The relationship between parenting factors and trait anxiety: Mediating role of cognitive errors and metacognition

The relationship between parenting factors and trait anxiety: Mediating role of cognitive errors and metacognition

Journal of Anxiety Disorders 22 (2008) 722–733 The relationship between parenting factors and trait anxiety: Mediating role of cognitive errors and m...

241KB Sizes 5 Downloads 71 Views

Journal of Anxiety Disorders 22 (2008) 722–733

The relationship between parenting factors and trait anxiety: Mediating role of cognitive errors and metacognition Bridie Gallagher *, Sam Cartwright-Hatton University of Manchester, Manchester, United Kingdom Received 12 March 2007; received in revised form 20 July 2007; accepted 23 July 2007

Abstract Research examining parenting factors in the development of anxiety has focused largely on the concepts of parental warmth and overcontrolling or intrusive parenting, This study investigated the relationship between these factors, and also parental discipline style and anxiety using self-report methodology with a sample of 16–18 year olds. In order to try to explain the relationship between parenting and anxiety, measures of cognition were also included. A multiple regression was conducted including all parenting factors as predictors of trait anxiety. The regression was a modest fit (R2 = 22%) and the model was significant (F(4, 141) = 9.90, p < 0.0001). Only the effect of Over-reactivity was significant, (t = 3.72, p < 0.0001). Furthermore, Over-reactive discipline was significantly associated with increased cognitive distortions (r = 0.361 p < 0.0001) and metacognition (r = 0.396 p < 0.0001). Both cognitive distortions and metacognition were found to partially mediate the relationship between discipline style and trait anxiety. The implications of these findings and areas for future research are discussed. # 2007 Elsevier Ltd. All rights reserved. Keywords: Trait anxiety; Parenting; Discipline styles; Metacognition; Cognitive distortions

1. Introduction Parenting has been the focus of much of the research into the development of childhood anxiety disorders and features prominently in many of the models proposed to explain their development in childhood and adolescence (e.g., Chorpita & Barlow, 1998; Manassis & Bradley, 1994; Rubin & Mills, 1991). This focus has arisen, in part, because of the evidence that anxiety is aggregated through families. Parents of clinically anxious children are more than twice as likely as parents of control * Corresponding author at: University of Manchester, School of Psychological Sciences, Zochonis Building, Oxford Road, Manchester M13 9PL, United Kingdom. Tel.: +44 161 275 2597. E-mail address: [email protected] (B. Gallagher). 0887-6185/$ – see front matter # 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.janxdis.2007.07.006

children to have a diagnosable anxiety disorder. What is more, parents of children with externalizing problems have rates of anxiety disorder closer to that of parents of controls, indicating the specific heritability of anxiety, (Last, Hersen, Kazdin, Orvaschel, & Perrin, 1991). The children of anxiety disordered parents have also been found to have significantly higher rates of anxiety disorders, with these children being reported to be seven times more likely to receive a diagnosis, (Turner, Beidel, & Costello, 1987). A review of behavioral genetics research employing twin samples indicates clearly that there is a genetic influence on childhood anxiety. However, it accounts for approximately one third of the variance, (Eley, 2001). The review also concluded that shared environment and individual specific factors contribute significantly to the heritability of anxiety in childhood and adolescence. As

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

parents are the primary influence on a child’s environment, much of the research investigating the development of childhood anxiety has focused on parenting styles and behaviors. 1.1. Care and Control: the most important parenting constructs? The majority of the extant research into the role of parenting in the etiology and maintenance of anxiety has focused on the concepts of Care and Control. Care refers to how warm and accepting parents are towards their child and Control refers to the limits that parents set for their child and how much they intrude in activities that the child is capable of undertaking independently. The focus on these two concepts is due, in part, to the prominence of the attachment literature, which suggests that insecure-resistant attachment, characterized by over involvement in combination with rejecting or inconsistent care giving, is associated with anxiety, (Bowlby, 1982). The link between insecure resistant attachment status in early childhood, and anxiety in late adolescence has been demonstrated longitudinally (Warren, Huston, Egeland, & Sroufe, 1997), providing strong support for the importance of over involved and rejecting or inconsistent care giving. The development of the Parental Bonding Instrument (PBI Parker, Tupling, & Brown, 1979) a retrospective measure of perceived parental care and control, was developed from the attachment literature, and facilitated research into these constructs. Numerous retrospective self-report studies using the PBI (and equivalent tools) have shown anxious adults to report their parenting as more controlling and less warm, (see Gerlsma, Emmelkamp, & Arrindell, 1990 for a review). 1.2. Parental discipline styles and childhood anxiety A review of parenting factors and internalizing disorders by Rapee (1997) suggests that factor analysis indicates three robust parenting factors: Care; Control; and discipline style. However, discipline style as a parenting factor has been largely ignored in populations with internalizing disorders. A small body of research has recently emerged, suggesting that parental discipline styles may be an important factor in the maintenance of anxiety disorders. In particular, the reduction of harsh and punitive discipline styles in standard behavioral parent training groups has been shown to reduce symptoms of internalizing problems as effectively as it reduces externalizing problems (Cart-

