decision- hypothesized to directly affect children’s emotional
making, overprotection, or instruction to children on and behavioral regulation (e.g., Darling & Steinberg,
how to think or feel (Barber, 1996; Steinberg, Elmer, 1993).
& Mounts, 1989). Although the construct of control How is parenting style differentiated from specific
is a broad one and encompasses numerous facets of parenting behaviors? Parenting style is traditionally
parent–child interactions, factor-analytic research assessed with paper-and-pencil measures that re-
suggests that these facets represent a single, higher- quire the respondent to evaluate global patterns of
order construct that cohere together into a meaning- parenting style over long or unspecified periods of
ful pattern of behavior (e.g., Schwarz, Barton-Henry, time (Holden & Edwards, 1989). Specific parenting
& Pruzinsky, 1985). The different manifestations of practices or behaviors are generally measured with
parental control each involve encouragement of observational approaches or time-delimited self-
children’s dependence on parents, which is hypo- report measures of parenting behaviors in particular
thesized to affect children’s perceptions of mastery situations (such as daily diaries; e.g., Repetti, 1996).
over the environment. Lack of mastery is posited to Laboratory observational paradigms permit the as-
contribute to high trait anxiety by creating a cogni- sessment of parental conversational warmth, con-
tive bias characterized by perceiving events as out trol, and modeling of anxiety during interactions
of one’s control (Chorpita & Barlow, 1998). with children in a specific kind of situation, namely,
Parental granting of autonomy is viewed as the a novel and potentially stressful setting (i.e., due to
opposite of excessive parental control (e.g., Matta- the use of instructions and videocameras). Although
nah, 2001). it is possible to use both self-report and observa-
The third category, modeling of anxious behavior, tional methods to measure either parenting style or
has been less thoroughly studied. It refers to de- parenting behaviors, a more limited measurement
scribing problems to children as irresolvable or approach has characterized the parenting and
dangerous, encouraging (rewarding) children to view childhood anxiety literature: Self-report measures
problems in a catastrophic manner, and extin- have generally been used to assess parenting style,
guishing or punishing children’s expressions of whereas observational approaches have been used
coping thoughts and problem-solving strategies to assess specific parenting practices or behaviors.
(Capps & Ochs, 1995; Whaley et al., 1999). Children Because parenting style and parenting behavior may
of parents who frequently model anxious behavior play different roles in the development and main-
may come to believe that there is no way of coping tenance of childhood anxiety, our review distin-
effectively with problems and are not likely to guishes between these two parenting constructs;
develop strategies that can reduce anxiety (Whaley however, it is important to note that these two con-
et al., 1999). structs are confounded with self-report and obser-
vational methodology, respectively, in this literature.
Parenting style vs. behavior
Limitations of traditional models
Various researchers have contrasted parenting style
of the parenting–anxiety linkage
with specific parenting practices or behaviors (e.g.,
Darling & Steinberg, 1993; Maccoby, 1992). Par- Traditional models of childhood anxiety sought to
enting ‘style,’ defined as a global set of parental at- explain the development of anxiety in terms of single
titudes, goals, and patterns of parenting practices, is main effects (e.g., parental control, inherited traits).
hypothesized to create an emotional climate for the But these models have failed to explain more than a
parent–child relationship. Children’s openness to fraction of the variance associated with anxiety
socialization by parents may be affected by global symptoms or disorders in children (cf. Vasey &
parenting style. However, parenting style has been Dadds, 2001). For instance, genetic traits and overly
136 Jeffrey J. Wood et al.
controlling parenting are both statistically asso- at least four plausible (and not mutually exclusive)
ciated with childhood anxiety in numerous studies, pathways linking it with childhood anxiety: (a) some
but neither factor accounts for most of the variability parenting styles or behaviors may directly cause or
in children’s anxiety symptoms (e.g., Eley, 2001; elicit childhood anxiety; (b) children’s manifestations
Mattanah, 2001). The failure of these traditional of fear and anxiety may elicit particular patterns of
models has led some to move towards a view that the parenting style or behavior; (c) genetic similarity
ontogeny of childhood anxiety is a complex, multi- between children and their parents may act as a
determined process (e.g., Craske, 1999; Vasey & ‘third variable’ accounting for both parenting as well
Dadds, 2001). as childhood anxiety; or (d) genetic traits, parenting,
children’s anxiety symptoms, and other risk or pro-
tective factors may reinforce or moderate each other
Contemporary models of childhood anxiety in a feedback loop. A major objective of this paper is
and parenting to assess the evidence for these pathways by exam-
ining the current state of knowledge on the direction
The rise of the developmental psychopathology per-
of effects linking parenting and the development of
spective (e.g., Cicchetti & Cohen, 1995; Masten &
childhood anxiety.
Braswell, 1991) has provided researchers with a new
conceptual framework for the study of childhood
anxiety (e.g., Manassis & Bradley, 1994; Vasey & The development and maintenance
Dadds, 2001). Proponents of this perspective have of childhood anxiety
synthesized the empirical findings from various
The developmental psychopathology perspective not
fields (e.g., genetics, clinical, and developmental
only emphasizes that specific risk factors can play
psychology) and have posited that both external (fa-
multiple roles in the development of psychopathology,
milial, social-environmental) and internal (genetic,
but it also posits that some risk factors help maintain
cognitive) risk and protective factors contribute to
psychopathology once it has developed. For example,
the process and outcome of childhood anxiety. Par-
genetic traits and temperament may play an import-
enting is posited to take on multiple roles depending
ant role in the development of social anxiety and
upon the context in which it occurs (e.g., stage of
withdrawal (e.g., Rubin & Stewart, 1996), but par-
development, presence or absence of other risk or
enting could play an important role in maintaining
protective factors). Thus, when attempting to single
this problem (e.g., by permitting school avoidance).
out the effects of parenting style or behavior on
Thus, an important distinction is made between
childhood anxiety, research must be guided by two
the roles specific risk factors play in the develop-
concepts. The first concept, multifinality, suggests
ment versus maintenance of psychopathology.
that a single risk factor (e.g., controlling parenting
Consistent with these principles, Craske (1999)
style) can have a variety of outcomes – including
has proposed a theoretical model that helps specify
anxiety, other psychological problems, or successful
the roles that parenting may play in the development
adaptation – depending upon the context in which it
and maintenance of childhood anxiety. Drawing on
operates. The second concept, equifinality, suggests
emotion theory and learning theory, this model
that there are multiple pathways to the same anxiety
suggests that parenting may be related to childhood
disorder (e.g., social phobia) and that a single risk
anxiety in at least two ways. First, general parenting
factor such as parenting cannot universally account
across contexts (i.e., parenting style, see above) is
for the development of a given disorder. These con-
hypothesized to provide an environmental context
cepts suggest that it is important to identify the
that influences the development of trait anxiety. For
consistency and specificity with which different
instance, frequent parental criticism could increase
parenting styles or behaviors are associated with
a child’s wariness and influence perceptions of self
childhood anxiety.1
and the world in a negative manner (e.g., cognitive
If parenting is hypothesized to take on multiple
features of trait anxiety). Second, among children
roles in the development of childhood anxiety, what
with high trait anxiety, specific parenting practices or
are these roles? When considering the role that
behaviors that promote or reinforce children’s ex-
parenting may play as a specific risk factor, there are
periences of anxiety in specific situations contribute
to the development of a particular anxiety disorder
1
A number of additional theoretical approaches have sought to by centering beliefs about threat and physiological
explain anxiogenesis (e.g., neural and animal models) and, arousal upon a specific theme or class of stimuli.
more specifically, the role parenting behavior plays in the de- Within Craske’s (1999) model, the role of parenting
velopment of childhood anxiety (e.g., attachment theory). A behavior in the development of particular child
comprehensive review of these and other theories of childhood
anxiety disorders is more specific than in the devel-
anxiety is beyond the scope of this paper. We have chosen to
opment of elevated trait anxiety. Although general
focus on models stemming from developmental psychopa-
thology, emotion theory, and learning theory, given the degree patterns of parenting style may exert a nonspecific
of specificity and applicability of these models to the topic of influence on children’s trait anxiety, situationally-
our review. and behaviorally-specific parenting behaviors are
hypothesized to account for the development of predictors of childhood anxiety were not included.
specific anxiety disorders. Although attachment status may arguably reflect an
aspect of the parent–child relationship, traditional
attachment measures are based on observations of a
Four objectives of the review
child’s behavior with a parent, rather than a parent’s
Though the new models posit that parenting can play behavior with a child. Similarly, the family environ-
multiple roles in the development and maintenance ment (e.g., family cohesiveness and adaptability)
of childhood anxiety, it is unclear whether there is involves a complex interactive process among all
empirical support for these models. Our aim, there- family members and is not a direct measure of par-
fore, is to examine the existing literature to deter- enting style or behavior. Therefore, studies such as
mine which aspects of these models have research Stark, Humphrey, Laurent, Livingston, and Chris-
support. We address four key theoretical questions topher (1993), employing measures of the global
in our review: (a) How consistent is the evidence family environment, were excluded. Also excluded
suggesting that specific parenting styles or behav- were studies including triadic family conversations
iors are risk factors for anxiety development or in which mother-to-child and father-to-child inter-
maintenance, either as a main effect or in conjunc- actions were not differentiated from mother-to-father
tion with other risk factors, (b) to what degree do interactions, such as in Dadds, Barrett, Rapee, and
specific parenting styles or behaviors serve as risk Ryan (1996).
factors specific to childhood anxiety, as opposed to We employed hand searches, literature trails, and
other forms of child psychopathology, (c) what is the the PsychInfo computer database, which indexes
direction of effects linking parenting and childhood (with key terms) and abstracts all articles from 1887
anxiety, and (d) does the evidence suggest that to the present. The search covered a time period from
general parenting styles are linked with chil- 1990 to 2002. We used the following 12 anxiety-
dren’s trait anxiety, whereas specific parenting related key terms and synonyms: Internaliz- or Anxi-
behaviors are linked with specific child anxiety or Fear- or Obses- or Compul- or OCD or Panic or
disorders? Phobi- or Worr- or Inhibit- or Shy- or Somat-. These
terms were cross-referenced with the following par-
enting-related key terms: Father- or Maternal or
Mother- or Parent- or Paternal or Rearing or Socializ-.
Parenting and childhood anxiety:
These steps produced an initial pool of 5402 articles,
empirical findings
which were reduced in a stepwise fashion using title,
In order to address these four questions, the recent abstract, method section, and result section, to pro-
empirical literature, including studies published duce a pool of 21 studies published between 1991
between 1990 and 2002, was reviewed. The review is and 2001 that met inclusion requirements, none of
presented in two sections. The first section focuses which were included in the Rapee (1997) or Masia
on the empirical evidence for the three traditional and Morris (1998) reviews. The age of the children
parenting categories: parental acceptance, control, ranged between 4 and 18 years (see Table 1).
and modeling of anxiety. To highlight potential dif-
ferences between measures of parenting style and
Measure definitions
measures of parenting behaviors, the results of self-
report versus observational studies are discussed Guided by theory (e.g., Maccoby, 1992), factor-ana-
separately in the first section. The second section lytic findings (e.g., Schwarz et al., 1985), and other
focuses on the four theoretical questions discussed reviews and meta-analyses of parenting (Rapee,
above by highlighting studies that have addressed 1997; Rothbaum & Weisz, 1994), we grouped
these questions. measures of parenting style and behavior into three
dimensions: acceptance versus criticism and rejec-
tion, control versus granting of autonomy, and
Study selection
modeling of anxious behavior (see above for con-
Studies included in this review: (a) included a struct definitions). Coding of these categories was
measure of the current (rather than retrospectively performed by two judges (JW and BM) and interrater
reported) parenting style or behavior of one parent reliability was acceptable ðj ¼ :83Þ. Measures of
towards a target child, or separate measures of both anxiety included continuous or diagnostic meas-
parents’ style or behavior toward the child, (b) in- ures of anxiety symptoms (including shyness) or
cluded a measure of current childhood anxiety (in- disorders.
cluding shyness; excluding behavioral inhibition), (c)
tested the relation between parenting style or be-
Review format
havior and childhood anxiety, and reported test
statistics, and (d) reported a mean age of 18 years or This review highlights studies that make use of
younger for the participating children. Studies of separate informants for each construct under
attachment status and the family environment as investigation. Studies that examine the association
Table 1 Summary of previous studies of parenting and childhood anxiety 138
Hypothesis support
Parenting Anxiety Parenting Anxiety Mother’s Father’s

