A Multi-Study Analysis of Conceptual and Measurement Issues Related to Health Research on Acculturation in Latinos

Abstract

Purpose: The aims of the study were to evaluate the orthogonality of acculturation for Latinos. Design: Regression analyses were used to examine acculturation in two Latino samples ( N = 77; N = 40). In a third study ( N = 673), confirmatory factor analyses compared unidimensional and bidimensional models. Method: Acculturation was assessed with the ARSMA-II (Studies 1 and 2), and language proficiency items from the Children of Immigrants Longitudinal Study (Study 3). Results: In Studies 1 and 2, the bidimensional model accounted for slightly more variance ( R2Study 1 = .11; R2Study 2 = .21) than the unidimensional model ( R2Study 1 = .10; R2Study 2 = .19). In Study 3, the bidimensional model evidenced better fit (Akaike information criterion = 167.36) than the unidimensional model (Akaike information criterion = 1204.92). Discussion/Conclusions: Acculturation is multidimensional. Implications for Practice: Care providers should examine acculturation as a bidimensional construct.