Screening for Obsessive and Compulsive Symptoms: Validation of the Clark-Beck Obsessive-Compulsive Inventory

Screening for Obsessive and Compulsive Symptoms: Validation of the Clark-Beck Obsessive-Compulsive Inventory
[Article]
Clark, David A.1,6; Antony, Martin M.2,3; Beck, Aaron T.4; Swinson, Richard P.2,3; Steer, Robert A.5
1Department of Psychology, University of New Brunswick, Fredericton, New Brunswick, Canada
2Anxiety Treatment and Research Centre, St. Joseph's Healthcare, Hamilton, Ontario, Canada
3Department of Psychiatry and Behavioural Neurosciences, McMaster University, Hamilton, Ontario, Canada
4Department of Psychiatry, University of Pennsylvania
5Department of Psychiatry, School of Osteopathic Medicine, University of Medicine and Dentistry of New Jersey.
6Correspondence concerning this article should be addressed to David A. Clark, Department of Psychology, University of New Brunswick, Bag Service #45444, Fredericton, New Brunswick, E3B 6E4, Canada. E-mail: clark@unb.ca
This research was supported by a grant from the Foundation of Cognitive Therapy and Research awarded to David A. Clark. We thank Cory Newman, Gregory Brown, and Margaret Richter for providing access to some of the data in the validation study. Michael Kyrios and Thomas Unger assisted in data collection for the pilot study. We are grateful to Andrea Liss, Laura Rocca, Noam Lindenboim, Adrienne Wang, Marvin Claybourn, and Nick Lowther for their assistance with data collection and entry. Gratitude is expressed to Darcy Santor, who ran the item response analyses. A copy of the Clark-Beck Obsessive-Compulsive Inventory and manual can be purchased from the Psychological Corporation, 55 Academic Court, San Antonio, TX 78204-2498; Web site: www.psychcorp.com .
David A. Clark and Aaron T. Beck are authors of the Clark-Beck Obsessive-Compulsive Inventory, which is published by the Psychological Corporation.
Received Date: March 3, 2003; Revised Date: October 19, 2004; Accepted Date: February 8, 2005
Abstract
The 25-item Clark-Beck Obsessive-Compulsive Inventory (CBOCI) was developed to assess the frequency and severity of obsessive and compulsive symptoms. The measure uses a graded-response format to assess core symptom features of obsessive-compulsive disorder (OCD) based on Diagnostic and Statistical Manual of Mental Disorders (4th ed.; American Psychiatric Association, 1994) criteria and current cognitive-behavioral formulations. Revisions were made to the CBOCI on the basis of psychometric and item analyses of an initial pilot study of clinical and nonclinical participants. The construct validity of the revised CBOCI was supported in a subsequent validation study involving OCD, nonobsessional clinical, and nonclinical samples. A principal-factor analysis of the 25 items found 2 highly correlated factors of Obsessions and Compulsions. OCD patients scored significantly higher on the measure than nonobsessional anxious, depressed, and nonclinical samples. The questionnaire had strong convergent validity with other OCD symptom measures but more modest discriminant validity.
________________________________________


Obsessive-compulsive disorder (OCD) is a primary anxiety disorder characterized by persistent, recurrent, and uncontrolled obsessions and compulsions that are time consuming or that cause marked distress or impairment in daily functioning (American Psychiatric Association, 1994). Obsessions are intrusive, recurrent, and persistent unwanted thoughts, images, or impulses that are experienced as unacceptable, upsetting, and uncontrollable, and they usually give rise to subjective resistance (Rachman & Hodgson, 1980). Compulsions, on the other hand, are repetitive, intentional behavioral or mental responses that are subjectively experienced as an urge to act, are performed according to certain rules or in a stereotypic fashion, and are intended to reduce anxiety or prevent the anticipated negative consequences associated with an obsession (Hollander & Wong, 2000; Rachman & Shafran, 1998).
OCD has a lifetime prevalence of 1% to 2% in the general population (Antony, Downie, & Swinson, 1998). Moreover elevated rates of obsessive-compulsive (OC) symptoms and subclinical OCD can be found in the general population as well as in other clinical disorders such as depression, generalized anxiety disorder, tic disorders, and OC spectrum disorders (see reviews by D. A. Clark, 2004; Gibbs, 1996; Hollander & Wong, 2000; Leckman, 1993; O'Connor, 2001). There is also a high comorbid rate of depression and other anxiety disorders in OCD (Brown, Campbell, Lehman, Grisham, & Mancill, 2001; Crino & Andrews, 1996) that complicates the detection and measurement of symptoms. Given the broad and variable distribution of obsessive and compulsive symptoms in both clinical and nonclinical populations, a brief symptom-screening instrument would be a useful clinical tool for providing the initial detection of OC symptoms.
At present, various instruments are available to assist in the diagnosis and measurement of OCD. The most widely accepted measure is the Yale-Brown Obsessive-Compulsive Inventory (YBOCS), which consists of a 64-item checklist of past and current obsessions and compulsions followed by a 10-item clinician rating scale that assesses the severity of obsessions and compulsions independent of the content or number of symptoms (Goodman, Price, Rasmussen, Mazure, Delgado, et al., 1989; Goodman, Price, Rasmussen, Mazure, Fleischmann, et al., 1989). The YBOCS has high interrater reliability for the 10 core severity items (Goodman, Price, Rasmussen, Mazure, Fleischmann, et al., 1989; Nakagawa, Marks, Takei, De Araujo, & Ito, 1996; Woody, Steketee, & Chambless, 1995), although support for the internal consistency of the YBOCS Obsessions and Compulsions subscales has been mixed (Amir, Foa, & Coles, 1997; Goodman, Price, Rasmussen, Mazure, Fleischmann, et al., 1989; Richter, Cox, & Direnfeld, 1994; Steketee, Frost, & Bogart, 1996; Woody et al., 1995). It has good convergent validity and sensitivity to treatment effects, but discriminant validity may be weaker (see reviews by Antony, 2001; Feske & Chambless, 2000; Taylor, 1995, 1998; Taylor, Thordarson, & Söchting, 2002).
Although the YBOCS is considered the “gold standard” for measuring OC symptoms (Steketee, 1994), it has a number of significant limitations. The instrument is time-consuming to administer and requires trained interviewers (Taylor, 1995). The factorial validity of the YBOCS has not been clearly supported (Amir et al., 1997), it has weak discriminant validity as evidenced by significant correlations with depression measures, and at times its correlations with other standardized self-report OCD measures have been unexpectedly low (e.g., Goodman, Price, Rasmussen, Mazure, Delgado, et al., 1989; Obsessive Compulsive Cognitions Working Group, 2003; Woody et al., 1995). For these reasons, other measures of OC symptoms are needed to supplement the YBOCS.
A number of self-report symptom measures have also been developed to assess the frequency and severity of obsessions and compulsions (for a comprehensive review, see Antony, 2001; D. A. Clark, 2004; Feske & Chambless, 2000; Steketee, 1993, 1994; Taylor, 1995, 1998; Taylor et al., 2002). Three of these instruments are of particular relevance to the development of the current measure.
Hodgson and Rachman (1977) constructed the 30-item Maudsley Obsessional Compulsive Inventory (MOCI) to assess the presence of different types of obsessive and compulsive complaints. The MOCI is actually a symptom checklist (Taylor, 1998) consisting of total score and four factorially derived subscales: Checking, Washing, Slowness/Repetition, and Doubting/Conscientiousness. It has strong convergent and factorial validity (e.g., Emmelkamp, Kraaijkamp, & van den Hout, 1999). Positive findings have supported the discriminant and criterion-related validity of the MOCI (see reviews by Feske & Chambless, 2000; Taylor, 1998). The questionnaire, however, has several serious shortcomings including (a) some items that do not directly assess OC symptoms, (b) a dichotomous response format that limits its ability to quantify symptom severity, and (c) an overrepresentation of washing and checking compulsions coupled with too few items on obsessional rumination (Taylor, 1995). A 52-item revision of the MOCI was undertaken to address many of these criticisms (Thordarson et al., 2004).
