The purpose of this study was to test training techniques designed (1) to induce conservation of discontinuous quantity in children, and (2) to induce ability to take different social roles. Also tested was this hypothesis: successfully training children in conservation will improve their ability to take different social roles, and conversely, successfully training children to take different social roles will induce conservation in them. A heterogeneous sample population of 103 children (aged 45 to 64 months) was given a battery of tests to measure grasp of correspondence, conservation, physical perspective, and social role play. Children were placed in 8 heterogeneous groups. Each of seven of these groups was given a 6-week training program involving a different combination of three basic conditions: reversibility-reciprocity, physical perspective-taking, and social role-play. The eighth group (control) was given no training. Subjects were posttested on the pretest battery. Results indicate that reversibility-reciprocity training does induce conservation of discontinuous quantities in children, and that improvement in children's social role-taking ability is more closely associated with reversibility-reciprocity training than with other kinds of training. [Not available in hard copy due to marginal legibility of original document.] (Author/WY)