at SciVerse ScienceDirect

Journal of Psychiatric Research 47 (2013) 1462e1468

Contents lists available

Journal of Psychiatric Research

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

An examination of low distress tolerance and life stressors as factorsunderlying obsessions

Richard J. Macatee, Daniel W. Capron, Norman B. Schmidt, Jesse R. Cougle*

Department of Psychology, Florida State University, P.O. Box 3064301, Tallahassee, FL 32306, USA

a r t i c l e i n f o

Article history:Received 6 March 2013Received in revised form25 June 2013Accepted 27 June 2013

Keywords:Emotion regulationOCDObsessionsDistress toleranceLongitudinalStressful events

* Corresponding author. Department of Psychology,W. Call Street, Tallahassee, FL 32306, USA. Tel.: þ1 857739.

E-mail address: cougle@psy.fsu.edu (J.R. Cougle).

0022-3956/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.jpsychires.2013.06.019

a b s t r a c t

A growing body of research has linked poor distress tolerance (DT) to obsessions, but not other OCsymptom domains. However, limited research has been conducted with clinical samples. Further, there isa dearth of research regarding the moderating influence of DT on the contribution of stress to OCsymptoms. In Study 1, we sought to test the specificity of the link between poor DT and greater ob-sessions relative to other OC symptom domains in a clinical sample. In Study 2, we conducted a longi-tudinal investigation with a non-clinical sample examining DT and daily stressors in the prediction ofdaily obsessions. For Study 1, 22 outpatients with an OCD diagnosis and 37 healthy controls completedmeasures of DT, depression, and OC symptoms. For Study 2, 102 undergraduates completed measures ofDT at baseline and daily assessments of OC symptoms and stressors twice weekly for one-month. InStudy 1, OCD diagnosis was not a significant predictor of DT, though greater obsessions, but not other OCsymptoms, were uniquely associated with lower DT. In Study 2, lower baseline DT predicted greater dailyobsessions among those experiencing greater daily negative life events, though this relationship wasabsent among those with elevated DT. The specific association between DT and obsessions was replicatedin a clinical sample. Further, results suggest that low DT increases obsessions in the context of life stress.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

ObsessiveeCompulsive Disorder (OCD) is a heterogeneouscondition characterized by severe impairment (Torres et al., 2006).There is evidence to suggest that the various OCD subtypes differ interms of their maintaining factors and demonstrate distinct pat-terns of neural activation (Mataix-Cols et al., 2004; McKay et al.,2004). Some researchers have argued that more attention needsto be given to the distinct subtypes of OCD if we are to improve ourtreatments for the disorder (Sookman et al., 2005).

Evidence has accumulated linking low distress tolerance tochronic, intrusive unwanted thoughts (i.e., obsessions). Distresstolerance (DT), a facet of emotion regulation, is conceptualized asan individual difference variable reflecting the capacity to experi-ence and tolerate aversive emotional states (Leyro et al., 2010). In aseries of studies using non-clinical samples, Cougle et al. (2011)found lower DT to be concurrently and prospectively predictive ofobsessions, but not other OC symptoms (e.g., washing, checking,ordering), after controlling for negative affect. Furthermore, in a

Florida State University, 11070 645 8729; fax: þ1 850 644

All rights reserved.

separate non-clinical sample, greater obsessions, but not other OCsymptoms, was associatedwith lower DT indexed using in-vivo andbehavioral measures (Cougle et al., in press). In another investiga-tion utilizing a non-clinical sample, DTwas also uniquely associatedwith obsessions, even after controlling for negative affect andpathological worry (Cougle et al., 2012). Only one study known tothe authors measured DT in a clinical OCD sample. Hezel et al.(2012) found individuals with OCD to have lower DT than ahealthy control group, but they did not control for co-occurringdepressive symptoms, which have also been linked to low DT andobsessions (Magidson et al., 2013; Ricciardi and McNally, 1995).Finally, they did not examine different OC symptoms with a vali-dated measure, precluding analyses of the relationship between DTand obsessions specifically.

Indirect evidence for the role of DT in obsessions comes from arecent treatment study for obsessions. The authors found stressmanagement training (SMT) to be as effective as cognitive therapyin the treatment of obsessions, though SMTwas previously found tobe minimally effective for OCD patient groups with diverse symp-tom presentations (e.g., washing, checking, obsessing, ordering)(Whittal et al., 2010). This suggests that SMT’s efficacy may belimited to OCD patients with symptom presentations primarilycharacterized by obsessions (Lindsay et al., 1997; Simpson et al.,2008). The SMT intervention was predicated on the idea that

Table 1Study 1: Descriptives between groups.

VariableOCD(n ¼ 22)M or %

Comparison(n ¼ 37)M or %

Age 28.00 42.46BDI 25.86 7.12DTS 2.78 3.62OCIR Check 7.23 1.08OCIR Obsess 7.48 .89OCIR Order 8.30 1.03OCIR Wash 6.73 .62Gender (Female) 70.0% 40.5%Race (Minority) 26.1% 35.1%Married 26.1% 24.3%MDD Diagnosis 26.1% 0.0%

Note. OCD¼ ObsessiveeCompulsive Disorder. BDI¼ Beck Depression Inventory e II.DTS ¼ Distress Tolerance Scale. OCIR ¼ Obsessive Compulsive Inventory e Revised.