Another self-report questionnaire that is frequently used to measure obsessive and compulsive symptoms is the 60-item Padua Inventory (PI) developed by Sanavio (1988). The questionnaire consists of four factorially derived subscales and a total score: Impaired Control Over Mental Activities, Contamination, Checking, and Urges and Worries of Losing Control of Motor Behavior. The PI, however, is highly correlated with measures of worry (Freeston et al., 1994), so a 41-item (van Oppen, Hoekstra, & Emmelkamp, 1995) and a 39-item (Burns, Keortge, Formea, & Sternberger, 1996) version were proposed that eliminated items that overlapped with worry. Good reliability and validity have been reported for these revisions, especially the Washington State University Revision developed by Burns et al. (1996; for reviews, see Antony, 2001; Feske & Chambless, 2000). However the PI and its revisions have low discriminate validity, and they fail to assess important characteristics of OCD such as symptom duration, interference, resistance, and uncontrollability. As well, certain types of obsessions and compulsions (e.g., mental rituals, sexual intrusive thoughts, neutralizing responses) are not well represented (Feske & Chambless, 2000; Foa, Kozak, Salkovskis, Coles, & Amir, 1998).
Foa et al. (1998) developed a 42-item questionnaire, the Obsessive-Compulsive Inventory (OCI), to assess a broader range of obsessive and compulsive symptom content in both clinical and nonclinical samples. The OCI scales have good internal consistency and test-retest reliability, and OCD patients score significantly higher on all subscales except Hoarding than other anxiety disorders and nonclinical controls (Foa et al., 1998). It has strong correlations with the MOCI and the Compulsive Activity Schedule, but the correlation coefficients with the YBOCS are low. A brief 18-item version of the OCI (OCI-R) was recently published, and the authors recommend that clinicians use the OCI-R rather than the OCI (Foa et al., 2002). However, the OCI-R may not provide an adequate assessment of obsessions, given that only 3 items deal with obsessional symptoms. As well, the instrument has a very high correlation with the MOCI (r =.85).
Diagnosis and assessment of OCD presents special challenges because the assessment process can activate incapacitating obsessional symptoms such as indecision, perfectionism, and doubting that heighten the respondent's anxiousness. This can lead to noncompliance or even to outright refusal to participate in the assessment process. Thus, brief OC measures are needed that can be included in a broad, integrated assessment battery of psychopathology that strikes a balance between increased administrative efficiency and diagnostic accuracy. In addition, a new OC symptom screener might have greater clinical utility if it more equally represented obsessive and compulsive content and severity across the primary Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV;American Psychiatric Association, 1994) symptom domains and incorporated recent cognitive-behavioral concepts of OCD.
A new 25-item questionnaire called the Clark-Beck Obsessive-Compulsive Inventory (CBOCI) was developed (a) to provide a brief screener for the frequency and severity of obsessive and compulsive symptoms; (b) to complement the Beck Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996) and Beck Anxiety Inventory (BAI), so that a battery of highly compatible but discriminating symptom measures can be assembled; (c) to assess the core symptom features of OCD as indicated in DSM-IV and current cognitive-behavioral theories of OCD; and (d) to assess obsessive and compulsive frequency, severity, and content in a manner that is relevant to both clinical and nonclinical population. The psychometric properties of the instrument were investigated in a pilot study, then in a larger validation study consisting of both clinical and nonclinical samples that completed standard measures of OC symptoms, anxiety, depression, and worry.
Study 1: Questionnaire Development and Revision
Development of the CBOCI began with a 27-item version of the questionnaire. Item construction was based on a review of published theoretical, diagnostic, and assessment literature on OCD; a survey of existing measurement instruments; and the consensus of an expert panel. The 27-item CBOCI was administered to four samples, and psychometric analyses led to further revisions that resulted in the final 25-item version. A brief summary of the pilot study results is presented below. For more details of the pilot study and item development, see the published manual (D. A. Clark & Beck, 2002).
Method
Participants
The pilot study sample consisted of 56 individuals with a DSM-IV Axis I principal diagnosis of OCD, 38 nonobsessional psychiatric outpatients, 35 nonclinical community adults, and 403 undergraduate students. Individuals in the OCD group were drawn from a variety of sites: the Anxiety Treatment and Research Centre at St. Joseph's Healthcare (Hamilton, Canada; n = 37), the Royal Melbourne Hospital (Melbourne, Australia; n = 7), the Center for Cognitive Therapy at the University of Pennsylvania (n = 2), and independent practice settings (n = 10). The 38 nonobsessional depressed or anxious patients were obtained from the Center for Cognitive Therapy, University of Pennsylvania. Individuals were excluded if they had a primary, secondary, or tertiary OCD diagnosis. Diagnosis for most of the clinical participants was based on the Structured Clinical Interview for DSM-III-R (SCID; Spitzer, Williams, & Gibbon, 1987), DSM-IV(SCID-IV; First, Gibbon, Spitzer, & Williams, 1996), or the Anxiety Disorders Interview Schedule for DSM-IV (Brown, Di Nardo, & Barlow, 1994). The 35 nonclinical community adults were drawn from community and church groups and from acquaintances of graduate students under David A. Clark's supervision. The 403 students were primarily first-year undergraduates of the University of New Brunswick enrolled in an introductory psychology class.
Development of the CBOCI
David A. Clark developed an initial 28-item pool based on a review of current research on diagnosis and symptom presentation in OCD as well as on DSM-IV diagnostic criteria (American Psychiatric Association, 1994). Items assessing each symptom construct were written with four response statements (0–3) reflecting increasing levels of symptom frequency or severity. The item response format was modeled after the BDI-II. In addition, a definition and examples of obsessions and compulsions were provided based on the DSM-IV and the YBOCS. The instructions were similar to the BDI-II, with respondents requested to select “the one statement in each group that best describes your thoughts, feelings or behavior during the past two weeks including today.”
The 28 items were sent to an expert panel of 10 internationally recognized OCD researchers, who provided quantitative and qualitative feedback on the items. This resulted in a 27-item pilot version of the questionnaire that assessed 15 features of obsessions and 12 characteristics of compulsions. The first 15 items were summed to yield an Obsessions score, the next 12 items were summed to yield a Compulsions score, and a Total Score was derived from the sum of the two subscales.
Measures
The BDI-II is a 21-item self-report measure designed to assess the frequency of depressive symptoms (Beck et al., 1996). Numerous studies have supported the reliability and validity of the BDI in clinical and nonclinical samples (Beck, Steer, & Garbin, 1988). The more recent version of the questionnaire (BDI-II) shares many of the same psychometric properties of the BDI and correlates .93 with the older measure (Dozois, Dobson, & Ahnberg, 1998).
The BAI is a 21-item self-report questionnaire that assesses the severity of anxious symptoms (Beck & Steer, 1993). The psychometric properties of the BAI have been supported in clinical and nonclinical samples (D. A. Clark, Steer, & Beck, 1994; Hewitt & Norton, 1993; Steer, Ranieri, Beck, & Clark, 1993). Steer et al. (1993) reported a moderately high correlation between the BAI and BDI (r =.61).
The Penn State Worry Questionnaire (PSWQ) is a 16-item questionnaire that assesses a general tendency to worry independent of the content of worry (Meyer, Miller, Metzger, & Borkovec, 1990). The PSWQ has been used extensively in clinical and nonclinical samples, and its construct validity has been supported (see Molina & Borkovec, 1994, for review). However the PSWQ Total Score is moderately correlated with self-report measures of OCD like the PI (Freeston et al., 1994) and the MOCI (Brown, Moras, Zinbarg, & Barlow, 1993).
