The purpose of this study was to demonstrate the use of confirmatory factor analysis to investigate the existence of item bias in an affective scale measuring self concept. The LISREL IV program was used to detect test item bias and factorial invariance. The ethnic groups compared in the study were Black, Hispanic, and White. The obtained factor structure was compared statistically across the three groups. The results of the study indicated a need to consider affective instruments, as well as cognitive instruments, in the study of item bias. If bias were found to exist in an affective scale, recognition of the effects of that bias and the influence on the interpretation of the resulting score would be important. LISREL analysis provided a method to investigate differential response patterns from affective scales simultaneously that could not be objectively analyzed by classical factor analysis. (DWH)