723

wright-Hatton, McNally, White, & Verduyn, 2005). Furthermore, a pilot study found that a behavioral parent training group aimed at reducing harsh and inconsistent discipline, successfully reduced anxiety in a clinically anxious sample of young children (Cartwright-Hatton, McNally, & White, 2005). Therefore, it appears that the relationship between parental discipline styles and anxiety warrants further attention. Inconsistent and harsh discipline behaviors have previously received little attention in relation to internalizing problems. However, a generally inconsistent style of parenting has been postulated as important in both the attachment literature (Bowlby, 1982) and Chorpita and Barlow’s (1998) model of the development of anxiety. Chorpita and Barlow suggest that inconsistent parenting responses lead to a decreased sense of mastery and control, as the child cannot reliably elicit a positive or negative response from their parent. This leads to a belief that they have little control over their environment. In support of this contention, inconsistent parenting behavior has been found to be associated with anxiety in a correlational study (Kohlman, Schumacher, and Streit, 1988). Additionally, we hypothesize that harsh and punitive responses may have a similar negative influence on the child’s beliefs about themselves and their capabilities. It is, therefore, theoretically plausible that inconsistent and harsh responses in discipline situations, may be just as important to children with anxiety as those with behavioral difficulties (Webster-Stratton, 1998). Clinically, a link between discipline styles and anxiety would be an important finding. Two studies, (Cartwright-Hatton, McNally, & White, 2005; Cartwright-Hatton, McNally, White, et al., 2005), make use of existing behavioral parent training technologies to instill gentle and consistent discipline styles in parents of children with internalizing symptoms. These techniques have traditionally been used with parents of children with externalizing problems, where they have met with great success (Webster-Stratton, 1998). These (and similar) parenting programs provide parents with clear, well-operationalized strategies directed at managing specific behaviors, rather than trying to address broader parenting style constructs, such as care and control. This is a significant advantage for delivery. Evidence-based technologies to address parenting that is controlling or lacking in warmth do not currently exist; therefore, an association between anxiety and parental discipline styles could have useful theoretical and clinical implications. The present study examined the relationship between anxiety and parental warmth, control and discipline

724

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

style. By examining the relative contribution of these three parenting styles, it was intended to compare the relative contribution of discipline styles to the broader constructs of care and control. Although there is now consistent evidence that some parenting styles are associated with anxiety, little attention has been paid to the reasons why parenting might have its reported effects. In this study, therefore, we also investigated how the parenting behaviors under investigation might be important in the development of anxiety. 1.3. Parenting and cognition Vasey and Dadds (2001) in their discussion of etiological models of the development of anxiety, advocate the investigation of the relationships between possible contributing variables and anxiety, in order to understand the complex pattern of multidirectional, multifinal variables involved in its development. Many of the models that attempt to explain the development of anxiety postulate that suboptimal parenting, (generally, this has been defined as a parenting style characterized by low care or acceptance and high levels of rejection and overcontrol) leads to anxiety through the development of dysfunctional cognitive schema that are biased towards threat and negative outcomes, or insecure internal working models (Chorpita & Barlow, 1998; Manassis & Bradley, 1994; Rubin & Mills, 1991; Vasey & Dadds, 2001). For example, Vasey and Dadds postulate that overprotective parenting, rather than making the child feel safer, alerts them to possible dangers and makes them believe they are vulnerable (in need of protection) and may lead to a cognitive bias towards threat. Attachment theorists also postulate that insecure attachment leads to biased cognitive appraisals of the child’s environment through the development of felt insecurity in ‘Internal Working Models’ (Manassis & Bradley, 1994). Despite the fact that numerous models now posit a mediating role for negatively biased cognition in the relationship between parenting and anxiety, there is little empirical research to support them. Despite this lack of research, initial evidence suggests that an association between parenting and cognition does exist. Self-reported negative cognition has been found to be associated with both current and retrospective self-reported parenting. For example, preliminary evidence suggests that worry is related to poor attachment and also to parental rejection and anxious rearing behaviors in primary school aged children, (Muris, Meesters, Merckelbach, & Hulsenbeck, 2000). Similarly, in an undergraduate sample,

negative automatic thoughts have been shown to be positively associated with low maternal care and overcontrol (Ingram, Overbey, & Fortier, 2001). 1.4. Empirical evidence for cognition as a mediator between parenting and anxiety A preliminary attempt to examine the relationship between parenting, anxiety and the mediating role of cognitive distortion was carried out by McGinn, Cukor and Sanderson (2005). Using a small clinical sample of adults, the Young Schema Questionnaire (YSQ, Young, 1998) was used to measure cognitive distortions, and a significant relationship was found between anxiety and the ‘‘hypervigilance’’ schema domain. However, contrary to previous findings, they found no relationship between anxiety and care or control as measured by the PBI. Therefore, mediational analyses could not be carried out. The authors suggest that the lack of a relationship between parenting and anxiety in this study may be due to their use of the Beck Anxiety Inventory (BAI), which is a physiologically focused measure of anxiety. McGinn et al. (2005) did find that reporting negative cognitive schema mediated the relationship between abusive or neglectful parenting and depression, but not low care or high control. The mediational role of cognition between parental maltreatment and depression has been demonstrated and replicated, (Hankin, 2005) and there is some evidence that low parental care leads to depressotypic attitudes and attributions, which in turn leads to increased depression (Whisman & Kwon, 1992). There is, therefore, evidence to suggest that the relationship between suboptimal parenting and emotional disorder is mediated by cognitive factors, however this has not, as yet, been demonstrated in relation to anxiety. This appears to be partly due to methodological problems with the measurement of anxiety itself and also the specific cognitive distortions associated with anxiety. For example, McGinn et al. (2005) suggest that not only was their measure of anxiety insufficient, but that the YSQ may measure depressotypic (rather than anxiotypic) cognitions. For the current study, we were interested in addressing the methodological problems encountered by McGinn et al. (2005) and in examining those cognitive styles that are implicated most strongly in the maintenance of anxiety, and which are most amenable to standard clinical interventions (e.g., Cognitive Behavior Therapy). Therefore, in the present study, measures of Cognitive Distortion, Metacognition, and Thought Control Strategies were used to examine the