Two self-report questionnaires of obsessive and compulsive symptoms were also included in the assessment battery. The Padua Inventory—Washington State University Revision (PI-WSUR) and the self-report version of the YBOCS were discussed in the introduction.
Procedure
Most of the OCD participants completed a battery of questionnaires at home that included the CBOCI, BDI-II, BAI, PSWQ, and self-report YBOCS, whereas a few participants also completed either the 60-item PI or the PI-WSUR. The OCD patients were at various stages of treatment when they completed the questionnaire battery. Analysis revealed that generally the OCD patients from different sites were quite similar on the main dependent variables of the pilot study, with the exception of those measured on the BAI. Therefore, we felt justified in combining the OCD patients into a single diagnostic group (n = 56).
The CBOCI was administered to the 38 nonobsessional psychiatric patients within the first or second session of treatment. Scores on the BDI-II and BAI were also available from the pretreatment intake assessment. The 35 nonclinical community adults completed a packet of questionnaires anonymously at home that included a demographic sheet, the CBOCI, BDI-II, BAI, PSWQ, PI-WSUR, and self-report YBOCS. There was no screen for past or present psychiatric disorders. Some of the participants in the student sample also completed self-report YBOCS (n = 159) and the 60-item PI (n = 161) in addition to the CBOCI, BDI-II, and BAI. PI-WSUR scores were calculated from the 39 items that correspond to that revision of the instrument.
Results
Internal Consistency
The CBOCI subscales had satisfactory internal consistency. For the 15-item CBOCI Obsessions subscale, [alpha] =.85 for the OCD sample (n = 51) and [alpha] =.89 for the total sample (n = 494). For the 12-item CBOCI Compulsions subscale, [alpha] =.84 for the OCD sample (n = 54) and [alpha] =.89 for the total sample (n = 499).
Group Differences
A one-way multivariate analysis of variance (MANOVA) with the four samples and gender as between-groups factors was performed on CBOCI Obsessions, Compulsions, BDI-II, and BAI. Only the main effect of group was significant, Wilk's [lambda] =.58, F(4, 12) = 22.97, p <.01, [eta]2 =.17. Follow-up F tests revealed significant group differences on all four scales, CBOCI Obsessions, F(3, 460) = 53.81, p <.01, [eta]2 =.26; CBOCI Compulsions, F(3, 460) = 59.99, p <.01, [eta]2 =.28; BDI-II, F(3, 460) = 39.93, p <.01, [eta]2 =.21; BAI, F(3, 460) = 17.57, p <.01, [eta]2 =.10. Student Newman-Keuls post hoc comparisons indicated that the OCD patients scored significantly higher than all other groups on CBOCI Obsessions and Compulsions.
A series of one-way analyses of variance (ANOVAs) revealed significant between-groups differences on CBOCI Total Score, F(3, 514) = 77.64, p <.01, [eta]2 =.31; YBOCS Total Score, F(2, 234) = 95.32, p <.01, [eta]2 =.45; and PI-WSUR Total Score, F(2, 239) = 37.67, p <.01, [eta]2 =.24. As expected, post hoc comparisons indicated that the OCD sample scored significantly higher than the community adults and students on the total scores of the CBOCI, YBOCS, and PI-WSUR. Furthermore, the OCD group was significantly higher than the nonobsessional psychiatric patients on CBOCI Total Score. The group differences on the CBOCI are in the predicted direction and therefore provide support for the criterion-related validity of the measure.
Correlations With Other Measures
The CBOCI Obsessions and Compulsions subscales were highly correlated in the OCD (r =.67, p <.01) and nonclinical (r =.61, p <.01) samples. Support for the concurrent validity of the CBOCI was evident as all three scales (CBOCI Obsessions, Compulsions, and Total Score) were highly correlated with the self-report YBOCS and PI-WSUR Total Scores in both OCD and nonclinical samples (rs ranged from .48 to .80). However the CBOCI scales were also moderately correlated with the BDI-II and BAI (rs of .26 to .54). The CBOCI Total Score and PSWQ Total Score were significantly correlated in the OCD (r =.68, p <.01) and nonclinical (r =.38, p <.05) samples.
Item Analysis
To assess the discriminability of individual questionnaire items, we conducted three types of item analyses. A series of one-way ANOVAs performed on the 27 CBOCI items indicated that the OCD patients scored significantly higher than the nonobsessional psychiatric patients on all items except obsessional impulses, doubting, perfectionism, and indecision. As well, the OCD patients did not score significantly higher than the nonclinical samples on the obsessional impulse item. A principal-components analysis with varimax rotation performed on the total sample (N = 478) revealed that all but 5 CBOCI items loaded >=.40 on a single factor (frequency of moral/religious obsessions:.33, obsessional impulses:.19, frequency of washing compulsions:.32, frequency of precision/symmetry:.39, and hoarding:.32).
The third set of analyses involved calculating response-characteristic curves to determine how well each item discriminated at varying levels of symptom severity as a function of the probability of endorsing an item option statement (Santor, Ramsay, & Zuroff, 1994). If the probability of endorsing higher item response levels changes as a function of OC symptom severity, then the item is a good discriminator (Ramsay, 1993; Santor, Zuroff, Ramsay, Cervantes, & Palacios, 1995). Response characteristic curves were calculated first for the OCD sample (n = 56) and then for the student sample (n = 403). A second item response analysis plotted the expected item score against the CBOCI Total Score for the OCD, nonobsessional psychiatric, community adult, and student samples.
Inspection of the response-characteristic curves revealed that 18 CBOCI items did not show optimal discrimination at increasing levels of symptom severity because the OCD patients and students rarely endorsed the fourth option. This indicates that the threshold for these items may have been set too high. Analysis of the expected item score also indicated that 4 items (frequency of harm/aggression obsessions, frequency of moral/religious obsessions, obsessional impulses, and hoarding) had even more serious problems because the “0” option was more likely endorsed at all levels of severity.
Overall, the findings from the pilot study suggested that the CBOCI Obsessions and Compulsive subscales and Total Score had acceptable internal consistency as well as good convergent and criterion-related validity. However, it was also clear that discriminant validity was weaker and that many of the items should be rewritten to improve their ability to more accurately discriminate at varying levels of OC symptom severity. Thus, an extensive revision of the questionnaire was conducted on the basis of these findings. Two items were deleted, obsessional impulse and hoarding, because they consistently failed to differentiate OC symptoms. In addition, 13 items were rewritten to improve their sensitivity to symptom severity and their ability to discriminate OCD patients. The revised 25-item CBOCI consists of a 14-item Obsessions subscale and an 11-item Compulsions subscale.
Study 2: Validation of the Revised Measure
A second study was undertaken to determine the psychometric properties of the revised CBOCI. This also provided an opportunity to improve on the research design of the pilot study. A larger sample of OCD patients was recruited from a single treatment setting, an expanded sample of nonobsessional psychiatric patients was obtained that allowed direct comparisons with specific diagnostic groups, and a subset of the student sample was retested to investigate the test-retest reliability of the CBOCI.
Method
Participants
The sample consisted of 83 OCD patients, 43 non-OCD patients with other anxiety disorders, 32 nonobsessional depressed patients, 26 nonclinical community adults, and 308 undergraduate students. None of the individuals in this sample participated in the pilot study. The OCD group was drawn from the Anxiety Treatment and Research Centre at St. Joseph's Healthcare (Hamilton, Canada). The patient edition of the SCID-IV Research Version (First et al., 1996) was used to establish the principal DSM-IV Axis I diagnosis. Eighty-two individuals in the OCD group had a principal diagnosis of OCD, and 1 was diagnosed with an anxiety disorder (OCD) in partial remission. Sixty-eight OCD patients (81.9%) had one or more additional diagnoses primarily consisting of depression and/or other anxiety disorders.
