Development of a scrupulosity severity scale using the PennsylvaniaInventory of Scrupulosity-Revised

Leslie J. Shapiro a,n, Jason W. Krompinger a,b, Christina M. Gironda a, Jason A. Elias a,bQ1

a OCD Institute, McLean Hospital, 115 Mill St. Belmont, MA 02478, United StatesQ3 b Harvard Medical School, Boston, MA, United States

a r t i c l e i n f o

Article history:Received 8 March 2013Received in revised form25 July 2013Accepted 1 August 2013

Keywords:ScrupulosityOCDAssessmentFactor analysisReceiver-operator characteristic

a b s t r a c t

Presently, the only clinically valid assessment of scrupulosity, an OCD subtype, is the PennsylvaniaInventory of Scrupulosity-Revised (PIOS-R; (Olatunji, Abramowitz, Williams, Connolly, & Lohr (2007).Journal of Anxiety Disorders, 21(6), 771–787). To date, no study has evaluated the factor structure anddiagnostic utility of this measure in a severe psychiatric sample. A clinical sample of 417 residential OCDpatients with and without primary scrupulosity was assessed using the PIOS-R. A confirmatory factoranalysis revealed that the previously-observed two-factor PIOS-R structure exhibited a good fit withthese data. A receiver-operator characteristic (ROC) analysis indicated that the PIOS-R could reliablyclassify patients with clinically significant scrupulosity among the residential sample, with a score of24 (out of 60) indicating the threshold of scrupulosity severity for which targeted treatment iswarranted. These results indicate that the PIOS-R is a useful and appropriate measure for use inevaluating scrupulosity in patients with severe OCD.

& 2013 Elsevier Ltd. All rights reserved.Q4

1. Introduction

Scrupulosity is an oft-clinically observed but somewhat under-studied obsessive–compulsive disorder (OCD) subtype consisting ofreligious and/or moral obsessions followed by perfectionistic “aton-ing” rituals (i.e. prayer, confession, mentally “fixing” thoughts).Although these rituals may be inspired by religious practice tradi-tional to the patient′s faith, the manner in which they arecompleted results in significant functional impairment. Further,the function of such behaviors becomes less about maintainingthe patient′s faith and connectionwith their God and instead servesonly to reduce an unwanted internal experience, i.e., anxiety orguilt. For example, an individual with scrupulosity might recite aprayer in a repetitive, ritualistic manner whenever a thought orimage deemed to be ‘unacceptable’ enters his mind so as to reducedistress by ostensibly eliminating the likelihood that his thoughtreflects and underlying intent to sin. Studies evaluating the extentto which scrupulosity is related to various dysfunctional cognitivesets confirm the over-importance of thoughts, moral thought-action-fusion, perfectionism, and overinflated sense of responsibil-ity as prominent in those with scrupulosity (Abramowitz, Deacon,Woods, & Tolin, 2004; Tek & Ulug, 2001).

The literature provides numerous clinical (Abramowitz,Huppert, Cohen, Tolin, & Cahill, 2002; Goodman et al., 1989;Greenberg & Huppert, 2010; Inozu, Karanci, & Clark, 2012;Mancini & Gangemi, 2004; Tek & Ulug, 2001; Zohar, Goldman,Calamary, & Mashiah, 2005) and case studies (Hunter, 2000;Weisner & Riffel, 1961), which speaks to the prevalence of thiscondition. There is evidence to suggest that the presence ofreligious obsessions can contribute to “treatment refractory”presentations of OCD (Eğrilmez, Gülseren, Gülseren, & Kültür,1997; Mahgoub & Abdel-Hafeiz, 1991; Tezcan & Millet, 1997).However, due to its subtle nature, scrupulosity is often overlookedin the presence of the more observable OCD symptoms (e.g.contamination/washing, safety and fear of harm/checking, andnot-just-right-experiences/superstitious rituals) and is underap-preciated in terms of its clinical impact. The ability to identifyscrupulosity severity in OCD patients promises to improve treat-ment outcome and may contribute to relapse prevention.