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

relationships between cognition and parenting, and to explore the role of cognition in mediating the relationship between parenting and offspring anxiety. Cognitive distortions, metacognition, and thought control strategies were selected as the putative mediator variables, as they have all been empirically demonstrated to have a strong association with anxiety and anxiety disorders. Cognitive distortions (or errors) refer to the patterns of negative thinking that have been associated with emotional disturbances in Beck’s (1991) cognitive model of psychopathology. Four main different distortions or errors have been delineated: overgeneralizing (believing that a negative event is indicative of all similar or future events, e.g., ‘‘I did not do well in that test so I will always fail at school work’’); selective abstraction, (focusing on a negative event to the exclusion of other information, e.g., ‘‘I missed the opportunity to score so the whole game was a disaster’’); personalizing (believing that negative events are due to personal failings, e.g., ‘‘my best friend is talking to someone else because I am boring,’’) and catastrophizing (expecting the worst possible outcome, e.g., ‘‘my vision is blurry, I have a brain tumor.’’) Anxious individuals usually display a range of cognitive distortions; they expect more negative outcomes (Butler & Mathews, 1983) and have been found to catastrophise more about a novel worry (Davey & Levy, 1998) than control participants. Cognitive distortions have been found to be robustly associated with anxiety across a number of disorders (Harvey, Watkins, Mansell, & Shafran, 2004), and have also been reported to be associated with anxiety in adolescents (Leung & Wong, 1998) as well as adults. Metacognition, (beliefs about worry,) focuses less on the content of the individuals cognitions and refers instead to their beliefs about their cognitions. Therefore, while the content of two individuals’ worry may be the same, for example their parents’ safety, their metacognitive beliefs about this may differ, causing one individual greater worry and distress. Positive metacognitive beliefs about worry (e.g., ‘‘worrying prevents bad things from happening to me’’ and ‘‘if I worry I will always be prepared’’) are thought to increase the use of worrying as a strategy, leading, in some cases, to overuse. The development of negative metacognitive beliefs about worry (e.g., ‘‘worrying will make me go mad’’ or ‘‘worrying is going to make me ill’’) increase levels of distress, and are thought to lead some individuals to engage in counterproductive attempts at thought suppression. High levels of both positive and negative metacognitive beliefs have been shown to be present in a number of anxiety disorders (Harvey et al.,

725

2004), and are also associated with trait anxiety and worry in community samples (Cartwright-Hatton & Wells, 1997). Distorted metacognitive beliefs have also been clearly demonstrated in adolescents as young as 13 (Cartwright-Hatton et al., 2004). Thought control is important in the cognitive model of worry, because the way in which the individual attempts to control worry and negative thoughts is also important in the development of anxiety. Trying to suppress negative thoughts, i.e., actively trying not to think about a particular thought, often has a paradoxical effect, (Wegner, Schneider, Carter, & White, 1987). However, how people try to prevent themselves thinking about a thought has been shown to be important (Wells & Davies, 1994). For example in the Wells & Davies’ study it was shown that using self-punishment or focusing on alternative worries as strategies for dealing with negative thoughts are associated with higher trait anxiety, whereas, distraction onto a more pleasant topic was not. The importance of believing that thoughts must be controlled, and the use of maladaptive thought control strategies, has been found to be associated with Obsessive Compulsive Disorder (OCD, Tolin, Worhunsky, & Maltby, 2006), Generalized Anxiety Disorder (GAD, Coles & Heimberg, 2005) and Post Traumatic Stress Disorder (PTSD, Roussis & Wells, 2006). 1.5. Summary of rationale The contribution of parental warmth and control to the development of anxiety has been investigated extensively in both self-report and observational studies. Discipline styles, despite being identified as a robust parenting factor (Rapee, 1997), have been largely ignored in the anxiety literature. Recent evidence suggests that tackling parents’ harsh, punitive and inconsistent discipline may be an effective treatment for young children with anxiety disorders (Cartwright-Hatton, McNally, & White, 2005). In order to examine the usefulness of discipline style as a predictor of anxiety it is important to look at how much predictive power it can add to the more extensively researched parenting factors of warmth and control. It is also pertinent to models of the development of anxiety to investigate by what mechanism parenting behavior increases the risk of offspring anxiety. One of the possible pathways postulated by existing models is the development of negative cognitions and a negative cognitive style. However, empirical evidence to support this contention is lacking and previous attempts to demonstrate a mediating role for cognition in the development of anxiety have been unsuccessful

726

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

(McGinn et al., 2005). A possible explanation for the lack of relationship found between variables in the McGinn study is the use of measurement tools and cognitions not suited to investigating anxiety. This study therefore aims to address gaps in the existing literature in order to achieve a better understanding of which parenting factors are important in the development of anxiety and how these factors have their effect. 1.6. Aims and objectives The present study investigates the relationship between parenting and trait anxiety, examining the constructs of care/warmth, control/overprotection, rejection and parental discipline styles. The relationship between these parenting behaviors and cognitive distortions, metacognitive beliefs, and thought control strategies, are also examined. 2. Method 2.1. Participants One hundred and sixty eight adolescents were recruited from a Sixth Form College in Greater Manchester and an Undergraduate Psychology course at the University of Manchester. Participation was open to those aged between 16 and 18 years old, (M = 17.23, S.D. = 0.86). No exclusion criteria were applied. Brief demographic information is summarized in Table 1. 2.2. Procedure Students at the Sixth Form College were approached by the psychology teacher and asked to complete seven short questionnaires in a number of allocated sessions at the college. Of a possible 112 participants, 75 completed all seven questionnaires. Two hundred and ninety two undergraduates were invited by e-mail to take part in the study online. Ninety-three Undergraduates who were 18 years old completed the same battery of questionnaires online in exchange for course credits. Respondents were asked to answer all parenting measures in reference to their primary caregiver.