R.J. Macatee et al. / Journal of Psychiatric Research 47 (2013) 1462e1468 1463

obsessions worsen under conditions of negative affect and stress(Horowitz, 1985). SMT teaches patients skills to cope with stressand it is possible that SMT increased patients’ abilities to toleratenegative emotions and reduced obsessions through this pathway,whereas these abilities may not be as relevant to other OC pre-sentations in which overt compulsions are dominant.

There are a number of gaps in the research on DTand obsessionsthat should be addressed. First, the majority of extant studies haveused non-clinical samples. Research is needed to establish a linkbetween poor DT and obsessions in clinical samples. Although onestudy measured DT in a clinical OCD sample, the authors did notcontrol for co-occurring depressive symptoms nor did theyexamine different OC symptoms with a validated measure (Hezelet al., 2012).

Second, because stressors are known to increase unwanted,distressing intrusive thoughts, it is important to examine the po-tential context-sensitivity of DT (Leyro et al., 2010). Specifically, itmay be that low DT contributes to greater obsessions largely in thecontext of life stress. Stress-induced intrusive thoughts may pro-vide the ‘raw material’ that becomes obsessions in individuals withlow DT, whereas individuals more tolerant of distress may be moreresilient to such stress-related intrusions. Prospective studies thatexamine daily obsessions and life stressors are necessary to clarifythe role of DT in the prediction of symptoms.

The following two studies were conducted to address theaforementioned gaps in the literature. Specifically, both studiesaimed to more rigorously test the cross-sectional and prospectiverelationship between obsessions and DT in clinical and non-clinicalsamples after accounting for relevant covariates (i.e., depressivesymptoms, life stressors). In Study 1, we sought to test the speci-ficity of the link between poor DT and greater obsessions relative toother symptom domains in a clinical OCD sample. Additionally, wecontrolled for co-morbid depressive symptoms and used a self-report measure of OC symptoms to examine relations betweenDT and specific OC symptoms. We predicted that after controllingfor co-occurring depressive symptoms, DT would be uniquelyassociated with obsessions but not OCD diagnosis. In Study 2, weconducted a longitudinal investigation in an unselected studentsample examining DT and daily stressful events in the prediction ofdaily obsessions. A non-clinical sample was selected because abroad range of OC symptom scores was needed to test our hy-potheses that would be less likely to be found in a clinical OCDsample (see Huppert et al., 2005). Further, prior research hasdemonstrated that obsessions do occur in nonclinical individuals(Rachman & de Silva, 1978). We sought to clarify the role of DT andits potential interactionwith daily stressful events in the predictionof daily obsessions. Given that the occurrence of stressors has beenlinked to increased intrusive thoughts, it is important to ensure thatthe relationship between DT and obsessions is not accounted for bythe presence of co-occurring stressors. Further, we sought toinvestigate the hypothesis that low DT may predispose individualsto obsessions when they experience more intrusive thoughts (i.e.,under conditions of stress). We predicted that (1) lower DT wouldbe concurrently related to greater obsessions, even after controllingfor number of stressors in the past month, and (2) baseline DTwould interact with daily stressful events such that stressors wouldbe more impactful in creating obsessions for individuals with lowDT relative to those with high DT.

2. Study 1 methods

2.1. Participants

The sample in this study consisted of 22 adult OCD outpatientsreceiving psychological services at an outpatient anxiety clinic and

37 individuals with no diagnoses sampled from the community (seeTable 1). In the OCD group, co-morbidities included major depres-sive disorder (N ¼ 6), social anxiety disorder (N ¼ 8), generalizedanxiety disorder (N ¼ 7), and posttraumatic stress disorder (N ¼ 2).Theno-diagnosis samplewasnot recruited specifically for this study,butwas composed of individuals presenting for various studieswhodid not meet criteria for any Axis I or Axis II diagnoses.

In the present sample, participants’ ages ranged from 18 to 76(M ¼ 36.92, SD ¼ 17.30). The sample was evenly distributed acrossgender (female ¼ 51.7%). All participants agreed to participate inthe IRB-approved research. The procedure consisted of firstcompleting the Structured Clinical Interview for DSM-IV Axis IDisorders (SCID-I) (First et al., 1995) administered by clinical psy-chology graduate students. All graduate students completed aSCID-I training course, and diagnoses were confirmed on a weeklybasis with a Ph.D. level supervisor. Similar training procedures havebeen used in other laboratory projects (Timpano and Schmidt,2013) and resulted in high inter-rater reliability. Primary di-agnoses were determined based upon the individual’s self-reportedreason for seeking treatment as well as the results of the SCID. Afterthe SCID, participants completed self-report measures (Table 2).