Half (n = 21) of the nonobsessional anxious group was drawn from the Hamilton site and the remainder (n = 22) was recruited from the Center for Cognitive Therapy, University of Pennsylvania, Philadelphia. Participants drawn from the Philadelphia site were diagnosed with the Outpatient Version of the SCID-IV (First et al., 1996). The diagnostic composition of the sample consisted of panic disorder with or without agoraphobic avoidance (n = 20, 46.5%), social phobia (n = 10, 23.3%), specific phobia (n = 2, 4.7%), generalized anxiety disorder (n = 8, 18.6%), and other anxiety disorder (n = 3, 7%). Twenty-eight (65.1%) participants had one or more additional diagnoses. Individuals with a secondary or tertiary diagnosis of OCD were excluded from the study.
The depressed sample consisted of 31 participants from the Philadelphia site and 1 patient with major depression from the Hamilton setting who did not have a comorbid diagnosis of OCD. The principal Axis I diagnosis of the sample included major depression (n = 22, 68.7%); dysthymic disorder (n = 4, 12.5%); bipolar disorder, with depressed mood the most recent episode (n = 5, 15.6%); and depressed mood due to a general medical condition (n = 1, 3.1%). Eighteen (56.3%) had one or more additional diagnoses.
Eleven (42.3%) of the nonclinical community adult sample came from the Hamilton setting, and 15 (57.7%) were recruited in Fredericton, Canada. The Hamilton controls were screened for disorders via telephone interview based on the SCID-IV screening questions, whereas a written version of this screener was included in the questionnaire packet given to the Fredericton sample.
The questionnaire battery was completed by 359 introductory psychology undergraduates at the University of New Brunswick, Canada. Students who indicated they received psychiatric or psychological treatment were excluded from further analysis (except for the confirmatory factor analysis [CFA]), leaving a final student sample of 308.
Table 1 presents the demographic characteristics across the four samples. Chi-square analyses of gender, [chi]2(4, N = 488) = 6.05, ns, and ethnicity [chi]2(4, N = 478) = 7.19, ns, were not significant, although the groups did differ in relationship (married vs. single) status, [chi]2(4, N = 485) = 131.71, p <.01, [eta]2 =.38, and age, F(4, 484) = 178.19, p <.01, [eta]2 =.60. A significantly greater proportion of the students were single, whereas a higher percentage of the community adults were married. The students were significantly younger than all other groups, whereas the OCD group was significantly younger than the two clinical controls and the community adult group. There was no significant difference between the OCD, anxious, depressed, or community adult groups in proportion with or without exposure to postsecondary education, [chi]2(3, N = 165) = 5.50, ns.

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 1 Demographic Characteristics of the Validation Sample
________________________________________
Measures
The same measures that are described in the pilot study were used in the present investigation.
Procedure
Individuals in the OCD group completed the CBOCI, self-report YBOCS, BDI-II, BAI, PI-WSUR, and the PSWQ at intake after the diagnostic interview. Participation was on a voluntary basis, with the assessment packet completed at home and returned to the clinic on a subsequent visit. Participants from the Center for Cognitive Therapy in Philadelphia also completed the CBOCI, BDI-II, and BAI at intake. Independent t tests indicated that individuals with an anxiety diagnosis from the Hamilton site were significantly older, t(41) = 2.2, p <.05, and had a higher BAI score, t(40) = 2.8, p <.01, than the Philadelphia patients. There were no significant differences on the CBOCI or BDI-II. The university student sample completed the entire questionnaire battery in large groups for course credit.
The 11 individuals from the Hamilton site who were included in the nonclinical community adult sample were recruited from the community via advertisements and word-of-mouth for another research project conducted at the Anxiety Treatment and Research Centre, whereas the 15 Fredericton community adults were drawn from friends and family of students taught by David A. Clark. Independent t tests failed to reveal significant differences between the two sets of nonclinical community groups on any of the questionnaires or demographic variables, with the exception that the Fredericton adults were significantly older than the Hamilton respondents, t(24) = 4.23, p <.01.
Results
Factorial Validity
The factor structure of the CBOCI was first investigated in a combined clinical group that included the OCD, anxious, and depressed samples (n = 160 after listwise deletion). The subject:variable ratio was too low to base the analysis only on the OCD sample, and the subject:variable ratio in the combined sample was also too low to conduct a CFA. However, the inclusion of nonobsessional anxious and depressed participants increased the scoring range, which is more conducive to factor analysis.
An initial principal-components analysis was performed to determine the number of factors to extract. On the basis of Cattell's (1966) scree test and Kaiser's coefficient alpha of generalizability (Kline & Barrett, 1983), it was decided that two factors should be retained. The first five consecutive eigenvalues were 12.19, 2.40, 1.25, 1.20, and 0.95, and the coefficient alphas of generalizability were .96, .61, and .21 for the first three components. Clearly, the third component did not attain an adequate level of internal consistency to warrant retention.
On the basis of the combined clinical sample, we next performed an iterated principal-factor analysis using the squared multiple correlations of the items as the initial communality estimates, and the two extracted factors were rotated to a Promax (k = 4) oblique criterion. Kaiser's Measure of Sampling Adequacy was .94 (Dziuban & Shirkey, 1974), a value that Kaiser and Rice (1974) described as “marvelous” for factor analytic purposes. A common factor model was appropriate because there were many very small values in the anti-image correlation matrix.1
Table 2 presents the Promax-rotated principal-factor pattern for the 25 CBOCI items. Based on the pattern of salient (>=.35) standardized-regression coefficients, two highly correlated (r =.67, p <.01) dimensions were identifiable; the first represented Obsessions and the second described Compulsions. All but one of the CBOCI Obsessions items (dirt/contamination) and one of the Compulsions items (avoidance) loaded specifically on their respective factor. Overall, the factor pattern provided empirical support for the rationally determined item composition of the Obsessions and Compulsions subscales.

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 2 Promax-Rotated Principal-Factor Pattern Standardized Regression and Orthogonalized Schmid-Leiman Coefficients of the Clark-Beck Obsessive-Compulsive Inventory (CBOCI) Items Based on the Combined Clinical Sample
________________________________________
Schmid-Leiman analysis
The high positive correlation between the two factors indicated that the Obsessions and Compulsions dimensions might be measuring a common second-order factor of symptom distress. To investigate this possibility, we performed a Schmid-Leiman transformation (Gorsuch, 1983; Loehlin, 1987) to assess the common and specific dimensions of the CBOCI. The same type of principal-factor analysis that was performed with the 25 CBOCI items was now performed with the two first-order factor loadings constrained to be equal. The standardized regression coefficient for both of these first-order factors on the second-order factor was .82, with a squared-communality estimate of .67 for each factor.
The Schmid-Leiman-transformed coefficients of the 25 CBOCI items on the orthogonalized second-order factor and the two first-order factors are also displayed in Table 2. The second-order factor explained approximately 39% of the total and 68% of the common or shared variance after the Schmid-Leiman transformation. The percentages of orthogonalized or unique variance contributed by the Obsessions and Compulsions factors were, respectively, 17% and 15%. As Table 2 indicates, only Item 3 (religious/moral/sexual) was not salient (>=.35) on the second-order factor or either of the first-order factors. Nine of the obsessions items (64%) were salient on the orthogonalized Obsessions factor, and 8 of the Compulsions items (73%) were salient on the orthogonalized Compulsions factor. Three obsessions items (doubting, perfectionism, and indecision) and 3 compulsion items (internal neutralizing, avoidance, and distress of compulsions) were shared variance items exclusively. One obsessions item (dirt/contamination) continued to show specificity with the Compulsions rather than the Obsessions orthogonalized factors.
CFA
Because there was a unselected sample of 343 undergraduates (after listwise deletion), the ratio of the number of participants to the number of structural equation parameters (343/51 = 6.7) was considered acceptable. Thus, a CFA was performed to determine whether the present scoring scheme for the Obsessions and Compulsions subscales would be appropriate for undergraduate students. The first 14 CBOCI items used in scoring the Obsessions scale were hypothesized to load on one factor, and the remaining 11 items were hypothesized to load on a second factor. The two factors were permitted to be correlated with each other. All of the error and uniqueness terms for the two factors and the 25 CBOCI items were assumed to be random. The CFA was performed with the SAS CALIS procedure.
