lable at ScienceDirect

Behaviour Research and Therapy 48 (2010) 1097e1104

Contents lists avai

Behaviour Research and Therapy

journal homepage: www.elsevier .com/locate/brat

Fear and perceived uncontrollability of emotion: Evaluating the uniquecontribution of emotion appraisal variables to prediction of worryand generalised anxiety disorder

Lexine A. Stapinski a,*, Maree J. Abbott b, Ronald M. Rapee a

aCentre for Emotional Health, Macquarie University, Sydney, NSW, Australiab School of Psychology, University of Sydney, Sydney, NSW, Australia

a r t i c l e i n f o

Article history:Received 27 August 2009Received in revised form15 July 2010Accepted 27 July 2010

Keywords:Generalised anxiety disorderGADWorryEmotionFear of emotionEmotion regulationPerceived controlThreatIntolerance of uncertaintyMeta-cognitionMeta-worry

* Correspondence to: Lexine A. Stapinski, DepartmUniversity, Sydney, NSW 2109, Australia. Tel.: þ612 9

E-mail address: lexine.stapinski@mq.edu.au (L.A. S

0005-7967/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.brat.2010.07.012

a b s t r a c t

Novel theoretical frameworks place the symptom profile of generalised anxiety disorder (GAD) withinthe context of dysfunctional emotional processes. It is suggested that fear and intolerance of emotionsexacerbate subjective distress and motivate the use of maladaptive coping strategies, such as worry. Todate, studies evaluating these models have suffered two key limitations. Firstly, few studies haveinvolved treatment-seeking samples, and secondly, none have evaluated the unique variance attributableto emotion appraisal variables above and beyond previously established predictors of worry and GAD.The present study begins to address these limitations by assessing the contribution of fear and perceiveduncontrollability of emotions in predicting worry and clinical GAD status after controlling for varianceattributable to depressive symptoms, meta-cognitive beliefs, intolerance of uncertainty, and perceptionsof external threat. Supporting current models, results showed that perceived control over emotionalreactions was a unique predictor of GAD diagnostic status and both clinical and non-clinical worry.

� 2010 Elsevier Ltd. All rights reserved.

Introduction

Characterised by persistent anxiety and chronic uncontrollableworry, generalised anxiety disorder (GAD) represents a consider-able public health concern, with an estimated lifetime prevalenceof 5.1% in the general population (Wittchen, Zhao, Kessler, & Eaton,1994). The cost of GAD to the individual and to society is substan-tial, particularly in view of its chronic course (Wittchen & Hoyer,2001), and the significant impairment to occupational, social andday-to-day functioning. Despite this, the development of effica-cious treatments for GAD has lagged behind other anxiety disorders(Brown, Barlow, & Liebowitz, 1994), and until recently, cognitivebehavioural treatment approaches for GAD were relatively non-specific, with no unifying theoretical foundation (Craske, 1999).Consequently, the past two decades have seen a timely increase in

ent of Psychology, Macquarie850 1801.tapinski).

All rights reserved.

studies seeking to enhance theoretical models and treatment forthis disorder. These advances have led to the delineationof important cognitive mechanisms (e.g., Dugas, Gagnon,Ladouceur, & Freeston, 1998; Wells, 1999), and novel conceptualframeworks that place GAD in the context of dysfunctionalemotional processes (Borkovec & Roemer, 1995; Mennin,Heimberg, Turk, & Fresco, 2002).

Consistent with conceptualisations of other anxiety disorders,biases in the interpretation and detection of threat are central tomodels of GAD (Borkovec,1994; Rapee,1991). A considerable bodyofresearch indicates that GAD is associated with preferential attentiontowards threat and inflated interpretation of subjective personal riskboth in terms of the probability and cost of danger occurring (seeCraske, 1999). Barlow and colleagues have extended this threatmodel by underscoring the importance of perceptions of control(e.g., Chorpita & Barlow,1998; Craske, Rapee, Jackel, & Barlow,1989).From this control perspective, the core of GAD and other emotionaldisorders is not the inflated perception of negative events per se, butthe perceiveduncontrollability of these experiences, and consequentamplification of emotional distress (Craske et al., 1989).

L.A. Stapinski et al. / Behaviour Research and Therapy 48 (2010) 1097e11041098

Additional psychological mechanisms have been identified asrelevant to the specific symptom profile associated with GAD.Several investigations indicate that worry and GAD are stronglyassociated with intolerance of uncertainty, which is defined as thetendency to exhibit negative cognitive, emotional and behaviouralreactions to uncertain situations and events (Buhr & Dugas, 2006;Dugas, Freeston, & Ladouceur, 1997; Dugas et al., 1998). Given theconsiderable degree of uncertainty present in everyday life,intolerance of uncertainty is thought to contribute to the symp-toms of heightened distress and worry observed in GAD (Buhr &Dugas, 2006). The construct of intolerance of uncertainty mayalso encompass aspects of perceived uncontrollability and inflatedthreat perception, as these conditions are likely to create a senseof unease and uncertainty regarding the possibility of futurethreat.

Several models suggest that worry itself may be self perpet-uating, due to maladaptive appraisal of one’s own cognitiveprocesses (Davey, Tallis, & Capuzzo, 1996; Wells, 1999). Morespecifically, the selection of worry as a coping strategy may beinfluenced by maladaptive beliefs about the positive conse-quences of worrying, while appraisals focused on the potentialnegative impact of worrying amplify subjective distress by trig-gering “meta-worry” (worry about worry; Wells, 1999). Severalstudies provide support for the relationship between meta-cognitive beliefs and worry proneness (Cartwright-Hatton &Wells, 1997; Wells & Papageorgiou, 1998), with some studiessupporting the specificity of these appraisals to GAD (Wells &Carter, 2001).

A number of groups have pointed to the importance ofdysfunctional emotional processes in conceptualising GAD andchronic worry (Borkovec & Roemer, 1995; Mennin et al., 2002;Roemer & Orsillo, 2002). The emotion dysregulation model ofGAD suggests these individuals have limited access to regulationstrategies, and experience emotions as threatening, difficult tounderstand, uncontrollable and overwhelming. Consequently,individuals with GAD may be prone to maladaptive coping strate-gies, particularly worry, in an effort to dampen or avoid negativeaffect and associated perceived negative outcomes (Borkovec,Alcaine, & Behar, 2004; Mennin et al., 2002). This aversion toaffective experience is thought to apply across a range of discreteemotional experiences, including positive emotional experience,although it may be most evident for anxious affect given itsprominence in GAD (Mennin, 2004).

