Validity and reliability evidence for the Rosenberg self-esteem scale with adults in Canada and the United States

Abstract

Although the Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1989) is the most widely used measure of self-esteem, the evidential basis to support the interpretation and use of the RSES scores with community adults is significantly lacking. Rosenberg (1965) defined self-esteem as the self-appraisal of worthiness of an individual’s socialized identity. In order to support this interpretation of the RSES for use with community adults within a research context, the dimensionality of the RSES scores was examined and the expected relationships with theoretically more related (convergent) and less related (discriminant) constructs were provided as validity evidence in a sample of 1,083 community adults. Utilizing confirmatory factor analysis (CFA), a series of models were tested based on the following emergent themes from the literature: (1) unidimensional, (2) bidimensional with distinct latent variables, representing both positive and negative self-esteem, or (3) unidimensional with method effects due to negatively worded (i.e., reverse-coded) items, positively worded items, or both. Results suggest that once method effects are taken into account, due to negatively worded (i.e., reverse-coded) items, a unidimensional model fit the data best. Convergent and discriminant evidence was examined along a theoretically and empirically expected continuum of constructs from highest to lowest expected absolute magnitude of validity coefficients: self-liking, global self-esteem, self- compete…