Factorial validity and invariance of the center for epidemiologic studies depression (CES-D) scale in a sample of black and white adolescent girls

Meaningful comparison of depression symptoms requires that the measurement instrument has equivalent measurement properties among racial and ethnic groups. We tested the factorial validity and invariance of the Center for Epidemiologic Studies-Depression (CES-D) Scale among Black (n=610) and White (n=452) adolescent girls in the 12th grade. The invariance analyses were conducted by using LISREL 8.70 with Maximum likelihood estimation and Satorra-Bentler scaled chi-square statistics and standard errors. The hypothesized second-order model (first-order factors: depressed affect, somatic and retarded activity, interpersonal, and positive affect; second-order factor: depression) demonstrated good overall fit in both groups. Comparison of nested models supported the between-group invariance of the overall factor structure, first- and second-order factor loadings, first-order factor variances, and the second-order factor variance. Item uniquenesses were not invariant. Our results support the hypothesis that a meaningful comparison of composite CES-D scores can be made between Black and White girls in the 12th grade.