Study N Age Ethnicity Informant Informant Dimension Type Parenting Parenting
Dumas et al. (1995) 84a 4.1 1 O T Control Trait 1 of 1

O T Acceptance Trait 1 of 1
Gruner et al. (1999) 121 10.4 1 C C Modeling Trait 1 of 1 1 of 1
C C Control Trait 1 of 1 1 of 1
C C Acceptance Trait 1 of 2 1 of 2
Jeffrey J. Wood et al.
Hernandez-Guzman & Sandez-Sosa (1996) 3432 15–18 2 C C Acceptance Cog, Mot, Som 3 of 3
Hibbs et al. (1993) 101a 12–16 NA P Dx. Int. Acceptance OCD 1 of 1b
Hudson & Rapee (2001) 75a 10.7 NA O Dx. Int. Control Anx. Dx. 1 of 1
O Dx. Int. Acceptance Anx. Dx. 1 of 1
Hummel & Gross (2001) 30a 11.9 1 O C Acceptance Social 2 of 2b
Kliewer & Kung (1998) 98 10.7 NA C C+P Acceptance Trait 0 of 1
Mattanah (2001) 91 9.7 1 C T Control Trait 0 of 1 0 of 1
C T Acceptance Trait 0 of 1 0 of 1
Mills & Rubin (1998) 35a 6–10 1 O T+P Control Shyness 2 of 3
Muris et al. (1996) 64 12 1 P C Acceptance Fears 0 of 2 0 of 2
C C Acceptance Fears 0 of 2 0 of 2
Muris et al. (2000) 159 10.8 1 C C Modeling Worry 1 of 1 1 of 1
C C Control Worry 1 of 1 0 of 1
C C Acceptance Worry 1 of 2 1 of 2
Papini et al. (1991) 231 12.8 1 C C Acceptance Trait 1 of 1 1 of 1
Papini & Roggman (1992) 47 12.6 1 C C Acceptance Social 3 of 3 2 of 3
C C Acceptance Physical 2 of 3 2 of 3
Pedersen (1994) 573 15–19 NA C C Acceptance Unspec. 1 of 1 1 of 1
C C Control Unspec. 0 of 1 0 of 1
Rubin et al. (1999)c 60 2 1 P P Control Shyness 1 of 3 0 of 4
P P Control Shyness 3 of 6 1 of 4
Rubin et al. (2001) 188 4 1 O P+O Control Shyness 1 of 2
Scott et al. (1991) 2699 11–20 3 C T Acceptance Unspec. 1 of 1b
C P Acceptance Unspec. 1 of 1b
C C Acceptance Unspec. 1 of 1b
P T Acceptance Unspec. 0 of 1b
P P Acceptance Unspec. 1 of 1b
P C Acceptance Unspec. 1 of 1b
Siqueland et al. (1996) 44a 11 1 C Dx. Int. Acceptance Anx. Dx. 1 of 1
C Dx. Int. Control Anx. Dx. 0 of 1
P Dx. Int. Acceptance Anx. Dx. 0 of 1
P Dx. Int. Control Anx. Dx. 0 of 1
O Dx. Int. Acceptance Anx. Dx. 0 of 1
O Dx. Int. Control Anx. Dx. 1 of 1
Stevenson-Hinde & Glover (1996) 126a 4–4.5 NA O M+T+O Acceptance Shyness 2 of 8
Tesser & Forehand (1991) 147 13.2 NA C T Acceptance Trait 0 of 2 0 of 2
between two self-report measures from the same
Parenting
in years (either mean or a range). For ethnicity, 1 ¼ primarily Caucasian, 2 ¼ primarily Latino/Latina, 3 ¼ multiple ethnicities/nationalities, NA ¼ not available. For Parenting and Anxiety
Informant, Dx. Int. ¼ Diagnostic interview, C ¼ Child, P ¼ Parent, T ¼ Teacher, O ¼ Observer (multiple informants aggregated into a summary score are indicated by a ‘+’). For Parenting
Unspecified measure of anxiety, Cog, Mot, Som ¼ cognitive, motoric, and somatic anxiety. Under Hypothesis Support is reported the number of significant findings linking child anxiety
Dimension, Modeling ¼ Modeling of anxious behaviors. For Anxiety Type, Anx. Dx. ¼ children with any (or multiple) anxiety disorders, rather than a specific anxiety disorder, Unspec. ¼
Father’s
with the specified parenting dimension of the number of statistical tests performed. Multiple lines for one study reflect different informants or different aspects of parenting that were
NOTE: All significant effects were in the expected direction except for Rubin et al. (2001) in which both effects were significant but one was in the opposite direction. Age is reported
source (e.g., Papini & Roggman, 1992) are at risk
The first row for Rubin et al. (1999) concerns analyses in which one parent rated child anxiety and the other parent rated parenting. The second row concerns analyses in which
Hypothesis support
for obtaining inflated correlations due to shared
method variance (Campbell & Fiske, 1959). To
provide a more conservative test that eliminates
these potentially confounding effects, it is prefer-
able for each construct (i.e., parenting, childhood
Parenting
Mother’s
3 of 6
2 of 2
1 of 2
anxiety) in correlational studies to be rated by
different informants. However, studies that used
only one informant for both constructs were also
included in Table 1.
To provide an estimate of the consistency of ef-
fects across analyses and studies, we report the
Anx. Dx.
Anx. Dx.
Anx. Dx.
Anxiety
ratio of significant statistical tests (i.e., tests with a

Type
p value of .05 or lower) to total tests conducted in

all of the multiple-informant studies. In addition,
the percent of variance in anxiety accounted for by
the parenting variables in these studies is dis-
cussed. Standard formulas were used to convert
Acceptance
Dimension
Modelling
Parenting
mean differences to correlations (and, then, to

Control
percents of variance explained, i.e., r 2 ) in group-

comparison studies.
Studies of parental acceptance

Informant
Dx. Int.
Dx. Int.
Dx. Int.
Anxiety
Child-report studies. Eleven studies were identified

Ratings of mothers and fathers, separately, were combined into a composite measure of parenting.
that used children’s reports of accepting parental

style (see Table 1). Of these, five included ratings of
children’s anxiety from teachers, parents, or clini-
cians. For instance, in a study of 2699 adolescents
(age 11–20 years) recruited from community high
Informant
Parenting
schools located in a variety of countries (including

Australia, China, United States, Germany, Japan,
O
O
O
and Taiwan), adolescents who reported more

parental nurturance (i.e., acceptance) tended to be
rated by their parents and teachers as less anxious
than did adolescents who reported less nurturance
(Scott, Scott, & McCabe, 1991; see Table 1). In this
Ethnicity
Sample divided into anxious group and normal control group.
study, the effects accounted for 1% and 3% of the

1
variance in parents’ and teachers’ ratings of child-

hood anxiety, respectively, corresponding to
Cohen’s (1988) criteria for a ‘small’ effect. A notable
weakness of this study was the use of brief
unstandardized measures of parenting and anxiety,
10.7
Age
with some alpha coefficients as low as .44 (range

.44–.77). Using a sample of 17 clinic-referred anxi-
the same parent rated both constructs.
ety-disordered children (with one of several diag-

noses) and 27 matched control children, Siqueland,
34a
Kendall, and Steinberg (1996) found children’s rat-

N
ings of maternal acceptance accounted for 21% of

measured in the same study.
the variance in diagnostic status (a ‘large’ effect).

