A sample of 234 fifth- and 259 sixth-grade students scaled the items of the Piers-Harris, Tennessee, Coopersmith, and Lipsett self-concept measures. The scaling of the Piers-Harris and the Tennessee inventories was examined in reference to their subscales. The present technique placed items on a bivariate plane of two orthogonal dimensions (Love/Hate and Dominance/Submission). The technique evidenced little discriminant ability among the subscales of either measure. The results suggest that subscales do not reflect separate components of self-concept. The use of these measures may provide an assessment of the self-concept; but the subscales connot be interpreted as empirical support for the currently accepted idea that self-concept factors have been accurately identified through the analysis of these measures. Attempts were made to suggest an alternative method for the analysis of the individual items and demonstrate a technique useful during the construction of self-concept measures. (Author/DWH)