To date, the most well-studied and psychometrically soundmeasure of scrupulosity is the PIOS (Abramowitz, Huppert, Tolin, &Cahill, 2002 Q2; since revised to 15 items; PIOS-R; Olatunji, Abramowitz,Williams, Connolly, & Lohr, 2007). During its initial development,exploratory factor analysis revealed that this measure consists ofitems that load on one of two factors—dubbed by its developers ‘fearof sin’ and ‘fear of God’ (Abramowitz et al., 2002). Evaluation of thepsychometric properties using undergraduate samples revealed thatthe PIOS-R evidenced good convergent and discriminant validity aswell as sound internal stability. This indicated that the PIOS is fit for

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

Contents lists available at ScienceDirect

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

Journal of Obsessive-Compulsive and Related Disorders

2211-3649/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.jocrd.2013.08.001

n Corresponding author. Tel.: +1 6178 552 924; fax: +1 617 855 25 83.E-mail address: ljshapiro@partners.org (L.J. Shapiro).

Please cite this article as: Shapiro, L. J., et al. Development of a scrupulosity severity scale using the Pennsylvania Inventory ofScrupulosity-Revised. Journal of Obsessive-Compulsive and Related Disorders (2013), http://dx.doi.org/10.1016/j.jocrd.2013.08.001i

Journal of Obsessive-Compulsive and Related Disorders ∎ (∎∎∎∎) ∎∎∎–∎∎∎

use as a research tool, and has since exhibited validity in a cross-cultural and religion analysis (Inozu et al., 2012). Along with asses-sing scrupulosity by religion (Abramowitz et al., 2004; Fergus &Valentiner, 2012) the PIOS-R has been used to assess the rolescrupulosity might play in a variety of other mental health issuessuch as terror management (Fergus and Valentiner, 2012), over-active conscience, motivation (Cohen, Hall, Koenig, & Meador, 2005),gender and religious affiliation in older adults, devotion (Cosgrove,2010), mediation of thought-control (Moore & Abramowitz, 2007),and relation between fearfulness, disgust sensitivity and religiousobsessions in a non-clinical sample (Olatunji, Tolin, Huppert, & Lohr,2005).

Given the increasingly evident prevalence of scrupulosity, it isimportant to further evaluate the extent that the putative con-struct validity of the most commonly used tool to evaluate it holdsin OCD populations. To our knowledge, no study has evaluated thePIOS-R in a clinical population. Further, no study has evaluated theextent that this measure may be helpful in identifying patientswith OCD whose scrupulosity merits clinical attention. The currentstudy sought to address this gap in the literature by using thePIOS-R to assess a sample of patients with severe, treatment-refractory OCD currently in a residential treatment facility. Ourgoal was to evaluate the factor structure of the PIOS-R, thendetermine the diagnostic utility of the measure by evaluatingwhether it can accurately identify patients with ‘clinically signifi-cant’ levels of scrupulosity, as determined by attending clinicians.

2. Method

2.1. Participants

The clinical sample comprised 417 patients from the Obsessive–CompulsiveDisorder Institute (OCDI) at McLean Hospital in Belmont, MA, with treatment-resistant OCD requiring partial or residential-hospital level of care. Patients in thissample received their OCD and scrupulosity diagnoses via extensive structuredinterview, behavioral assessments by their treatment team consisting of a psychia-trist, behavioral therapist, and social worker. Such diagnoses were typicallyconvergent with diagnoses made by the patients’ outpatient providers prior toadmission to the Institute. Because this study sought to evaluate the factorstructure and diagnostic utility of the PIOS-R in a naturalistic sample, no patientswere excluded on the basis of Axis I or Axis II comorbidity.

The OCDI offers treatment groups that are tailored towards addressing specificsymptom clusters for subsets of the census that are suffering with those particularsymptoms. The ‘scrupulosity group’ (run by author LS) is an example of sucha group. This group focuses on numerous topics, including identifying thedifferences between healthy faith and life-interfering scrupulous symptoms, anddiscussing the rationale for and methods to approach implementing exposures. Forthe purposes of this study, our clinical sample was separated into ‘scrupulous’(N¼95) versus ‘non-scrupulous’ (N¼322) OCD on the basis of assignment to thescrupulosity group by their behavior therapist. Behavioral therapists madea determination to assign a patient to the scrupulosity group based on a semi-structured clinical interview performed at intake. Several factors played a role inmaking the decision to assign to the scrupulosity group, most notably: whenpatients’ obsessions and/or rituals carried religious themes, e.g., fears of eternaldamnation, praying rituals; when it was apparent that guilt, or avoidance thereof,was a significant factor in motivating OCD behavior; or when patients’ approachesto religious activity was perfectionistic in nature and resulted in significantfunctional impairment. Additionally, because author LS served as the group leader,we were in a position to confirm whether the group assignment determined by thebehavior therapist was appropriate by way of repeated assessments with the PIOS-R in-group. Table 1 describes the characteristics of each sample.