2.2. Self-report measures

2.2.1. Obsessive Compulsive Inventory e Revised (OCIR)The OCIR is an 18-item self-report questionnaire of common

OCD symptoms (Foa et al., 2002). It is made up of six subscalesrelated to symptom categories of OCD: obsessions, washing,checking, ordering, neutralizing, and hoarding. Prior work indicatesthat the subscales of the OCIR are valid measures of OCD subtypes(Huppert et al., 2007); thus, subscales were used for that purpose inthe present investigation. The neutralizing and hoarding subscaleswere excluded from analyses because of psychometric limitationsof the neutralizing subscale and evidence suggesting that hoardingbe considered separately from OCD (Wu and Watson, 2003;Huppert et al., 2007; Mataix-Cols et al., 2010; Pertusa et al.,2010). Further, there is some evidence that the hoarding subscale ofthe OCI-R does not have a latent dimensional structure (Olatunjiet al., 2008), whereas other measures of hoarding symptomswere found to have a latent dimensional structure (Timpano et al.,2013), suggesting that the OCI-R hoarding subscale has psycho-metric limitations. In the present sample, the OCIR had excellentinternal consistency across the subscales (a’s ¼ .91e.95).

2.2.2. Distress Tolerance Scale (DTS)The DTS is a 15-item self-report questionnaire designed to

measure individual differences in the ability to experience and

Table 2Study 1: Correlational analyses.

Measure 1 2 3 4 5

1. BDI e

2. DTS �.41* e

3. OCIR Checking .54* �.27* e

4. OCIR Obsessions .74* �.42* .77* e

5. OCIR Ordering .69* �.30* .79* .71* e

6. OCIR Washing .51* �.36* .67* .49* .75*

*p < .05. BDI ¼ Beck Depression Inventory e II. DTS ¼ Distress Tolerance scale.OCIR ¼ Obsessive Compulsive Inventory e Revised.

R.J. Macatee et al. / Journal of Psychiatric Research 47 (2013) 1462e14681464

withstand negative psychological states (Simons and Gaher, 2005).Respondents are asked to indicate how strongly they agree withstatements regarding times when they have felt distressed or upsetbased on a 5-point Likert type scale ranging from (1) strongly agreeto (5) strongly disagree, with lower scores indicating lower DT. Inthe present sample, the DTS had excellent internal consistency(a ¼ .91).

2.2.3. Beck Depression Inventory e II (BDI)The BDI is a widely used 21-item scale of depressive symptoms

(Beck et al., 1988). In the present sample, the BDI had excellentinternal consistency (a ¼ .95).

3. Study 1 results

Results of correlational analyses are presented in Table 2. OCDand control group participants were compared on baseline de-mographics. Analyses indicated significant group differences ongender and age and thus both variables were added as covariates toall analyses. Due to concerns about restricted range and statisticalpower, groups were combined (N ¼ 59) and multiple regressionwas conducted. Examination of DTS scores revealed no violations ofnormality or homoscedasticity of variance assumptions. The BDIwas included as a covariate to control for co-occurring depressivesymptoms. Results revealed a significant effect of sex (b ¼ �.16,p < .01) and BDI (b ¼ �.41, p < .001) on DTS scores, but the effect ofOCD diagnosis was non-significant (b ¼ .06, p ¼ .36).

Next, we sought to examine the unique associations betweenDTS scores and the four OCI-R subscales (see Table 3). Four multipleregression analyses were performed, with age, gender, and BDIincluded in Step 1 and OCIR subscales entered individually in Step2. Consistent with prediction, OCIR obsessions was a significantpredictor of DTS but the other OCIR subscales did not predictunique additional variance in DT.

Finally, we sought to examine the unique associations betweenDTS subscale scores and the OCI-R obsessions subscale. Four mul-tiple regression analyses were performed, with age, gender, BDI,and OCIR obsessions entered as predictors, and a DTS subscale (i.e.,

Table 3Study 1 e OCI-R subscales predicting Distress Tolerance Scale scores.

Step and predictor B SE B b R2 D R2

Step 1: Age �.004 .008 �.06Gender .36 .27 .17BDI �.04 .01 �.44 .20

aStep 2: OCI-R obsessions �.11 .05 �.45 .26 .07*Step 2: OCI-R washing �.04 .04 �.17 .21 .02Step 2: OCI-R checking �.03 .04 �.12 .19 .01Step 2: OCI-R ordering �.01 .05 �.02 .20 .00

*p< .05. BDI¼ Beck Depression Inventory; OCI-R¼ Obsessive Compulsive Inventorye Revised.

a Four separate regressions were conducted with each OCI-R subscale enteredseparately into Step 2.

tolerance, appraisal, regulation, absorption) selected as thedependent variable. None of the subscales violated normality orhomoscedasticity of variance assumptions. OCIR obsessions was asignificant predictor of DTS-Tolerance (b¼�.55, p¼ .014) and DTS-Absorption (b ¼ �.44, p ¼ .051), but did not predict unique addi-tional variance in DTS-Appraisal or DTS-Regulation (p’s > .15).

4. Study 2 methods

4.1. Participants

An unselected student sample provided informed consent toparticipate in this IRB-approved research and then completed thebaseline assessment in a lab. Participants were informed that theywould be receiving an e-mail at 6 PM every Tuesday and Thursdayfor the next four weeks. They were told that the e-mail wouldcontain a link to a questionnaire they were to fill out that night (i.e.,before 6 AM the next morning) in reference to their experiencesthat day. Eighty-eight (86.3%) participants completed all eightquestionnaires over the four week period.