Preliminary empirical support for the importance ofemotional appraisal and regulation deficits in GAD has begun toemerge. Analogue GAD samples report heightened intensity ofnegative affect relative to non-anxious and socially anxiousparticipants (Mennin, Heimberg, Turk, & Fresco, 2005; Turk,Heimberg, Luterek, Mennin, & Fresco, 2005). Compared to non-anxious participants, analogue and treatment-seeking GADparticipants experience more difficulty identifying, describingand accepting their emotions, and report greater fear of bothpositive and negative emotional experiences (McLaughlin,Mennin, & Farach, 2007; Mennin et al., 2005; Roemer et al.,2009; Roemer, Salters, Raffa, & Orsillo, 2005; Salters-Pedneault,Roemer, Tull, Rucker, & Mennin, 2006). Furthermore, empiricaldata has been put forward to support the association betweenfear of emotions and worry (Roemer et al., 2005), and thepremise that worry serves as a cognitive avoidance strategy thatdampens emotional arousal through distraction from more dis-tressing cognitions (Borkovec & Hu, 1990; Borkovec & Roemer,1995; Vrana, Cuthbert, & Lang, 1986).

Although initial self report studies support the associationbetween emotion dysfunction and GAD, more rigorous methodol-ogies are needed to confirm the importance of these processes in

the aetiology and maintenance of this disorder. Clearly, there is anoverlap between processes implicated by the emotion regulationmodel, and mechanisms emphasised in alternate cognitive modelsof GAD. Aversion to negative affect is likely related to negativebeliefs about the uncontrollability and consequences of worry asimplicated in Wells’ (1999) meta-cognitive model of GAD.Furthermore, fear of emotional experience could be considereda natural consequence of the chronic apprehension resulting frompersistent threat detection and intolerance of uncertainty. Toprovide solid support for the emotion dysregulation model, it isimportant to determine whether dysfunctional emotionalprocesses share a unique association with GAD that is not betterexplained by these other cognitive factors. Preliminary support hasbeen demonstrated for the independent contribution of emotionalnon-acceptance and dysregulation in predicting worry andanalogue GAD status above and beyond the degree of negativeaffectivity experienced over the past month (Salters-Pedneaultet al., 2006). More recently, Roemer et al. (2009) found that selfreported emotion regulation difficulties accounted for variance inanalogue GAD symptom severity independently of variance sharedwith mindful attention and awareness, depression and anxietysymptoms.While these findings are promising, to date no study hasevaluated the contribution of emotional processes in predictingGAD independently of previously established cognitive mecha-nisms such as inflated threat perception, intolerance of uncertaintyand meta-cognitions.

Emotion appraisal and regulation models of GAD are animportant avenue of investigation, particularly given the directimplications for treatment of this disorder. Indeed, recentadvances in therapeutic interventions for this disorder haveunderscored the role of emotional processes and incorporatecomponents targeting emotional exposure, regulation andacceptance (e.g., Mennin, 2005; Newman, Castonguay, Borkovec,& Molnar, 2004; Roemer & Orsillo, 2003). Given the considerableinfluence of emotion dysregulation conceptualisations of GAD, itis perhaps surprising that few studies have evaluated thesemodels within a treatment-seeking sample. With the exceptionof a few smaller studies (Study 2, Mennin et al., 2005; Study 2,Roemer et al., 2005; Study 2, Roemer et al., 2009), most inves-tigations have involved university students screened for GADcriteria on the basis of a self report diagnostic measure (GAD-Q-IV; Newman et al., 2002). Clearly, more empirical data are neededto evaluate and clarify aspects of emotional dysfunction relevantto the psychopathology of GAD.

The present study begins to address these limitations byexploring the relationship between appraisal of emotional experi-ence and GAD symptomatology in a large treatment-seekingsample. Specifically, it was predicted that compared to non-anxiouscontrols; GAD participants would report less perceived controlla-bility and greater fear of their emotions. An additional aim was toextend previous research by delineating the unique contributionsof these constructs after controlling for cognitive processes previ-ously implicated in models of GAD. Furthermore, the relationshipbetween subjective appraisal of emotional experiences and thetendency to worry excessively was assessed. The emotion dysre-gulation model proposes that fear and negative appraisal ofemotions lead to the adoption of worry as a coping mechanism todampen or avoid intense affect (Borkovec et al., 2004; Mennin et al.,2002), thus these variables were expected to account for uniquevariance above and beyond previously established predictors ofworry. Given the considerable diagnostic overlap between GAD anddepression (see Kessler, 1997), depressive symptoms were alsocontrolled statistically in order to identify appraisal processescharacteristic of GAD independently of those associated witha depressive personality style.

L.A. Stapinski et al. / Behaviour Research and Therapy 48 (2010) 1097e1104 1099

Method

Participants

Two groups of participants were recruited to the study. Theclinical group were 126 participants meeting diagnostic criteria forgeneralised anxiety disorder (Diagnostic and Statistical Manualof Mental Disorders, 4th ed., text rev.; American PsychiatricAssociation, 2000). All clinical participants were seeking treat-ment for GAD at the Macquarie University Centre for EmotionalHealth. The non-anxious control (NAC) group consisted of 79volunteers recruited through advertisements in the local commu-nity who were paid a small sum as reimbursement for their timeand travel. Diagnostic assessment of all participants was based ona structured interview using the Anxiety Disorders InterviewSchedule for DSMeIV (ADIS-IV; DiNardo, Brown, & Barlow, 1994).All diagnostic interviews were conducted by graduate studentswho had been trained by clinical psychologists experienced in theassessment and treatment of anxiety disorders. Although diag-nostic reliability was not determined for the entire sample, datafrom our clinic indicates good diagnostic reliability for adultanxiety and depression using the ADIS-IV (Rapee, Abbott, Baillie, &Gaston, 2007).

For participants in the clinical group, co-morbid Axis I diagnoseswere accepted, with the exception of schizophrenia, provided thatGAD was the primary diagnosis as determined by a clinicianseverity rating encompassing impairment and subjective distress.Only NAC participants who did not meet criteria for GAD, or anyother clinical disorder, were included in the study. Two NACparticipants met criteria for an Axis I diagnosis as assessed by theADIS-IV and were excluded from the study. Three clinical partici-pants and one NAC participant were excluded due to missing dataon one or more key variables, leaving 76 participants in the NACgroup and 123 participants in the GAD group.

