Psychometric evaluation and comparison of three retrospective, multi-item measures of childhood sexual abuse

Abstract

Objective: To evaluate and compare the psychometric qualities of three retrospective, multi-item measures of childhood sexual abuse (CSA): the CSA Frequency Scale, CSA Count Index, and CSA Multiple Characteristics Index.

Methods: Two samples of women 20-50 years old who experienced CSA were recruited from a family practice clinic (N=132) and the community (N=19). The measures were designed to represent the three most commonly used approaches for multi-item measurement of past CSA, as determined by a review of literature. Items were derived from a self-administered instrument from a larger study. The CSA Frequency Scale was evaluated for test-retest reliability, internal consistency, and criterion-related validity. The CSA Count and CSA Multiple Characteristics Indexes were evaluated for content specification, indicator specification, indicator collinearity, external validity, criterion-related validity, and test-retest reliability. Criterion-related validity variables were (a) physical and psychosocial symptomatology and (b) depression.

Results: Upon evaluation, all three measures met criteria for satisfactory quality, with one exception: external validity for both indexes. When compared across test-retest reliability and criterion-related validity results, none of the measures proved superior. Intraclass correlation coefficients for test-retest reliability ranged from .92 to .94; correlations with the criterion-related validity variables ranged from r=.36 to .41.

Conclusions: The results support the continued use of established scales similar to the CSA Frequency scale. In addition, the results suggest further development and standardization of indexes similar to the CSA Count and Multiple Characteristics Indexes is warranted.