The association between perceptions of parent worry and generalised anxiety disorder symptoms in a community sample of adolescents

Abstract

Parental rearing behaviours, including modelling of anxiety, have been associated with corresponding anxiety symptoms in children and adolescents. However, there is a paucity of research investigating the relationship between worry-related learning experiences from parents and general worry or generalised anxiety disorder (GAD) symptoms in children and adolescents. The purpose of the present study was to examine this association. It was predicted that worry-related learning experiences from parents would be a significant predictor of adolescent GAD symptoms. To evaluate this hypothesis, 348 adolescents from public and private secondary schools in the Southeast were surveyed. Consistent with previous research, the results of this study indicated that observation of parent worry was significantly and positively associated with adolescent GAD symptoms. Overall, these results suggest that parental modelling of worry may be related to the development and maintenance of pathological worry in youth. The findings of this study serve to inform prevention and treatment programmes targeting general worry symptoms in youth.

Keywords::

• worry
• anxiety
• generalised anxiety disorder
• youth
• adolescent
• parent

Introduction

Generalised anxiety disorder (GAD) is characterised by excessive and uncontrollable worry and has a relatively high lifetime prevalence rate of approximately 4% (American Psychiatric Association, 2000; Grant et al., 2005; Rapee, 2002). Childhood and adolescence is a common time of onset and GAD has been found to be a common diagnosis in children referred to anxiety clinics (Kendall et al., 1997; Laugesen, Dugas, & Bukowski, 2003; Merikangas et al., 2009). Despite the relatively high prevalence rate and early onset of GAD, relatively few studies have focused on the developmental risk factors related to GAD and pathological worry in children and adolescents (Borkovec, Shadick, & Hopkins, 1991; Cartwright-Hatton, McNicol, & Doubleday, 2006).

Anxiety disorders, including GAD, have been shown to aggregate within families (Coelho, Cooper, & Murray, 2007; Manassis, Hudson, Webb, & Albano, 2004). Consequently, a large body of research has focused on the role of the parent–child relationship in the development and maintenance of child anxiety. Perhaps the most commonly studied parenting variable is control/overprotection, which has consistently been associated with child and adolescent anxiety (for reviews see McLeod, Wood, & Weiz, 2007; Van der Bruggen, Stams, & Bögels, 2008). More specific to GAD, initial studies have found parental control to be associated with GAD symptoms in youth (Hale, Engles, & Meeus, 2006; Muris, 2002; Muris, Meesters, Merckelbach, & Hulsenbeck, 2000; Muris & Merckelbach, 1998). However, it is noteworthy that, when compared with other anxiety disorders, relatively few studies have examined this association. Further, another limitation to the research in this area is that control may account for a relatively small amount of the variance in child and adolescent anxiety symptoms (see McLeod et al., 2007).

In addition to the parenting behaviours discussed above, learning experiences from parents including parent modelling of anxiety symptoms have also been associated with child and adolescent anxiety (Burnstein & Ginsburg, 2010; Fisak & Grills-Taquechel, 2007; Mineka & Zinbard, 2006; Muris, Steerneman, Merckelbach, & Meesters, 1996). Specifically, there is evidence to suggest that parental learning experiences may foster the development of anxious cognitions, avoidant behaviours, and anxiety symptoms in children (Barrett, Rapee, Dadds, & Ryan, 1996; Burnstein & Ginsburg, 2010; Muris et al., 1996). Further, there is evidence to suggest that specific sets of learning experiences relate to particular anxiety disorders. More specifically, parental modelling of panic symptoms and related learning experiences has been found to be associated with the development of panic symptoms (Ehlers, 1993; Stewart et al., 2001; Watt & Stewart, 2000; Watt, Stewart, & Cox, 1998). Further, parental modelling of social anxiety and related parenting behaviours have been associated with social phobia symptoms (Bruch & Heimberg, 1994; Bruch, Heimberg, Berger, & Collins, 1989; Caster, Inderbitzen, & Hope, 1999; Harvey, Ehlers, & Clark, 2005). Finally, a number of studies suggest that modelling may be associated with the development of specific phobias (see King, Eleonora, & Ollendick, 1998).

