The Social Physique Anxiety Scale: Cross Validation, Factorial Invariance, and Latent Mean Structure

Abstract

Recently Motl and Conroy (2000) established the factorial validity and invariance of a 7-item, unidimensional model to the Social Physique Anxiety Scale (SPAS). This study was a test of the cross-validity of the 7-item, unidimensional model to the SPAS across 2 samples of women. The factorial invariance and latent mean structure were examined for the 7-item, unidimensional model to the SPAS across men and women. Data (N = 1,053) were retrieved from 2 published studies (Eklund, Mack, & Hart, 1996; Eklund, Kelley, & Wilson, 1997), and consisted of 3 samples of women (ns = 140, 380, and 380) and 1 sample of men (n = 153). Confirmatory factor analysis (CFA) was used to indicate that the unidimensional model to the 7-item SPAS demonstrated invariance of the variance-covariance matrices, factor structure, factor loadings, factor variance, and item uniquenesses across 2 samples of women and across men and women. CFA was used to indicate that women had a higher latent mean score on the 7-item SPAS than men. Therefore, evidence of "tight" cross-validity and factorial invariance across gender was provided for the unidimensional model to the 7-item SPAS. Expected gender differences in SPAS latent mean scores were supported. Researchers can confidently employ the 7-item SPAS in studies of social physique anxiety among teenagers and young adults.