RELIABILITY AND VALIDITY OF THE SYMPTOM C HECKLIST (SCL-90-R, DEROGATIS, 1986) IN STANDARDIZED VERSUS HOMOGENOUS ITEM-BLOCKED SEQUENCE

In 1992 Bullinger reported the need for economic means of assessing of psychological data in medical areas. The Symptom-Checklist (SCL-90-R; Derogatis, 1986) may be able to meet this requirement, because it is a multidimensional symptom self-report inventory comprised of 90 items , each rated on a five-point scale of distress, and it is scored on ni ne primary dimensions plus three global indices. Until now there has b een lack of investigations on the SCL-90-R in German populations, espe cially in clinical and non-clinical groups. Also, questions about the stability and validity of the original factor structure need to be ans wered. Based on responses of 1006 members of a German university, in s tudy I the reliability and validity of the SCL-90-R were analyzed. The usefulness of the instrument was demonstrated, and an experimental ei ght-factor-solution is presented. Empirical results of studies II and III are presented on the widespread practice of extracting subscales f rom extensive symptom inventories and presenting their items as more o r less homogeneous item-blocks in questionnaire applications (Krampen et al., 1992). Studies II and III were conducted in different contexts (study II, controlled group testing; study III, noncontrolled individ ual testing) and in different groups (study II, 123 student nurses, st udy III, 143 university students), in which the standard item arrangem ent (A) was contrasted with the presentation of the items of each orig inal subscale (B) or of each experimental subscale (C), which matches the extraction of subscales from more extensive inventories. The resul ts confirm that there are significant effects of item blocking on the reliability, the means, and the intercorrelation of the subscales. The admissibility of subscale extraction is doubtful for the SCL-90-R bec ause the reliability and validity parameters of the scales change mark edly.