Surprisingly few studies have examined the degree to which parental modelling of worry is associated with GAD symptoms in children and adolescents. Among the few studies in this area, Muris and Merckelbach (1998) explored this association by using the Anxious Rearing Scale of the Egna Minnen Beträffande Uppfostran (EMBU). The EMBU is a 10-item scale designed to measure anxious parenting and consists of five items directly relating to parent worry (Grüner, Muris, & Merckelbach, 1999; Muris & Merckelbach, 1998). Using the anxious rearing scale, Muris and Merckelbach (1998) found an association between anxious rearing behaviours and GAD symptoms in a non-clinical sample of children. Follow-up studies have found similar associations between self-report of parent anxious rearing and current worry symptoms in both child and adolescent samples (Muris, 2002; Muris et al., 2000). Using a different methodology, Fisak, Holderfield, Douglas-Osborn, and Cartwright-Hatton (in press) developed the Parent Worry Questionnaire, a parent-report measure designed to assess parent worry in relation to their children. The Parent Worry Questionnaire was found to partially mediate the association between parent and child anxiety. Further, parent worry was found to be a stronger predictor of child anxiety than parent anxiety symptoms.

Overall, although several studies have examined the role of parenting behaviours and anxiety symptoms in youth, relatively few studies have focused on this association in relation to GAD symptoms. There is a particular paucity of research examining the degree to which parent modelling of worry symptoms is associated with child and adolescent GAD symptoms. The majority of studies that have examined this association have relied on the anxious rearing questionnaire of the EMBU. Although this measure does appear to include items that assess parent worry, the anxious rearing scale does not exclusively focus on the construct of parent worry (see Muris & Merckelbach [1998] for the items on the anxious rearing scale). It appears that the anxious rearing subscale addresses anxious parenting as a more general construct, and only five of the 10 items in the anxious rearing subscale measure relate directly to parent worry. For example, the item ‘your parents are anxious people and therefore you are not allowed to do as many things as other children’ appears to assess parental control. Further, most of the research in this area has focused on child self-report and only one study to date has examined the association between perceived parent worry and GAD symptoms in an adolescent sample (Muris, 2002).

In response to the limitations, the purpose of the current study was to examine the potential relationship between adolescent perceptions of parent worry (i.e. parent modelling of worry) and current adolescent worry symptoms. As there are no existing measures focusing exclusively on the observation of parent worry symptoms, the Parent Worry Scale (PWS) was developed. It was hypothesised that the PWS would be significantly and positively associated with adolescent GAD symptoms. It was also hypothesised that this association would remain significant after controlling for other perceived learning experiences, including parent modelling of social fears and modelling of panic symptoms/symptoms of autonomic arousal. Support for these hypotheses would suggest parental modelling of worry may play a role in the development and maintenance of GAD symptoms in youth.

Method

Participants

Participants were 348 adolescents (226 female, 114 male, and eight not reported) recruited from a convenience sample of public and private secondary schools in a medium-sized city in the southeastern United States. A variety of secondary schools throughout the city were invited to participate in order to maximise diversity of the sample. The mean participant age was 16.84 (standard deviation = 0.73, range 15–18 years). The ethnic distribution is as follows: Caucasian (72.1%), African American (7.2%), Asian American (6.9%), Hispanic American (3.7%), Middle Eastern American (3.7%), and Multiracial/Other (6.3%). The annual family income, as reported by participants, was as follows: $150,000 or above (15.8%), $100,001–150,000 (17.2%), $75,001–100,000 (23.6%), $50,001–75,000 (23.0%), $25,000–50,000 (14.9%), and $25,000 or below (2.6%).

Design and procedures

Researchers introduced the study in the classroom and explained the requirement of parental consent. Students were asked to bring parental consent forms home to their parents, and if they were interested in participating to return the signed forms to their teacher. Parental consent forms were then collected from prospective sites and lists of students who had returned parental consent forms were compiled. Researchers returned to the schools and students with parental consent were invited to participate. Student assent was explained and was required for participation. Once signed assent forms were completed, they were collected by the researchers and each participant was given the self-report survey packets. Interested participants completed a survey packet in their classrooms. Researchers were available at all points throughout administration to ensure confidentiality, independent reporting, and to provide assistance when needed. Upon completion and collection of all survey packets, participants and teachers were debriefed and thanked for their participation.

Measures

Parent Worry Scale

The PWS was developed for the purpose of this study to measure perception of parental modelling of worry and other symptoms consistent with GAD. A number of items were based on symptoms listed in the Diagnostic Statistical Manual of Mental Disorders (4th edition – text revision; American Psychiatric Association, 2000) for GAD and the Anxiety Disorder Interview Schedule, a structured interview of anxiety symptoms (Brown, Di Nardo, & Barlow 2004). Additional items were developed by rewording items from the Penn State Worry Questionnaire, the Anxious Rearing Scale of the EMBU, the Modeling Subscale of the Child Learning of Anxiety Scale, and the Parent Experience of Child Illness Scale (Bonner et al., 2006; Fisak, Negy, & Ehrenreich, 2008; Meyer, Miller, Metzger, & Borkovec, 1990; Muris, Bögels, Meesters, & van der Kamp (1996)).

