Abstract P243: Measurement Properties of the Short Acculturation Scale: Preliminary Findings from the Hispanic Community Health Study/Study of Latinos

Abstract

Background: Extensive epidemiologic evidence suggests that greater acculturation levels are associated with higher risk of cardiovascular disease (CVD). However, there is little information on the reliability and validity of instruments to measure acculturation across Hispanic/Latino (H/L) groups who prefer to communicate in English or Spanish. Objective: To evaluate the psychometric properties of the Short Acculturation Scale (SASH) to determine whether it is a reliable measure in a large, diverse sample of H/Ls, and whether there is measurement invariance across language groups. Methods: Cross-sectional data were used from the multi-site HCHS/SOL (Bronx, Chicago, Miami and San Diego) and included Cuban, Dominican, Puerto Rican, Mexican, Central or South American background household resident adults (18 to 74 years old; n = 16,331). Households were selected using a stratified two-stage probability sampling design, and door-to-door recruitment with sampling weights calibrated to the 2010 US Population Census. A 10-item abbreviated version of the SASH was administered in English or Spanish per the participant’s preference. Standardized factor loadings were calculated for all item-level confirmatory factor analysis (CFA) and multi-group confirmatory factor analysis (MCFA) models. Results: The overall scale reliability was acceptable in the full sample (α =.90) and for both language versions (α =.78 for English, and α =.85 for Spanish). The Cronbach’s alphas were similar across H/L ethnic groups (ranging from α =.85 for South Americans to α =.89 for Mexicans). Data from the exploratory factor analysis suggested a 2 factor solution, with dimensions of language use and ethnic social relationships. In addition, CFA showed that the 2 factor solution was invariant across the Spanish and English groups. Conclusion: These preliminary findings suggest that the abbreviated SASH scale is reliable across language versions, and comparable (in terms of reliability and factors structure) across H/L ethnic groups.