A majority of the 11 studies used community-
based, school samples, and in these studies, 3 of 10
Whaley et al. (1999)
Table 1 Continued
statistical tests yielded significant relations when

children rated parental accepting style and a sepa-
rate judge rated childhood anxiety, all of which were
in the expected direction. However, the majority of
Study
effects were not statistically significant, suggesting

that children’s ratings of accepting parental style are
b
a
c
not reliably related to others’ reports of children’s coders using a global 5-point scale rated maternal
anxiety.2 warmth, positivity, and criticism (i.e., lack of ac-
ceptance). Anxious mothers of children with an
Parent-report studies. Four studies using parent anxiety disorder ðn ¼ 10Þ were rated as less warm
reports of an accepting parenting style and a separ- and positive, and more critical, than non-anxious
ate judge for ratings of childhood anxiety (i.e., chil- mothers of children who did not meet criteria for an
dren, teachers, etc.) were identified (see Table 1). anxiety disorder ðn ¼ 16Þ, and all of the effects were
Findings across these studies were inconsistent. For in the ‘large’ range (i.e., explaining more than 25% of
instance, in a study of 64 clinic-referred Dutch youth the variance). However, anxious mothers of children
(age 8 to 18 years) presenting at an outpatient clinic with no disorder ðn ¼ 8Þ did not differ from anxious
with a diverse range of psychiatric problems (e.g., mothers with anxious children on these dimensions.
anxiety disorders, disruptive behavior disorders, and A subsequent study, currently under review, inclu-
conditions not attributable to a mental disorder), the ded a group of non-anxious mothers with anxious
correlation between fathers’ and mothers’ self- children; the patterns found in Whaley et al. were
reported positive and negative behavior on the replicated for maternal warmth and positivity, but
EMBU (i.e., acceptance and aversiveness) and chil- criticism was entirely related to child anxiety status
dren’s ratings of their own anxiety symptoms was in this second study (Moore, Whaley, & Sigman,
not significant (Muris, Bogels, Meesters, van der under review). These two studies suggest that ma-
Kamp, & van Oosten, 1996). A strength of this study ternal anxiety rather than child anxiety could be an
was the use of standardized measures of each con- underlying factor accounting for reduced maternal
struct. In these studies, 2 of 9 correlations were warmth and positivity, but that maternal criticism
significant and both were in the expected direction. could play an important role in the development or
Hibbs, Hamburger, Kruesi, and Lenane (1993) maintenance of child anxiety disorders.
compared 49 children diagnosed with OCD with 41 Using the same coding system for warmth,
normal control children, on parental expressed Siqueland and colleagues (1996) compared 17 clin-
emotion (EE) as rated on a five-minute speech-sam- ically anxious children and 27 control children
ple test. Children with OCD were less likely than matched to the anxious children, and did not find
normal children to have both parents exhibit low EE group differences in ratings of maternal warmth.
(i.e., acceptance), and EE explained 5% of the vari- Another observational study grouped children into
ance in children’s diagnostic status. Of course, be- ‘low,’ ‘medium,’ and ‘high’ shyness categories on the
cause high EE was defined as high criticism or high basis of multiple informants and observations (Ste-
emotional overinvolvement, these results do not in- venson-Hinde & Glover, 1996). However, the sample
dicate which aspect of EE may have played a more size for the low shyness group was low for boys and
important role. In the Scott et al. (1991) study, par- girls (n ¼ 10 for each gender group). Even when
ent reports explained less than 1% of the variance in comparing only the medium-shy (ns ¼ 27 and 32,
children’s self-reports of anxiety. Thus, most effects respectively) and the high-shy children (ns ¼ 24 and
were not statistically significant, and the two that 9, respectively), observations of maternal positive
were fell in the ‘small’ range (Cohen, 1988). It should style during unstructured interactions at home, and
be noted that with the exception of Scott et al., the positive interactions during a structured drawing
other three studies in this category used clinical task at home, were significantly higher in the me-
convenience samples (i.e., Hibbs et al., 1993; Muris dium-shy group in only 2 of 6 tests of significance
et al., 1996; Siqueland et al., 1996). (the percent of variance explained could only be
computed for one of these two tests, and it was 9%, a
Observational studies. Five studies that used ob- ‘medium’ effect). There were no significant group
server ratings of accepting parental behavior were differences on maternal negative interactions during
identified (see Table 1). For instance, Whaley et al. the drawing task (0 of 2 tests). Hudson and Rapee
(1999) examined the family interaction patterns of (2001) compared 43 clinic-referred anxiety-dis-
anxious mothers and their 7–14-year-old children. ordered children (with one of several disorders) to 32
Anxious mothers were recruited from clinics and the control children recruited from the community.
community, and their diagnoses were confirmed Ratings of maternal negativity during laboratory-
with a structured diagnostic interview. Control based conversations revealed that mothers of anxi-
mothers were recruited from the community and did ous children were more negative (i.e., less accepting)
not meet criteria for any disorders. In this primarily than mothers of control children, accounting for 13%
middle-class and Caucasian sample, mothers were of the variance in diagnostic status (a medium ef-
asked to talk with their children about three topics fect). In a study of trait-anxious ðn ¼ 42Þ and nor-
(e.g., a parent–child conflict and something that mally developing ðn ¼ 42Þ preschoolers, identified by
made the child feel nervous) for 15 minutes. Trained scoring above or below anxiety cutoff scores on a
teacher rating scale, ratings of maternal positivity
2
Predictably, the single informant self-report studies yielded a (i.e., acceptance) during a parent–child problem-
high proportion of significant effects (see Table 1). solving game task were considerably higher for
mothers of normal as compared to anxious children, Asendorpf, 1999). Across the concurrent and pro-
and 55% of the variance in group status was ex- spective analyses, 1 of 7 tests (mother’s ratings) was
plained (Dumas, LaFreniere, & Serketich, 1995). In significantly and positively correlated with the other
each of these studies, children’s anxiety was as- parent’s ratings of children’s shyness; 7% of the
sessed by structured interviews or normed ques- variance in shyness was explained in this test, a
tionnaire measures, and behavioral observations small effect. There is insufficient data to draw con-
yielded acceptable inter-rater reliabilities. Therefore, clusions about the possible linkage between parent
this group of studies was characterized by superior reports of controlling style and childhood anxiety.
methodological attributes.
Overall, the five observational studies showed Observational studies. More consistent findings
significant relations between observer ratings of were obtained in the six studies using observer rat-
parental accepting behavior and self-reports or cli- ings of parental controlling behavior versus auton-
nician diagnoses of childhood anxiety in 8 of 18 omy-granting (see Table 1). In the Whaley et al.
statistical tests, all of which were in the expected (1999) study, observers rated maternal psychologi-
direction. This body of evidence is moderately con- cal granting of autonomy during 15 minute conver-
sistent with the proposition that variations in par- sations on a 5-point scale, based on mothers’
ental acceptance, as rated by observers, are solicitation of children’s opinions, acceptance of
associated with variations in childhood anxiety. differences of opinion, and other aspects of demo-
However, there is also evidence suggesting that ma- cratic parent–child interactions. Anxious mothers of
ternal anxiety could act as a third variable explaining clinically anxious children were rated as less grant-
away this correlation, at least for maternal accepting of autonomy than anxious mothers of nondiag-
ance. In contrast, maternal criticism could be nosed children and non-anxious mothers of
directly linked with children’s anxiety. nondiagnosed children. This pattern of findings was
replicated in the subsequent study (Moore et al.,
Studies of parental control under review), suggesting that maternal control
is related to child anxiety status rather than to
Child-report studies. Results were inconclusive for
mother’s anxiety status. In the Siqueland et al.
studies of children’s reports of controlling parenting
(1996) study, observers utilized the same coding
style (see Table 1). All but two studies were based on
system as in the Whaley et al. study, and also rated
children’s ratings of both constructs, presenting the
mothers of clinically anxious children as less grant-
problem of respondent variance. In one of the studies
ing of autonomy than mothers of nondiagnosed
in which separate judges (clinicians) rated childhood
children. Mills and Rubin (1998), Rubin, Chea, and
anxiety and children rated parental control, there
Fox (2001), Dumas et al. (1995), and Hudson and
was no significant association between group status
Rapee (2001) also reported significant effects using
(anxiety-disordered versus normal) and children’s
observational measures. As with the observational
ratings of maternal control (Siqueland et al., 1996;
studies of parental acceptance, this group of studies
see Table 1). In the second study, normal fourth-
was characterized by sound methodological charac-
grade children’s ratings of fathers’ (but not mothers’)
teristics such as the use of standardized measure-
granting of autonomy were positively associated with
ment instruments and reliable observational rating
teacher’s ratings of childhood anxiety, accounting for
systems.
9% of the variance – a medium effect (Mattanah,
In these six studies, 8 of 10 statistical tests yielded
2001). In these two studies, 1 of 3 statistical tests
significant relations in the expected direction, with
was significant (in the expected direction); due to the
all effects except one in the medium (i.e., 9% or more
limited evidence pertaining to childhood anxiety and
variance in anxiety explained) or large (i.e., 25% or
controlling parental style as rated by children, no
more of the variance) range. The one effect in the
definitive conclusions appear to be warranted con-
small range (3% of the variance in shyness) was re-
cerning this aspect of the literature.
ported by Rubin et al. (2001) in a study of 188 nor-
Parent-report studies. Two studies employed parent mal 4-year-olds recruited from the community. In
reports of controlling parental style and a separate this study, mother’s oversolicitous (i.e., controlling)
rater of childhood anxiety (see Table 1). In the behavior during a free play interaction was positively
Siqueland et al. (1996) study, described above, there associated with a composite measure of children’s
was no difference between the anxiety-disordered shyness based on mother’s report and behavioral
and control groups on mothers’ ratings of maternal observation. One statistical test, also reported by
control on the Children’s Reports of Parental Be- Rubin et al., yielded a significant effect in the op-
havior Inventory. The second study involved 60 posite direction – more parental oversolicitous be-
young children recruited from the community, and havior during a structured play period was
attained Q-sort ratings of parental control and mul- associated with less child shyness. However, the
tiple questionnaire ratings of children’s shyness authors had specifically predicted this finding for the
from both parents when the children were age 2 particular observational task in question (a highly
years and again at age 4 (Rubin, Nelson, Hastings, & structured manipulative task that would have been
difficult to accomplish for preschoolers without Evidence of main effects vs. interactions with other
considerable parental direction). The issue of situa- risk factors. Most studies examined parenting
tional context in determining the effect of parental style or behavior in isolation, without considering
behavior on children’s anxiety is discussed below. potential interaction effects with other risk factors.
Overall, these findings suggest that children who are Observer ratings of parental control showed the most
clinically anxious or shy are likely to have mothers consistent linkages with childhood anxiety in these
who are relatively more controlling and less likely to studies, and specifically, with children’s anxiety-
grant autonomy during parent–child conversations. disordered diagnostic status (in the case of three
studies), shyness (in the case of two studies), and
Studies of parental modeling of anxious behaviors teacher-rated trait anxiety (one study). Observa-
tional ratings of parental acceptance also yielded a
Child-report studies. No child-report studies were modestly consistent pattern of significant effects (in
identified that used separate informants. Nonethe- 8 of 18 statistical tests in five studies). Of course, the
less, two recent studies using separate samples of use of cross-sectional methodology has only estab-
normal Dutch school children as single informants lished that these parenting behaviors may play a role
found evidence that children’s reports of ‘anxious as a risk factor for childhood anxiety in a simple,
rearing’ styles by both mothers and fathers, as additive manner (i.e., independent of context).
exemplified by questionnaire items such as ‘your Relatively few studies examined the role of context
parents warn you against all possible dangers,’ were (e.g., the presence of other risk factors) in moder-
associated with children’s reports of greater anxiety ating or explaining the relation between parenting
symptoms (Gruner, Muris, & Merckelbach, 1999; and childhood anxiety. As described above, Whaley
Muris, Meesters, Merckelbach, & Hulsenbeck, 2000). et al. (1999) found that maternal anxiety disorders
These findings suggest that further research may be of proved to be an important contextual factor ac-
value. In these studies, 4 of 4 statistical tests were counting for the apparent association between ma-
significant, suggesting that children who view their ternal acceptance and children’s anxiety disorder
parents as engaging in less modeling of anxious diagnostic status. Specifically, when mothers were