2.2. Measures

PIOS-R. The PIOS-R is a 15 item self-report questionnaire designed to assess thedegree of scrupulosity-related symptomatolgy in the respondent with one subscalemeasuring “fear of sin” and another measuring “fear of God” on a 5-point scaleranging from 0 (never) to 4 (constantly). Subjects indicate their (non)religiousaffiliation as well as the strength of religious belief on a scale from 1 (not at alldevoted) to 5 (very strongly devoted). There have been numerous previous studiesthat reflect the generally excellent psychometric properties of this measure(Abramowitz et al., 2002; Nelson, Abramowitz, Whiteside, & Deacon, 2006;

Olatunji et al., 2007). Exploratory factor analyses of this measure using primarilyundergraduate samples have yielded a two-pronged factor structure: ‘fear of sin (10items),’ and ‘fear of God (5 items).’

In their analysis of the PIOS-R using an unselected college student population,Olatunji and colleagues found that the measure exhibited excellent internalconsistency (α¼0.943, inter-item correlations 0.27–0.78). Further, a confirmatoryfactor analysis revealed that the two-factor solution exhibited a good fit to theirdata (RMSEA¼0.06, CFI¼0.97, SRMR¼0.05; Olatunji et al., 2007). Finally, this studyshowed that the PIOS-R was highly correlated with OC symptoms, state anxiety,trait anxiety, negative affect, disgust sensitivity, and specific fears. Results fromtheir analysis also revealed that the relationship between OCD and scrupulosity, asmeasured by the PIOS-R, is not merely epiphenomena of trait anxiety or negativeaffect (Olatunji et al., 2007). This study found that the total score and the twofactors were moderately to strongly associated with theoretically relevant con-structs (e.g., anxiety symptoms, negative affect) and weakly related to constructsthat may be considered theoretically distinct from scrupulosity (e.g., positive affect;Olatunji et al., 2007).

Yale-Brown Obsessive Compulsive Scale (Y-BOCS, (Goodman et al., 1989). TheY-BOCS is a ten item measure that assesses the severity of OCD symptoms byevaluating time spent managing, distress and functional impairment caused by,and unsuccessful efforts to eliminate obsessions and compulsions. The Y-BOCS isthe most commonly used measure for the assessment of OCD symptoms (e.g.,Storch et al., 2010 and many others).

2.3. Procedure

Patients were administered the PIOS-R upon admission to the Institute as partof a standard self-report assessment battery. Although the current study evaluatesonly PIOS-R scores at admission, the PIOS-R was also administered to all patientsmonthly, and at discharge.

2.4. Data analysis strategy

The current study has two primary objectives: to determine the extent that thefactor structure of the PIOS-R holds in a naturalistic sample and to evaluate theextent that the PIOS-R can identify individuals struggling with clinically significantlevels of scrupulosity.

We used LISREL (version 8.80 student) to perform a confirmatory factoranalysis on the full clinical sample, the scrupulous patients, and the non-scrupulous patients using maximum likelihood estimates. The two factor modelwas estimated by loading the ten fear of sin items and five fear of God itemsa priori. When evaluating the full sample and the non-scrupulous patients, weevaluated goodness of fit of the two-factor structure using the standardized rootmean square residual (SRMR), root-mean-square error of approximation (RMSEA),and the comparative fit index (CFI), in accordance with recommendations by Huand Bentler (1999). Good model fit was defined by the following criteria:RMSEAo¼ .06; SRMRo¼ .08; and CFI4¼ .95 (Hu & Bentler, 1999). The use ofmultiple indices provides a conservative and reliable evaluation of model fitrelative to the use of a single-fit index. When evaluating the scrupulous patients,we used only the SRMR and CFI, given that the RMSEA tends to reject the model toooften yet the SRMR and CFI are understood to performwell when the sample size issmall (No250; Hu & Bentler, 1999; Tabachnick & Fidell, 2007; Hooper, Coughlan,& Mullen, 2008). In order to evaluate the diagnostic utility of the PIOS-R, weutilized a receiver-operator characteristic (ROC) analysis. ROC analysis uses theassociation between sensitivity (1—false negative rate) and specificity (1—falsepositive rate) to estimate the area under the curve (AUC) to indicate how wellscores on a measure distinguish between positive (in this case, a ‘diagnosis’ ofclinically significant scrupulosity) and negative (no significant scrupulosity) casesbased on the diagnosis by expert clinicians. A value of 1.0 indicates perfect

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

Table 1Participant characteristics.