Parents Attitudes Toward Children Rearing Scale

The Parents Attitudes Toward Children Rearing Scale (PACR) is a 19-item self-report measure designed to assess learning experiences related to symptoms of social fears/phobia (Bruch & Heimberg, 1994; Bruch et al., 1989). Participants are asked to rate how frequently each item occurs using a five-point Likert scale. Consistent with previous research, items were reworded to reflect current adolescent perceptions rather than retrospective reports of learning experiences (Caster et al., 1999). Criterion validity has been established for the PACR, as the measure has been found to be significantly associated with social anxiety symptoms (for example, Bruch & Heimberg, 1994; Caster et al., 1999). Cronbach's alpha in the current sample was α = 0.71.

Penn State Worry Questionnaire for Children

The Penn State Worry Questionnaire for Children (PSWQ-C) is a self-report measure designed to assess symptoms of GAD and pathological worry (Chorpita, Tracey, Brown, Collica, & Barlow, 1997). Participants are asked to rate how frequently they have experienced each symptom using a four-point Likert scale. Higher total scores reflect higher levels of generalised anxiety symptoms. The PSWQ-C has demonstrated good test–retest reliability; with Cronbach's alpha values ranging from α = 0.81 to α = 0.90 (Chorpita et al., 1997). The measure has demonstrated good convergent validity, with a significant positive correlation with the Worry/Oversensitivity subscale of the Revised Children's Manifest Anxiety Scale, and the discriminant validity of the PSWQ-C has been supported (Chorpita et al., 1997; Reynolds & Richmond, 1985). Cronbach's alpha in the current sample was α = 0.91.

The Learning History Questionnaire – Revised

The Learning History Questionnaire – Revised (LHQ-R) is a self-report measure designed to assess learning experiences and panic-related symptoms (Ehlers, 1993; Watt & Stewart, 2000). Items were reworded to reflect current experiences rather than retrospective reports of symptoms. Respondents are asked to rate how frequently they have been exposed to each learning experience using a five-point Likert scale. The first section is comprised of 22 items assessing parental reinforcement/encouragement of panic symptoms (encouragement). In the second section, participants are asked 20 questions regarding parental modelling of panic symptoms. Scores on both the encouragement and modelling scales reflect the amount and degree of panic-related learning experiences. The validity of the LHQ-R has been supported, as items have been shown to be associated with higher levels of panic symptoms and anxiety sensitivity (Watt & Stewart, 2000). The LHQ-R has yielded adequate internal consistency with Cronbach alpha values of α = 0.91 for the encouragement scale and α = 0.96 for the modelling scale (Watt & Stewart, 2000). Based on the current sample, internal consistency was adequate, with obtained Cronbach's alpha values of α = 0.88 and α = 0.89 for the encouragement and modelling scales, respectively. For the purpose of this study, only the modelling subscale was used (Learning History Questionnaire – Modeling [LHQ-M]).

Screen for Child Anxiety Related Emotional Disorders – Revised

The Screen for Child Anxiety Related Emotional Disorders – Revised (SCARED-R) is a 41-item self-report measure designed to assess symptoms of common anxiety disorders, including GAD, panic disorder, separation anxiety disorder, and social phobia (Muris, Merckelbach, Ollendick, King, & Bogie, 2002; Muris, Merckelbach, Schmidt, & Mayer, 1999). The frequencies of symptoms are rated on a three-point Likert scale. The measure has demonstrated strong psychometric properties in non-clinical adolescent samples (Muris et al., 2002). In particular, convergent validity has been established with other commonly utilised measures of anxiety symptoms in youth and Cronbach's alpha values for the subscales range from α = 0.91 to α = 0.76. For the purpose of the current study, only the Screen for Child Anxiety Related Emotional Disorders Generalised Anxiety subscale was utilised (SCARED-GAD).

Results

Factor analysis of the PWS

The PWS was subjected to an exploratory factor analysis via the principal component analysis extraction method. An obtained value of 0.94 on the Kaiser–Meyer–Olkin measure of sampling accuracy suggested that the correlation matrix was adequate for factor analysis. Based on obtained eigenvalues and the scree plot, a robust, single-factor solution was retained. The eigenvalue for this factor was 12.77, accounting for 39.90% of the variance. Items with loadings of 0.4 or greater on the obtained factor were retained, and based on this criterion 28 of the 32 were retained. This 28-item version of the PWS was utilised in subsequent analyses. Cronbach's alpha for the PWS was α = 0.95.