The pack of questionnaires for all participants included: 2.3. Measures 2.3.1. State-Trait Anxiety Inventory-Trait version (STAI-T, Spielberger, Gorsuch, & Lushene, 1970) The STAI Y-2 is a 20-item self-report measure, which assesses trait anxiety. It asks the respondent to describe how they feel generally, by responding to a number of statements on a four-point scale from ‘‘almost never’’ to ‘‘almost always.’’ Items include ‘‘I feel pleasant’’ and ‘‘I feel inadequate.’’ The STAI has been widely used and shown to demonstrate adequate validity and reliability, Cronbach’s alpha is reported as 0.90. The STAI trait version was used in this study as a general measure of trait anxiety. 2.3.2. Parental Bonding Instrument (PBI, Parker et al., 1979) The PBI is a 25-item self-report measure that aims to assess individuals’ perceptions of their parent’s behavior towards them during the first 16 years of their life. It focuses on the dimensions of Care (12 items) and Control (13 items). All items are rated on a four-point scale from ‘‘very like’’ to ‘‘very unlike.’’ The Care dimension ranges from rejection and coldness (low scores) to affection and warmth (high scores) and includes items such as ‘‘seemed emotionally cold to me’’ and ‘‘spoke to me with a warm and friendly voice.’’ The Control dimension ranges from the promotion of independence, to intrusive behaviors and the promotion of dependence. This includes items such as ‘‘invaded my privacy’’ and ‘‘let me dress in any way I wanted.’’ The PBI has been used in many studies of parenting and has demonstrated adequate split-half reliability: Care 0.74, Overprotection 0.88, (Gerlsma et al., 1990; Parker et al., 1979). The PBI was included in this study as a retrospective measure of parental overcontrol and warmth. 2.3.3. The Childrearing Styles Questionnaire, (CSQ) The CSQ is a 30-item self report measure, adapted from the Arnold Parenting Scale, (Arnold, O’Leary,

Table 1 Demographic information Age of respondent

Sex of respondent

Primary care giver

16 years

17 years

18 years

Female

Male

Mother

Father

Other

31%

8%

61%

85%

15%

89%

10%

1%

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

Wolff, & Acker, 1993) which asks about the respondent’s perceptions of their childrearing experiences before the age of 12 years old. The adaptations phrase the items in the past tense for use with adults and adolescents. Each item gives a situation, then a positive and a negative anchor with a seven-point scale. The respondent is asked to choose which behavior is closest to what happened to them when they were a child. The CSQ correlates with observed parenting behavior, suggesting validity, and Cronbach’s alpha is reported at 0.84, (Arnold et al., 1993). The measure includes three subscales derived through factor analysis; Laxness, Verbosity, and Overreactivity. The Laxness scale measures parent’s ability to discipline their child effectively and stick to punishments and rewards that they have promised. Verbosity measures a parent’s propensity to use lengthy explanations and get into arguments with their child. The Over-reactivity subscale measures punitive, harsh, or inconsistent discipline and includes items such ‘‘When I was misbehaving . . .. My parents almost always used bad language or cursed, (negative anchor) . . .. My parents rarely used bad language and cursed, (positive anchor).’’ The Cronbach’s alpha of this subscale is reported at 0.82. Only the Over-reactivity subscale and the total score of dysfunctional parenting were included in the analysis. 2.3.4. My memories of upbringing short version, (EMBU-s, Arrindell et al., 1999) The EMBU-s is a 23-item questionnaire developed from the EMBU, an 81-item Swedish self-report measure. It assesses perceptions of the parental rearing behaviors of one’s parents before the age of 16. Each item presents a behavior or attitude and responses are on a four-point scale from ‘‘no, never’’ to ‘‘yes, most of the time.’’ The measure is comprised of three subscales: Rejection; Warmth; and Overprotection, with seven, six, and nine items respectively. Rejection measures criticism and coldness towards the child. The Warmth scale measures praise, acceptance, and affection. The Overprotection scale measures parents’ limit setting and use of aversion in controlling their child’s behavior. The EMBU-s has demonstrated adequate validity and reliability, with Cronbach alpha ranging from 0.72 to 0.85 for subscales, cross-culturally (Arrindell et al., 1999). The EMBU-s was included in this study as a retrospective measure of overprotection and rejection. 2.3.5. Cognitive Error Questionnaire, (CEQ, Lefebvre, 1981) The CEQ is a 24-item self-report questionnaire, which assesses cognitive distortions or errors in thinking. Each item contains a short vignette about