Within the GAD group, the mean clinical severity rating forgeneralised anxiety disorder was 6.2 (SD¼ .86), on a 9-point scalewhere a higher score indicates greater severity. The majority ofparticipants (73.8%) also met diagnostic criteria for an additionalAxis I diagnosis. Co-morbid Axis I diagnoses included social phobia(50.8%), major depressive disorder (24.6%), specific phobia (18.0%),panic disorder (8.2%), and other anxiety disorders (4.1%). The meanage of the GAD and NAC participants was 36.6 years (SD¼ 12.2) and35.7 years (SD¼ 14.1) respectively, and the difference in the meanages was not statistically significant, F(1,195)¼ .23, p> .05.Nonparametric analyses revealed no significant differencesbetween groups in terms of gender, c2 (1, N¼ 197)¼ 2.32,p> .05, income, c2 (4, N¼ 176)¼ 3.82, p> .05, ethnicity, c2

(1,N¼ 197)¼ 1.73, p> .05, education, c2 (2,N¼ 195)¼ 5.68, p> .05,or employment, c2 (7, N¼ 197)¼ 5.13, p> .05. Across groups,the sample was predominantly female (71.6%), Anglo-Australian(80.7%), with tertiary qualifications (79.7%), and income below AUD$40,000 (62.4%). Analyses indicated that clinical groups signifi-cantly differed in terms of marital status, c2 (1, N¼ 197)¼ 15.03,p< .001; around half of GAD participants were married or in a defacto relationship (54.9%), whereas themajority of NAC participantswere unmarried (73.3%).

Measures

Symptom measuresIn addition to a structured diagnostic assessment, the Penn State

Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, & Borkovec,1990) was used to assess worry severity. This 16 item measure isdesigned to capture the excessive and uncontrollable nature ofpathological worry, where higher scores indicate more severe

worry. Numerous studies have demonstrated sound psychometricproperties both in terms of the validity and reliability of the PSWQ(Brown, Antony, & Barlow, 1992; Davey, 1993; Fresco, Mennin,Heimberg, & Turk, 2003; Meyer et al., 1990). Good internalconsistency was also demonstrated in the present sample (a¼ .97).

Symptoms of anxiety, depression and stress were measuredusing the 21 item Depression Anxiety and Stress Scales (DASS-21;Lovibond & Lovibond, 1993), where higher scores indicated moresevere symptoms experienced over the past week. This scale hasdemonstrated good psychometric properties in both non-clinicaland clinical samples (Antony, Beiling, Cox, Enns, & Swinson, 1998;Lovibond & Lovibond, 1995). All subscales demonstrated goodinternal consistency in the present sample (Depression: a¼ .91,Anxiety: a¼ .83, Stress: a¼ .92).

Threat appraisalThe Probability Cost and Coping Questionnaire (PCCQ) was used

to assess participants’ probability, cost and coping estimates inrelation to possible future threat. The questionnaire consists of 11brief scenarios describing hypothetical future catastrophes acrossa broad range of situations and concerns (e.g., “You break up withyour partner/spouse”, “One of your close family members is fatallyinjured in an accident”). For each hypothetical scenario, partici-pants were asked to rate on a 9-point scale: (1) the likelihood of theevent happening to them in the near future, (2) the consequences(how bad or distressing) it would be if the event were to occur, and(3) how difficult it would be to cope should the event occur. A totalscore for this measure comprised a total of all items, such thathigher scores indicated higher perceptions of the likelihood,consequences and difficulty coping with future threat. Full scaledevelopment and psychometrics for this measure will be reportedelsewhere (Abbott, Kemp, Rapee, & Stapinski, in preparation).Convergent validity of the PCCQ in the current sample was sup-ported by moderate correlations with symptoms of anxiety andstress as assessed by the DASS-21 in both clinical (DASS-A: r¼ 0.38;DASS-S: r¼ 0.42) and non-anxious (DASS-A: r¼ 0.35; DASS-S:r¼ 0.29) groups. Reliability analyses revealed high internalconsistency (a¼ .93) for the scale in the current sample, andadequate testeretest reliability was demonstrated in a subset of theclinical group (n¼ 19) over a 12 week period (r¼ 0.73).

Intolerance of uncertaintyThe Intolerance of Uncertainty Scale (IUS; Freeston, Rheaume,

Letarte, Dugas, & Ladouceur, 1994) is a 27-item measure thatassesses aspects of intolerance of uncertainty such as the percep-tion that uncertainty is unacceptable, immobilising, stressful,frustrating, and reflects badly on the person (e.g., “Uncertaintymakes me uneasy, anxious or stressed”, “When it’s time to act,uncertainty paralyses me”). Higher scores on this questionnaireindicate greater intolerance of uncertainty. Good internal consis-tency and testeretest reliability have been demonstrated for thismeasure (Buhr & Dugas, 2002). Comparable internal consistencywas also demonstrated in the present sample (a¼ .97).

Meta-beliefs about worryThe short form of the Meta-Cognitions Questionnaire (MCQ-30;

Cartwright-Hatton &Wells, 1997) was used to assess processes andknowledge involved in the appraisal and management of cogni-tions. This measure evaluates meta-cognitions across a number ofdomains considered relevant to psychopathology, where higherscores reflect greater endorsement of meta-cognition beliefs: (1)positive beliefs about worry (e.g., “Worrying helps me solve prob-lems”), (2) negative beliefs about worry concerning uncontrolla-bility and danger (e.g., “My worrying is dangerous for me”), (3)perceived cognitive competence (e.g., “Mymemory can misleadme

Table 1Group differences in mean scores on symptom measures and predictor variables.

Variable GADMean (SD)

NACMean (SD)

hp2 F

Symptom measuresPSWQ 66.82 (8.95) 32.64 (9.64) .77 F(1,195)¼ 638.67***

DASSDepression 16.47 (9.59) 3.44 (4.19) .39 F(1,195)¼ 123.75***

Anxiety 14.61 (7.55) 2.91 (3.19) .46 F(1,194)¼ 161. 73***

Stress 25.06 (8.79) 6.99 (5.16) .57 F(1,194)¼ 261.08***

Predictorsa

PCCQ 153.93 (29.96) 109.48 (32.43) .16 F(1,194)¼ 35.79***

IOU 81.33 (22.93) 42.30 (10.98) .23 F(1,194)¼ 56.58***

MCQ 72.53 (13.97) 49.59 (11.53) .17 F(1,194)¼ 39.95***

ACS 3.90 (.76) 2.56 (.70) .17 F(1,194)¼ 40.43***

ACQ-R 27.84 (9.84) 53.21 (8.71) .45 F(1,194)¼ 156.08***

Note: hp2 refers to partial eta squared, an estimate of effect size: small¼ .01; mod-erate¼ .06; large¼ .14.

a DASS depressions scores were included as a covariate for all analyses ofpredictor variables. All comparisons are significant at p< .001 after employingBonferroni correction to account for number of tests.

L.A. Stapinski et al. / Behaviour Research and Therapy 48 (2010) 1097e11041100

at times”), (4) general negative beliefs about the consequences ofnot controlling thoughts (e.g., “If I could not control my thoughts, Iwould not be able to function”) and (5) cognitive self-consciousness(e.g., “I think a lot about my thoughts”). Previous research indicatesthis measure has sound psychometric properties, and is positivelyassociated with worry proneness (Cartwright-Hatton & Wells,1997). Good internal consistency for the total scale was alsodemonstrated in the current sample (a¼ .93).