behavior also experience less anxiety themselves. clinically anxious, they were less warm and positive
Parent-report studies. We were unable to find any during interactions with their children, whether or
parent-report studies of parental modeling of an- not their children had an anxiety disorder. Papini,
xious behaviors and children’s anxiety symptoms. Roggman, and Anderson (1991) examined the mod-
erating role of pubertal status in the relation between
Observational studies. Whaley et al. (1999) found parental acceptance and adolescents’ trait anxiety in
that anxious mothers of clinically anxious children a group of 231 seventh-graders, and found no evi-
were more likely than control mothers of non- dence of an interaction effect. Finally, as mentioned
diagnosed children to discuss problems with their above, Rubin et al. (2001) found that the context of
children in catastrophic terms that emphasized the interaction situation itself moderated the relation
one’s lack of control over the problem, or one’s lack of between maternal oversolicitous parenting and
ability to cope effectively with the problem. In fact, children’s shyness. Although oversolicitous, excess-
47% of the variance in children’s diagnostic status ively controlling parenting during a free play period
was explained by maternal catastrophizing. In the was linked with more shyness in children, the exact
subsequent study (Moore et al., under review), a same kind of parenting in a structured play task was
main effect for maternal anxiety was found, as well linked with less shyness, suggesting that the inter-
as an interaction effect suggesting that non-anxious actional context in which parental control emerges
mothers were more likely to catastrophize with clin- may have important implications for possible effects
ically anxious children than with non-anxious chil- on children’s regulation of shyness or social anxiety.
dren. These findings suggest one possible Parents who are highly controlling during activities
mechanism through which poor anxiety regulation that could be directed by their children, such as free
skills may be passed from parents to their anxious play, unintentionally deprive their children of control
children: Parents who model poor coping strategies, experiences in social contexts that are important for
such as catastrophizing and avoidance, are more the development of confidence and assertiveness.
likely to have children who lack the ability to regulate Inspection of the demographic characteristics of
fear and anxiety effectively. these studies provides little evidence of other mod-
erating contextual processes. For example, mothers’
and fathers’ parenting were each associated with
Evidence for contemporary models
patterns of childhood anxiety in a comparable
of the ontogeny of childhood anxiety
manner (see Table 1). Although the majority of the
Four main issues concerning the linkage between studies used primarily Caucasian samples, signifi-
childhood anxiety and parenting style or behavior cant effects were also obtained in a sample of Mex-
were raised in the introduction and are now ican children (Hernandez-Guzman & Sanchez-Sosa,
addressed. 1996) and a sample of Asian children (Scott et al.,
1991). Child age also did not appear to moderate the predicting a time 2 outcome) and none controlled for
pattern of findings, as significant effects were re- initial levels of anxiety when predicting later anxiety.
ported for samples ranging from preschool age to In the study of 60 two-year olds conducted by Rubin
adolescence (e.g., Pedersen, 1994; Stevenson-Hinde et al. (1999; described above), a pattern emerged
& Glover, 1996). So although few studies examined indicating that early shyness predicted subsequent
the moderating effects of context, it is clear that such parental control, rather than the reverse. Mothers
studies can illustrate the mechanisms that influence and fathers who rated their children as more shy at
how much parenting style or behavior may act as a age two were significantly more likely to rate their
risk factor for childhood anxiety. own level of control as higher than other parents’
when children were age four, in 4 of 5 statistical tests
Specificity of parenting to anxiety vs. other forms of (however, 3 of these effects were based on a single
psychopathology. Four studies included a non- informant’s ratings of both constructs). In contrast,
anxious clinical comparison group as well as a nor- initial parental control was not significantly associ-
mal control group. Such studies are useful for ated with shyness at age four. The Pedersen (1994)
identifying whether particular parenting styles or study found that normal adolescents’ ratings of
behaviors are specific to child anxiety in particular, mothers’ and fathers’ acceptance (but not control)
or are common to child psychopathology in general. predicted their self-reports of anxiety symptoms in
In the Dumas et al. (1995) study, described above, the follow-up assessment approximately six months
an additional group of teacher-identified aggressive later. Of course, method variance could have ac-
preschoolers ðn ¼ 42Þ were compared to the anxious counted for this finding, since one informant was
sample ðn ¼ 42Þ, and observational ratings showed employed to rate both constructs. Finally, although
that mothers of aggressive children were more pos- Papini and Roggman (1992) obtained three sets of
itive and less controlling with their children than repeated measures from their sample of 47 12-year-
were mothers of anxious children. olds between sixth and seventh grades, they only
The remaining three studies found nonsignificant reported concurrent associations among the meas-
results. In a study of 573 Norwegian adolescents ures. However, they did examine patterns of change
recruited from the community and grouped into in the magnitude of the correlations and found evi-
normal, anxious, delinquent, and anxious/delin- dence of an increase of the relation between parental
quent groups on the basis of self-report measures, acceptance and children’s anxiety (a negative as-
teens in the three ‘problem’ groups rated fathers sociation) over time. These researchers hypothesized
and mothers lower in acceptance than did teens in that the transition into junior high may have mag-
the normal group (Pedersen, 1994). However, there nified the impact of a lack of parental acceptance on
were no significant differences among the three children’s emotional well-being. It is unfortunate
‘problem’ groups on acceptance, suggesting that that longitudinal associations were not also exam-
low acceptance was a risk factor for psychopathol- ined. Few implications about the direction of effects
ogy in general, rather than just for anxiety. Sim- linking parenting style or behavior and childhood
ilarly, Hudson and Rapee (2001; sample described anxiety are evident in this limited set of longitudinal
above) found no difference between their anxiety- findings.
disordered group ðn ¼ 43Þ and oppositional-defiant
disordered (ODD) comparison group ðn ¼ 20Þ on Parenting determinants of trait anxiety vs. specific
observed maternal acceptance and control, even symptoms or disorders. Craske’s (1999) model
though both groups differed from the control group suggests that general parenting style across situa-
on these measures, with one exception (the ODD tions may contribute nonspecifically to child trait
group and control group did not differ on accept- anxiety. The correlational studies reviewed above
ance). In the Hibbs et al. (1991) study, in addition to provided little evidence suggesting that children’s
the OCD and normal groups described above, chil- and parent’s reports of global parenting acceptance
dren with disruptive behavior disorders ðn ¼ 34Þ (across situations and time) were linked with chil-
were included as a clinical comparison group, and dren’s trait anxiety, as measured by self-report
they did not differ from the OCD group in rates of measures such as the Revised Children’s Manifest
parental high EE. Thus, although nonsignificant Anxiety Scale (RCMAS; Reynolds & Richmond,
results from 3 of 4 studies do not provide conclusive 1978). None of the six statistical tests in the three
evidence, a tentative synthesis of these studies multiple-informant self-report studies that ex-
suggests that a lack of parental acceptance and amined this linkage were significant (i.e., Kliewer &
excessive parental control may be risk factors for Kung, 1998; Mattanah, 2001; Tesser & Forehand,
various forms of psychopathology, and may not be 1991). Additionally, there is too little evidence to
specific to anxiety. evaluate the association between self-reported glo-
bal parental controlling style or modeling of anxious
Direction of effects. Only three of the reviewed behavior and children’s trait anxiety. Only one
studies were longitudinal. Two of these provided study obtained self-reports of parental controlling
prospective analyses (e.g., a time 1 measure style and children’s trait anxiety from separate
informants, and a significant effect (in the expected critical parenting behaviors in novel conversational
direction) was found for fathers’, but not mothers’, contexts, such as those involved in laboratory tasks,
parenting in this study (Mattanah, 2001). No stud- could be particularly influential in shaping chil-
ies met this basic criterion for parental modeling. dren’s patterns of sociability and shyness. But it is
Therefore, a conservative view would suggest that equally possible that shy, inhibited children influ-
there is very little evidence for a linkage between an ence their parents’ patterns of communication with
accepting parental style and children’s trait anxiety, them in these contexts over the years.
and there is also insufficient data to adequately test
the association between parental controlling style or
modeling and trait anxiety. Summary of findings
No studies have examined the role of situationally- Several studies suggested that the context in which
specific parenting practices or behaviors as a con- parenting behaviors occur – including the particular
tributing factor to the development of particular nature of the situation as well as the parent’s own
anxiety disorders in children with high trait anxiety symptoms of anxiety – plays an important role in
(i.e., Craske’s [1999] second pathway). However, moderating or explaining linkages between parent-
three studies compared children who were already ing and childhood anxiety. Three studies indicated
diagnosed with one of several anxiety disorders to that parental warmth and control are not specifically
normal control children using behavioral observa- related to anxiety problems in children, but rather to
tion methodology (i.e., Hudson & Rapee, 2001; risk for psychopathology in general. This illustrates
Siqueland et al., 1996; Whaley et al., 1999). Each of the concept of multifinality proposed by develop-
these studies demonstrated that parental controlling mental psychopathologists, or the notion that the
conversational behavior was more prominent in same risk factor (i.e., parenting) may have multiple
parent–child dyads that contained a child with an outcomes (i.e., different kinds of psychopathology).
anxiety disorder. Whaley et al. (1999) also found that Very few longitudinal examinations were conducted
parents of anxiety-disordered children were less that might shed light on the direction of effects
warm, more critical, and more likely to model anxi- linking parenting behavior and childhood anxiety,
ous behavior than were parents of normal control and preliminary results appeared consistent with
children during laboratory-based conversations. It is parenting as either a cause or an effect of children’s
unclear whether discomfort with the novel situation manifestations of anxiety. There was very little evi-
precipitated these parental interactional behaviors, dence supporting the contention that general par-
or whether a long-standing pattern of these parent- enting style was related to children’s trait anxiety.
ing behaviors in novel situations may have directly However, observed parental controlling behaviors
contributed to the development or maintenance of during parent–child interactions was consistently
the children’s specific patterns of anxiety. linked with shyness and child anxiety disorder sta-
Six studies examined parenting correlates of chil- tus across studies. The effect sizes in these latter
dren’s shyness and social anxiety (see Table 1). Al- studies were almost all in the medium or large range,
though these studies did not diagnose the children in indicating that at a minimum, parental control dur-
question, their focus on a specific constellation of ing parent–child conversations in novel situations
anxiety symptoms merits comment. Four of these represents a clinically significant area of intervention
studies employed observational methodology that and further research.
may reflect on parenting behavior in novel interac-
tion situations. In a sample of 30 sixth-graders who
were selected for either scoring in the clinical range
Limitations of the extant literature
on a self-report measure of social anxiety ðn ¼ 15Þ or
for scoring in the normal range on this measure Although the models of Craske (1999) and Vasey and
ðn ¼ 15Þ, Hummel and Gross (2001) found a strong Dadds (2001) are useful for conceptualizing the dif-
association between observed parental acceptance ferent roles that parenting style and behavior play in
during conversations related to a puzzle task com- the development of childhood anxiety and its disor-
pleted in children’s homes and social anxiety group- ders, these models are relatively new and have not
status (7–85% of the variance in group status was guided much empirical research to date. Further-
explained by mothers’ and fathers’ acceptance more, previous research does not provide a satis-
across analyses). Mills and Rubin (1998) also factory test of most aspects of these models. Four
grouped 6–10-year-old children ðn ¼ 35Þ into high- characteristics of the empirical literature limit the
shy and low-shy groups and found evidence of more conclusions and action implications that can be
observed maternal control in the high-shy group. drawn from the results: (a) the studies generally
Rubin et al. (2001) and Stevenson-Hinde and Glover employed nonrepresentative samples, (b) the studies
(1996), both discussed above, also found some evi- relied primarily upon self-report methodology, (c)
dence of greater parental control and less parental nearly all of the studies were cross-sectional, and (d)
acceptance during videotaped interaction tasks most of the studies employed global, rather than
with shy, socially inhibited children. Controlling or specific, measures of parenting.
Additionally, the studies using diagnostic group-