727

an everyday event then a distorted thought that the respondent might have in this situation. The respondent is then asked how similar this thought is to how they might think in this situation. Each item is rated on a five-point scale from ‘‘not at all like I would think’’ to ‘‘exactly like I would think.’’ All items are summed to give a Cognitive Error Score. The measure is also said to contain four subscales, each with six items: catastrophizing; overgeneralizing; personalizing; selective abstraction. However, the four-factor structure has not been confirmed by factor analysis. The scale has good internal consistency with Cronbach’s alpha reported to range from 0.89 to 0.92 (Lefebvre, 1981) and has been shown to be a valid measure of dysfunctional thinking, (Schutte & Malouff, 1995). The CEQ was included as a general measure of cognitive distortion. 2.3.6. The Metacognitions Questionnaire-30 (MCQ-30: Wells & Cartwright-Hatton, 2004) The MCQ-30 measures individual differences in metacognitive beliefs. It consists of 30-items that make up five replicable factors. It is derived from a 65-item version of the MCQ with the same five-factor structure, (Cartwright-Hatton & Wells, 1997). The five factors measure beliefs regarding: Positive beliefs about worry (e.g., ‘‘Worrying helps me avoid problems in the future’’); Uncontrollability and Danger (e.g., When I start worrying I can’t stop’’); Cognitive Confidence (e.g., ‘‘I do not trust my memory’’); Superstitious beliefs and Personal Responsibility (e.g., ‘‘If I did not control a worrying thought and then it happened it would be my fault’’), and Cognitive Consciousness (‘‘I constantly study my thoughts’’). Respondents are asked to respond to each statement on a four-point scale from ‘‘Agree very much’’ to ‘‘Do not agree.’’ Adequate Cronbach’s alpha, (ranging from 0.72 for Cognitive Self-consciousness to 0.89 for uncontrollability and danger) and test–retest reliability have been demonstrated, (Cartwright-Hatton & Wells, 1997). 2.3.7. Thought Control Questionnaire, (TCQ, Wells & Davies, 1994). The TCQ is a 30-item self-report measure, which was developed to assess the frequency of five strategies that people use to control their thoughts. The five strategies are: distraction; punishment; reappraisal; social control, and worry. Respondents are asked to respond to each item on a four-point likert scale ranging from ‘‘never’’ to ‘‘almost always.’’ The scale has adequate internal consistency and test-retest reliability, with Cronbach alphas ranging from 0.64 (punishment) to 0.83 (reappraisal). The TCQ was included as a

728

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

general measure of dysfunctional thought control strategies. 3. Results 3.1. Preliminary analyses Prior to analysis, the data were examined for correct entry, missing values, and acceptable skew and kurtosis for the planned parametric analysis. A KolmogorovSmirnov test indicated that scores on the CEQ, STAI, PBI Care, and EMBU rejection were not normally distributed. Scores on the CEQ and STAI were transformed into natural logarithms and analyses were conducted on these scores. Transformations of the PBI and EMBU scores did not improve skewness or kurtosis. However, in a sample of this size, Tabachnick and Fidell (2001) suggest that parametric analyses can tolerate this degree of skewness and kurtosis. Missing or incomplete data on one or more measures was handled using pair-wise or list-wise (when more than two cases was missing) deletion (Tabachnick & Fidell, 2001). Means and standard deviations are presented in Table 2. 3.2. The relationship between reported parenting and anxiety symptoms Pearson correlations were calculated between the STAI and total/subscale scores of all parenting measures. The correlations are presented in Table 3. Because of the use of multiple correlations, a Bonferroni correction was calculated, and an adjusted alpha level was adopted (adjusted a = 0.05/7 = 0.007). Using this alpha level, significant positive correlations were found between anxiety and all subscales of the PBI and EMBU-s, and the total score of the CSQ and its Table 2 Means and standard deviations for all completed measures

PBI care PBI control EMBU rejection EMBU warmth EMBU overprotection CSQ total CSQ Over-reactivity STAI CEQ MCQ-30 TCQ

M

S.D.

27.75 12.32 10.56 18.86 19.27 3.33 3.22 43.97 4.82 60.72 63.68

7.17 7.40 4.20 4.07 4.80 0.52 0.96 10.33 1.59 13.92 8.78

Over-reactivity subscale. All correlations were modest. Over-reactivity as measured by the CSQ, demonstrated the strongest correlation with trait anxiety, r = 0.38, p < 0.01. The data were then further analyzed using a forced entry multiple regression, entering as regressors the four most strongly correlated subscales from the parenting measures. These were Over-reactivity, PBI Care, EMBU Overprotection, and EMBU Rejection. The dependent variable was anxiety (STAI). The regression was a modest fit (R2 = 22%) and the overall model was significant, F(4, 141) = 9.90, p < 0.0001. Only the effect of Over-reactivity as measured by the CSQ was significant, t = 3.72, p < 0.0001. 3.3. Associations between reported parenting and dysfunctional cognition Pearson correlations were employed to analyze the relationship between dysfunctional parenting as measured by the Over-reactivity scale of the CSQ and the three measures of dysfunctional cognition: cognitive errors (CEQ); metacognition (MCQ); and thought control (TCQ). The CEQ and MCQ showed modest, positive correlations with Over-reactive Parenting, r = 0.36 and r = 0.40 respectively, p’s < 0.001. The TCQ showed a small, nonsignificant, positive correlation with Over-reactivity, r = 0.12, p = 0.147. 3.4. The role of cognition in the relationship between reported parenting and anxiety The significant relationships indicated by the correlations between parenting as measured by the CSQ Over-reactivity subscale, and Anxiety as measured by the STAI, provided the basis for a mediational analysis. To establish mediation the following conditions must be met: the independent variable (Parental Over-reactivity) must predict the dependent variable (Anxiety); the independent variable must also predict the mediator variable, (Cognitive Errors or Metacognition); the mediator (Cognitive Errors or Metacognition) must predict the dependent variable (Anxiety), and the association between the independent variable (Overreactivity) and dependent variable (Anxiety), must be eliminated or attenuated when the mediator variable (Cognitive Errors or Metacognition) is added to the model, (Baron & Kenny, 1986). Two mediational analyses were employed. The first examined the potential mediating role of cognitive errors, while the second examined the potential mediating role of metacognition.