Fear and perceived uncontrollability of emotional experiencesTwo self report measures were used to assess dimensions of

affect appraisal central to the emotion dysregulation model of GAD.The Affect Control Scale (ACS; Williams, Chambless, & Ahrens,1997) was used to assess fear, and fear of loss of control, overemotional experiences. This 42 item measure assesses fear acrossfour domains of emotional experience: anger, anxiety, depressionand positive emotions (e.g., “Depression is scary tomee I am afraidI could get depressed and never recover”; “It scares me when I feelshaky”). The total score of this measure is the average of all items,where higher scores are indicative of greater fear of emotionalexperiences. The ACS has demonstrated convergent validity withother measures of perceived emotional control, and good temporalstability and internal consistency in an undergraduate sample(Williams et al., 1997). Comparable internal consistency wasdemonstrated for the total scale in the current sample (a¼ .95).

Perceptions of control over emotions were assessed using thereactions subscale of the Anxiety Control Questionnaire (ACQ;Rapee, Craske, Brown, & Barlow, 1996). Higher scores on this 14item subscale indicate greater perceived control over internalemotional reactions (e.g., “When I am put under stress, I am likelyto lose control”; “My emotions seem to have a life of their own”).Research suggests this measure has good reliability and validitywithin clinical and non-clinical samples (Rapee et al., 1996; Zebb &Moore, 1999). Good internal consistency for this subscale wasdemonstrated in the current sample (a¼ .89).

Data scoring and analysis strategy

Missing values on all self report measures were replaced withthe scalemean, provided that at least 80% of items on that scalewerecompleted. Initially, one-way ANOVAs were conducted to examinegroup differences on symptom measures and predictor variables.A Bonferroni correction was applied to avoid error rate inflationacross multiple tests of significance. Principal analyses involvedlogistic regression and hierarchical linear regression procedures toevaluate the unique contribution of fear and perceived uncontrol-lability of emotions in predicting clinical GAD status and worryseverity (PSWQ) above and beyond established predictors.Depression symptom scores (DASS-D) were first entered intoregression models to take into account variance attributable todepressogenic attributional styles. Given the established impor-tance of inflated threat appraisals (PCCQ), intolerance of uncertainty(IOU) and meta-cognitive appraisal (MCQ) in the maintenance ofGAD and worry, variance attributable to these constructs wascontrolled for in the second block of the regressionmodel.Measuresassessing fear and perceived uncontrollability of emotions (ACS andACQ-R) were entered into the last block, to determine theimprovement in prediction following inclusion of these measures.This analysis strategy allowed for a highly conservative test of thecontribution of emotional appraisal processes to GAD and worry.

Results

Prior to analysis, all variables were screened for outliers andnormality of distribution. Significant multivariate outliers were

identified for one participant in the GAD group, and one participantin the NAC group (standard residuals> 3.5). Due to the sensitivityof regression procedures to outlying data points, these participantswere removed from all analyses (see Tabachnick & Fidell, 2001).

Mean scores on symptom measures and predictor variables byclinical group status are shown in Table 1. As would be expected,symptom severity was significantly higher for GAD participants asindexed by trait worry, anxiety, depression and stress scores. Pearsoncorrelations between each of the predictor variables for each groupare presented in Table 2. As shown, there were significant intercor-relations between most measures. Given the number of relatedmeasures included in analyses, it was necessary to assess for multi-collinearity within each regression model. High multicollinearity isgenerally indicated by tolerance levels less than 0.2 and VIF levelsgreater than 5 (Tabachnick & Fidell, 2001). Examination of theseindices showed acceptable levels of multicollinearity for all regres-sion models. The ratio of observations to number of predictor vari-ableswasgreater than10:1 for all analyses, indicating that the samplesize was adequate for regression analysis procedures (Peduzzi,Concato, Kemper, Holford, & Feinstein, 1996). Logistic regressionanalyses are not sensitive to unequal sample sizes across categories(Spicer, 2004), and thus the difference in participant numbersbetween GAD and NAC groups was not considered problematic.

Group comparisons

All analyses included DASS-D scores as a covariate in order toassess group differences on predictor variables while controlling fordepressive symptoms. The first prediction was that GAD partici-pants would report greater fear and perceived uncontrollability ofemotions compared with NAC participants. As can be seen in Table1, this hypothesis was supported by the data. Mean fear of emotion(ACS) scores were significantly higher for GAD participants thanNAC participants. GAD participants also endorsed significantlylower perceived control over internal experiences (ACQ-R).Consistent with previous research GAD participants displayedsignificantly lower tolerance for uncertainty (IOU), stronger meta-cognitive beliefs (MCQ), and inflated appraisals of threat (PCCQ)compared to NAC participants.

Prediction of GAD

A hierarchical logistic regression analysis evaluated the uniquecontribution of fear and perceived uncontrollability of emotions in

Table 2Intercorrelations between predictor key variables for GAD and NAC participants.

Scale 1 2 3 4 5 6 7

GAD participants (n[ 122)1. PSWQ e �.28** .49*** .50*** .45*** .33*** �.46***2. DASS-D e .21* .42*** .48*** .51*** �.24**3. PCCQ e .48*** .48*** .46*** �.41***4. MCQ e .64*** .55*** �.38***5. IOU e .56*** �.34***6. ACS e �.52***7. ACQ-R e

NAC participants (n[ 75)1. PSWQ e .34** .55*** .56*** .40*** .34** �.57***2. DASS-D e .26* .44*** .49*** .43*** �.41***3. PCCQ e .32** .38** .27** �.46***4. MCQ e .63*** .38*** �.42***5. IOU e .41*** �.42***6. ACS e �.55***7. ACQ-R e

*p< .05, **p< .01, ***p< .001.

Table 4Hierarchical multiple regression predicting worry severity (PSWQ) in GAD and NACparticipants.

Predictors entered DR2 b in final model B (SE) in final model

GAD participants (n[ 122)Model 1 .08**DASS-D .10 .09 (.08)

Model 2 .26***PCCQ .26** .08 (.03)MCQ .25* .16 (.06)IOU .13 .05 (.04)

Model 3 .07**ACS �.19 �2.29 (1.23)ACQ-R �.30** �.27 (.08)

NAC participants (n[ 75)Model 1 .12**DASS-D .02 .05 (.23)

Model 2 .35***PCCQ .32** .10 (.03)MCQ .38** .32 (.09)IOU �.09 �.08 (.10)

Model 3 .06*ACS �.04 �.51 (1.44)ACQ-R �.31** �.35 (.12)

Note. See Results section for discussion of hierarchical regression strategy.*p< .05, **p< .01, ***p< .001.