The sample (N ¼ 102) was 72.5% female and ranged in age from18 to 35 years (M ¼ 19.51, SD ¼ 2.69). The sample consisted ofdiverse ethnic groups: Caucasian (62.7%), Hispanic (22.5%), African-American (11.8%), Asian (1.0%), and Other (2.0%).

4.2. Baseline self-report measures

4.2.1. Distress Tolerance Scale (DTS)This scale, described in Study 1, was administered at baseline

and demonstrated excellent internal consistency (a ¼ .91).

4.2.2. Positive and Negative Affect Schedule (PANAS)The PANAS consists of two 10-item mood scales and was

developed to measure positive and negative affect (Watson et al.,1988). Only the negative affect subscale was used in the presentstudy (a ¼ .85). Participants are asked to rate the extent to whichthey generally experience each particular negative emotion withreference to a five-point scale ranging from 1 ¼ ‘very slightly or notat all’ to 5 ¼ ‘very much.’

4.2.3. Negative Life Events Questionnaire (NLEQ)The NLEQ was developed specifically for use with college stu-

dents and contains several categories of negative life events (NLE)(e.g., school, work, family, friends, romantic, etc.) (Saxe andAbramson, 1987). Items were rated on a 0e4 scale (0 ¼ Never,4 ¼ Always) on how frequently they had occurred during the pastfour weeks and summed to form a general negative life eventsscore. The NLEQ has been shown to be reliable and valid (Alloy andClements, 1992).

4.2.4. Obsessive Compulsive Inventory e Revised (OCI-R)The subscales, described in Study 1, showed adequate to good

internal consistencies in the present sample (a’s: Ordering ¼ .81,Obsessing ¼ .85, Washing ¼ .69, Checking ¼ .60) and were used toassess OC symptom dimensions in the present study.

4.3. Daily diary measures

A twice-weekly daily diary questionnaire was created for use inthe present study.

Participants completed the questionnaire online eight separatetimes over a four week period. To avoid overburdening participants,the questionnaire was constructed with brief measures of theconstructs of interest (see below). Participants were instructed torespond to each item with reference to their experiences that day.

Table 4Items from the assessment of daily obsessive compulsive symptoms.

1. I checked things quite a bit.2. I felt compelled to arrange my possessions until it felt “just right.”3. I sometimes had towash or cleanmyself simply because I felt contaminated.4. I found it difficult to control my own thoughts.5. I had to check things (e.g., gas or water taps, doors, etc.) several times.6. I felt compelled to arrange objects so that they were balanced and evenly

spaced.7. I was concerned about contamination (touching dirty things, germs, or

chemicals).8. I was upset by unpleasant thoughts that came into mymind against my will.9. I had to do things several times before I thought they were properly done.10. I worried about germs.11. I spent time straightening and arranging objects in my home.12. I washed my hands quite a bit.13. I got nasty thoughts and had difficulty in getting rid of them.14. I washed and cleaned quite a bit.15. I was concerned about germs and disease.

Note: OC Subscales: Checking ¼ 1, 5, 9; Contamination ¼ 3, 7, 10, 12, 14, 15;Ordering ¼ 2, 6, 11; Obsessing ¼ 4, 8, 13.

Table 5Study 2: Descriptives for assessments of daily obsessive compulsive symptoms andcorrelations with baseline OCD symptom domains (N ¼ 613).

R.J. Macatee et al. / Journal of Psychiatric Research 47 (2013) 1462e1468 1465

4.3.1. OCD symptomsBecause there are no validatedmeasures of daily OCD symptoms

known to the authors, a questionnaire assessing daily OCD symp-toms was constructed. The measure of daily OCD symptoms wascreated with items taken or modified from validated OCD symptommeasures (Foa et al., 2002; Hodgson and Rachman, 1977; Burnset al., 1996; Radomsky and Rachman, 2004; Watson and Wu,2005; Thordarson et al., 2004). Items were taken or modifiedfrom other measures because some OCI-R itemswere not amenableto assessing daily compulsion frequency (e.g., “I get upset if objectsare not arranged properly”). Additionally, items assessingcontamination fear were added to themeasure given that the OCI-Rwashing subscale items that were retained primarily reflectedwashing compulsions. The complete measure was composed of 15items making up four scales assessing ordering, contamination,checking, and obsessing symptoms (see Table 4). Respondents wereasked to indicate the extent to which each symptom was experi-enced throughout the day using a 0 ¼ ‘Not at all’ to 4 ¼ ‘Extremely’scale. The scales showed adequate to good internal consistency inthe present sample (a’s: Ordering ¼ .85, Washing ¼ .86,Checking ¼ .75, Obsessing ¼ .76). To demonstrate validity of thesedaily assessments, between-person variance in daily OC symptomsattributable to baseline OCI-R subscale scores was calculated bytransforming t-values into correlation coefficients (Kashdan andSteger, 2006). As shown in Table 5, each baseline OCI-R subscalecorrelated most strongly with its corresponding daily OC symptommeasure.1

4.3.2. Daily stressors (DISE)Participants completed the Daily Inventory of Stressful Events

(DISE) to assess the occurrence of specific types of stressors(Almeida et al., 2002). Participants responded to seven yes/noquestions regarding various kinds of stressful events that may haveoccurred that day, including arguments, potential arguments thatwere let go to avoid disagreement, work/school stress, home stress,discrimination events, friend/relative stress, and other stress notcaptured by the aforementioned categories. Each ‘yes’ was addedtogether to create a total daily stressors variable (M ¼ 1.15,SD ¼ 1.34). Among all of the diary entries, 55.6% indicated that theyhad experienced at least one stressful event that day.