Descriptive statistics

Bivariate correlations among study variables and key demographic variables (i.e. participant age and gender) are provided in Table 1. Age and gender were both found to be significantly associated with the PSWQ-C and SCARED-GAD. Consequently, age and gender were entered as covariates in subsequent analyses. Descriptive statistics for study variables are provided in Table 2.

Table 1 Correlation matrix of study variables.

	PSWQ-C	SCARED-GAD	PWQ	PACR	LHQ-M	Participant gender	Participant age
PSWQ-C	–
SCARED-GAD	0.815*	–
PWQ	0.405*	0.409**	–
PACR	0.167**	0.216**	0.562**	–
LHQ-M	0.288**	0.297**	0.361**	0.138*	–
Participant gender	0.258**	0.259**	0.088	− 0.073	0.01	–
Participant age	− 0.119^*	− 0.136*	− 0.038	0.043	− 0.117	− 0.141**	–

Note: PSWQ-C, Penn State Worry Questionnaire for Children; SCARED_GAD, Generalized Anxiety Subscale of the Screen for Child Anxiety Related Emotional Disorders PWQ, Parent Worry Questionaire; PACR, Parents Attitudes Toward Children Rearing Scale; LHQ-Modeling, Modeling Subscale of the Learning History Questionnaire. *p < 0.05, **p < 0.01.

Table 2 Means, standard deviations, and reliabilities of study variables.

Variable	Cronbach's alpha	Mean	Standard deviation
PSWQ-C	0.91	15.28	8.81
PWS	0.95	44.33	20.04
PACR	0.71	26.11	8.62
LHQ-M	0.89	35.99	12.32

Note: PWS, the Parent Worry Scale.

PWS, adolescent worry, and GAD symptoms

A hierarchical regression was conducted to examine the degree to which the PWS predicts adolescent worry after controlling for demographic variables, and other categories of anxious parenting. Participant gender and age were entered as covariates in the first step of the equation. The PACR and the LHQ were entered into the second step of the regression equation, and the PWS was entered in the third step of the equation. The PSWQ-C was the designated criterion variable. All steps were significant, and the addition of the PWS to the third step of equation let to a significant improvement in the model, F(1, 242) = 20.44, p < 0.001, ΔR² = 0.064, and in the final model, the PWS (β = 0.33) was the most robust predictor of PSWQ-C scores. The modelling scale of the LHQ (β = 0.17) was also significantly associated with worry. In contrast, PACR was not found to be associated with worry (see Table 3).

Table 3 Hierarchical regression for variables predicting scores on the PSWQ-C.

Variable	Rtotal	ΔR^2	β
Step 1	0.28***
Age			− 0.04
Gender			0.27***
Step 2	0.42***	0.10***
Age			− 0.009
Gender			0.28***
PACR			0.27*
LHQ-M			0.15**
Step 3	0.49***	0.06***
Age			0.001
Gender			0.22***
PACR			− 0.02
LHQ-M			0.17**
PWS			0.33***

Note: *p < 0.05, **p < 0.01, ***p < 0.001.

The above analysis was repeated with the GAD subscale of the SCARED entered as the criterion variable. Again, all steps were significant, and the addition of the PWS led to a significant improvmeent in the model, F(1, 238) = 14.71, p < 0.001, ΔR² = 0.05. The results of the hierarchical regression were similar to pattern obtained above. Again, the PWS (β = 0.29) was the most robust predictor of SCARED-GAD scores. Further, the LHQ-M (β = 0.17) was significantly associated with SCARED-GAD scores, and the PACR was not significantly associated with GAD symptoms (see Table 4).

Table 4 Hierarchical regression for variables predicting scores on the SCARED-GAD.

Variable	Rtotal	ΔR^2	β
Step 1	0.25***
Age			− 0.08
Gender			0.23***
Step 2	0.43***	0.12***
Age			− 0.05
Gender			0.24***
PACR			0.19**
LHQ-M			0.26***
Step 3	0.48***	0.05***
Age			− 0.04
Gender			0.19**
PACR			− 0.05
LHQ-M			0.17**
PWS			0.29***

Note: *p < 0.05, **p < 0.01, ***p < 0.001.