L.A. Stapinski et al. / Behaviour Research and Therapy 48 (2010) 1097e1104 1101

predicting clinical GAD status (GAD versus NAC). A goodmodel fit was observed on the basis of the first model, c2

(1, N¼ 197)¼ 120.63, p< .001, indicating that symptoms ofdepression reliably contributed to the prediction of GAD. Based ondepression scores alone, a correct classification rate of 84.8% wasobserved. Comparison of log-likelihood ratios indicated that addi-tion of the established predictors (PCCQ, MCQ, IOU), reliablyimproved classification, c2 (3, N¼ 197)¼ 43.54, p< .001. Classifi-cation on the basis of these predictors increased to 88.3%.

Prediction of clinical status was further improved with theaddition of measures assessing fear and perceived uncontrollabilityof emotional experiences, c2 (2, N¼ 197)¼ 32.79, p< .001. Thefull regression model demonstrated good model fit, c2 (6,N¼ 197)¼ 196.97, p< .001, with correct classification of 92.9%.However, examination of the relative contribution of each variable(shown in Table 3) revealed that only perceived controllability ofinternal reactions (ACQ-R) and Intolerance of Uncertainty (IOU)significantly contributed to classification of GAD status within thefull model.

Prediction of clinical and non-clinical worry

To evaluate predictions of the emotion regulation model, therelationship between the constructs of interest and worry severity(as assessed by PSWQ) was examined using a hierarchical multipleregression approach for GAD and NAC groups separately. Results ofthis analysis are presented in Table 4.

In the GAD group, depression scores significantly contributed tothe prediction of worry, F(1, 120)¼ 10.46, p¼ .002, and addition ofthe established predictors (PCCQ, MCQ, IOU) further improved the

Table 3Logistic regression results (odds ratios) for full model predicting GAD versus NACstatus.

Predictors B (SE) Wald df Odds ratio Odds ratio 95%CI

Lower Upper

DASS-D .12 (.07) 3.24 1 1.13 0.99 1.30PCCQ .00 (.01) 0.00 1 1.00 0.98 1.03MCQ .01 (.03) 0.18 1 1.01 0.95 1.08IOU .06 (.03) 4.97* 1 1.06 1.01 1.11ACS �.50 (.67) 0.54 1 0.61 0.17 2.25ACQ-R �.19 (.05) 16.80*** 1 0.82 0.75 0.90

Note. See Results section for discussion of logistic regression strategy.*p< .05, **p< .01, ***p< .001.

predictive power of the model, F(3, 117)¼ 15.36, p< .001. Aftercontrolling for variance attributable to these variables, the emotionappraisal measures demonstrated a unique contribution to theprediction of pathological worry, F(2, 115)¼ 6.27, p¼ .003. The fullmodel accounted for 41% of the variance in worry scores. Onlyscores on the MCQ, PCCQ and ACQ-R significantly contributed tovariance in worry scores within the full model.

The same pattern emerged in the NAC group. Depression scoressignificantly added to the prediction of worry, F(1, 73)¼ 9.75,p¼ .003, with prediction further improved with the addition of theestablished predictors, F(3, 70)¼ 15.43, p< .001. Inclusion of ACSand ACQ-R scores demonstrated an independent contribution tothe prediction of worry, F(2, 68)¼ 4.32, p¼ .017, with the full modelaccounting for 53% of the variance in worry scores. Within the finalmodel, scores on the MCQ, PCCQ and ACQ-R were uniquely asso-ciated with worry beyond their shared variance with otherpredictors.

Discussion

The present study sought to evaluate predictions of theemotional dysregulation model of GAD in a treatment-seekingsample. Compared to non-anxious control participants, GADparticipants reported less perceived control and greater fear oftheir emotional experiences. The between group comparisons alsoreplicated previous studies demonstrating the importance ofinflated threat perception, intolerance of uncertainty, and meta-cognitive appraisals in the conceptualisation of this disorder.

To extend previous research, the unique contribution of fear andappraisal of emotional experience above and beyond establishedpredictors of GAD and pathological worry was assessed. Measuresassessing fear and perceived controllability of emotional experi-ences significantly improved the classification of clinical GAD statusabove and beyond established predictors (inflated threat appraisal,intolerance of uncertainty and meta-cognitive appraisal) anddepressive symptoms. Although the improvement in predictionwas modest (4.6%), it should be noted that the analytic procedureemployed was highly conservative, with four established predic-tors, all with strong associations with GAD, prioritised in theregression model. Despite considerable intercorrelation between

L.A. Stapinski et al. / Behaviour Research and Therapy 48 (2010) 1097e11041102

all variables of interest, the present results suggest that subjectiveemotional experience and appraisal is independently associatedwith GAD symptomatology and impairment. The unique contri-bution of each predictor within the final model was also evaluated,revealing that only perceived emotional control and intolerance ofuncertainty independently predicted clinical status. Althoughintolerance of uncertainty and perceived emotional control sharecommon variance with other predictors, they appear to be betteroverall explanatory constructs for this disorder.

Emotion dysregulation models of GAD posit that fear andintolerance of emotional experiences motivate the use of worry asa control strategy to dampen physiological arousal. Analysesexamined the relationship between emotion appraisal processesand clinical and non-clinical worry while controlling for varianceshared with depression, inflated threat perception, meta-cognitiveappraisal and intolerance of uncertainty. Supporting a dimensionalview of worry, the same pattern of results emerged for clinical andnon-clinical groups. In both analyses, the addition of emotionalvariables improved prediction of worry. Within the final model,perceived emotional control, meta-cognitive appraisals and infla-ted threat appraisal showed a significant unique relationship withboth clinical and non-clinical worry. Based on these results, itwould appear that the worry process is associated with theperception of threat at various levels of experience. As has beensuggested, worry may be triggered by the perception of external orinternal danger, and further exacerbated by meta-cognitiveappraisals of theworry process itself (Wells, 2004). The finding thatperceived control over emotions independently predicted worryseverity supports current models suggesting worry may be auto-matically or strategically selected in an attempt to regain control ofunpleasant internal experiences (e.g., Borkovec et al., 2004; Menninet al., 2002).

