The purposes of this study were to see if the items from the Piers Harris Self Concept Scale and the Coopersmith Self Esteem Inventory had construct and predictive validity. Items used in this study were 50 items from the Coopersmith Self-Esteem Inventory and 80 items from the Piers Harris Self-Concept Scale. Construct measures were obtained using 18 items from the Children's Semantic Differential providing measures of Self-Esteem, Idealized-Self and Empathic-Self-Concept. Ratings were obtained on all items on a five point scale from a sample of 381 children in grades 4-6 in suburban Boston schools. Each of the two tests were subjected to separate Principal Components Factor Analysis and Varimax Rotation. The three semantic differential measures were used in each of the two analyses. A multiple regression was made of the Piers; Harris and Coopersmith items that loaded substantially with the construct measures. The results of this analysis yielded only one item out of the 130 total for the Piers Harris and the Coopersmith which was a significant predictor of the Semantic Differential construct measures. This result questions the predictive and construct validity of the Piers Harris and Coopersmith measures. (Contains 3 tables.)