Testing for measurement invariance across gender in the 12-item CES-D: An investigation among a nationally representative sample of African Americans.

Abstract

Previous studies have conducted tests of measurement invariance (MI) when comparing depressive symptoms across race groups. However, investigations that explore potential within-group differences remain scant. Thus, the current study sought to uncover whether noninvariance exists when using the 12-item CES-D to measure depressive symptoms in African American women and men. Responses from a nationally representative sample of 1,271 African American men and 2,299 African American women who participated in the National Survey of American Life (2001-2003) were analyzed. Exploratory and confirmatory factor analysis were used to first assess dimensionality of the 12-item CES-D. Next MI tests were performed to determine whether the factor structure differed for men and women. The 2-factor model was retained, and partial metric invariance was achieved. Noninvariance was found in 2 loadings and some of the thresholds. Results suggest that noninvariance may help explain disparate rates of depressive symptoms among African American men and women. Additional consideration should be given before comparing latent mean scores across groups. (PsycInfo Database Record (c) 2021 APA, all rights reserved).