Representativeness of samples
ings of children were at risk for (a) obtaining ex-
One limitation was the homogeneity of the samples. aggerated results due to the use of extreme groups
Most of the studies employed samples that were and (b) obtaining non-generalizable results due to
predominantly Caucasian, and few studies exam- the use of convenience samples. Thus, studies of
ined the relationship between parenting style or be- anxiety-disordered children may have little applica-
havior and childhood anxiety among different ethnic tion to the effects of parenting style and behavior
and cultural groups (see Scott et al., 1991 for a not- on variations of anxiety symptoms in the general
able exception). Ethnic differences in parenting may population.
therefore limit the generalizability of the findings In addition to the four studies of clinically diag-
from the reviewed studies. For example, there is de- nosed children discussed above, four additional
bate over whether Baumrind’s (1971) parenting studies using continuous measures of child anxiety
typologies (e.g., authoritative, authoritarian, and or shyness created arbitrary groups of children (e.g.,
permissive) apply equally to other ethnic and racial with a median split; Dumas et al., 1995; Hummel &
groups (e.g., Dornbusch, Ritter, Leiderman, Roberts, Gross, 2001; Mills & Rubin, 1998; Stevenson-Hinde
& Fraleigh, 1987; Steinberg, Dornbush, & Brown, & Glover, 1996). The use of median splits may have
1992); concepts such as ‘strictness’ and ‘control’ artificially reduced meaningful variance in anxiety,
may have different meanings for people of different leading to reduced statistical power.
cultures (e.g., overinvolvement, parental concern, or
organizational control). Therefore, it is unclear whe-
Cross-sectional designs
ther the findings reviewed above apply to children
from different cultures. Cross-sectional designs are generally employed to
establish basic relations between target constructs
(e.g., childhood anxiety and parenting). Such de-
Reliance on self-report measures
signs cannot, however, establish the direction of ef-
The majority of the reviewed studies used self-report fects. Given the prevalence of cross-sectional
data as the sole measure of parenting, and the va- designs, and the lack of prospective or experimental
lidity of this type of data has been questioned. Child- studies, it appears that there may have been a mis-
report measures may make unrealistic demands on conception in the field about the kind of information
long-term memory and information-processing skills provided by a cross-sectional design. The use of
by asking children to make generalizations about cross-sectional designs has demonstrated that a re-
patterns of past interactions (Holden & Edwards, lation between specific parenting behaviors and
1989). Furthermore, children who are anxious at the childhood anxiety disorders or shyness may exist,
time of assessment may provide biased reports of but the direction of this relationship cannot be elu-
their parents’ behavior, possibly resulting in more cidated with this design. Without empirical data
negative appraisals of parenting. A social desirability supporting the direction of effects, the applicability
bias may limit the accuracy of parent-report data as of this body of research to theory development or
well. As noted above, measures of parenting style clinical use is limited.
were confounded with self-report methodology, as
were measures of parenting practices or behaviors
Global parenting measures
with observational methodology. Given the limita-
tions of typical self-report parenting measures, the The broad nature of the existing theoretical categ-
role of parenting style in contributing to child trait ories of parenting style and behavior (e.g., accept-
anxiety may have been underestimated in these ance and control) limit the direct clinical, theoretical,
studies. and empirical applications to be derived from re-
A limitation of the self-report anxiety ques- search findings. Establishing a general relation be-
tionnaires employed in many studies is that these tween global parental accepting style and current
measures (e.g., RCMAS; Reynolds & Richmond, child anxiety symptoms advances the field little be-
1978) have limited validity information and have not cause such findings do not help refine clinical
generally been found to discriminate between chil- practice or theoretical models. For example, ques-
dren with other clinical problems, such as attention tionnaire items designed to measure an accepting
deficit/hyperactivity disorder (cf. Reiss, Silverman, parental style refer to speaking to one’s child in a
& Weems, 2001). Current self-report child anxiety warm and friendly voice. But this parenting ap-
measures with better validity data were not em- proach may not always protect children from anxi-
ployed in most of the reviewed studies, but should be ety. Speaking to a child in a warm and friendly voice
considered for future research (e.g., the Multi- for a prolonged period immediately following the
dimensional Anxiety Scale for Children [MASC]; child’s avoidance of a feared situation may reinforce
March, 1998; March, Parker, Sullivan, Stallings, & the anxious behavior (i.e., by acting as a reward for
Conners, 1997; Wood, Piacentini, Bergman, avoidance). As Rubin et al. (2001) demonstrated,
McCracken, & Barrios, 2002). high levels of parental control in certain interaction
situations could inhibit children’s sociability, which parenting style and behavior may contribute
whereas in other situations, parental control may to children’s trait anxiety and anxiety disorders.
provide children with needed structure, suggesting Some theoretical works have highlighted specific
that interactional contexts can influence the impact conditioning mechanisms that might link parenting
of parenting behavior on children’s anxiety. Greater behavior to the development of anxious states and
specificity of parenting behaviors may help to further patterns of avoidance in children (e.g., Ollendick,
define the role of parent–child interactions in influ- Vasey, & King, 2001). However, there are few theory-
encing childhood anxiety, as well as provide clini- derived hypotheses proposed in the literature about
cians and parenting specialists with more practical specific situations in which particular parenting
advice for parents. The use of broad, generic par- behaviors might impact children’s experiences of
enting categories such as ‘acceptance’ in past re- anxiety.
search may have resulted from the lack – until very
recently – of theoretical models that provide a Two heuristics to guide research and clinical
framework for conceptualizing the role of specific interventions with anxious children. We propose
parenting behaviors in particular contexts as a two heuristics for the development of research and
causal factor in childhood anxiety. clinical hypotheses intended to identify specific
parenting behaviors that might elicit or maintain
manifestations of child anxiety in specific situations.
Summary of limitations
Because the reviewed studies consistently found
The methodological characteristics of past studies observed parental control to be related to childhood
limit conclusions about the nature of the relation- anxiety disorders and shyness, these heuristics ex-
ship between parenting style or behavior and child- plore two specific ways in which controlling, over-
hood anxiety. The use of nonrepresentative samples, solicitous, or overinvolved parenting may contribute
self-report measures, and cross-sectional designs to particular child anxiety symptoms.
constrains our understanding of the generalizability, We offer one heuristic based on human and animal
validity, and direction of effects of previous findings. models of control and mastery as determinants of
Perhaps most significantly, the reliance on nonspe- anxiety (e.g., Chorpita & Barlow, 1998; Ray & Sa-
cific theories and measures of ‘parenting style’ may polsky, 1992): When parents fail to provide children
have contributed to a lack of precision in previous with the opportunity to engage in age-appropriate
studies, leading to the relatively inconclusive state of self-help behaviors (e.g., dressing and grooming) and
the research literature. However, the consistent to experience developmentally appropriate parent–
findings in the observational research on specific child boundaries at home (see Sroufe, Jacobvitz,
parenting behaviors, and particularly parental con- Mangelsdorf, DeAngelo, & Ward, 1985), children
trol, provide an important basis for conducting may not develop a sense of control, mastery, and
future research. autonomy. Instead, children may experience an in-
creasing sense of dependence on the parents. Among
children with high trait anxiety, this dependence and
Future directions the pattern of contingencies that reinforce it might
result in separation anxiety.
The limitations of the extant literature and the
This heuristic provides a framework for generating
questions raised about contextual factors, the di-
specific, testable hypotheses. Specifically, children
rection of effects, and the specificity of particular
who are below age-norms in adaptive functioning
parenting behaviors to childhood anxiety necessitate
and independence from parents may be at risk for
further evaluation and refinement of our theoretical
developing a sense of helplessness, which has been
models. In this section, we present suggestions for
associated with increased anxiety (cf. Chorpita &
refining and extending models of the relation be-
Barlow, 1998). For example, when children are at the
tween parenting and childhood anxiety, as well as
age of middle-school entry (i.e., 11 years old), it is
recommendations for research designs that could
normative in American culture for children to dress,
clarify issues of timing and the direction of effects.
bathe, and sleep independently and privately, and
for physical affection between parents and children
to be less intense (e.g., less frequent lap-sitting). An
Further specification of theories and hypotheses
11-year-old who has not reached these age-norms
Greater theoretical specificity is needed to guide re- might experience lowered self-efficacy and a less-
search on the role that parents may play in the de- ened sense of mastery when comparing his or her
velopment of child anxiety disorders, as well as to repertoire of adaptive and independent behaviors to
inform intervention and prevention programs tar- those of his/her peers. Parental participation in the
geting childhood anxiety. Contemporary models of child’s self-help routines and initiation of (or re-
anxiety development (i.e., Chorpita & Barlow, 1998; sponsiveness to) frequent, intense physical affection
Craske, 1999; Vasey & Dadds, 2001) are useful for might reinforce and maintain the child’s immature
conceptualizing the multiple pathways through level of functioning, and thus, his/her lowered
self-efficacy and mastery – even if the parent’s be- most social contexts, such as those described by
havior was not the initial cause of these routines or Rubin et al. (1997), may be very slow to habituate to
the child’s anxiety. Because the child’s anxiety about novel situations and thus may extinguish alternative
his/her lack of mastery may be temporarily dimin- parenting behaviors (such as pushing for and rein-
