An aptitude treatment interaction was predicted in which 30 retarded adolescents were expected to profit most from associative training and 30 intellectually average third graders most from conceptual training. Ss received either associative training designed to strengthen direct associations among pairs of conceptually related members of a free recall list or members and their corresponding categories. Associative training facilitated free recall of trained items at both intellectual levels. For clustering data a treatment by groups interaction was found in which nonretarded Ss benefited relatively more than retardates from associative training. (Author)