Supporting emotion dysregulation frameworks, perceivedcontrol over internal sensations was the only construct to emergeas a significant unique predictor of both GAD and worry. Whilecognitive mechanisms (meta-cognitive appraisals and inflatedthreat appraisal) appear more strongly related to the cognitiveaspects of GAD (i.e. worry), perceived emotional control appears toshows a more generalised association with worry and the broadersymptom profile of GAD. It may be that this sense of lack of controlover internal emotional states amplifies the experience of negativeaffectivity (Barlow, 2000; Craske et al., 1989), contributing to thechronic restlessness and tension that characterises this disorder.Furthermore, this subjective loss of control might render individ-uals with GAD prone to maladaptive control strategies, such asworry, aimed at regaining emotional equilibrium. At present thisconceptualisation of GAD is speculative, and in need of additionalempirical support, particularly in terms of establishing the direc-tion of causality between variables. Nonetheless, this frameworkprovides a promising explanation for the current pattern of results,and important implications for treatment of this disorder. Accord-ing to Barlow and colleagues, early life experiences of uncontrol-lability, trauma or parental overcontrol produce a psychologicalvulnerability to interpret future experiences, including internalreactions, as uncontrollable (Chorpita & Barlow, 1998; Rapee, 1997).Applying this perspective to GAD, therapeutic interventions mightbe beneficial to the extent that they shift expectations regardingone’s capacity to control affective experiences. On the other hand,interventions that challenge the drive for control and encourageexperiential acceptance may assist in reducing the distress associ-ated with low perceived control (e.g., Roemer & Orsillo, 2002).Future investigations will be needed to evaluate the influence ofcurrent therapeutic approaches on perceived emotional controlla-bility, and assess the relationship between control perceptions andimprovements in GAD symptomatology.

While perceived emotional control emerged as an independentpredictor, fear of emotions as assessed by the ACS did not,indicating the latter construct may be less important in terms ofpredicting worry and GAD status. This finding was unexpected,and may relate to the degree of overlap between this measureand measures assessing alternate cognitive and emotionalprocesses. Inspection of the data also revealed higher variability forACS scores relative to other measures which may have contributedto the poorer predictive ability demonstrated in the current study.It is important to note that variables that do not make a uniquecontribution in the context of alternative predictors may still haveclinical relevance. Structural equation analyses will be advanta-geous in future investigations to model how these underlyingcognitive and emotional constructs interact to contribute to GADsymptomatology.

Although not the focus of the present study, the differentpattern of results obtained when predicting clinical GAD statuscompared to prediction of worry severity warrants comment.While intolerance of uncertainty emerged as a significant uniquepredictor of GAD, this measure did not significantly predict worryseverity when considered within the full model. The reversepattern was observed for measures assessing meta-cognitiveappraisal and external threat perception, which accounted forunique variance in clinical and non-clinical worry but did notsignificantly add to the classification of GAD. Although worry is thecentral characteristic of GAD, diagnosis is not made solely on thebasis of worry, and it has been suggested that worry and GADshould be considered related, but distinguishable, constructs(Ruscio, 2002). In fact, the experience of excessive worry is notunique to GAD, and is manifest in other anxiety disorders (Brownet al., 1992), as well as non-clinical groups (Ruscio, 2002). Giventhe chronic course and relatively early onset of GAD (Kessler,Keller, & Wittchen, 2001; Wittchen & Hoyer, 2001), some havesuggested this disorder may be better conceptualised as a person-ality disturbance (Beck, Stanley, & Zebb,1996; Sanderson &Wetzler,1991), with studies suggesting individuals diagnosed with GADdiffer from non-anxious controls on personality attributes such asdependence, suspiciousness, neuroticism, interpersonal sensitivity,hypervigilance and reassurance seeking (Gasperini, Battaglia,Diaferia, & Bellodi, 1990; Gomez & Francis, 2003). The contrastingpattern of results observed in the present study point to thepotential value of further exploring the distinction between GADand worry. While meta-cognitive beliefs and inflated threatperception emerged as important cognitive processes related to thegeneration and maintenance of worry, intolerance of uncertaintymight be conceived as an important trait dimension within thebroader personality profile of GAD. Indeed, Laugesen, Dugas, andBukowski (2003) suggest that some individuals are predisposedto intolerance of uncertainty from childhood, and describe thisconstruct as a cognitive schema or filter through which individualswith GAD view theworld. Intolerance of uncertainty is also likely toencompass additional processes of relevance, such as inflatedthreat appraisal and perceived uncontrollability, given that theseprocesses likely contribute to a sense of uncertainty regarding thepotential for imminent danger. However, at present these distinc-tions are speculative, as firm conclusions cannot be formed fromcomparisons across the different analytic approaches employed inthis study to evaluate predictors of worry severity and clinicalstatus. Replication and extension will be needed to clarify therelative importance of these psychological mechanisms to worry ascompared to the clinical profile of GAD.

The relatively large treatment-seeking sample employed in thepresent study is a significant strength, as was the evaluation ofsubjective emotional appraisals within the context of previouslyestablished predictors of GAD and worry. Furthermore, all analyses

L.A. Stapinski et al. / Behaviour Research and Therapy 48 (2010) 1097e1104 1103

controlled for variance attributable to depression. This statisticalcontrol strengthens the current conclusions, suggesting theseemotional processes are specifically relevant to GAD and worry,rather than indicative of a depressive personality profile andcorresponding attributional style. On the other hand, the currentconclusions are tempered by the cross-sectional study design andreliance on self report measures. Studies incorporating prospectiveor experimental designs are needed to provide more directevidence that maladaptive emotional appraisals precede andcontribute to chronic worry and symptomatology of GAD. It shouldalso be noted that the current findings do not speak to the speci-ficity of these processes to GAD. Future studies incorporatingclinical control groups are needed to identify disorder-specificversus transdiagnostic patterns of emotional responding.

Additional caveats of the current research warrant consider-ation. Firstly, the large proportion of unmarried participants in thenon-anxious group (73%) was not representative of the broaderpopulation and may limit the generalisability of these findings.Secondly, examining the prediction of worry separately in the GADand non-anxious groups reduces the range of scores and may leadto an underestimation of the association between variables. Finally,it is important to note that this study focussed on investigating oneaspect of emotional dysregulation, negative appraisal and fear ofemotional experiences. Other aspects of emotional dysregulationsuch as poor acceptance, heightened intensity, and the use ofmaladaptive management strategies are important avenues forfuture investigation, particularly in view of preliminary evidencethat the latter two show disorder-specificity to GAD (Mennin,Holaway, Fresco, Moore, & Heimberg, 2007). These limitationsnotwithstanding, the current results add to the existing literature inproving support for the role of emotional dysfunction modelsin enhancing theoretical explanations of GAD. In particular, thepresent study indicates that perceptions of emotional control mightbe a promising target for therapeutic intervention and furtherresearch efforts to unravel psychological mechanisms underpin-ning GAD. Given that emotional regulation models of GAD are intheir infancy, it will be important to delineate additional appraisalprocesses that may contribute to the aversion to and avoidance ofemotional experience.

Acknowledgements

This research was supported by a research training fellowshipfrom the New SouthWales Institute of Psychiatry to the first author.The authors would like to acknowledgeMike Jones for his statisticaladvice, Ileana Hatton for her assistance with data collection, andtwo anonymous reviewers for their comments on an earlier draft ofthis manuscript.