ished when the parents are present to assist and forcing children’s social approaches) that could po-
comfort the child, the child may develop a specific tentially lead to increased child sociability over time.
anxiety about separation from the parents. Parental Therefore, parental anxiety and the pervasiveness of
encouragement of normative self-help skills and children’s inhibition may provide important contexts
parent–child boundaries may therefore represent for the emergence of excessively responsive parental
a fruitful area for further research, as well as an reactions to children’s anxiety that eventually con-
important area of assessment and intervention in tribute to the development of a specific set of fears
clinical work with anxious children. and patterns of avoidance in children.
A second heuristic is suggested in the writings of In comparison to the relatively nonspecific tradi-
Kenneth Rubin and his colleagues (e.g., Rubin, tional parenting style categories that may have few
Hastings, Stewart, Henderson, & Chen, 1997). practical implications for the treatment of childhood
These researchers propose that excessive parental anxiety, the identification of specific parenting be-
responsiveness and oversolicitousness in a situation haviors contributing to particular manifestations of
that initially elicits anxious responses in a child can fear and anxiety in children could inform our re-
reinforce child anxiety and sensitize children to the search and clinical practice by identifying important
situation, preventing exposure and habituation. This mechanisms of change in child anxiety development
process may occur in novel contexts such as the first and maintenance that are partially or wholly under
day(s) of kindergarten, at which time oversolicitous parental control. The two heuristics proposed above
parental reactions to initial child shyness might are intended to extend theories of child anxiety
perpetuate social anxiety over time. For instance, development by identifying particular parenting
parents might stay with their children at school in behaviors that may influence children’s feelings
response to early signs of child wariness and dis- of mastery and their ability to regulate anxiety in
tress, unintentionally rewarding the anxious behav- specific situations.
ior with parental attention and simultaneously
interfering with (a) trial-and-error learning of social Issues of timing and sequencing. Few theoretical
approach behaviors and (b) habituation to the novel formulations have specified the time frame in which
social context of being separated from parents and a specific parenting behavior might lead to increased
confronted with peer social interactions. Conversely, childhood anxiety. Although oversolicitous parenting
parents who encourage their children to participate responses to children’s expressions of anxiety may
in social activities and remove rewarding conse- reward anxiety (i.e., through operant mechanisms)
quences for initial shyness or distress may help to while preventing exposure and habituation (i.e., via
inoculate children with high trait anxiety against classical conditioning mechanisms), the timing of
becoming phobic of social interaction in such situa- this process is not specified. Based on operant and
tions. There are probably many other specific situa- classical learning mechanisms, manifestations of
tions in which parental regulation of contingencies childhood anxiety could be shaped by the onset and
and opportunities for habituation could affect their repetition of parenting behavior in a relatively short
shy or inhibited child’s development of specific time frame (i.e., following several ‘trials’ over the
anxious and avoidant responses (e.g., parental re- course of minutes, hours, or days). In contrast,
actions to children’s reticence about spending the theoretical models that focus on cognitive mediators
night at a friend’s house). Observational studies of imply a longer time frame. Experiences with exces-
parenting behaviors in developmentally significant sive parental control may lead to a loss of perceived
naturalistic situations, such as in the classroom mastery and personal control in children and, in
during the first week of kindergarten – rather than in turn, increased childhood anxiety (Chorpita & Bar-
the laboratory – may help further our understanding low, 1998). But, changes in perceptions of personal
of the ontogeny of specific anxious reactions in control could take months or years to occur.
children who are already high in trait anxiety. As Before such theories can be tested, it will be ne-
noted by Whaley et al. (1999), parental anxiety may cessary to specify the temporal processes involved
be one factor contributing to specific patterns of and posit whether the expected effect would be on
parenting behavior. Parents with high trait anxiety or children’s state anxiety (i.e., anxiety on a given day
previous experiences with traumatic events might be or week), situationally specific anxiety (i.e., consis-
especially likely to respond in an oversolicitous, tently experienced anxiety or avoidance associated
protective manner to manifestations of anxiety and with one or several specific situations that persists
distress in their children. The temporary reduction of over time), or trait anxiety (i.e., an increase in non-
children’s anxiety in response to such parenting specific anxiety and negative affect across situations,
behavior may, in turn, reinforce the parent. Fur- persisting over months or years). For example, a
thermore, children who are profoundly inhibited in single instance of parental criticism could lead to a
brief increase in state anxiety on the same day, with hypotheses concerning the role of child anxiety in
no long-term implications. In sum, the field would eliciting specific types of parenting behavior. Several
benefit from further refinement of existing models studies have experimentally evaluated the effects of
to include a more definitive statement about the hyperactive or noncompliant child behavior on the
expected duration of time between a given parenting subsequent behavior of parents (e.g., Pelham et al.,
behavior (or set of behaviors) and the onset of 1998). For instance, Pelham and colleagues have
increased child state or trait anxiety. investigated whether exposure to a compliant versus
noncompliant child confederate in a laboratory in-
teraction task led to differential short-term patterns
Clarifying the direction of effects
of alcohol consumption in parents of externalizing
To evaluate the possible pathways linking parenting children. A similar approach might examine the ef-
behavior and childhood anxiety, we need to move fects of specific anxious child behaviors on the
from basic correlational research to methodologies interactional style of parents of anxiety-disordered
that can reveal more about the direction of effects. children. For instance, parents might be randomly
Because a specific sequence of events is a necessary, assigned to teach an ‘anxious’ or ‘normal’ child
but not sufficient, precondition to establishing a di- confederate a specific skill, and then to play freely
rectional pathway (e.g., children’s separation anxiety with the child. The child confederate might be as-
eliciting parental overprotective behavior), utilizing signed to either act in a shy, indecisive, and/or
repeated measures of parenting behavior and child- clingy manner, or in a sociable and confident man-
hood anxiety at theoretically meaningful time-inter- ner. Group differences between the parents in their
vals can help test initial hypotheses about the interactional styles with the child confederates could
direction of effects. Second, experimental methods provide valuable experimental evidence about the
can be employed to directly evaluate the effects of potential role that child anxious behavior may play
manipulating parenting behavior on childhood in eliciting specific parenting behaviors such as
anxiety, and vice versa. excessive control and intrusiveness.
Longitudinal approach. Repeated-measures longi- Intervention design. Intervention designs can help
tudinal research could help elucidate the temporal clarify the direction of effects between family inter-
sequence of events linking systematic changes of action patterns and childhood anxiety. Such re-
parenting behavior and situational or global changes search designs seek to alter family interaction
in childhood anxiety. Ideally, such research designs patterns to assess the resulting impact on children’s
would employ more than two time-points of data for developmental outcomes (Cowan & Cowan, 2002).
both specific parenting behavior in particular situa- Though intervention designs cannot provide in-
tions (e.g., as suggested above, oversolicitous par- formation regarding the initial causes of childhood
enting behavior in the kindergarten classroom) and anxiety, such designs can test whether altering cur-
specific child anxiety symptoms. Statistical modeling rent family interaction patterns affects current
(e.g., hierarchical linear modeling) of the change in childhood anxiety, which may have implications for
the trajectory of parenting on the change in the tra- the role that parents can play in maintaining child-
jectory of anxiety symptoms, and vice versa, could hood anxiety disorders.
then be implemented. Although some longitudinal In an intervention design, subjects are randomly
studies have been conducted (e.g., Papini & Rogg- assigned to either a condition that alters family in-
man, 1992), we are not aware of any that have tested teraction patterns, or a condition that does not
the effects of change in parenting behavior on change (Cowan & Cowan, 2002; Brent & Kolko, 1998).
in childhood anxiety. One challenge facing re- Children’s behavior and family interaction patterns
searchers using this approach is identifying an ap- are measured before and after the intervention.
propriate time interval during which to administer Strong evidence exists for family interactions play-
repeated measures. The time interval should be long ing a maintaining role in children’s anxiety if (a) the
enough to permit a stable change in parenting be- family intervention improves the child outcome
havior, but not so long that multiple changes would measure more than does the child intervention, (b)
have occurred. Intervals of less than one year are the family intervention improves family interactions
advisable and the selection of developmentally sig- more than does the child intervention, and (c)
nificant time points (e.g., entry into preschool, kin- improvements in the child outcome measure are
dergarten, or middle-school), may help ensure that mediated by improvements in family interactions.
meaningful change in parenting behavior will be Of course, even results that meet these three con-
observed. Of course, though suggestive, longitudinal ditions do not ‘prove’ that family interactions
designs do not provide evidence of cause-and-effect maintain or elicit a given child outcome. However,
relationships (Cowan & Cowan, 2002). such results provide more convincing evidence that
family interactions could play a role in the main-
Experimental design. Short-term experimental tenance of child anxiety disorders than do correla-
designs may provide weight to the rationale for tional results.
To date, four studies have compared ‘family-