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

729

Table 3 Associations (Pearson’s correlation coefficients) between parenting constructs and anxiety

STAI Trait anxiety Correlation Sig (2 tailed) N

Care (PBI)

Overcontrol (PBI)

Rejection (EMBU)

Warmth (EMBU)

Overprotection (EMBU)

Over-reactivity (CSQ)

Total CSQ score

0.291 0.000 158

0.219 0.005 159

0.263 0.001 151

0.238 0.003 154

0.240 0.003 153

0.380 0.000 156

0.250 0.002 152

3.5. Over-reactive Parenting, Cognitive Errors and Trait Anxiety The association between Over-reactive Parenting and Anxiety was positive and significant (b = 0.712, p < 0.0001). The association between Parenting and Cognitive Error was also positive and significant (b = 4.616, p < 0.0001). The association between Cognitive Errors and Anxiety was positive and significant (b = 0.073 p < 0.0001). The introduction of the hypothesized mediator, Cognitive Error, reduced the association between Parenting and Anxiety by 37% (b = 0.448). According to the Sobel test, this reduction is significant (3.65, p < 0.001), indicating that Cognitive Errors partially mediated the association between Over-reactive Parenting and Anxiety (Fig. 1). 3.6. Over-reactive Parenting, metacognition and trait anxiety The association between Over-reactive Parenting and Anxiety was positive and significant (b = 0.712, p < 0.0001) as was the association between Parenting and Metacognition (b = 43.12, p < 0.0001). The relationship between Metacognition and Anxiety was also positive and significant (b = 0.009, p < 0.0001). The introduction of the hypothesized mediator, Metacognition, reduced the association between Parenting and Anxiety by 42%, (b = 0.409). According to the Sobel

test, this reduction was significant (4.276, p < 0.001), indicating that metacognition is a partial mediator in the relationship between parenting and anxiety, (see Fig. 2). 4. Discussion Models of the development of anxiety are as yet poorly defined and this study adds to the existing literature with two findings. First, it suggests that parental discipline styles contribute to the development or maintenance of anxiety. While the constructs of Warmth, Overprotection, and Rejection also predicted a small amount of the variance in anxiety, these effects were eclipsed by the effect of over-reactive parental discipline. Second, the findings provide empirical support for an association between harsh and punitive discipline styles and dysfunctional cognition, namely cognitive distortions and metacognition. Furthermore, it appears that this dysfunctional cognitive style partially mediates the relationship found between dysfunctional parenting and anxiety. These findings will now be discussed in turn. 4.1. Parental discipline styles and anxiety The findings suggest that reporting having received a harsh, punitive, and inconsistent style of discipline before the age of twelve is closely associated with trait anxiety in late adolescence. Although Warmth,

Fig. 1. Mediational model of Over-reactive Parenting, Cognitive Errors and Trait Anxiety.

730

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

Fig. 2. Mediational model of Over-reactive Parenting, Metacognition and Trait.

Overprotection, and Rejection contributed a small proportion of variance in anxiety, this was eclipsed by the contribution of over-reactive parenting, which was the only significant predictor in the regression model. This finding questions the focus of much of the previous research on the constructs of Care and Control and suggests that research into parenting factors needs to be extended to investigate the role of discipline behaviors, particularly consistency and harshness in behavior management strategies. As discipline style explains more of the variance in current anxiety than warmth, control or rejection, increased research focus in this area may be warranted. Clinically, the finding that parental discipline style is closely associated with anxiety is interesting, and suggests that interventions focusing on changing parents’ discipline behaviors should be effective in ameliorating anxiety symptoms. This result is consistent with the finding that an intervention aimed at reducing harsh and inconsistent discipline improved internalizing symptoms in 18–42-month-old children, (Cartwright-Hatton, McNally, & White, 2005). Protocols for parental management of the type of discipline behaviors operationalized in the CSQ (the instrument used to measure discipline style in this study) have already been developed and evaluated for use with children with externalizing problems, (e.g., WebsterStratton, 1998); and are widely used in this context. Therefore, the technology is already available and accessible. 4.2. Why is discipline style important in the development of anxiety? The second finding of this study is that retrospectively reported punitive and inconsistent discipline styles are associated with both increased cognitive errors and more maladaptive metacognitive beliefs. Moreover, these cognitive factors appear to partially

explain the relationship between over-reactive parenting and trait anxiety. This finding has given an insight into the understanding of how discipline styles contribute to the development of anxiety. The roles of cognitive and metacognitive errors will now be considered in turn. 4.3. Cognitive errors The findings suggest that parenting characterized by harsh, punitive, or inconsistent discipline behaviors contributes to the development of cognitive distortions. The cognitive distortions measured in this study by the CEQ include: catastrophizing, selective abstraction, overgeneralizing, and personalizing. These findings support the contention that if parents ‘‘Over-react’’ and respond harshly to their child’s misdemeanours, this is likely to bias the child’s interpretations of their environment in a negative direction. The child may become prone to thinking that ambiguous situations are going to end badly, to focusing on negative events and to taking personal responsibility for them. Mathews and Mackintosh (2000) have demonstrated that inducing cognitive biases in individuals can lead to them becoming more anxious, providing evidence for the contention that cognitive distortions, which affect interpretation of events, may place the individual at increased risk of becoming anxious. 4.4. Metacognitive beliefs The findings also suggest that the offspring of overreactive parents are more likely to hold dysfunctional metacognitive beliefs about worry. The association between a perceived, harsh, punitive upbringing, and certain meta-cognitive beliefs is unsurprising. First, the relationship between harsh, over-reactive parenting, and positive beliefs about worry might arise as the individual comes to rely on worry as a tool for

