Gender Differences in the Coopersmith Self-Esteem Inventory The Incidence of Differential Item Functioning

In a study of 802 16-year-old pupils, Francis ( 1998) found that males scored significantly higher than females on the school short-form of the Coopersmith Self-Esteem Inventory. He concluded that this instrument may be biased in favor of males. The original data employed by Francis have now been re-examined for the incidence of differential item functioning (DIF) by the Mantel-Haenszel, logistic regression, and SIBTEST statistical procedures. The present study has confirmed that 11 items exhibit DIF, on six of which females outperform males and on five of which males outperform females, but no statistically significant collective DIF was found for the set of the 11 items that exhibited DIF. It has been concluded, therefore, that the differences in levels of self-esteem reported by Francis (1998) are the result of genuine gender differences in self-esteem and not to any artifact of test construction. Individual examination of the items that show DIF supports the widely held view that, in general, among males self-esteem tends to be generated personally, whereas among females self-esteem is more dependent on interactions with others.