Identification Of Differential Item Functioning Of Self-Esteem In A Sample Of High Achieving Minority Students

No Thumbnail Available

Links to Files


Author/Creator ORCID






Doctor of Philosophy

Citation of Original Publication


This item is made available by Morgan State University for personal, educational, and research purposes in accordance with Title 17 of the U.S. Copyright Law. Other uses may require permission from the copyright owner.


Increasingly, noncognitive variables, such as motivation, self-efficacy and self-esteem, are utilized in academics, employment and accountability. Unlike cognitive variables such as the SAT or GRE, noncognitive variables are not extensively evaluated to ensure measurement invariance. When groups (i.e., gender, ethnicity, SES) of equal ability have varying likelihoods of endorsing or selecting items, the items are said to display differential item functioning (DIF). The purpose of the present study was to determine if DIF was present in items from the Rosenberg Self-Esteem Scale (RSES). The Rasch measurement model was utilized to answer four research questions. Data was found to fit the Rasch model, with fit statistics meeting expectations. Results of the study indicated the presence of DIF by gender, race/ethnicity and the interaction of gender and race/ethnicity. One item from the RSES displayed DIF by gender, with males having more difficulty endorsing the item “I feel I do not have much to be proud of.” Two items displayed DIF by race/ethnicity. African American respondents had less difficulty endorsing “I feel that I'm a person of worth, at least on an equal plane with others” and “On the whole, I am satisfied with myself” than American Indian and Hispanic American respondents, respectively. Five out of the six RSES items displayed DIF by race/ethnicity x gender resulting in 18 significant DIF comparisons. The items “I feel I do not have much to be proud of” and “I am able to do things as well as most others people” resulted in the most DIF. Results of the racking analysis indicated that the six RSES items displayed stability (i.e., little to no changes in item endorsability) between two time points. Results indicate further research is needed to ensure measurement invariance for noncognitive measures. Intracultural comparisons in noncognitive assessment may result in differences due to ethnic, cultural and gender related influences.