Discussion

The focus of the current study was to examine the degree to which perceptions of parent modelling of worry is associated with adolescent GAD symptoms. As predicted, the PWS was a significant predictor of adolescent GAD symptoms. These findings add to the research literature, as the present study appears to be the first to focus exclusively on the association between the observations of parent worry and GAD symptoms in youth. The results of the current study are consistent with previous research conducted by Muris and colleagues on the association between anxious parenting and child and adolescent GAD symptoms (Grüner, Muris, & Merckelbach, 1999; Muris, 2002; Muris et al., 2000). It is noteworthy that although the current study is similar to the previous research conducted by research Muris and colleagues, the present study differs in that the PWS was specifically designed to assess the construct of parent worry.

Support for the incremental validity of the PWS was also found, as the PWS predicted variance in adolescent worry beyond the variance accounted for by parent modelling of social anxiety and panic symptoms. This finding, along with the finding that the PWS was the most robust predictor of adolescent GAD symptoms, provides initial support for the specificity of parent worry as a risk factor related to development and maintenance of GAD. Although more research is needed, examination of parent modelling of worry may be a viable direction for researchers interested in the aetiology of GAD.

The current findings are consistent with a more general pattern across several anxiety disorders, in which parent modelling of specific fears is associated with the presence of corresponding anxiety symptoms in their offspring. For example, parent modelling of social fears has been found to be associated with social anxiety symptoms (Caster et al., 1999; Harvey et al., 2005), and parent modelling of panic symptoms and autonomic arousal has been found to be associated with symptoms of panic disorder (Ehlers, 1993; Stewart et al., 2001; Watt & Stewart, 2000; Watt et al., 1998).

It is also noteworthy that the PWS included a variety of worry symptoms, including perceived parent worry regarding their adolescent and general parent worry. Based on factor analysis, these items loaded on a single factor. These preliminary findings suggest that perceived parent worry is a general construct. It is possible to interpret this finding in that parent worry is not specific to the child. In other words, parents who worry excessively about their child or adolescent may also experience worry about other topics unrelated to their child. These findings also suggest that parent worry in general may influence the adolescent. The degree to which parents experience general worry may impact the adolescent, yet these worries may not necessarily be concerns specific to the adolescent. This finding is consistent with previous research, in which the Parent Worry Questionnaire (which measures parent report of their worry in relation to their child) was found to form a single factor (Fisak et al., In press).

Although the current study provides a unique contribution to the research literature regarding the aetiology of GAD, several limitations and directions for future research should be mentioned. First, this study relied solely on adolescent self-report, which may have led to biased recall of learning experiences. Consequently, it is recommended that follow-up studies include the use of multiple informants. Another recommended approach is to utilise other methodology, including examination of the interaction of parent–child dyads. Second, although the current study implies that parent modelling of anxiety may influence levels of adolescent worry, it is not possible to ascertain the directionality of the association between parent and child worry. Specifically, it is possible that anxious children and adolescents induce or elicit worry-related content from their parents. It is also possible the association may be reciprocal. As a result, additional research is needed (e.g. longitudinal studies) to determine the direction of this association. Third, females were slightly over-represented in the sample, as the female sample was approximately 66% of the sample. This gender discrepancy may suggest the current findings may be slightly less generalisable to males. Finally, the current study is based on a non-clinical sample and may not accurately represent patterns seen in clinical populations of children and adolescents with anxiety disorders. Consequently, a recommended direction for future research is to replicate this study utilising clinical samples.

In conclusion, the current study provides a substantial contribution to the understanding of the pathways related to the development of pathological worry in children and adolescents, as this appears to be among the first studies to provide empirical support for the association between parent modelling of worry and adolescent GAD symptoms. These findings serve to inform both practitioners and researchers interested in the aetiology and treatment of GAD and other disorders involving worry. For example, interventions to reduce modelling of parent worry may enhance the effectiveness of existing GAD treatment programmes in youth.

Additional information

Notes on contributors

Brian Fisak

Brian Fisak, PhD, earned his doctorate in clinical psychology from the University of Central Florida, and he is currently an assistant professor in the Department of Psychology at the University of North Florida.

Angela Mann

Angela Mann, MA, BCBA is a doctoral student in the School Psychology Program at the University of South Florida and a practicing behavior analyst.

Kristin Heggeli

Kristin Heggeli, BSc, graduated with her bachelor's degree in psychology from University College London, and she is currently a clinical research coordinator in the Mayo Clinic Neuroscience Research Department in Jacksonville Florida.