Variable Scrupulous sample Non-scrupulous sample

N % PIOS-RM SD n % PIOS-RM SD

Sample size 95 22.8 322 77.2Gender (male) 61 64.2 161 50Race (white) 84 88.4 299 92.9Religious affiliationCatholic 38 40 37.54 17.6 116 36 17.5 13.25Jewish 6 6.3 25.20 22.29 55 17.1 14.60 11.31Protestant 14 14.7 36.75 14.7 39 12.1 16.7 14.8All other religions 37 39 112 34.8Age (years) 34.1 13.2 35.1 14.8Y-BOCS total score 25.9 6.5 25.5 6.8PIOS-R total score 34.4 16.6 17.1 14.2

L.J. Shapiro et al. / Journal of Obsessive-Compulsive and Related Disorders ∎ (∎∎∎∎) ∎∎∎–∎∎∎2

Please cite this article as: Shapiro, L. J., et al. Development of a scrupulosity severity scale using the Pennsylvania Inventory ofScrupulosity-Revised. Journal of Obsessive-Compulsive and Related Disorders (2013), http://dx.doi.org/10.1016/j.jocrd.2013.08.001i

diagnostic prediction, whereas a value of.50 indicates that predictive utility is nobetter than chance. We evaluated the accuracy of the PIOS-R in distinguishing bothscrupulous from non-scrupulous patients. Second, we sought to use the ROCanalysis to establish a PIOS-R cutoff score that would reflect optimal diagnosticaccuracy to identify those with clinically significant scrupulosity.

3. Results

3.1. Internal consistency of the PIOS-R

In our evaluation the whole sample (N¼417), the PIOS-Rexhibited excellent internal consistency, α¼0.95, inter-item corre-lations ranging from 0.29 to 0.82. When taken separately, the Fearof Sin subscale, α¼0.95, inter-item correlations ranging from .53 to.85, and the Fear of God subscale, α¼0.86, inter-item correlationsranging from .42 to .85, each also exhibited good to excellentinternal consistency.

3.2. Confirmatory factor analysis—One factor solution

We initially performed three separate confirmatory factor analyseson our data in order to evaluate the extent that they would fit aone factor solution. The first CFA was on the data gleaned from theentire sample (N¼417). Goodness of fit indices generally indicated amoderate to substandard fit, χ2 (90)¼878.5, po.001; RMSEA¼0.17(90% CI¼0.16, 0.18); CFI¼0.95; SRMR¼0.07). Next, we tested thesingle factor model using the data gleaned only from the ‘scrupulous’OCD patients (N¼95). Again, fit indices reflected a modest fit,χ2 (90)¼284.8, po.001; CFI¼0.93; SRMR¼0.08. Finally, we tested thesingle factor model using data gleaned only from the ‘non-scrupulous’OCD patients (N¼322). Once again, fit indices were unimpressive,

χ2 (90)¼841.8, po.001; RMSEA¼0.19 (90% CI¼0.18, 0.20); CFI¼0.92,SRMR¼0.09).

3.3. Confirmatory factor analysis—Two factor solution

We then performed three separate confirmatory factor analysesinvestigating the extent that our hypothesized two-factor solutionfit the data. The first CFA (Fig. 1; N¼417) evaluated the extent thatthe two factor model fit the data gleaned from the entire clinicalsample. Aside from RMSEA, goodness of fit indices indicated agenerally good fit for the two factor model, and were markedlyimproved over the one-factor solution, χ2 (89)¼368.4, po .001;RMSEA¼0.09 (90% CI¼0.08, 0.10); CFI¼0.98; SRMR¼ .04. Next, wetested the two factor model using PIOS-R data gleaned from the‘scrupulous’ OCD patients. Fit indices again indicated a generallygood fit and improvement over the one-factor solution, and arehighly similar to those of Fig. 1, χ2 (89)¼172.3, po .001; CFI¼0.97;SRMR¼ .06. Finally, we evaluated the two factor model using datafrom the ‘non-scrupulous’ clinical sample. Once again, fit indicessuggested a good fit and improvement over the one-factor solu-tion, aside from RMSEA, χ2 (89)¼360.7, po .001; RMSEA¼0.10(90% CI¼0.06, 0.11); CFI¼0.97; SRMR¼ .05. These results indicatethat the two-factor structure of the PIOS-R generally holds in ournaturalistic clinical sample.