References

Abbott, M. J., Kemp, N., Rapee, R. M., & Stapinski, L. A. Development and validationof the probability, cost and coping questionnaire to evaluate threat perceptionin generalised anxiety disorder, in preparation.

American Psychiatric Association (APA). (2000). Diagnostic and statistical manual ofmental disorders (4th ed., text rev.). Washington, DC: Author.

Antony, M. M., Beiling, P. J., Cox, B. J., Enns, M. W., & Swinson, R. P. (1998).Psychometric properties of the 42-item and 21-item versions of the depressionanxiety stress scales in clinical groups and a community sample. PsychologicalAssessment, 10, 176e181.

Barlow, D. H. (2000). Unraveling the mysteries of anxiety and its disorders from theperspective of emotion theory. American Psychologist, 55, 1248e1263.

Beck, J. M., Stanley, A., & Zebb, B. J. (1996). Characteristics of generalized anxietydisorder in older adults: a descriptive study. Behaviour Research and Therapy, 34,225e234.

Borkovec, T. D. (1994). The nature, functions, and origins of worry. In G. C. Davey,& F. Tallis (Eds.), Worrying: Perspectives on theory, assessment and treatment(pp. 5e33). England: John Wiley & Sons.

Borkovec, T. D., Alcaine, O., & Behar, E. (2004). Avoidance theory of worry andgeneralized anxiety disorder. In R. G. Heimberg, C. L. Turk, & D. S. Mennin (Eds.),Generalized anxiety disorder: Advances in research and practice (pp. 77e108).New York, NJ: Guildford Press.

Borkovec, T. D., & Hu, S. (1990). The effect of worry on cardiovascular response tophobic imagery. Behaviour Research and Therapy, 28, 69e73.

Borkovec, T. D., & Roemer, L. (1995). Perceived functions of worry among general-ized anxiety disorder subjects: distraction from more emotionally distressingtopics? Journal of Behavior Therapy and Experimental Psychiatry, 26, 25e30.

Brown, T. A., Antony, M. M., & Barlow, D. H. (1992). Psychometric properties of thePenn State worry questionnaire in a clinical anxiety disorders sample. BehaviourResearch and Therapy, 30, 33e37.

Brown, T. A., Barlow, D. H., & Liebowitz, M. R. (1994). The empirical basis ofgeneralized anxiety disorder. American Journal of Psychiatry, 151, 1272e1280.

Buhr, K., & Dugas, M. (2002). The intolerance of uncertainty scale: psychometricproperties of the English version. Behaviour Research and Therapy, 40, 931e946.

Buhr, K., & Dugas, M. J. (2006). Investigating the construct validity of intolerance ofuncertainty and its unique relationship with worry. Journal of Anxiety Disorders,20, 222e236.

Cartwright-Hatton, S., & Wells, A. (1997). Beliefs about worry and intrusions: themeta-cognitions questionnaire and its correlates. Journal of Anxiety Disorders, 11,279e296.

Chorpita, B. F., & Barlow, D. H. (1998). The development of anxiety: the role ofcontrol in the early environment. Psychological Bulletin, 124, 3e21.

Craske, M. G. (1999). Anxiety disorders: Psychological approaches to theory andtreatment. Boulder, CO: Westview Press.

Craske, M. G., Rapee, R. M., Jackel, L., & Barlow, D. H. (1989). Qualitative dimensionsof worry in DSM-III-R generalized anxiety disorder subjects and nonanxiouscontrols. Behaviour Research and Therapy, 27, 397e402.

Davey, G. C. (1993). A comparison of three worry questionnaires. Behaviour Researchand Therapy, 31, 51e56.

Davey, G. C., Tallis, F., & Capuzzo, N. (1996). Beliefs about the consequences ofworrying. Cognitive Therapy and Research, 20, 499e520.

DiNardo, P. A., Brown, T. A., & Barlow, D. H. (1994). Anxiety disorders interview schedulefor DSM-IV. San Antonio, TX: The Psychological Corporation, Harcourt Brace.

Dugas, M. J., Freeston, M. H., & Ladouceur, R. (1997). Intolerance of uncertainty andproblem orientation in worry. Cognitive Therapy and Research, 21, 593e606.

Dugas, M. J., Gagnon, F., Ladouceur, R., & Freeston, M. H. (1998). Generalized anxietydisorder: a preliminary test of a conceptual model. Behaviour Research andTherapy, 36, 215e226.

Freeston, M. H., Rheaume, J., Letarte, H., Dugas, M. J., & Ladouceur, R. (1994). Why dopeople worry? Personality and Individual Differences, 17, 791e802.

Fresco, D. M., Mennin, D. S., Heimberg, R. G., & Turk, C. L. (2003). Using the PennState worry questionnaire to identify individuals with generalized anxietydisorder: a receiver operating characteristic analysis. Journal of Behavior Therapyand Experimental Psychiatry, 34, 283e291.

Gasperini, M., Battaglia, M., Diaferia, G., & Bellodi, L. (1990). Personality featuresrelated to generalized anxiety disorder. Comprehensive Psychiatry, 31, 363e368.

Gomez, R., & Francis, L. M. (2003). Generalised anxiety disorder: relationships withEysenck’s, Gray’s and Newman’s theories. Personality and Individual Differences,34, 3e17.

Kessler, R. C. (1997). The prevalence of psychiatric comorbidity. In S. Wetzler, &W. C. Sanderson (Eds.), Treatment strategies for patients with psychiatriccomorbidity (pp. 23e48). New York, NJ: John Wiley & Sons.

Kessler, R. C., Keller, M., & Wittchen, H. (2001). The epidemiology of generalizedanxiety disorder. Psychiatric Clinics of North America, 24, 19e39.

Laugesen, N., Dugas, M. J., & Bukowski, W. M. (2003). Understanding adolescentworry: the application of a cognitive model. Journal of Abnormal ChildPsychology, 31, 55e64.

Lovibond, P. F., & Lovibond, S. H. (1995). The structure of negative emotional states:comparison of the depression anxiety stress scales (DASS) with the Beckdepression and anxiety inventories. Behaviour Research and Therapy, 33, 335e343.

Lovibond, S. H., & Lovibond, P. F. (1993). Manual for the depression anxiety stressscales (2nd ed.). Sydney: Psychology Foundation.

McLaughlin, K. A., Mennin, D. S., & Farach, F. J. (2007). The contributory role ofworry in emotion generation and dysregulation in generalized anxiety disorder.Behaviour Research and Therapy, 45, 1735e1752.

Mennin, D. S. (2004). Emotion regulation therapy for generalized anxiety disorder.Clinical Psychology and Psychotherapy, 11, 17e29.

