Two intermediate elementary schools in the Bay Shore District of Bay Shore, New York, were studied to assess the impact upon student achievement of a program in cooperative learning applied to mathematics, reading, and writing. The schools include the Gardiner Manor and South Country Schools; in the 1986-87 school year, 510 students attended the former, and 449 attended the latter. Twenty-four third-, fourth-, and fifth-grade teachers and their classrooms participated in the study. Assessments were also made to determine the effects of such programming on students' attitudes, self-esteem, and gender and race relations. Demographic analyses were conducted. The study adopted a pre-test/post-test quasi-experimental design. Data collection instruments included the Iowa Tests of Basic Skills, Coopersmith Self-Esteem Inventory, a modified version of "Who Are Your Friends?", student writing samples, and various attitude inventories. Unlike previous studies, cooperative learning in this instance did not prove to be any more effective than traditional educational strategies in increasing students' achievement, enhancing race and gender relations, or improving youngsters' attitudes toward school. The complexity of the experimental design, its compressed nature, and the effectiveness of the existing traditional curriculum may explain this result. Positive gains found for Hispanic students should be investigated further. A 115-item reference list is presented. (TJH)