The chi-square test for the CFA model was significant, [chi]2(274, N = 343) = 698.02, p <.001, indicating there was residual variance that still remained to be explained. Bentler's comparative fit index, the adjusted goodness-of-fit index, and Tucker-Lewis index were, respectively, .80, .81, and .78. These values are <.90, which according to Hatcher (1994), suggests that the two-factor model is an inadequate fit for this data set. However, the root-mean-square error of approximation was .07 and <.08, suggesting an acceptable error of approximation. We conclude that for research purposes, the first 14 CBOCI items can be summed to derive the Obsessions subscale and the last 11 items can be summed to produce the Compulsions subscale in college student samples.
Reliability Analyses
Internal consistency
Table 3 presents Cronbach's alpha coefficients and the mean corrected item-total correlations of the three CBOCI scale scores for the OCD, combined clinical, and student samples as well as McDonald's (1999) omega coefficient for the Total Score. McDonald's omega coefficient is the ratio of the variance that is explained by the common construct assumed to be underlying a scale to the total amount of variance that is explained by the scale. Omega is an index that may be used to evaluate whether a set of items is sufficiently homogeneous to warrant summing its item ratings to yield a total score.2 Because omega is only relevant for estimating the homogeneity of a total scale, omega coefficients were not calculated for the CBOCI subscales. Generally the omega coefficients support the homogeneity of the CBOCI Total Score in all three samples, although the coefficient is somewhat lower in the OCD and student samples. This suggests that the homogeneity of the CBOCI Total Score may be lower in samples with a more restricted scoring range such as the OCD or nonclinical sample. On the other hand, the mean corrected-item total correlations for the OCD and combined clinical samples are well above .30, and the alpha coefficients are in the mid- .80s or above .90. The range of the latter alpha coefficients is considered by Cicchetti (1994) to reflect good to excellent internal consistency for clinical purposes. The students' alpha coefficients are lower, especially for the CBOCI Obsessions subscale.

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 3 Homogeneity and Internal Consistency Estimates for the Three CBOCI Scale Scores by Type of Sample
________________________________________
Test-retest stability
Sixty-seven students completed the same questionnaire battery a second time after a 1 month interval (M = 31.5 days, range 10–45 days). After deletion of missing data, the final sample consisted of 55 students. The CBOCI Obsessions (r =.69, p <.001), Compulsions (r =.79, p <.001), and Total Score (r =.77, p <.001) showed a moderate level of temporal stability. There was a significant decline in the mean CBOCI Obsessions, t(54) = 4.66, p <.001, Compulsions, t(54) = 3.79, p <.001, and Total, t(54) = 5.01, p <.001, scores. The average CBOCI Obsessions score decreased by 2.22 (SD = 3.64), Compulsions decreased by 1.70 (SD = 3.34), and the Total Score declined by 3.93 (SD = 5.83). However, length of time between test administrations did not correlate with magnitude of decrease in CBOCI scoring. The CBOCI stability coefficients were comparable to those achieved by other measures in the questionnaire battery such as the BDI-II (r =.81, p <.001), PSWQ (r =.74, p <.001), and BAI (r =.65, p <.001). However, in comparison to CBOCI Total Score, the PI-WSUR Total Score achieved an unusually high test-retest correlation (r =.93, p <.001), whereas the YBOCS Total Score produced a much lower test-retest coefficient (r =.52, p <.001).
Criterion-Related Validity: Group Comparisons
Table 4 presents means and standard deviations for the five samples on the CBOCI scales and other questionnaire measures included in the study.3 A one-way MANOVA performed on CBOCI Obsessions and Compulsions was significant, Wilk's [lambda] =.49; F(8, 962) = 51.88, p <.001, [eta]2 =.30, for both Obsessions, F(4, 482) = 85.29, p <.001, [eta]2 =.41, and Compulsions, F(4, 482) = 96.84, p <.001, [eta]2 =.45. Scheffé's post hoc comparisons indicated that the OCD group scored significantly higher than all other groups on both CBOCI subscales. On the Obsessions subscale, the depressed group scored significantly higher than the remaining samples, the anxious and student groups followed with similar scores, and the community adults were significantly lower than all other participants. On the CBOCI Compulsions subscale, the depressed, anxious, and student groups were not significantly different from each other, whereas the community adults were again significantly lower than all groups. A one-way ANOVA on the CBOCI Total Score was highly significant, F(4, 485) = 113.73, p <.001, [eta]2 =.48, with the OCD group scoring significantly higher than all other groups; the depressed, anxious, and student groups scoring at a similar level; and the community adults responding significantly lower than any other participants.4

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 4 Means, Standard Deviations, and Sample Size of the Dependent Measures for the Five Samples
________________________________________
As expected, significant group differences were evident on the YBOCS Total Score, F(3, 395) = 174.04, p <.001, [eta]2 =.57; PI-WSUR Total Score, F(3, 420) = 51.39, p <.001, [eta]2 =.27; PSWQ, F(3, 425) = 42.59, p <.001, [eta]2 =.23; BDI-II, F(4, 477) = 63.56, p <.001, [eta]2 =.35; and the BAI, F(4, 477) = 25.87, p <.001, [eta]2 =.18. The OCD sample scored significantly higher than all other participants on the YBOCS, PI-WSUR, and PSWQ. Together, the depressed and OCD groups were significantly higher on BDI-II, whereas the anxious and OCD groups had significantly elevated scores on the BAI. Once again, the expected pattern of group differences on the CBOCI scales supports the criterion-related validity of the measure. Moreover, the CBOCI and YBOCS total scores clearly accounted for more between-groups variance than the PI-WSUR, PSWQ, BAI, and BDI-II.
Convergent and Discriminant Validity
Table 5 presents the zero-order correlation matrix for the CBOCI and other symptom measures. Correlation coefficients for the OCD sample (n = 64 after listwise deletion) are presented above the diagonal, whereas correlations for the nonclinical sample (adults and students combined) are found below the diagonal (n = 282 after listwise deletion). Correlations for the nonobsessional depressed and anxious samples were not calculated because of an insufficient sample size (n = 19 after listwise deletion).

________________________________________

[Help with image viewing]
[Email Jumpstart To Image]
Table 5 Zero-Order Correlation Matrix Based on the OCD Group and Combined Nonclinical Samples in the Validation Study
________________________________________
The CBOCI subscales and Total Score were highly correlated with the YBOCS subscales and PI-WSUR Total Score in both the OCD and nonclinical samples, thus supporting the instrument's concurrent validity. However a z test comparison of correlated correlation coefficients (Meng, Rosenthal, & Rubin, 1992) indicated that the CBOCI Total Score correlation with the YBOCS Total Score was not significantly greater than the PI-WSUR correlation with the YBOCS in both the OCD (z = 0.18, ns) and nonclinical (z = 1.20, ns) samples.
The CBOCI scales were also moderately correlated with self-report measures of depression and anxiety. However, z test comparisons indicated that the CBOCI Total Score was significantly more correlated with the PI-WSUR Total Score than with the BAI Total Score in both the OCD (z = 2.21, p <.05, zdiff =.31) and nonclinical (z = 3.30, p <.001, zdiff =.24) samples. Both of these effect sizes are considered medium (Cohen, 1992). The YBOCS Total Score, on the other hand, was significantly more correlated with the PI-WSUR Total Score than with the BAI Total Score in the OCD sample (z = 2.95, p <.01, zdiff =.38) but not in the nonclinical group (z = 1.87, p =.06, zdiff =.13). The PI-WSUR and CBOCI did not differ significantly in their correlation with the BAI Total Score in the OCD (z = 0.15, p =.88, zdiff =.02) or the nonclinical (z = 1.38, p =.17, zdiff =.08) samples. These findings indicate that the CBOCI is more closely related to OC symptoms than it is to general anxiety symptoms, thus supporting its discriminant validity. There is also evidence that the CBOCI might have slightly better discriminant validity than the YBOCS but not than the PI-WSUR in nonclinical samples. However, the CBOCI correlations with the BDI-II were as high as its correlations with the YBOCS or PI-WSUR.
