The psychoeducational assessment of children commonly involves measuring self-concept. Although Hispanic children comprise a large number of special education referrals, there is little research ruling out bias in the construct validity of existing self-concept measures with this population. Although Hispanic children were adequately represented in the norming sample of the Multidimensional Self Concept Scale (MSCS), a recently developed measure of children's self-concept, there has been no examination of bias in construct validity of the MSCS between White and Hispanic student populations. The purpose of this study was to examine whether the factor structure of the MSCS was invariant across White and Hispanic student populations. In this study, we first confirmed Bracken's (1992) six-domain model of the MSCS with a sample of 458 White fifth-, eighth-, and twelfth-grade students. Next, the covariance matrix of MSCS scores from a sample of 246 English-speaking Hispanic fifth-, eighth-, and twelfth-grade students was tested for goodness-of-fit with the structural model derived using the White sample. The MSCS self-concept model was found to be consistent with Bracken's (1992) theoretical model in a sample of White students and was invariant across an English-speaking Hispanic sample.