Shorter communication
Psychometric properties of the OCI-R in a college sample

https://doi.org/10.1016/j.brat.2003.08.002Get rights and content

Abstract

Two studies examined the psychometric properties of the Obsessive–Compulsive Inventory—Revised (OCI-R; Psychol. Assessment 14 (2002) 485) in a nonclinical student sample. In Study 1, we investigated the factor structure and internal consistency of the OCI-R using a sample of 395 undergraduate students. At a second testing session 1 month later, 178 students completed the OCI-R. Test–retest reliability was examined using data from 94 students who completed the OCI-R in both sessions. Convergent validity was also assessed with the Maudsley Obsessive–Compulsive Inventory (MOCI). In Study 2, we further investigated the convergent and divergent validity of the OCI-R using a new sample of 221 students who completed a battery of measures of obsessive–compulsive symptoms, worry, and depression. There was a significant order effect for both the OCI-R and the MOCI: means of each measure were significantly lower when presented second. Despite the order effect, statistical analyses indicated that the OCI-R has adequate test–retest reliability for the full scale and subscale scores, solid factor structure, and high internal consistency. Convergent validity with other measures of obsessive–compulsive symptoms was moderate to excellent, and divergent validity was good. The results indicate that the OCI-R is a short, psychometrically sound self-report measure of obsessive–compulsive symptoms.

Introduction

The Obsessive–Compulsive Inventory—Revised (OCI-R; Foa et al., 2002) is a shortened version of the Obsessive–Compulsive Inventory (OCI; Foa, Kozak, Salkovskis, Coles, & Amir, 1998) that measures obsessive–compulsive symptoms. In addition to providing a total score, the OCI-R yields six subscales: washing, checking, ordering, obsessing, hoarding, and neutralizing. Compared to the OCI, the OCI-R is significantly shorter, eliminates the frequency scale which appears to be redundant, has subscales of three items each which are summed (making them easy to compare), and retains excellent psychometric properties in a mixed sample of patients with obsessive–compulsive disorder, other anxiety disorders, and non-patients.

Although much research on OCD focuses on clinical populations, it is increasingly common for this research to extend to nonpatient samples to investigate constructs that are important to the theoretical understanding of OCD (for reviews, see Burns, Formea, Keortge and Sternberger, 1995, Gibbs, 1996). For research on OCD to benefit from studies in nonclinical populations, it is crucial for measures that assess OCD symptoms to have sound psychometric properties within both clinical and nonclinical populations. For example, Simonds, Thorpe, and Elliott (2000) found that the original version of the OCI had excellent psychometrics in a population of college students—comparable to the properties reported in clinical populations by Foa et al. (1998). However, Wu and Watson (2002) investigated the factor structure of the OCI in a sample of college students and reported that a different, five-factor structure than that theorized by Foa et al. (1998) better fit the data, though goodness-of-fit indices were still not excellent according to most standards (e.g., Hu, & Bentler, 1999).

Although some data from nonanxious controls were reported for the OCI-R in the Foa et al. (2002) study, the psychometric performance of the OCI-R was not fully examined in a nonclinical student sample. The present study sought to examine the psychometric properties of the OCI-R in a population of college students.

Section snippets

Participants

Two large introductory psychology courses at the University of Delaware participated in a subject pool, completing a number of self-report measures of different psychological constructs. From this pool of subjects, 395 students had usable data for the current study. This sample consisted of 242 females, 146 males, and five students who did not identify their gender. A second sample of 178 students (106 women and 72 men) were administered self-report measures 1 month later. Of these students, 94

Study 2

In Study 1, the correlation between the OCI-R and MOCI total scores was rather low, in contrast to Simonds et al.’s (2000) finding that the original version of the OCI was highly correlated with the MOCI in a sample of college students. Additionally, Foa et al. (2002) report a high correlation between the OCI-R and the MOCI in 34 nonanxious control subjects. Given the discrepant findings in Study 1, we sought to evaluate the convergent validity of the OCI-R in a new sample using two measures of

Discussion

Overall, the OCI-R appears to have excellent psychometric properties. Foa et al. (2002) demonstrated this within a large clinical sample and within a small sample of nonanxious controls. The present study extends these findings by demonstrating that the OCI-R continues to have good reliability and a solid factor structure in large samples of nonanxious college students. The OCI-R has good convergent and divergent validity with other measures related to anxiety and depression.

Although the

References (14)

There are more references available in the full text version of this article.

Cited by (0)

View full text