CitationBenson, Mark J.; Faas, Caitlin; & Kaestle, Christine E. (2010). Constructing reliable multi-item scales: Wave I prediction and Wave IV outcomes. 2010 Add Health Users Conference. Bethesda, MD: University of North Carolina at Chapel Hill, Carolina Population Center.
AbstractMulti-item scales hold particular value in assessing the complex contextual variables in adolescents’ lives. Using a systematic procedure, we indentified 30 underlying multi-item scales from Wave I for use in predicting Wave IV outcomes. Stage 1 centered on descriptive analyses of conceptually similar items; stage 2 produced results from a series of principal components factor analyses; and stage 3 used confirmatory factor analysis (CFA) and structural equation modeling to identify correlated errors and patterns across scales. The study procedures yielded 30 multi-item scales from Wave I based on parent, youth, or interviewer reports. Alphas exceeded .80 for 9 scales and surpassed .70 for 26 scales. The 30 scales included measures of: family processes (7), peer relations (2), school performance (4), neighborhood quality (2); internal processes (5), physical well-being (2), sexuality (4), externalizing behavior (4), and social impression (1). The full presentation provides a detailed handout on the conceptual underpinnings for each scale, the syntax for scale construction, internal consistency estimates, results from CFA, and findings illustrating links between Wave I scales and Wave IV outcomes.
Reference TypeConference proceeding
Book Title2010 Add Health Users Conference
Author(s)Benson, Mark J.
Kaestle, Christine E.