Psychometric Properties and Confirmatory Factor Analysis of the Obsessive Compulsion Symptoms Scale in A Sample of Kuwait University Students

Abstract

The aim of the present study was to test the reliability and validity of Mo’men & Abu Hendi’s (2006) Obsessive Compulsion Scale (OCS). The scale consist of 83 items divided into 14 subscales. A sample of (446) male and female undergraduate students from Kuwait University, with mean age of 21.28 (±1.66) responded to the OCS. Alpha coefficients ranged from (0.49 0.86). They were acceptable for only 6 subscales out of the 14 subscales. Retest reliabilities were acceptable only in five subscales. The exploratory factor analysis revealed (21) factors with eigenvalues ≥ 1, but all of these factors – except the first one – did not contain three items with loading ≥ 0.3. Results also showed that most of the OCS items loaded onto a single factor that fulfilled all the criteria that is suggested in the present study. The confirmatory factor analysis was used twice: On the first implementation, the model for the original study was used. Four measures were tested to assess the goodness of fit as follows: the Comparative Fit Index, the Root Mean Square Errors of Approximation, the Normal Fit Index, and the Goodness of Fit. In applying these measures, it was found that the original study model with 14 subscales cannot be accepted. The confirmatory factor analysis was applied again but with a single-factor model. The results revealed acceptable measures of goodness of fit. It was concluded that the best fitted model of the OCS was the single-factor model.