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

predicting future stressful events, such as were common in their childhood. Likewise, worry may have come to be used to avoid problems, in order that harsh punishments for mistakes and failures could be avoided (‘worrying helps me to avoid problems in the future’). The relationship between over-reactive parenting and negative beliefs about the uncontrollability and danger in worry, might have arisen as a function of general beliefs about uncontrollability and danger, in the world of a child who is frequently in receipt of harsh and unpredictable parenting (‘My worrying thoughts persist, no matter how I try to stop them’). Likewise, the relationship between poor cognitive confidence and over-reactive parenting might have arisen as a function of the poor self-esteem and general confidence in children who have lacked early warmth and encouragement or received frequent criticism (‘I have a poor memory’). Finally, the relationship between overreactive parenting and superstitious beliefs about punishment and responsibility is likely to have arisen in children who are frequently punished, (I will be punished for not controlling certain thoughts’), and who have developed superstitious explanations for outcomes, perhaps as a result of being held responsible for their actions when this was not appropriate (‘If I did not control a worrying thought, and then it happened, it would be my fault’). 4.5. Limitations A number of limitations to this study should be acknowledged. The finding that cognitive/metacognitive style is only a partial mediator of the relationship between over-reactive parenting and anxiety indicates that the relationship is explained by other factors in addition to cognition. However, this is unsurprising, as the development of anxiety is widely held to be a multifactorial process, (Vasey & Dadds, 2001). The influence of shared genetic predisposition (Eley, 2001) and the transmission of fear information (Field, Argyris, & Knowles, 2001; Field & Lawson, 2003; Field, Hamilton, Knowles, & Plews, 2003) for example, provide other potential pathways between parenting behavior and anxiety. This sample was heavily biased in gender, with 85% females, and only included young people who have continued their education after 16 years, limiting its generalizability. The tertiary college that was used has a wide catchment area with an intake that is ethnically and economically representative of the local community. The undergraduate students who took part, however, are not representative of their local commu-

731

nity. A wider sample would be required to replicate these initial findings. The use of retrospective measures of parenting behaviors has also been criticized, chiefly because of the potential biases in recall that this introduces. The use of a sample of 16–18-year olds, however, limits problems of recall by reducing the time that has elapsed since the individual experienced the parenting behaviors, while allowing assessment of complex cognitive factors that might not be achievable in a younger sample. It has also been suggested that because of general cognitive biases, individuals who report more anxiety may also interpret their parents’ behaviors as more negative. However, Brewin, Andrews, and Gotlib (1993) have reported that there is little evidence to suggest that psychiatric status is associated with poorer recall of early experiences. 5. Conclusion These findings suggest that parental discipline styles characterized by harsh, punitive, or inconsistent discipline are associated with increased anxiety. Care and control in parenting explained a small proportion of the variance, but discipline style was the only significant predictor, suggesting the need for a shift in research focus. Fortuitously, treatments that tackle harsh and inconsistent parental discipline are already available, and this also offers a unique opportunity to experimentally investigate this relationship further. The finding that a negative cognitive style mediates the relationship found between discipline style and anxiety aids understanding of why parenting is important in the development of anxiety. It also provides preliminary empirical support for theoretical models of anxiety that posit that the relationship between suboptimal parenting and trait anxiety is explained by a negative cognitive style. References Arnold, D. S., O’Leary, S. G., Wolff, L. S., & Acker, M. M. (1993). The Parenting Scale: A measure of dysfunctional parenting in discipline situations. Psychological Assessment, 5, 137–144. Arrindell, W. A., Sanavio, E., Aguilar, G., Sica, C., Hatzichristou, C., Eisemann, M., et al. (1999). The development of a short form of the EMBU: Its appraisal with students in Greece Guatamala, Hungary and Italy. Personality and Individual Differences, 27, 613–628. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.

732

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733

Beck, A. T. (1991). Cognitive therapy: A 30-year retrospective.. American Psychologist, 46, 368–375. Bowlby, J. (1982). Attachment and loss—retrospect and prospect. American Journal of Orthopsychiatry, 52, 664–678. Brewin, C. R., Andrews, B., & Gotlib, I. H. (1993). Psychopathology and early experience—a reappraisal of retrospective reports. Psychological Bulletin, 113, 82–98. Butler, G., & Mathews, A. (1983). Cognitive processes in anxiety. Advances in Behavior Therapy Research, 5, 51–62. Cartwright-Hatton, S., Mather, A., Illingworth, V., Brocki, J., Harrington, R., & Wells, A. (2004). Development and preliminary validation of the Meta-cognitions Questionnaire—Adolescent version. Journal of Anxiety Disorders, 18, 411–422. Cartwright-Hatton, S., McNally, D., & White, C. (2005a). A new cognitive behavioral parenting intervention for families of young anxious children: A pilot study. Behavioral and Cognitive Psychotherapy, 3, 243–247. Cartwright-Hatton, S., McNally, D., White, C., & Verduyn, C. (2005b). Parenting Skills Training: An effective intervention for internalizing symptoms in younger children? Journal of Child and Adolescent Psychiatric Nursing, 18, 45–52. Cartwright-Hatton, S., & Wells, A. (1997). Beliefs about worry and intrusions: The Meta-Cognitions Questionnaire and its correlates. Journal of Anxiety Disorders, 11, 279–296. Chorpita, B. F., & Barlow, D. H. (1998). The development of anxiety: The role of control in the early environment. Psychological Bulletin, 124, 3–21. Coles, M. E., & Heimberg, R. G. (2005). Thought control strategies in generalized anxiety disorder. Cognitive Therapy and Research, 29, 47–56. Davey, G. C. L., & Levy, S. (1998). Catastrophic worrying: Personal inadequacy and a perseverative iterative style as features of the catastrophizing process. Journal of Abnormal Psychology, 107, 576–586. Eley, T. C. (2001). Contributions of behavioral genetics research: Quantifying genetic, shared environmental and nonshared environmental influences. In: M. W. Vasey & M. Dadds (Eds.), The developmental psychopathology of anxiety (pp. 45–59). Oxford: University Press. Field, A. P., Argyris, N. G., & Knowles, K. A. (2001). Who’s afraid of the big bad wolf: a prospective paradigm to test Rachman’s indirect pathways in children. Behavior Research and Therapy, 39, 1259–1276. Field, A. P., Hamilton, S. J., Knowles, K. A., & Plews, E. L. (2003). Fear information and social phobic beliefs in children: a prospective paradigm and preliminary results. Behavior Research and Therapy, 41, 113–123. Field, A. P., & Lawson, J. (2003). Fear information and the development of fears during childhood: effects on implicit fear responses and behavioral avoidance. Behavior Research and Therapy, 41, 1277–1293. Gerlsma, C., Emmelkamp, P. M. G., & Arrindell, W. A. (1990). Anxiety, depression, and perception of early parenting—a metaanalysis. Clinical Psychology Review, 10, 251–277. Hankin, B. L. (2005). Childhood maltreatment and psychopathology: Prospective tests of attachment, cognitive vulnerability, and stress as mediating processes. Cognitive Therapy and Research, 29, 645–671. Harvey, A., Watkins, E., Mansell, W., & Shafran, R. (2004). Cognitive Behavioral Processes Across Psychological Disorders: A transdiagnostic approach to research and treatment. Oxford: Oxford University Press.