3.4. Receiver operator characteristic analysis

We conducted ROC analyses for the PIOS-R total and subscalescores to determine which best distinguished patients with clinicallysignificant scrupulosity from those with clinically significant OCD but

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566 Fig. 1. CFA for Whole Clinical Sample. CFA¼Confirmatory factor analysis; N¼417; Chi square¼403.22; df¼89; p-value¼0.00000; RMSEA¼0.092.

L.J. Shapiro et al. / Journal of Obsessive-Compulsive and Related Disorders ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

Please cite this article as: Shapiro, L. J., et al. Development of a scrupulosity severity scale using the Pennsylvania Inventory ofScrupulosity-Revised. Journal of Obsessive-Compulsive and Related Disorders (2013), http://dx.doi.org/10.1016/j.jocrd.2013.08.001i

without scrupulosity-related symptoms. AUC estimates for the totalPIOS-R plus the two subscales were .78 (PIOS-R Total), .78 (PIOS-RSin), and .76 (PIOS-R God; Fig. 2). These data indicate that the PIOS-Rtotal score, in particular, discriminates individuals with clinicallysignificant scrupulosity quite well from others with OCD. Accord-ingly, we elected to use total scores to examine the diagnosticaccuracy of the PIOS-R.

Next, we examined the accuracy of different PIOS-R cutoffscores in correctly classifying patients as having clinically signifi-cant scrupulosity. We evaluated diagnostic accuracy by calculatingthe sensitivity and specificity of various PIOS-R total scores. Acutoff score of 24 provided the best balance between sensitivityand specificity. This cutoff score correctly classified about 73% ofclinically significant scrupulosity patients (sensitivity) and 71% ofother OCD patients (specificity; Table 2).

4. Discussion

Scrupulosity appears to be an under recognized OCD subtypeand, as a result, may account for untreated residual symptoms thatmay lead to relapses. Our CFA shows that a two-factor solution forthe PIOS-R demonstrates adequate fit in a sample of patients withclinically significant OCD. This indicates that the PIOS-R appears tobe measuring the same latent variables in our clinical sample ashave been measured in non-clinical samples. The results of ourROC analysis indicate that the PIOS-R is suitable for identifying

clinically significant scrupulosity in this sample. Scores above 23seem to indicate meaningful interference from scrupulosity factorsthat may warrant considerable attention in treatment.

The PIOS-R is a reliable and valid assessment tool for fleshing outreligious/moral obsessions that may be concurrent with the moreobvious OCD symptoms. Increased use of the PIOS-R with OCDpatients may afford clinicians a more thorough assessment ofoverall symptomatology. For example, the common obsessive fearof contaminating others may have a scrupulosity underpinningwhen the sufferer fears that compliance with response preventionwill result in retribution from God. Quantifying a clinically mean-ingful scrupulosity score will help clinicians identify a differential inOCD symptoms, particularly in patients that were ostensibly treat-ment refractory. A scrupulosity-focused approach for such identi-fied patients may, then, improve treatment outcome. Future studiesshould evaluate the extent that scrupulosity, as measured by thePIOS-R, is related to treatment outcome across all OCD subtypes.

Further, developing a more fine-grained understanding ofsymptom severity level, as reflected by PIOS-R scores, is necessaryin order to further inform the extent to which scrupulosity may bethe primary OCD symptom for given patients, and enable treat-ment to be conducted accordingly. Additionally, guilt related tounderlying “religious” and/or “moral” symptoms is also considereda neglected treatment factor and that warrants further study(Mancini & Gangemi, 2004; Shapiro & Stewart, 2010).

