Seventy-five 2nd graders were randomly assigned to cooperative learning groups in which oral discussion was structured or unstructured or to an individualistic learning group so that daily achievement, postinstructional achievement, and retention could be compared. Groups were stratified for sex and ability level. Results show that Ss in cooperative groups performed significantly higher on the accuracy of daily work than did Ss working individually. In addition, the high-, medium-, and low-ability Ss in the structured oral discussion cooperative condition scored higher on the postinstructional and retention tests than did Ss in the other 2 conditions. Ss in the unstructured oral discussion cooperative condition scored higher on these tests than did Ss who learned individually. Findings suggest that group-to-individual transfer takes place within cooperative learning groups and that orally summarizing the material being learned and the monitoring of others' summaries contribute to the efficacy of cooperative learning.