The Reliability and Factorial Validity of a Multidimensional Measure of Psychological Symptomatology in South Africa

Abstract

The Psychiatric Symptom Index (PSI) was originally developed and validated on a large sample of Chicago residents as a 29-item self-report measure of psychological distress (Ilfeld, 1976b). It has a number of advantages that make it suitable for use in community surveys: relative brevity, a clear and meaningful multidimensional structure, and promising initial evidence for reliability and validity. Since there appears to have been little cross-validation of the PSI, the present study set out to assess its reliability and factorial validity in a large, nationally representative sample of white South Africans ( N = 782). The findings indicated good internal consistency reliability, and excellent factorial validity for the four subscales of anger, anxiety, depression, and cognitive disturbance. The PSI scales also showed an appropriate degree of independence from each other except for depression and anxiety where the intercorrelation was somewhat high. It was found that shorter depression and anxiety subscales would be adequately reliable and reduce this overlap. Finally, normative data are reported for use with white South African samples.