5. Study 2 results

5.1. Relationships between baseline DT, OC symptoms, and pastmonth NLE

Baseline DT was significantly correlated with obsessions,ordering, checking, and washing symptoms (r ¼ �.51, p < .001;r¼�.31, p< .01; r¼�.23, p< .05; r¼�.21, p< .05). Obsessions andchecking symptoms were significantly related to past month NLE(r ¼ .35, p < .001; r ¼ .24, p < .05), though the correlation betweenNLE and ordering symptoms approached significance (r ¼ .18,p¼ .08). Because DTwas also significantly related to pastmonthNLE(r¼�.45, p< .001), partial correlationswere conducted betweenDTand obsession, ordering, checking, and washing symptoms, con-trolling for past month NLE; obsession and ordering symptomsremained significantly correlated with DT (r ¼ �.42, p < .001 and

1 Given that the OCI-R washing subscale only assesses washing compulsions andour measure of daily contamination symptoms assesses both washing behavior andcontamination fear, we conducted additional validity analyses using the contami-nation subscale of the Vancouver ObsessiveeCompulsive Inventory (Thordarsonet al., 2004), which was also administered at baseline. The VOCI contaminationsubscale was more strongly correlated with our daily measure of contaminationsymptoms (r ¼ .29, p < .01) than was the OCI-R washing subscale (r ¼ .22, p < .05).

r ¼ �.26, p < .01, respectively), whereas associations betweenwashing symptoms (r ¼ �.19, p ¼ .06) and checking symptoms(r ¼ �.14, p ¼ .16) with DT were at a trend or were non-significant.

5.2. Relationships between baseline DT, daily stressors, and dailyobsessions

In order to examine the effects of baseline DTS and dailystressors on daily ratings of obsessions, a hierarchical linear modelwas constructed using HLM 7.0 software (Raudenbush and Bryk,2002). For the present analyses, equations were constructed suchthat Level 1 modeled repeated measurements within individualsand Level 2 modeled baseline differences between individuals. AllLevel 1 predictors were groupmean centered, with the exception ofthe time variable which was entered uncentered. Level 2 predictorswere grand mean centered. Interaction terms were constructedcross-level. Differences in daily obsessions at Level 1 were modeledand then predicted by Level 1 variables, Level 2 variables, andinteraction terms that were added as fixed effects.

An unconditional, random ANOVA model was examined first inorder to partition variance in daily obsessions into Level 1 and Level2. Substantial variability between subjects was found for daily ob-sessions, as the intraclass coefficient was .30, indicating that hier-archical modeling of the datawas necessary (c2¼ 349.99, p< .001).

Next, the time variable was added in order to determinewhether daily obsessions significantly changed throughout thecourse of the study and whether time should be consequentlyincluded in the final model. A KolmogoroveSmirnov test indicatedthat the daily obsessions variable was not normally distributed(D ¼ .473, p < .001), and so robust standard errors were used in allsubsequent analyses (Garson, 2012). Results indicated that the

Measure M SD OCI-Robsess

OCI-Rcheck

OCI-Rwash

OCI-Rorder

1. Daily Obsessions 0.41 1.23 .43*** .34*** .13 .142. Daily Checking 0.59 1.26 .20* .47*** .14 .27*3. Daily Contamination 0.92 2.16 .09 .22* .22* .23*4. Daily Ordering 0.48 1.31 .15 .36*** .16 .33***

Note: OCI-R ¼ Obsessive Compulsive Inventory e Revised.*p < .05; **p < .01; ***p < .001.

Table 6Hierarchical linear model testing DT as a vulnerability factor in the prediction of daily obsessions measured bi-weekly e including daily number of stressful events.

Coefficient SE t-ratio Approximate df p value

Fixed effectsIntercept of Daily Obsessions .653 .124 5.27 97 <.001Baseline variablesNA .028 .017 1.63 97 .11DTS �.007 .009 �.75 97 .46

Daily variablesTime �.053 .019 �2.82 510 <.01Stressors .170 .049 3.50 510 <.001

Interaction effectsStressors*Baseline DTS �.010 .005 �1.95 510 .052

SD Variance component df Chi-square p value

Random effectsBetween-subjects residual .671 .450 97 344.42 <.001Within-subjects residual .995 .991

Note: DTS ¼ Distress Tolerance Scale; NA ¼ Negative Affect.

R.J. Macatee et al. / Journal of Psychiatric Research 47 (2013) 1462e14681466

fixed effect of time was significant (t(512) ¼ �0.069, p < .01). Thus,time was included as a fixed effect in the final model.

To test a model in which baseline DT interacts with dailystressors to predict daily obsessions, Level 1 and Level 2 variableswere added as fixed effects in order to examine predictors of dailyobsessions (see Table 6). Trait negative affect measured at baselinewas also included in the model as a covariate to examine thespecificity of DT to obsessional symptoms. Level 2 DT and traitnegative affect were not found to significantly predict daily ob-sessions (coefficient ¼ �.007, p ¼ .46; coefficient ¼ .028, p ¼ .11),whereas a significant positive relationship between number ofdaily stressors and obsessions was found (coefficient ¼ .17,p < .001). Furthermore, level 2 DT interacted with number of dailystressors to predict obsessions (coefficient ¼ �.01, p ¼ .05).

