Translation, psychometric evaluation, and preliminary validation of a Zulu version of the SCL-90-R.

Abstract

This study is seen as a step towards addressing the acute shortage of psychometric instruments that can be validly and reliably used in the assessment of Zulu-speaking individuals. The Symptom Checklist-90-Revised (SCL-90-R) , a 90-item multidimensional self-report symptom inventory, was translated into Zulu. A multistage translation procedure, involving back-translation, decentering, and the committee approach was employed. The translated instrument was pretested on a group of Zulu farm workers (N = 12) and revisions made in order to improve its comprehensibility and acceptability to Zulu respondents. The psychometric equivalence of the Zulu and English versions of the SCL-90-R was investigated with bilingual Zulu students (N = 61) . Scale-level factor analysis yielded similar factor structures for both versions, and suggests that the Zulu SCL-90-R may be best utilised as a global measure of psychological distress. Scores at the scale- and item-level were reasonably comparable for the two language versions, although retest effects and apparent bilingual response sets suggest that the bilingual technique may not be a valid means of assessing translation equivalence. Acceptable test-retest reliability and internal consistency measures were obtained, indicating that the translation into Zulu did not adversely affect the reliability of the SCL-90-R. The concurrent validity of the Zulu SCL-90-R was investigated with samples of male psychiatric inpatients (N = 23) and nonpatients (N = 26). The global severity index of the SCL-90-R demonstrated moderate diagnostic efficiency, with a sensitivity of 70% and a specificity of 77%. These results suggest that the Zulu SCL-90-R may be validly utilised for the purpose of screening for mental illness.