Acknowledgments
focused’ with ‘child-focused’ cognitive behavioral
therapy (CBT) programs for clinically anxious chil- This paper was supported in part by a training grant
dren (Barrett, 1998; Barrett, Dadds, & Rapee, 1996; from the National Institute of Mental Health awarded
Mendlowitz et al., 1999; Spence, Donovan, & to Jeffrey Wood (National Research Service Award;
Brechman-Toussaint, 2000). The family treatment F31-MH64999) and a grant from the National Insti-
condition in each of these studies included specific tute of Mental Health awarded to Marian Sigman
parent-training modules intended to improve family (R03-MH63836). We thank Rosalie Corona, Matt
interaction patterns. In each study, there was evi- Yergovich, Brian Buzzella, and Amie Bettencourt for
dence of superior child anxiety outcomes among their valuable feedback and assistance on this paper.
children assigned to the family treatment condition
(although there was only a marginally significant
post-treatment difference between groups on anxi- Correspondence to
ety disorder status in the Spence et al. study). This
Jeffrey J. Wood, Psychology Clinic, 2191 Franz Hall,
evidence fulfills Cowan and Cowan’s (2002) first
Department of Psychology, University of Califor-
condition, demonstrating that a family intervention
nia, Los Angeles, California, 90095, USA. Email:
program alters a given child outcome measure in a
jeffwood@ucla.edu
more favorable manner than a non-family inter-
vention. Surprisingly, these studies have not tested
the second and third conditions of the model; no
References
research on the effects of these interventions on
family interaction processes exists. The next em- Barber, B. (1996). Parental psychological control:
pirical step is to test whether parent–child interac- Revisiting a neglected construct. Child Development,
tions are improved by family CBT. Clearly, an 67, 3296–3319.
expansion of the research program comparing Baumrind, D. (1971). Current patterns of parental
authority. Developmental Psychology Monographs, 4.
family CBT to child CBT represents a fruitful step
Barrett, P.M. (1998). Group therapy for childhood
for examining the role of parenting behavior
anxiety disorders. Journal of Clinical Child Psychol-
as a maintaining influence in childhood anxiety ogy, 27, 459–468.
disorders. Barrett, P.M., Dadds, M.R., & Rapee, R.M. (1996).
Family treatment of childhood anxiety: A controlled
trial. Journal of Consulting and Clinical Psychology,
Summary 64, 333–342.
Bell-Dolan, D., & Brazeal, T.J. (1993). Separation
Are parental warmth, control, and modeling of anxiety disorder, overanxious disorder, and school
anxious behaviors associated with the development refusal. Child and Adolescent Psychiatric Clinics of
and maintenance of childhood anxiety? The extant North America, 2, 563–580.
research offers no definitive conclusions, but a few Bowen, R.C., Offord, D.R., & Boyle, M.H. (1990). The
consistent patterns have emerged. Relatively greater prevalence of overanxious disorder and separation
observed parental control during parent–child in- anxiety disorder: Results from the Ontario Child
teractions in laboratory tasks was consistently Health Study. Journal of the American Academy of
linked with more child shyness and a higher risk for Chid and Adolescent Psychiatry, 29, 753–758.
meeting criteria for an anxiety disorder in children Brent, D.A., & Kolko, D.J. (1998). Psychotherapy:
Definitions, mechanisms of action, and relationship
and adolescents. This association tended to be of a
to etiological models. Journal of Abnormal Child
medium or large magnitude, with clinically signifi-
Psychology, 26, 17–25.
cant implications. Contextual determinants of the Campbell, D.T., & Fiske, D.W. (1959). Convergent and
relationship between parenting and child anxiety, discriminant validation by the multitrait-multi-
such as situational characteristics and parental method matrix. Psychological Bulletin, 56, 81–105.
anxiety, play important moderating roles. Recent Capps, L., & Ochs, E. (1995). Constructing panic: The
models of anxiety development provide a broad discourse of agoraphobia. Cambridge, MA: Harvard
conceptual framework for more programmatic theory University Press.
development and empirical evaluation of the relation Chorpita, B.F., & Barlow, D.H. (1998). The development
between specific parenting behaviors and particular of anxiety: The role of control in the early environ-
manifestations of childhood anxiety. Although the ment. Psychological Bulletin, 124, 3–21.
Cicchetti, D., & Cohen, D.J. (1995). Perspectives on
methodological characteristics of past studies limit
developmental psychopathology. In D. Cicchetti &
conclusions about nature of this relationship, ad-
D. Cohen (Eds.), Developmental psychopathology:
vances in the past decade in the field of develop- Vol. 1. Theory and methods (pp. 3–20). New York: Wiley.
mental psychopathology suggest that researchers in Cohen, J. (1988). Statistical power analysis for the
this area may soon be able to critically assess con- behavioral sciences (2nd edn). Hillsdale, NJ: Erlbaum.
temporary models of the effects of parenting style Cowan, P.A., & Cowan, C.P. (2002). What an interven-
and behavior on childhood anxiety. tion design reveals about how parents affect their
children’s academic achievement and behavior pro- Manassis, K., & Bradley, S.J. (1994). The development
blems. In J.G. Borkowski & S. Ramey (Eds.), Parent- of childhood anxiety disorders: Toward an integrated
ing and the child’s world: Influences on intellectual, model. Journal of Applied Developmental Psychology,
academic, and social-emotional development (pp.75– 15, 345–366.
97). Mahwah, NJ: Erlbaum. March, J.S. (1998). Multidimensional anxiety scale for
Craske, M.G. (1999). Anxiety disorders: Psychological children. North Tonawanda, NY: MHS.
approaches to theory and treatment. Boulder, CO: March, J.S., Parker, J.D.A., Sullivan, K., Stallings, P., &
Westview Press. Conners, K. (1997). The Multidimensional Anxiety
Dadds, M.R., Barrett, P.M., Rapee, R.M., & Ryan, S. Scale for Children (MASC): Factor structure, relia-
(1996). Family process and child anxiety and aggres- bility, and validity. Journal of the American Academy
sion: An observational analysis. Journal of Abnormal of Child and Adolescent Psychiatry, 36, 554–565.
Child Psychology, 24, 715–734. Masia, C.L., & Morris, T.L. (1998). Parental factors
Darling, N., & Steinberg, L. (1993). Parenting style as associated with social anxiety: Methodological limita-
context: An integrative model. Psychological Bulletin, tions and suggestions for integrated behavioral
113, 487–496. research. Clinical Psychology: Science and Practice,
Dornbusch, S., Ritter, P., Leiderman, P., Roberts, D., & 5, 211–228.
Fraleigh, M. (1987). The relations of parenting style to Masten, A.S., & Braswell, L. (1991). Developmental
adolescent school performance. Child Development, psychopathology: An integrative framework. In
58, 1244–1257. P.R. Martin (Ed.), Handbook of behavior therapy
Dumas, J.E., LaFreniere, P.J., & Serketich, W.J. and psychological science: An integrative approach
(1995). ‘Balance of power’: A transactional analysis (pp. 35–56). New York: Pergamon.
of control in mother–child dyads involving socially Mattanah, J.F. (2001). Parental psychological auton-
competent, aggressive, and anxious children. Journal omy and children’s academic competence and be-
of Abnormal Psychology, 104, 104–113. havioral adjustment in late childhood: More than just
Eley, T.C. (2001). Contributions of behavioral genetics limit-setting and warmth. Merrill-Palmer Quarterly,
research: Quantifying genetic, shared environmental, 47, 355–376.
and nonshared environmental influences. In M.W. Mendlowitz, S.L., Manassis, K., Bradley, S., Scapillato,
Vasey & M.R. Dadds (Eds.), The developmental D., Mietzitis, S., & Shaw, B. (1999). Cognitive-
psychopathology of anxiety (pp. 45–59). Oxford: Ox- behavioral group treatments in childhood anxiety
ford University Press. disorders: The role of parental involvement. Journal of
Gerlsma, C., Emmelkamp, P.M.G., & Arrindell, W.A. the American Academy of Child and Adolescent
(1990). Anxiety, depression and perception of early Psychiatry, 38, 1223–1229.
parenting? A meta-analysis. Clinical Psychology Re- Mills, R.S.L., & Rubin, K.H. (1998). Are behavioural and
view, 10, 251–277. psychological control both differentially associated
Gottman, J.M., Katz, L.F., & Hooven, C. (1997). Meta- with childhood aggression and social withdrawal?
emotion: How families communicate emotionally. Canadian Journal of Behavioural Science, 30, 132–
Mahwah, NJ: Erlbaum. 136.
Gruner, K., Muris, P., & Merckelbach, H. (1999). The Moore, P.S., Whaley, S., & Sigman, M. (2002). Child,
relationship between anxious rearing behaviors and parent, and transactional contributions to the devel-
anxiety disorders symptomatology in normal chil- opment and maintenance of anxiety disorders of
dren. Journal of Behavior Therapy and Experimental childhood. Manuscript in submission.
Psychiatry, 30, 27–35. Muris, P., Bogels, S., Meesters, C., der Kamp, N., & van
Hernandez-Guzman, L., & Sanchez-Sosa, J.J. (1996). Oosten, A. (1996). Parental rearing practices, fearful-
Parent–child interactions predict anxiety in Mexican ness, and problem behavior in clinically referred
adolescents. Adolescence, 31, 955–963. children. Personality and Individual Differences, 21,
Hibbs, E.D., Hamburger, S.D., Kruesi, M.J.P., & 813–818.
Lenane, M. (1993). Factors affecting expressed emo- Muris, P., Meesters, C., Merckelbach, H., & Hulsen-
tion in parents of ill and normal children. American beck, P. (2000). Worry in children is related to
Journal of Orthopsychiatry, 63, 103–112. perceived parental rearing and attachment. Behavior
Holden, G.W., & Edwards, L.E. (1989). Parental atti- Research and Therapy, 38, 487–497.
tudes toward child rearing: Instruments, issues, and Ollendick, T.H., Vasey, M.W., & King, N.J. (2001).
implications. Psychological Bulletin, 106, 29–58. Operant conditioning influences in childhood anxiety.
Hudson, J.L., & Rapee, R.M. (2001). Parent–child inter- In M.W. Vasey & M.R. Dadds (Eds.), The develop-
actions and anxiety disorders: An observational study. mental psychopathology of anxiety (pp. 231–252).
Behaviour Research and Therapy, 39, 1411–1427. Oxford: Oxford University Press.
Hummel, R.M., & Gross, A.M. (2001). Socially anxious Papini, D.R., Roggman, L.A., & Anderson, J. (1991).
children: An observational study of parent–child inter- Early adolescent perceptions of attachment to mother
action. Child and Family Behavior Therapy, 23, 19–41. and father: A test of the emotional-distancing and
Kliewer, W., & Kung, E. (1998). Family moderators of buffering hypotheses. Journal of Early Adolescence,
the relation between hassles and behavior problems 11, 258–275.
in inner-city youth. Journal of Clinical Child Psychol- Papini, D.R., & Roggman, L.A. (1992). Adolescent
ogy, 27, 278–292. perceived attachment to parents in relation to
Maccoby, E.E. (1992). The role of parents in the competence, depression, and anxiety: A longi-
socialization of children: An historical review. Devel- tudinal study. Journal of Early Adolescence, 12,
opmental Psychology, 28, 1006–1017. 420–440.
Pedersen, W. (1994). Parental relations, mental health, Schwarz, J.C., Barton-Henry, M.L., & Pruzinsky, T.
and delinquency in adolescents. Adolescence, 29, (1985). Assessing child-rearing behaviors: A compar-
975–990. ison of ratings made by mother, father, child, and
Pelham, W.E., Lang, A.R., Atkeson, B., Murphy, D.A., sibling on the CRPBI. Child Development, 56, 462–479.
Gnagy, E.M., Greiner, A.R., Vodde-Hamilton, M., & Siqueland, L., Kendall, P.C., & Steinberg, L. (1996).
Greenslade, K.E. (1998). Effects of deviant child Anxiety in children: Perceived family environments
behavior on parental alcohol consumption: Stress- and observed family interaction. Journal of Clinical
induced drinking in parents of ADHD children. Child Psychology, 25, 225–237.
American Journal on Addictions, 7, 103–114. Spence, S.H., Donovan, C., & Brechman-Toussaint, M.
Rapee, R.M. (1997). Potential role of childrearing prac- (2000). The treatment of childhood social phobia: The
tices in the development of anxiety and depression. effectiveness of a social skills training-based, cogni-
Clinical Psychology Review, 17, 47–67. tive-behavioral intervention, with and without par-
Ray, J.C., & Sapolsky, R.M. (1992). Styles of male social ental involvement. Journal of Child Psychology and
behavior and their endocrine correlates among high- Psychiatry, 41, 713–726.
ranking wild baboons. American Journal of Primatol- Sroufe, L.A., Jacobvitz, D., Mangelsdorf, S., deAngelo,
ogy, 28, 231–250. E., & Ward, M.J. (1985). Generational boundary
Reiss, S., Silverman, W.K., & Weems, C.F. (2001). dissolution between mothers and their preschool
Anxiety sensitivity. In M.W. Vasey & M.R. Dadds children: A relationship systems approach. Child
(Eds.), The developmental psychopathology of anxiety Development, 56, 317–325.
(pp. 92–111). Oxford: Oxford University Press. Stark, K.D., Humphrey, L.L., Laurent, J., Livingston,
Repetti, R.L. (1996). The effects of perceived daily social R., & Christopher, J. (1993). Cognitive, behavioral,
and academic failure experiences on school-age chil- and family factors in the differentiation of depressive
dren’s subsequent interactions with parents. Child and anxiety disorders during childhood. Journal of
Development, 67, 1467–1482. Consulting and Clinical Psychology, 61, 878–886.
Reynolds, C.R., & Richmond, B.O. (1978). ‘What I Think Steinberg, L., Dornbusch, S., & Brown, B.B. (1992).
and Feel’: A revised measure of children’s manifest Ethnic differences in adolescent achievement: An
anxiety. Journal of Abnormal Child Psychology, 6, ecological perspective. American Psychologist, 47,
271–280. 723–729.
Rothbaum, F., & Weisz, J.R. (1994). Parental caregiving Steinberg, L., Elmer, J.D., & Mounts, N.S. (1989).
and child externalizing behavior in nonclinical sam- Authoritative parenting, psychosocial maturity, and
ples: A meta-analysis. Psychological Bulletin, 116, academic success among adolescents. Child Develop-
55–74. ment, 60, 1424–1436.
Rubin, K.H., Cheah, C.S.L., & Fox, N. (2001). Emotional Stevenson-Hinde, J., & Glover, A. (1996). Shy girls and
regulation, parenting and display of social reticence boys: A new look. Journal of Child Psychology and
in preschoolers. Early Education and Development, Psychiatry, 37, 181–187.
12, 97–115. Tesser, A., & Forehand, R. (1991). Adolescent function-
Rubin, K.H., Hastings, P.D., Stewart, S.L., Henderson, ing: Communication and the buffering of parental
H.A., & Chen, X. (1997). The consistency and anger. Journal of Social and Clinical Psychology, 10,
concomitants of inhibition: Some of the children, all 152–175.
of the time. Child Development, 68, 467–483. Vasey, M.W., & Dadds, M.R. (2001). An introduction to
Rubin, K.H., Nelson, L.J., Hastings, P., & Asendorpf, J. the developmental psychopathology of anxiety. In
(1999). The transaction between parents’ perceptions M.W. Vasey & M.R. Dadds (Eds.), The developmental
of their children’s shyness and their parenting styles. psychopathology of anxiety (pp. 3–26). Oxford: Oxford
International Journal of Behavioral Development, 23, University Press.
937–957. Whaley, S.E., Pinto, A., & Sigman, M. (1999). Char-
Rubin, K.H., & Stewart, S.L. (1996). Social withdrawal. acterizing interactions between anxious mothers and
In E.J. Mash & R.A. Barkley (Eds.), Child psycho- their children. Journal of Consulting and Clinical
pathology (pp. 277–307). New York: Guilford. Psychology, 67, 826–836.
Scott, W.A., Scott, R., & McCabe, M. (1991). Family Wood, J.J., Piacentini, J.C., Bergman, R.L., McCrac-
relationships and children’s personality: A cross- ken, J., & Barrios, V. (2002). Concurrent validity of
cultural, cross-source comparison. British Journal of the Anxiety Disorders Section of the Anxiety Dis-
Social Psychology, 30, 1–20. orders Interview Schedule for DSM-IV: Child and
Schniering, C.A., Hudson, J.L., & Rapee, R.M. (2000). parent versions. Journal of Clinical Child and Adoles-
Issues in the diagnosis and assessment of anxiety cent Psychology, 31, 335–342.
disorders in children and adolescents. Clinical Psy-
chology Review, 20, 453–478. Manuscript accepted 15 August 2002

