We present a psychometric evaluation of the Center for Epidemiologic Studies Depression Boston Form (CES-D-B) for use with different Latino subgroups as there is inconsistency regarding its performance across subgroups of Latinos, a large and rapidly growing cultural group in the United States. We evaluated the reliability and structural validity of the scores generated by the CES-D-B using four distinct Latino samples residing in US: Mexicans, Venezuelans, Cubans, and "other Latinos" (total N = 1033). To further explore structural validity of CES-D-B scores, we conducted measurement invariance analyses across different countries of origin, gender groups, educational levels, and languages of assessment (English, Spanish). For all four samples, CES-D-B scores were highly reliable as indicated with the coefficients ranging from 0.82 to 0.88, and the factor structure provided an adequate fit to the data with the fit indices CFI/TLI ranging from 0.96 to 0.99, RMSEA estimates between 0.02 and 0.07, and SRMR estimates between 0.02 and 0.04. While measurement invariance analyses for different educational levels indicated scalar invariance across all samples, the same level of measurement equivalency was achieved only for Mexicans and Venezuelans with varying gender and languages of assessment. The findings indicated that CES-D-B scores are internally consistent, possess a strong four-factor structure, and have somewhat equivalent psychometric properties across diverse Latino groups. Findings from this study highlight the importance of considering gender and languages of assessment when assessing depressive symptoms of various Latino subgroups.
Read full abstract