To test the effect of daily stressors on daily obsessions atdiffering levels of DT, regression coefficients, coefficient variances,and covariances were entered into an online calculator to computesimple slopes of the relationship between daily stressors and dailyobsessions at low (�1 SD) and high (þ1 SD) levels of DT (Preacheret al., 2006). At low levels of DT, there was a significant positiverelationship between daily stressors and daily obsessions (simpleslope ¼ 0.29, p < .001). However, at high levels of DT, there was anon-significant relationship between daily stressors and obsessions(simple slope ¼ 0.05, p ¼ .61) (see Fig. 1).

To test the specificity of the above model to obsessions, alter-native models predicting daily ordering, contamination, andchecking symptoms were constructed. Neither DT nor its interac-tion with stressors was significantly predictive of daily ordering,contamination, or checking symptoms (p’s > .21).

Fig. 1. Testing the interaction between distress tolerance and daily stressors in theprediction of daily obsessions.

5.3. DT subscale analyses

To test models in which specific subscales of the DTS interactwith daily stressors to predict daily obsessions, four modelsidentical to the model shown in Table 6 were constructed, butthe DTS was replaced with a single subscale (i.e., tolerance, regu-lation, appraisal, absorption). Level 2 DTS-Tolerance and DTS-Absorption interacted with stressors to predict obsessions(coefficients ¼ �.04 and �.05 respectively, p’s < .05), whereas DTS-Appraisal and DTS-Regulation did not (p’s > .23). At low levels ofDTS-Tolerance and Absorption, there was a significant positiverelationship between daily stressors and obsessions (both simpleslopes¼ 0.73, p’s< .01), whereas a non-significant relationship wasfound at high levels of DTS-Tolerance and Absorption (simpleslopes ¼ �0.40, p ¼ .09 and .08, respectively).

6. Discussion

Overall, the results of these studies largely support our hy-potheses and replicate and extend prior findings on the relation-ship between DT and obsessions. In Study 1, individuals with anOCD diagnosis did not report significantly lower DT than a healthycontrol group after controlling for co-occurring depressive symp-toms, suggesting that prior findings of lower DT in patients withOCD may have been attributable to depressive symptoms (Hezelet al., 2012). However, greater obsessions (but not other OCsymptoms) were uniquely related to lower DT. These resultsreplicate findings of prior research revealing a specific relationshipbetween DT and obsessions, but they also extend such findings bydemonstrating the obsessions and DT association in a clinicalsample. In Study 2, lower DT was concurrently related to greaterobsessions, even after controlling for past month stressors.

R.J. Macatee et al. / Journal of Psychiatric Research 47 (2013) 1462e1468 1467

Furthermore, low DT at baseline predicted daily obsessions acrosstime, but only under stressful conditions. Low DT did not predictany other daily OC symptoms nor did it interact with stressors topredict other OC symptoms, suggesting that the interaction be-tween low DT and stressors is specific to obsessions. These resultsreplicate findings of prior research revealing a prospective role forDT in the prediction of later obsessions (Cougle et al., 2011), butthey also extend such findings by demonstrating the contextsensitivity of DT, suggesting that lowDT increases obsessions underconditions of heightened stress. Finally, subscale analyses acrossboth studies suggest that intolerance and attentional absorptionare the active facets of DT operative in obsessions.

Including the present study, low DT and obsessions have nowbeen linked in six separate non-clinical samples and one clinicalsample (Cougle et al., 2011, 2012, in press). Importantly, the asso-ciation between DT and obsessions was not accounted for bynegative affect, which is especially noteworthy given the associa-tion between DT and a wide array of psychopathology (Leyro et al.,2010). Furthermore, given that stressors are known to increaseintrusions, it was important to determine whether DT was stillassociated with obsessions after accounting for the occurrence ofstressful events. Results of the present investigation suggest thatlow DT is not redundant with stressors in the prediction of obses-sions, but appears to strengthen the positive association betweenstressful events and obsessions. Finally, subscale findings suggestthat only particular facets of DT are relevant to obsessions, specif-ically items related to the perception of distress as unbearable anddifficulties controlling attention while distressed.

The present investigation has a number of limitations. First,Study 2 used an unselected sample. Because of the nature of thesample, only 15.8% of diary entries contained at least someendorsement of obsessional symptoms. However, despite relativelylow endorsement of daily obsessions, robust effects were stillfound. Nevertheless, future investigations of this kind should useclinical samples. Secondly, the present investigation relied exclu-sively on self-report measures of DT. Future research shouldincorporate behavioral measures of DT to clarify the relationshipbetween the construct (e.g., perceived vs. objective DT), stressors,and obsessions. Thirdly, to our knowledge, the relationship be-tween DT and obsessions has only been investigated by oneresearch group; replication of these findings by other research sitesis needed. Fourthly, though the daily obsessions scale (the primaryoutcome of interest in Study 2) showed sound psychometricproperties, the correlation between the OCI-R washing subscaleand the daily contamination subscale was weak; this may be due tocontent differences between the two measures. Further examina-tion of the psychometric properties of these daily OC symptomassessments is needed. Fifthly, given controversy over the kind ofintrusions captured by the OCI-R obsessions subscale (Huppertet al., 2007), more sophisticated measurement approaches shouldbe utilized to distinguish repugnant obsessions from other un-wanted intrusive thoughts (e.g., worry) (Clark and Purdon, 1995).Lastly, the designs were correlational; studies that experimentallymanipulate DT are necessary to evaluate the causal role of DT inobsessions.