Another aspect of discriminant validity concerns the relationship between the CBOCI and PSWQ. Previous OCD measures have had difficulty distinguishing worry from obsessions (Freeston et al., 1994; Turner, Beidel, & Stanley, 1992). Partial correlations were calculated between the CBOCI scales, other measures of OC symptoms (PI-WSUR and YBOCS Total Scores), and worry (PSWQ). For the OCD sample, CBOCI Total Score continued to have a high correlation with PI-WSUR Total Score (r =.78, pr =.67, p <.001) and YBOCS Total Score (r =.78, pr =.71, p <.001) after covarying PSWQ. Partialing out the YBOCS Total Score resulted in a substantial reduction in the relationship of the CBOCI Total Score with the PSWQ (r =.60, pr =.45, p <.001). These findings were replicated in the student sample. Together, these results indicate that the CBOCI has a closer relationship with OC symptom measures than with measures of worry (see D. A. Clark & Beck, 2002, for additional partial correlations), although these results also confirm a close continuing relationship between worry and obsessionality.
General Discussion
The present study reported on the development and initial validation of a 25-item self-report measure of obsessive and compulsive symptoms intended to provide an accurate and reliable brief assessment that could be included in a more comprehensive diagnostic assessment of OCD. The reliability and validity of the CBOCI equals, or surpasses, lengthier OC symptom measures. The integrity of the questionnaire's two-dimensional structure was supported by factor analysis, and the measure distinguishes individuals with OCD from those with major depression, other anxiety disorders, and nonclinical controls. It correlates with other symptom measures like the YBOCS and PI-WSUR. The similar pattern of relations between the CBOCI and other measures in the nonclinical samples indicates that it has relevance for clinical and nonclinical populations, although more caution is needed when interpreting nonclinical scores. As a 25-item questionnaire that correlates moderately with the BDI-II and BAI, it clearly can be used along with these other measures. The more cognitive-behavioral items in the questionnaire, such as controllability and fixity of belief, also had high loadings on their respective dimensions. Thus, the strong cognitive-behavioral orientation to the questionnaire was empirically supported in the psychometric analyses.
An initial exploratory factor analysis of the 25 CBOCI items resulted in two highly correlated dimensions that corresponded very closely to the rationally derived Obsessions and Compulsions subscales. A subsequent factor analysis based on a Schmid-Leiman transformation revealed that two thirds of the variance in the CBOCI represented a second-order general factor, whereas the remaining one third of the variance was accounted for by specific first-order Obsessions and Compulsions factors. It is noteworthy that a significant amount of CBOCI variance is specific to OC symptoms and that the item loadings on the two first-order factors generally supports the 14-item Obsessions and 11-item Compulsions subscales. However the two-factor structure of the questionnaire had less support from a CFA performed on the student sample in which only one out of four indices suggested adequate fit.
The two-factor structure of the CBOCI gained further support from the reliability analyses summarized in Table 3. Cronbach's alpha indicated that the CBOCI Obsessions and Compulsions subscales were homogeneous in both clinical and nonclinical samples. However, the CBOCI Total Score omega coefficient was low in the student sample. Although there is no convention on what is an acceptable omega value, it is clear that caution must be exercised in using the CBOCI Total Score with a nonclinical sample. Given that OCD is defined by the presence of distinct obsessive and compulsive symptoms, it is likely that researchers and practitioners will find the two CBOCI subscales more clinically useful than the Total Score. The lower reliability of the CBOCI scales in the student sample and weak support from the CFA also suggest that a symptom measure like the CBOCI must be used cautiously in nonclinical samples that have a restricted scoring range.
It is clear from the Schmid-Leiman analysis that some CBOCI items are more specific to obsessions and compulsions, whereas others are more general in nature. Doubting, perfectionism, indecision, mental compulsions (internal neutralizing), avoidance of stimuli that trigger compulsions, and distress associated with compulsions had significant loadings only on the second-order factor. Although these items may assess more general phenomena, they should not be eliminated from the measure because they do assess important features of the obsessional state. However, other questionnaire items such as uncontrollability of obsessions, salience, effort to control, degree of insight, presence of cleaning or precision/symmetry rituals, and slowness had substantial loadings on the OC unique factors. In circumstances in which differentiating OC symptoms from other clinical presentations may be difficult, the clinician could examine how the patient responds to the OC-specific items to determine whether an obsessional state might be present.
Although the CBOCI has fewer items, its concurrent and discriminant validity is comparable to the YBOCS and PI-WSUR. In fact, the squared eta values associated with the ANOVAs indicated that the CBOCI and YBOCS total scores were able to differentiate OCD from nonclinical status better than the PI-WSUR or the nonobsessional symptom measures. The partial correlations indicate that the CBOCI is more relevant to OC symptoms than to worry. Nevertheless the CBOCI also appears to be sensitive to general anxiety, worry, and depression. It may be that obsessions, compulsions, depression, anxiety, and worry co-occur at such a high frequency that a sharp discrimination is impossible. Also, all of these symptom states have overlapping features that will inflate their interrelation, and all are likely linked to a common higher order construct such as negative affect or neuroticism (D. A. Clark et al., 1994; Watson & Clark, 1992). Thus, like the attempt to distinguish anxiety and depression, there may be a lower boundary to the discriminant validity of OC symptom measures (for discussion of the discriminability of anxiety and depression measures, see L. A. Clark & Watson, 1991).
The CBOCI was developed as a relatively quick self-report screener for obsessive and compulsive symptoms. It is one of the first self-report OCD questionnaires to have equally represented and empirically verified specific obsession and compulsion subscales. Moreover the cognitive-behavioral orientation to the CBOCI and its closer approximation to DSM-IV diagnostic criteria distinguish it from other OCD measures such as the PI-WSUR. The inclusion of symptom content items as well as more cognitive-behavioral items is a departure from the YBOCS. As a symptom screener, the 18-item OCI-R is more similar to the CBOCI than other OC questionnaires. Although it is difficult to assess the relative performance of these two measures without a direct comparison, the item content of the OCI-R and CBOCI is very different, with 17% of OCI-R items assessing obsessions compared with 56% of the CBOCI items. It would be interesting to compare the performance of these two brief OCD measures in a future study. In the meantime there are a number of features to the CBOCI that distinguish it from existing OCD symptom measures.
In terms of a clinical assessment strategy, the CBOCI could be administered as part of an initial suite of depression and anxiety symptom measures. Individuals who obtain a high score on the CBOCI should then be given a structured clinical interview to determine whether the diagnostic criteria for OCD are met. As with all self-report symptom measures, the CBOCI is not a diagnostic instrument but rather a measure of the frequency and severity of symptoms. The clinical utility of self-report symptom measures like the CBOCI can only be fully realized within the context of individual clinical interviews.
There are limitations to the current studies that should be addressed in any further research on the CBOCI. A larger OCD sample would be helpful so that exploratory and CFA can be conducted on an OCD-only sample. Test-retest reliability data are also needed on an OCD sample and the sensitivity of the instrument to treatment effects remains to be determined. Whether the CBOCI is equally sensitive to all subtypes of OCD should be addressed. Like most clinical instruments, a larger nonstudent community adult sample is needed to provide normative data on the CBOCI. Also, it would be important to determine the diagnostic sensitivity and specificity of the questionnaire in a large mixed clinical sample drawn from a single treatment setting. Finally, the current studies relied exclusively on self-report questionnaires. The relation of the CBOCI to interviewer-based instruments such as the standard YBOCS or the Compulsive Activity Checklist is unknown. Further research is also needed on OCD symptoms in depression. Although the depressed sample had elevated CBOCI scores, the failure to administer the full assessment protocol to the depressed sample limits the conclusions that can be drawn about the differentiation of obsessions and compulsions from major depression.
