Factorial Invariance and Self-Esteem: Reassessing Race and Sex Differences

Abstract

Either sound theoretical explanations or strong empirical support has been lacking in studies examining race and sex differences in self-esteem. Most theories fail to explain why blacks generally score higher than whites on self-esteem and many studies have failed to find differences in self-esteem between males and females. These problems suggest an alternative explanation for reported differences in self-esteem: scales purporting to measure this concept may be measuring something different within the groups being compared. This explanation is tested for the Rosenberg and Simmons' Self-Esteem Scale using a confirmatory factor model with structured means for groups separated on the basis of race and sex. The results indicate that for this particular scale, the items appear to reflect similar meaning across race and sex. Although differences in variation of the underlying factor and measurement errors were found between groups, this had little effect on mean differences in self-esteem: blacks were found to score higher than whites and males scored higher than females.