The results of the present investigation suggest important di-rections for future research. Interventions incorporating DT skillsmay bolster the effectiveness of current treatments for patientswith primary obsessions (Whittal et al., 2010). Treatment studieson obsessions should include measures of DT and stressors toexamine their relative contributions to symptom change. Futureresearch on DT, stressors, and obsessions should consider anecological momentary assessment methodology, which may allowfor a better understanding of DT and the temporal dynamics of therelationship between stressors and obsessions.

Role of funding

The authors have no funding sources to declare.

Contributors

The following individuals are contributors on this paper. Allhave contributed substantially and approve of this submittedversion of the paper. All are affiliated with Florida State University:

Daniel W. Capron and Norman B. Schmidt designed study one.Daniel W. Capron conducted the analyses for study one. Daniel W.Capron, Norman B. Schmidt, and Jesse R. Cougle commented uponand edited versions of the present manuscript.

Richard J. Macatee and Jesse R. Cougle designed study two.Richard J. Macatee conducted the analyses for study two and alsowrote the first draft of the present manuscript.

Conflicts of interest

The authors have no conflicts of interest to declare.

Acknowledgments

The authors have no acknowledgements for this paper.

References

Almeida DM, Wethington E, Kessler RC. The daily inventory of stressful events: aninterview-based approach for measuring daily stressors. Assessment 2002;9(1):41e55.

Alloy LB, Clements CM. Illusion of control: Invulnerability to negative affect anddepressive symptoms after laboratory and natural stressors. Journal ofAbnormal Psychology 1992;101:234e45.

Beck AT, Steer RA, Carbin MG. Psychometric properties of the Beck Depression In-ventory: twenty-five years later. Clinical Psychology Review 1988;8:77e100.

Burns GL, Keortge SG, Formea GM, Sternberger LG. Revision of the Padua Inventoryof obsessive compulsive disorder symptoms: distinctions between worry, ob-sessions, and compulsions. Behavior Research and Therapy 1996;34:163e73.

Clark DA, Purdon CL. The assessment of unwanted intrusive thoughts: a review andcritique of the literature. Behavior Research and Therapy 1995;33(8):967e76.

Cougle JR, Timpano KR, Fitch KE, Hawkins KA. Distress tolerance and obsessions: anintegrative analysis. Depression and Anxiety 2011;28:906e14.

Cougle JR, Timpano KR, Goetz AR. Exploring the unique and interactive roles ofdistress tolerance and negative urgency in obsessions. Personality and Indi-vidual Differences 2012;52(4):515e20.

Cougle JR, Timpano KR, Sarawgi S, Smith CM, Fitch KE. A multi-modal investigationof the roles of distress tolerance and emotional reactivity in obsessiveeecompulsive symptoms. Anxiety, Stress, and Coping, in press.

First MB, Gibbon ML, Spitzer RL. Structured clinical interview for DSM-IV-TR (SCID-I): user’s guide and interview, research version. New York: Biometrics ResearchDepartment, New York Psychiatric Institute; 1995.

Foa EB, Huppert JD, Leiberg S, Langner R, Kichic R, Hajcak G, et al. The obsessiveecompulsive inventory: development and validation of a short version. Psy-chological Assessment 2002;14:485e96.

Garson DG. Hierarchical linear modeling: guide and applications. Thousand Oaks,CA: Sage Publications; 2012. p. 66.

Hezel DM, Riemann BC, McNally RJ. Emotional distress and pain tolerance inobsessiveecompulsive disorder. Journal of Behavior Therapy and ExperimentalPsychiatry 2012;43(4):981e7.

Hodgson R, Rachman S. Obsessional compulsive complaints. Behavior Research andTherapy 1977;15:389e95.

Horowitz M. Intrusive and repetitive thoughts after experimental stress. Archives ofGeneral Psychiatry 1985;32:1457e63.

Huppert JD, Moser JS, Gershuny BS, Riggs DS, Spokas M, Filip J, et al. The rela-tionship between obsessiveecompulsive and posttraumatic stress symptomsin clinical and non-clinical samples. Journal of Anxiety Disorders 2005;19(1):127e36.

Huppert JD, Walther MR, Hajcak G, Yadin E, Foa EB, Simpson HB, et al. The OCI-R:validation of the subscales in a clinical sample. Journal of Anxiety Disorders2007;21(3):394e406.

Kashdan TB, Steger MF. Expanding the topography of social anxiety: an experience-sampling assessment of positive emotions, positive events, and emotion sup-pression. Psychological Science 2006;17(2):120e8.