Despite these areas for further investigation, we believe the current findings provide strong support for the construct validity of the CBOCI. The present analyses indicated that a highly focused 25-item symptom screener can be used to assess symptom severity in a complex clinical condition like OCD with as much discriminability as longer, more time-consuming measures. Only further research can establish whether the CBOCI is a significant improvement over existing OCD symptom measures.
References
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. [Context Link]
Amir, N., Foa, E. B., & Coles, M. E. (1997). Factor structure of the Yale-Brown Obsessive Compulsive Scale. Psychological Assessment, 9, 312–316. Ovid Full Text [Context Link]
Antony, M. M. (2001). Measures for obsessive-compulsive disorder. In M. M. Antony, S. M. Orsillo, & L. Roemer (Eds.), Practitioner's guide to empirically based measures of anxiety (pp. 219–243). New York: Kluwer Academic/Plenum Press Publishers. [Context Link]
Antony, M. M., Downie, F., & Swinson, R. P. (1998). Diagnostic issues and epidemiology in obsessive-compulsive disorder. In R. P. Swinson, M. M. Antony, S. Rachman, & M. A. Richter (Eds.), Obsessive-compulsive disorder: Theory, research and treatment (pp. 3–32). New York: Guilford Press. [Context Link]
Beck, A. T., & Steer, R. A. (1993). Manual of the Beck Anxiety Inventory. San Antonio, TX: Psychological Corporation. [Context Link]
Beck, A. T., Steer, R. A., & Brown, G. (1996). Manual for Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation. [Context Link]
Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77–100. Bibliographic Links [Context Link]
Brown, T. A., Campbell, L. A., Lehman, C. L., Grisham, J. R., & Mancill, R. B. (2001). Current and lifetime comorbidity of the DSM-IV anxiety and mood disorders in a large clinical sample. Journal of Abnormal Psychology, 110, 585–599. Ovid Full Text Bibliographic Links [Context Link]
Brown, T. A., Di Nardo, P. A., & Barlow, D. H. (1994). Anxiety disorders interview schedule for DSM-IV. Client Interview Schedule. Boulder, CO: Graywind. [Context Link]
Brown, T. A., Moras, K., Zinbarg, R. E., & Barlow, D. H. (1993). Diagnostic and symptom distinguishability of generalized anxiety disorder and obsessive-compulsive disorder. Behavior Therapy, 24, 227–240. Bibliographic Links [Context Link]
Burns, G. L., Keortge, S. G., Formea, G. M., & Sternberger, L. G. (1996). Revision of the Padua Inventory of obsessive compulsive disorder symptoms: Distinctions between worry, obsessions and compulsions. Behaviour Research and Therapy, 34, 163–173. [Context Link]
Cattell, R. B. (1966). The scree test for the number of factors. Multivariate Behavioral Research, 1, 245–276. Bibliographic Links [Context Link]
Cicchetti, D. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological Assessment, 6, 284–290. Ovid Full Text [Context Link]
Clark, D. A. (2004). Cognitive-behavioral therapy for OCD. New York: Guilford Press. [Context Link]
Clark, D. A., & Beck, A. T. (2002). Manual for the Clark-Beck Obsessive Compulsive Inventory. San Antonio, TX: Psychological Corporation. [Context Link]
Clark, D. A., Steer, R. A., & Beck, A. T. (1994). Common and specific dimensions of self-reported anxiety and depression: Implications for the cognitive and tripartite models. Journal of Abnormal Psychology, 103, 645–654. [Context Link]
Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100, 316–336. Ovid Full Text Bibliographic Links [Context Link]
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159. [Context Link]
Crino, R. D., & Andrews, G. (1996). Obsessive-compulsive disorder and Axis I comorbidity. Journal of Anxiety Disorders, 10, 37–46. Bibliographic Links [Context Link]
Dozois, D. J. A., Dobson, K. S., & Ahnberg, J. L. (1998). A psychometric evaluation of the Beck Depression Inventory-II. Psychological Assessment, 10, 83–89. Ovid Full Text [Context Link]
Dziuban, C. D., & Shirkey, E. C. (1974). When is a correlation matrix appropriate for factor analyses? Psychological Bulletin, 81, 358–361. Ovid Full Text [Context Link]
Emmelkamp, P. M. G., Kraaijkamp, H. J. M., & van den Hout, M. A. (1999). Assessment of obsessive-compulsive disorder. Behavior Modification, 23, 269–279. [Context Link]
Feske, U., & Chambless, D. L. (2000). A review of assessment measures for obsessive-compulsive disorder. In W. K. Goodman, M. V. Rudorfor, & J. D. Maser (Eds.), Obsessive-compulsive disorder: Contemporary issues in treatment (pp. 157–182). Mahwah, NJ: Erlbaum. [Context Link]
First, M. B., Gibbon, M., Spitzer, R. L., & Williams, J. B. W. (1996). User's guide for the Structured Clinical Interview for DSM-IV Axis I disorders—Research Version (SCID-I, Version 2.0, February 1996 final version). New York: Biometrics Research Department, NY State Psychiatric Institute. [Context Link]
Foa, E. B., Huppert, J. D., Leiberg, S., Langner, R., Kichic, R., Hajack, G., & Salkovskis, P. M. (2002). The Obsessive-Compulsive Inventory: Development and validation of a short version. Psychological Assessment, 14, 485–496. Ovid Full Text [Context Link]
Foa, E. B., Kozak, M. J., Salkovskis, P. M., Coles, M. E., & Amir, N. (1998). The validation of a new obsessive-compulsive disorder scale: The Obsessive-Compulsive Inventory. Psychological Assessment, 10, 206–214. Ovid Full Text [Context Link]
Freeston, M. H., Ladouceur, R., Rhéaume, J., Letarte, H., Gagnon, F., & Thibodeau, N. (1994). Self-report of obsessions and worry. Behaviour Research and Therapy, 32, 29–36. [Context Link]
Gibbs, N. A. (1996). Nonclinical populations in research on obsessive-compulsive disorder: A critical review. Clinical Psychology Review, 16, 729–773. [Context Link]
Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Delgado, P., Heninger, G. R., & Charney, D. S. (1989). The Yale-Brown Obsessive Compulsive Scale II: Validity. Archives of General Psychiatry, 46, 1012–1016. [Context Link]
Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleischmann, R. L., & Hill, C. L., et al. (1989). The Yale-Brown Obsessive Compulsive Scale I: Development, use, and reliability. Archives of General Psychiatry, 46, 1006–1011. [Context Link]
Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: Erlbaum. [Context Link]
Hatcher, L. (1994). A step-by-step approach to using the SAS system for factor analysis and structural equation modeling. Cary, NC: SAS Institute. [Context Link]
Hewitt, P. L., & Norton, G. R. (1993). The Beck Anxiety Inventory: A psychometric analysis. Psychological Assessment, 5, 408–412. Ovid Full Text [Context Link]
Hodgson, R. J., & Rachman, S. J. (1977). Obsessional compulsive complaints. Behaviour Research and Therapy, 15, 389–395. [Context Link]
Hollander, E., & Wong, C. M. (2000). Spectrum, boundary, and subtyping issues: Implications for treatment-refractory obsessive-compulsive disorder. In W. K. Goodman, M. V. Rudorfor, & J. D. Maser (Eds.), Obsessive-compulsive disorder: Contemporary issues in treatment (pp. 3–22). Mahwah, NJ: Erlbaum. [Context Link]
Kaiser, H. F., & Rice, J. (1974). Little Jiffy, Mark IV. Educational and Psychological Measurement, 34, 111–117. [Context Link]
