Measurement invariance of the CES-D-10 in urban adults across Latin America: A multi-country study

Depression prevalence varies substantially across Latin American countries (4.3–21.4%), yet evidence on measurement equivalence of depression screening instruments remains limited. This study evaluated the factorial structure and measurement invariance of the 10-item Center for Epidemiologic Studies Depression Scale (CES-D-10) across eleven Latin American countries using data from 10,207 adults aged 25–64 years participating in the Development Bank of Latin America (CAF) 2019 survey. Sequential multigroup confirmatory factor analysis was conducted using robust weighted least squares estimation with mean and variance adjustment (WLSMV) for ordinal indicators in urban samples from Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Mexico, Panama, Paraguay, Peru, and Uruguay. Competing factorial models were tested. Both a correlated two-factor model (depressive affect and positive affect) and a unidimensional model with correlated residuals between positively worded items demonstrated excellent and comparable fit, as indicated by the Comparative Fit Index. Measurement invariance testing supported configural (CFI = 0.972), metric, and scalar invariance according to established criteria (change in Comparative Fit Index [ΔCFI] ≤ 0.010; change in Root Mean Square Error of Approximation [ΔRMSEA] ≤ 0.015), whereas strict invariance was not achieved. These findings indicate that the CES-D-10 captures a comparable latent construct of depressive symptomatology and maintains equivalent item–factor relationships across Latin American countries. The attainment of scalar invariance supports meaningful cross-national comparisons of latent mean differences, reinforcing the suitability of the CES-D-10 for regional epidemiological surveillance and comparative research.