Leyro TM, Zvolensky MJ, Bernstein A. Distress tolerance and psychopathologicalsymptoms and disorders: a review of the literature among adults. PsychologicalBulletin 2010;136:576e600.

R.J. Macatee et al. / Journal of Psychiatric Research 47 (2013) 1462e14681468

Lindsay M, Crino R, Andrews G. Controlled trial of exposure and response pre-vention in obsessiveecompulsive disorder. The British Journal of Psychiatry1997;171:135e9.

Magidson JF, Listhaus AR, Seitz-Brown CJ, Anderson KE, Lindberg B, Wilson A, et al.Rumination mediates the relationship between distress tolerance and depres-sive symptoms among substance users. Cognitive Therapy and Research2013;37:456e65.

Mataix-Cols D, Frost RO, Pertusa A, Clark LA, Saxena S, Leckman JF, et al.Hoarding disorder: a new diagnosis for DSM-V? Depression and Anxiety2010;27:556e72.

Mataix-Cols D, Wooderson S, Lawrence N, Brammer MJ, Speckens A, Phillips ML.Distinct neural correlates of washing, checking, and hoarding symptom di-mensions in obsessiveecompulsive disorder. Archives of General Psychiatry2004;61(6):564e76.

McKay D, Abramowitz JS, Calamari JE, Kyrios M, Radomsky A, Sookman D, et al.A critical evaluation of obsessiveecompulsive disorder subtypes: symptomsversus mechanisms. Clinical Psychology Review 2004;24(3):283e313.

Olatunji BO, Williams BJ, Haslam N, Abramowitz JS, Tolin DF. The latent structure ofobsessiveecompulsive symptoms: a taxometric study. Depression and Anxiety2008;25:956e68.

Pertusa A, Frost RO, Fullana MA, Samuels J, Steketee G, Tolin D, et al. Refining thediagnostic boundaries of compulsive hoarding: a critical review. Clinical Psy-chology Review 2010;30(4):371e86.

Preacher KJ, Curran PJ, Bauer DJ. Computational tools for probing interactions inmultiple linear regression, multilevel modeling, and latent curve analysis.Journal of Educational and Behavioral Statistics 2006;31:437e48.

Rachman S, de Silva P. Abnormal and normal obsessions. Behavior Research andTherapy 1978;16(4):233e48.

Radomsky AS, Rachman S. Symmetry, ordering and arranging compulsive behavior.Behavior Research and Therapy 2004;42(8):893e913.

Raudenbush SW, Bryk AS. Hierarchical linear models: applications and data analysismethods. 2nd ed. Newbury Park, CA: Sage Publications; 2002.

Ricciardi JN, McNally RJ. Depressed mood is related to obsessions, but not tocompulsions, in obsessiveecompulsive disorder. Journal of Anxiety Disorders1995;9(3):249e56.

Saxe LL, Abramson LY. The negative life events questionnaire: reliability and validity1987. Unpublished manuscript.

Simpson HB, Foa EB, Liebowitz MR, Ledley DR, Huppert JD, Cahill S, et al.A randomized, controlled trial of cognitive-behavioral therapy for augmentingpharmacotherapy in obsessiveecompulsive disorder. American Journal of Psy-chiatry 2008;165:621e30.

Simons JS, Gaher RM. The Distress Tolerance Scale: development and validation of aself-report measure. Motivation and Emotion 2005;29:83e102.

Sookman D, Abramowitz JS, Calamari JE, Wilhelm S, McKay D. Subtypes of obses-siveecompulsive disorder: implications for specialized cognitive behaviortherapy. Behavior Therapy 2005;36(4):393e400.

Thordarson DS, Radomsky AS, Rachman S, Shafran R, Sawchuk CN, Hakstian AR. TheVancouver Obsessional Compulsive Inventory (VOCI). Behavior Research andTherapy 2004;42(11):1289e314.

Timpano KR, Broman-Fulks JJ, Glaesmer H, Exner C, Rief W, Olatunji BO, et al.A taxometric exploration of the latent structure of hoarding. PsychologicalAssessment 2013;25(1):194e203.

Timpano KR, Schmidt NB. The relationship between self-control deficits andhoarding: a multimethod investigation across three samples. Journal ofAbnormal Psychology 2013;122(1):13e25.

Torres AR, Prince MJ, Bebbington PE, Bhugra D, Brugha TS, Farrell M, et al. Obses-sive-compulsive disorder: prevalence, comorbidity, impact, and help-seeking inthe British National Psychiatric Morbidity Survey of 2000. American Journal ofPsychiatry 2006;163:1978e85.

Watson D, Clark LA, Tellegen A. Development and validation of brief measures ofpositive and negative affect: the PANAS scales. Journal of Personality and SocialPsychology 1988;54(6):1063e70.

Watson D, Wu KD. Development and validation of the Schedule of Compulsions,Obsessions, and Pathological Impulses (SCOPI). Assessment 2005;12(1):50e65.

Whittal ML, Woody SR, McLean PD, Rachman SJ, Robichaud M. Treatment of ob-sessions: a randomized controlled trial. Behavior Research and Therapy2010;48:295e303.

Wu KD, Watson D. Further investigation of the obsessiveecompulsive inventory:psychometric analysis in two non-clinical samples. Journal of Anxiety Disorders2003;17(3):305e19.