Kline, P., & Barrett, P. (1983). The factors in personality questionnaires among normal subjects. Advances in Behaviour Research and Therapy, 5, 141–202. [Context Link]
Leckman, J. F. (1993). Tourette's syndrome. In E. Hollander (Ed.), Obsessive-compulsive-related disorders (pp. 113–137). Washington, DC: American Psychiatric Press. [Context Link]
Loehlin, J. C. (1987). Latent variable models: An introduction to factor, path, and structural analysis. Hillsdale, NJ: Erlbaum. [Context Link]
McDonald, R. P. (1999). Test theory: A unified treatment. Mahwah, NJ: Erlbaum. [Context Link]
Meng, X-L., Rosenthal, R., & Rubin, D. B. (1992). Comparing correlated correlation coefficients. Psychological Bulletin, 111, 172–175. [Context Link]
Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy, 28, 487–495. [Context Link]
Molina, S., & Borkovec, T. D. (1994). The Penn State Worry Questionnaire: Psychometric properties and associated characteristics. In G. C. L. Davey & F. Tallis (Eds.), Worrying: Perspectives on theory, assessment and treatment (pp. 265–283). Chichester, England: Wiley. [Context Link]
Nakagawa, A., Marks, I. M., Takei, N., De Araujo, L. A., & Ito, L. M. (1996). Comparisons among the Yale-Brown Obsessive-Compulsive Scale, Compulsion Checklist, and other measures of obsessive-compulsive disorder. British Journal of Psychiatry, 169, 108–112. Bibliographic Links [Context Link]
Obsessive Compulsive Cognitions Working Group (2003). Psychometric validation of the Obsessive Beliefs Questionnaire and the Interpretation of Intrusions Inventory: Part I. Behaviour Research and Therapy, 41, 863–878. [Context Link]
O'Connor, K. P. (2001). Clinical and psychological features distinguishing obsessive-compulsive and chronic tic disorders. Clinical Psychology Review, 21, 631–660. Bibliographic Links [Context Link]
Rachman, S. J., & Hodgson, R. J. (1980). Obsessions and compulsions. Englewood Cliffs, NJ: Prentice Hall. [Context Link]
Rachman, S. J., & Shafran, R. (1998). Cognitive and behavioral features of obsessive-compulsive disorder. In R. P. Swinson, M. M. Antony, S. Rachman, & M. A. Richter (Eds.), Obsessive-compulsive disorder: Theory, research and treatment (pp. 51–78). New York: Guilford Press. [Context Link]
Ramsay, J. O. (1993). TESTGRAF: A program for the graphical analysis of multiple choice test and questionnaire data. McGill University, Unpublished computer program. [Context Link]
Richter, M. A., Cox, B. J., & Direnfeld, D. M. (1994). A comparison of three assessment instruments for obsessive-compulsive symptoms. Journal of Behavior Therapy and Experimental Psychiatry, 25, 143–147. [Context Link]
Sanavio, E. (1988). Obsessions and compulsions: The Padua Inventory. Behaviour Research and Therapy, 26, 169–177. [Context Link]
Santor, D. A., Ramsay, J. O., & Zuroff, D. C. (1994). Nonparametric item analyses of the Beck Depression Inventory: Examining gender item bias and response option weights. Psychological Assessment, 6, 255–270. Ovid Full Text [Context Link]
Santor, D. A., Zuroff, D. C., Ramsay, J. O., Cervantes, P., & Palacios, J. (1995). Examining scale discriminability in the BDI and CES-D as a function of depressive severity. Psychological Assessment, 7, 131–139. Ovid Full Text [Context Link]
Spitzer, R. L., Williams, J. B. W., & Gibbon, M. (1987). SCID: Structured Clinical Interview for DSM-III-R. New York: New York Psychiatric Institute. [Context Link]
Steer, R. A., Ranieri, W. F., Beck, A. T., & Clark, D. A. (1993). Further evidence for the validity of the Beck Anxiety Inventory with psychiatric outpatients. Journal of Anxiety Disorders, 7, 195–205. Bibliographic Links [Context Link]
Steketee, G. S. (1993). Treatment of obsessive compulsive disorder. New York: Guilford Press. [Context Link]
Steketee, G. S. (1994). Behavioral assessment and treatment planning with obsessive compulsive disorder: A review emphasizing clinical application. Behavior Therapy, 25, 613–633. Bibliographic Links [Context Link]
Steketee, G. S., Frost, R. O., & Bogart, K. (1996). The Yale-Brown Obsessive Compulsive Scale: Interview versus self-report. Behaviour Research and Therapy, 34, 675–684. [Context Link]
Taylor, S. (1995). Assessment of obsessions and compulsions: Reliability, validity, and sensitivity to treatment effects. Clinical Psychology Review, 15, 261–296. Bibliographic Links [Context Link]
Taylor, S. (1998). Assessment of obsessive-compulsive disorder. In R. P. Swinson, M. M. Antony, S. Rachman, & M. A. Richter (Eds.), Obsessive-compulsive disorder: Theory, research and treatment (pp. 229–257). New York: Guilford Press. [Context Link]
Taylor, S., Thordarson, D. S., & Söchting, I. (2002). Obsessive-compulsive disorder. In M. M. Antony & D. H. Barlow (Eds.), Handbook of assessment and treatment planning for psychological disorders (pp. 182–214). New York: Guilford Press. [Context Link]
Thordarson, D. S., Radomsky, A. S., Rachman, S., Shafran, R., Sawchuk, C. N., & Hakstian, A. R. (2004). The Vancouver Obsessional Compulsive Inventory (VOCI). Behaviour Research and Therapy, 42, 1289–1314. [Context Link]
Turner, S. M., Beidel, D. C., & Stanley, M. A. (1992). Are obsessional thoughts and worry different cognitive phenomena? Clinical Psychology Review, 12, 257–270. [Context Link]
van Oppen, P., Hoekstra, R. J., & Emmelkamp, P. M. G. (1995). The structure of obsessive-compulsive symptoms. Behaviour Research and Therapy, 33, 15–23. [Context Link]
Watson, D., & Clark, L. A. (1992). Affects separable and inseparable: On the hierarchical arrangements of the negative affects. Journal of Personality and Social Psychology, 62, 489–505. [Context Link]
Woody, S. R., Steketee, G., & Chambless, D. L. (1995). Reliability and validity of the Yale Brown Obsessive-Compulsive Scale. Behaviour Research and Therapy, 33, 597–605. [Context Link]
1Copies of the CBOCI item intercorrelation matrix for both the combined patient and student samples can be found in D. A. Clark and Beck (2002). [Context Link]
2A copy of the item means and standard deviations for each of the three samples reported in Table 3 is available from David A. Clark. [Context Link]
3A series of t tests investigated gender differences on the dependent measures within each sample. In the OCD sample, women scored significantly higher on CBOCI Compulsion, PI-WSUR, BDI-II, and BAI. There were no significant gender differences in the anxious or community adult samples. Women scored significantly higher than men on PSWQ in the student sample and significantly higher on BDI-II in the depressed group. Because no consistent gender differences emerged across samples, analyses were conducted on the total sample. [Context Link]
4Because the depressed sample did not complete the PI-WSUR or the YBOCS, and part of the community adult group did not have YBOCS scores, the CBOCI Total Score, PI—WSUR, and YBOCS were analyzed separately. [Context Link]
Keywords: obsessive-compulsive disorder; OCD; obsessions; compulsions; assessment of OCD; Clark-Beck Obsessive-Compulsive Inventory
________________________________________

Accession Number: 00012030-200506000-00003
Copyright (c) 2000-2007 Ovid Technologies, Inc.
Version: rel10.5.8, SourceID 1.13281.2.32.1.0.2.197.1.4.1.5