Psychometric Analysis of Racial Differences on the Maudsley Obsessional Compulsive Inventory

Abstract

Black university students scored significantly higher than White students on the Maudsley Obsessional Compulsive Inventory (MOCI). They tended to endorse more Cleaning and Checking subscale items in the pathological direction. Subsequent analyses examined whether this finding is a reflection of valid group differences in the prevalence of OCD or a psychometric artifact. Structured interviews were conducted to determine the correspondence of MOCI scores with OCD diagnoses. The race difference in endorsement frequency on the MOCI did not extend to OCD diagnoses. The MOCI scores showed modest predictive validity in Whites, but they did not predict interview-based diagnoses in Blacks. Multivariate item response theory was then employed to examine race differences in the Cleaning and Checking subscales. Equivalent item discrimination parameters fit the data for Black and White participants for both subscales. A more restrictive model in which relative item difficulties were also constrained to be equal for Black and White participants did not fit. This interaction between race and item difficulty suggests that the items do not have equivalent psychometric properties in Blacks and Whites.