Mennin, D. S. (2005). Emotion and the acceptance-based approaches to the anxietydisorders. In S. Orsillo, & L. Roemer (Eds.), Acceptance and mindfulness-basedapproaches to anxiety: Conceptualization and treatment (pp. 37e68). New York,NJ: Springer.

Mennin, D. S., Heimberg, R. G., Turk, C. L., & Fresco, D. M. (2002). Applying anemotion regulation framework to integrative approaches to generalized anxietydisorder. Clinical Psychology: Science and Practice, 9, 85e90.

Mennin, D. S., Heimberg, R. G., Turk, C. L., & Fresco, D. M. (2005). Preliminaryevidence for an emotion dysregulation model of generalized anxiety disorder.Behaviour Research and Therapy, 43, 1281e1310.

Mennin, D. S., Holaway, R., Fresco, D. M., Moore, M. T., & Heimberg, R. G. (2007).Delineating components of emotion and its dysregulation in anxiety and moodpsychopathology. Behavior Therapy, 38, 284e302.

Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development andvalidation of the Penn State worry questionnaire. Behaviour Research andTherapy, 28, 487e495.

L.A. Stapinski et al. / Behaviour Research and Therapy 48 (2010) 1097e11041104

Newman, M. G., Castonguay, L. G., Borkovec, T. D., & Molnar, C. (2004). Integrativetherapy for generalized anxiety disorder. In R. G. Heimberg, C. L. Turk, &D. S. Mennin (Eds.), Generalized anxiety disorder: Advances in research andpractice (pp. 320e350). New York, NJ: Guildford Press.

Newman, M. G., Zuellig, A. R., Kachin, K. E., Constantino, M. J., Przeworski, A.,Erickson, T., et al. (2002). Preliminary reliability and validity of the generalizedanxiety disorder questionnaire-IV: a revised self-report diagnostic measure ofgeneralized anxiety disorder. Behavior Therapy, 33, 215e233.

Peduzzi, P., Concato, J., Kemper, E., Holford, T. R., & Feinstein, A. R. (1996).A simulation study of the number of events per variable in logistic regressionanalysis. Journal of Clinical Epidemiology, 49, 1373e1379.

Rapee, R. M. (1991). Generalized anxiety disorder: a review of clinical features andtheoretical concepts. Clinical Psychology Review, 11, 419e440.

Rapee, R. M. (1997). Potential role of childrearing practices in the development ofanxiety and depression. Clinical Psychology Review, 17, 47e67.

Rapee, R. M., Abbott, M. J., Baillie, A. J., & Gaston, J. E. (2007). Treatment of socialphobia through pure self-help and therapist-augmented self-help. BritishJournal of Psychiatry, 191, 246e252.

Rapee, R. M., Craske, M. G., Brown, T. A., & Barlow, D. H. (1996). Measurement ofperceived control over anxiety-related events. Behavior Therapy, 27, 279e293.

Roemer, L., Lee, J. K., Salters-Pedneault, K., Erisman, S. M., Orsillo, S. M., &Mennin, D. S. (2009). Mindfulness and emotion regulation difficulties ingeneralized anxiety disorder: preliminary evidence for independent and over-lapping contributions. Behavior Therapy, 40, 142e154.

Roemer, L., & Orsillo, S. M. (2002). Expanding our conceptualization of and treat-ment for generalized anxiety disorder: integrating mindfulness/acceptance-based approaches with existing cognitive-behavioral models. ClinicalPsychology: Science and Practice, 9, 54e68.

Roemer, L., & Orsillo, S. M. (2003). Mindfulness: a promising intervention strategy inneed of further study. Clinical Psychology: Science and Practice, 10, 172e178.

Roemer, L., Salters, K., Raffa, S. D., & Orsillo, S. M. (2005). Fear and avoidance ofinternal experiences in GAD: preliminary tests of a conceptual model. CognitiveTherapy and Research, 29, 71e88.

Ruscio, A. M. (2002). Delimiting the boundaries of generalized anxiety disorder:differentiating high worriers with and without GAD. Journal of Anxiety Disor-ders, 16, 377e400.

Salters-Pedneault, K., Roemer, L., Tull, M. T., Rucker, L., & Mennin, D. S. (2006).Evidence of broad deficits in emotion regulation associated with chronic worryand generalized anxiety disorder. Cognitive Therapy and Research, 30, 469e480.

Sanderson, W. C., & Wetzler, S. (1991). Chronic anxiety and generalized anxietydisorder: issues in comorbidity. In R. M. Rapee, & D. H. Barlow (Eds.), Chronicanxiety: Generalized anxiety disorder and mixed anxiety-depression. New York, NJ:Guildford Press.

Spicer, J. (2004).Making sense of multivariate data analysis. Thousand Oaks, CA: Sage.Tabachnick, B., & Fidell, L. (2001). Using multivariate statistics (4th ed.). New York,

NY: Harper Collins.Turk, C. L., Heimberg, R. G., Luterek, J. A., Mennin, D. S., & Fresco, D. M. (2005).

Emotion dysregulation in generalized anxiety disorder: a comparison withsocial anxiety disorder. Cognitive Therapy and Research, 29, 89e106.

Vrana, S. R., Cuthbert, B. N., & Lang, P. J. (1986). Fear imagery and text processing.Psychophysiology, 23, 247e253.

Wells, A. (1999). A cognitive model of generalized anxiety disorder. BehaviorModification, 23, 526e555.

Wells, A. (2004). A cognitive model of GAD: metacognitions and pathological worry.In R. G. Heimberg, C. L. Turk, & D. S. Mennin (Eds.), Generalized anxiety disorder:Advances in research and practice (pp. 169e186). New York, NJ: Guildford Press.

Wells, A., & Carter, K. (2001). Further tests of a cognitive model of generalizedanxiety disorder: metacognitions and worry in GAD, panic disorder, socialphobia, depression, and nonpatients. Behavior Therapy, 32, 85e102.

Wells, A., & Papageorgiou, C. (1998). Relationships between worry, obsessive-compulsive symptoms and meta-cognitive beliefs. Behaviour Research andTherapy, 36, 899e913.

Williams, K. E., Chambless, D. L., & Ahrens, A. (1997). Are emotions frightening? Anextension of the fear of fear construct. Behaviour Research and Therapy, 35,239e248.

Wittchen, H. U., & Hoyer, J. (2001). Generalized anxiety disorder: nature and course.Journal of Clinical Psychiatry, 62, 15e21.

Wittchen, H., Zhao, S., Kessler, R., & Eaton, W. (1994). DSM-III-R generalized anxietydisorder in the national comorbidity survey. Archives of General Psychiatry, 51,355e364.

Zebb, B. J., & Moore, M. C. (1999). Another look at the psychometric properties of theanxiety control questionnaire. Behaviour Research and Therapy, 37, 1091e1103.