Parenting and Childhood Anxiety: Theory, Empirical Ndings, and Future Directions

Uploaded by

Copyright:

Available Formats

Parenting and Childhood Anxiety: Theory, Empirical Ndings, and Future Directions

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Parenting and Childhood Anxiety: Theory, Empirical Ndings, and Future Directions

Uploaded by

Copyright:

Available Formats

Journal of Child Psychology and Psychiatry 44:1 (2003), pp 134–151

Parenting and childhood anxiety: theory, empirical

Parenting Anxiety Parenting Anxiety Mother’s Father’s

Dumas et al. (1995) 84a 4.1 1 O T Control Trait 1 of 1

between two self-report measures from the same

ratio of significant statistical tests (i.e., tests with a

p value of .05 or lower) to total tests conducted in

mean differences to correlations (and, then, to

percents of variance explained, i.e., r 2 ) in group-

Studies of parental acceptance

Child-report studies. Eleven studies were identified

that used children’s reports of accepting parental

schools located in a variety of countries (including

and Taiwan), adolescents who reported more

Sample divided into anxious group and normal control group.

study, the effects accounted for 1% and 3% of the

variance in parents’ and teachers’ ratings of child-

with some alpha coefficients as low as .44 (range

ety-disordered children (with one of several diag-

Kendall, and Steinberg (1996) found children’s rat-

ings of maternal acceptance accounted for 21% of

the variance in diagnostic status (a ‘large’ effect).

statistical tests yielded significant relations when

effects were not statistically significant, suggesting

Additionally, the studies using diagnostic group-

To date, four studies have compared ‘family-

You might also like