Ingram, R. E., Overbey, T., & Fortier, M. (2001). Individual differences in dysfunctional automatic thinking and parental bonding: specificity of maternal care. Personality and Individual Differences, 30, 401–412. Kohlman, C. W., Schumacher, A., & Streit, R. (1988). Trait-Anxiety and parental child-rearing behavior: Support as a moderator variable. Anxiety Research, 1, 53–64. Last, C. G., Hersen, M., Kazdin, A., Orvaschel, H., & Perrin, S. (1991). Anxiety disorders in children and their families. Archives of General Psychiatry, 48, 928–934. Lefebvre, M. F. (1981). Cognitive distortion and cognitive errors in depressed psychiatric and low-back-pain patients. Journal of Consulting and Clinical Psychology, 49, 517–525. Leung, P. W. L., & Wong, M. M. T. (1998). Can cognitive distortions differentiate between internalizing and externalizing problems? Journal of Child Psychology and Psychiatry and Allied Disciplines, 39, 263–269. Manassis, K., & Bradley, S. J. (1994). The development of childhood anxiety disorders: Towards an integrated model. Journal of Applied Developmental Psychology, 15, 345–366. Mathews, A., & Mackintosh, B. (2000). Induced emotional interpretation bias and anxiety. Journal of Abnormal Psychology, 109, 602–615. McGinn, L. K., Cukor, D., & Sanderson, W. C. (2005). The relationship between parenting style. Cognitive style and anxiety and depression: Does increased early adversity influence symptom severity through the mediating role of cognitive style. Cognitive Therapy and Research, 29, 219–242. Muris, P., Meesters, C., Merckelbach, H., & Hulsenbeck, P. (2000). Worry in children is related to perceived parental rearing and attachment. Behavior Research and Therapy, 38, 487–497. Parker, G., Tupling, H., & Brown, L. B. (1979). Parental Bonding Instrument. British Journal of Medical Psychology, 52, 1–10. Rapee, R. M. (1997). Potential role of childrearing practices in the development of anxiety and depression. Clinical Psychology Review, 17, 47–67. Roussis, P., & Wells, A. (2006). Post-traumatic stress symptoms: Tests of relationships with thought control strategies and beliefs as predicted by the metacognitive model. Personality and Individual Differences, 40, 111–122. Rubin, K. H., & Mills, R. S. L. (1991). Conceptualizing developmental pathways to internalizing disorders in childhood. Canadian Journal of Behavioral Science, 23, 300–317. Schutte, N. S., & Malouff, (1995). Sourcebook of adult assessment strategies. New York: Plenum Press. Spielberger, C. D., Gorsuch, R. C., & Lushene, R. F. (1970). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Tabachnick, B. G., & Fidell, L. S. (2001). Using Multivariate Statistics (Fourth ed.). London: Allyn and Bacon. Tolin, D. F., Worhunsky, P., & Maltby, N. (2006). Are ‘‘obsessive’’ beliefs specific to OCD?: A comparison across anxiety disorders. Behavior Research and Therapy, 44, 469–480. Turner, S. M., Beidel, D. C., & Costello, A. (1987). Psychopathology in the offspring of anxiety disorder patients. Journal of Consulting and Clinical Psychology, 55, 229–235. Vasey, M., & Dadds, (2001). The Developmental Psychopathology of Anxiety. Oxford: University Press. Warren, S. L., Huston, L., Egeland, B., & Sroufe, L. A. (1997). Child and adolescent anxiety disorders and early attachment. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 637–644.

B. Gallagher, S. Cartwright-Hatton / Journal of Anxiety Disorders 22 (2008) 722–733 Webster-Stratton, C. (1998). Preventing conduct problems in head start children: Strengthening parenting competencies. Journal of Consulting and Clinical Psychology, 66, 715–730. Wegner, D. M., Schneider, D. J., Carter, S. R., & White, T. L. (1987). Paradoxical effects of thought suppression. Journal of Personality and Social Psychology, 53, 5–13. Wells, A., & Cartwright-Hatton, S. (2004). A short form of the metacognitions questionnaire: Properties of the MCQ-30. Behavior Research and Therapy, 42, 385–396.

733

Wells, A., & Davies, M. I. (1994). The Thought Control Questionnaire—a measure of individual-differences in the control of unwanted thoughts. Behavior Research and Therapy, 32, 871–878. Whisman, M. A., & Kwon, P. (1992). Parental representations, cognitive distortions, and mild depression. Cognitive Therapy and Research, 16, 557–568. Young, J. E. (1998). Young Schema-Questionnaire Short Form. New York: Cognitive Therapy Center.