There are some limitations to the current study that should beconsidered when interpreting these results. First, there is an issueregarding the ‘purity’ of our OCD sample. The naturalistic sampleused in the study was likely diagnostically heterogeneous, as istypically the case in the more severe cases of OCD (Pallanti, Grassi,Sarrecchia, Canisani, & Pellegrini, 2011). Because we did notconduct formal individual structured diagnostic assessments oneach patient in our sample, there exists the possibility that whatwe measured and interpreted to be symptoms of scrupulosity arein fact clinical manifestations of a comorbid diagnosis, such asmajor depressive disorder. Relatedly, we also did not collect dataon current medication regimens among our participants. A futurestudy would potentially benefit from including comorbid diag-noses and medication status as covariates at the model specifica-tion stage prior to running factor analyses. On the other hand,a primary goal of the current study was to evaluate the PIOS-R ina severe psychiatric sample. Controlling for these variables mayinterfere with that goal insofar as severe samples are, by defini-tion, largely heterogeneous. Second, the sample of patients withclinically significant scrupulosity in the current study was rela-tively small (N¼95), limiting to some extent the interpretationsgleaned from both the CFA and ROC analyses.

Additionally, there are likely variations in degree of scrupulos-ity as a function of religious group. The current study was limitedto only four identified groups (Protestant, Catholic, Jewish, and‘other.’). A replication and extension of the current study wouldbenefit from utilizing a larger patient population that includesrepresentative samples from a wider variety of religious groups.Finally, clinically significant scrupulosity was determined by wayof an assessment performed by a single clinician at intake. Futurestudies would benefit from a more comprehensive multimethod/multimodal assessment of scrupulosity symptoms in order tobetter establish a ‘gold standard’ to best accommodate the ROCanalysis.

References

Abramowitz, J., Deacon, B., Woods, C., & Tolin, D. (2004). Association betweenProtestant religiosity and obsessive–compulsive symptoms and cognitions.Depress Anxiety, 20(2), 70–76.

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

Fig. 2. ROC Analysis for comparing scrupulous and non-scrupulous samples. ROCanalysis¼receiver-operator characteristic analysis; AUC¼area under the curve.PIOS-R total: AUC¼0.780, Cutoff¼23.5; Sensitivity¼0.74; Specificity¼0.71. PIOS-RSin subtotal: AUC¼0.759. PIOS-R God subtotal: AUC¼0.778.

Table 2ROC Analysis for comparing scrupulous and non-scrupulous samples.

Measure AUC Cutoff Sensitivity Specificity

PIOS-R Total .78 23.5 .73 .71PIOS-R Sin .76PIOS-R God .78

Note. ROC analysis¼receiver-operator characteristic analysis; AUC¼area underthe curve.

L.J. Shapiro et al. / Journal of Obsessive-Compulsive and Related Disorders ∎ (∎∎∎∎) ∎∎∎–∎∎∎4

Please cite this article as: Shapiro, L. J., et al. Development of a scrupulosity severity scale using the Pennsylvania Inventory ofScrupulosity-Revised. Journal of Obsessive-Compulsive and Related Disorders (2013), http://dx.doi.org/10.1016/j.jocrd.2013.08.001i

Abramowitz, J., Huppert, J., Cohen, A., Tolin, D., & Cahill, S. (2002). Religiousobsessions and compulsions in a non-clinical sample: The Penn Inventory ofScrupulosity (PIOS). Behaviour Research and Therapy, 40(7), 825–838.

Cohen, A. B., Hall, D. E., Koenig, H. G., & Meador, K. G. (2005). Social versusindividual motivation: Implications for normative definitions of religiousorientation. Personality and Social Psychology Review, 9(1), 48–61.

Cosgrove, E. (2010). Religious devotion: a risk factor for mental illness? Examiningthe link between religiosity and obsessive–compulsive disorder.

Eğrilmez, A., Gülseren, L., Gülseren, S., & Kültür, S. (1997). Phenomenology ofobsessions in a Turkish series of OCD patients. Psychopathology, 30(2), 106–110.

Fergus, T. A., & Valentiner, D. P. (2012). Terror management theory and scrupulos-ity: An experimental investigation. Journal of Obsessive–Compulsive and RelatedDisorders, 1(2), 104–111.

Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleishman, R. L., Hill, C. L.,Heninger, G. R., & Charney, D. S. (1989). The Yale-Brown Obsessive CompulsiveScale. Archives of General Psychiatry, 46, 1012–1016.

Greenberg, D., & Huppert, J. D. (2010). Scrupulosity: A unique subtype of obsessive–compulsive disorder. Current Psychiatry Reports, 12, 282–289.

Hooper, D., Coughlan, J., & Mullen, M. R. (2008). Structural equation modelling:Guidelines for determining model fit. Electronic Journal of Business ResearchMethods, 6(1), 53–60.

Hu, L.-t., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structureanalysis: Conventional criteria versus new alternatives. Structural EquationModeling: A Multidisciplinary Journal, 6(1), 1–55.

Hunter, M. C. W. (2000). Robert Boyle, 1627-91 : Scrupulosity and science. Wood-bridge England ; Rochester, NY: Boydell Press.

Inozu, M., Karanci, A. N., & Clark, D. A. (2012). Why are religious individuals moreobsessional? The role of mental control beliefs and guilt in Muslims andChristians. Journal of Behavior Therapy and Experimental Psychiatry, 43(3),959–966.

Mahgoub, O., & Abdel-Hafeiz, H. (1991). Pattern of obsessive–compulsive disorderin eastern Saudi Arabia. British Journal of Psychiatry, 158, 840–842.

Mancini, F., & Gangemi, A. (2004). Fear of guilt from behaving irresponsibly inobsessive–compulsive disorder. Journal of Behavior Therapy and ExperimentalPsychiatry, 35(2), 109–120.

Moore, E. L., & Abramowitz, J. S. (2007). The cognitive mediation of thought-controlstrategies. Behaviour Research and Therapy, 45(8), 1949–1955.

Nelson, E., Abramowitz, J., Whiteside, S., & Deacon, B. (2006). Scrupulosity inpatients with obsessive–compulsive disorder: Relationship to clinical andcognitive phenomena. Journal of Anxiety Disorders, 20(8), 1071–1086.

Olatunji, B., Abramowitz, J., Williams, N., Connolly, K., & Lohr, J. (2007). Scrupulosityand obsessive–compulsive symptoms: Confirmatory factor analysis and validityof the Penn Inventory of Scrupulosity. Journal of Anxiety Disorders, 21(6),771–787.

Olatunji, B., Tolin, D. F., Huppert, J. D., & Lohr, J. M. (2005). The relation betweenfearfulness, disgust sensitivity and religious obsessions in a non-clinicalsample. Personality and Individual Differences, 38(4), 891–902.

Pallanti, S., Grassi, G., Sarrecchia, E., Canisani, A., & Pellegrini, M. (2011). Obsessive–compulsive disorder comorbidity: Clinical assessment and therapeutic implica-tions. Frontiers in Psychiatry, 2(70), 1–11.

Shapiro, L., & Stewart, E. (2010). Pathological guilt: A persistent yet overlookedtreatment factor in obsessive–compulsive disorder. Annals of Clinical Psychiatry,23(1), 63–70.

Storch, E. A., Rasmussen, S. A., Price, L. H., Larson, M. J., Murphy, T. K., & Goodman,W. K. (2010). Development and psychometric evaluation of the Yale—BrownObsessive–Compulsive Scale—2nd ed.. Psychological Assessment, 22(2),223–232. http://dx.doi.org/10.1037/a0018492.

Tabachnick, B. G., & Fidell, L. S. (2007). Using Multivariate Statistics (5th ed.). Boston:Pearson.

Tek, C., & Ulug, B. (2001). Religiosity and religious obsessions in obsessive–compulsive disorder. Psychiatry Research, 104(2), 99–108.

Tezcan, E., & Millet, B. (1997). Phenomenology of obsessive–compulsive disorders.Forms and characteristics of obsessions and compulsions in East Turkey.Encephale, 23(5), 342–350.

Weisner, W., & Riffel, P. (1961). Scrupulosity: Religion and obsessive–compulsivebehavior in children. Current Psychiatry and Therapy, 1, 39–42.

Zohar, A., Goldman, E., Calamary, R., & Mashiah, M. (2005). Religiosity andobsessive–compulsive behavior in Israeli Jews. Behaviour Research and Therapy,43(7), 857–868.

123456789

10111213141516171819202122232425262728

L.J. Shapiro et al. / Journal of Obsessive-Compulsive and Related Disorders ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5

Please cite this article as: Shapiro, L. J., et al. Development of a scrupulosity severity scale using the Pennsylvania Inventory ofScrupulosity-Revised. Journal of Obsessive-Compulsive and Related Disorders (2013), http://dx.doi.org/10.1016/j.jocrd.2013.08.001i