When Cooperative Learning Improves the Achievement of Students with Mild Disabilities: A Response to Tateyama-Sniezek

The article by Karen Tateyama-Sniezek in the February 1990 issue of Exceptional Children raises the question of whether cooperative learning can have a positive impact on the academic achievement of mainstreamed students with mild disabilities. This is an important question and one that in today's focus on mainstreaming takes on added significance. However, we would like to offer a response to the article that will clarify how cooperative learning can improve academic achievement of these students. It is our view that Tateyama-Sniezek did not provide an adequate balance among the various types of cooperative learning in the theoretical background of her article. There are many types of cooperative learning, some emphasizing students' academic achievement gains to a much greater extent than others (Sharan, 1980; Slavin, 1983, 1990). This overgeneralization is seen in the conclusion, which stated that in terms of the effectiveness of cooperative learning, "school achievement of handicapped students is in fact a goal secondary to improving [their] interactions and relationships" (Tateyama-Sniezek, 1990, p. 436). This statement ignores the differences in cooperative learning programs and the research that has been done on what common components of these programs have a positive impact on achievement. In a previous review of cooperative learning, Slavin (1983) found that two components were necessary for cooperative learning to be effective in producing achievement gains: group rewards and individual accountability. Certainly if these constructs are important for the academic achievement of students without disabilities, they should be considered in a review of cooperative learning and achievement for those with disabilities. To present the positive achievement potential of cooperative learning for students with disabilities, we would like to reconsider the table presented by Tateyama-Sniezek. However, we would like to make a few corrections and additions to the original table to present more information in a clear format. SIGNIFICANT RESULTS First, we present all comparison statistics (e.g., F ratios and t values) and their significance level. In the original article, the author made entries in the "Significance" column for those comparisons that were nonsignificant, leaving blanks for the significant effects. We believe that a more balanced presentation would include the significance levels of all comparisons of achievement in all studies. EFFECT SIZES We believe the table would benefit greatly from presenting the effect sizes of the studies-a methodology standard in meta-analyses, that is becoming common in narrative reviews as well. An effect size provides a measure of the magnitude of difference between the treatment groups in standard deviation units. Effect sizes can be computed from means and standard deviations of the treatment groups (mean treatment minus mean control divided by standard deviation of control), from the test statistic and group sizes, or other information (see Glass, McGaw, & Smith, 198 1). This methodology provides more information to the reader, particularly where small sample sizes may inhibit studies with large differences from attaining classical levels of significance. This often happens in studies of mainstreamed handicapped students, where there are typically only a few such students in each class. In some cases, the original articles did not provide enough information to compute effect sizes. For these studies, we entered a plus, minus, or zero to indicate group mean differences that were significantly positive, negative, or nonsignificant differences, respectively. We do not suggest that researchers throw out stringent requirements for significance, but rather that researchers and, particularly, reviewers include effect sizes as additional information (see Slavin, 1986). MEASURES OF ACHIEVEMENT One minor change in the selection of studies for inclusion in this review is that the achievement measures used as outcomes be given as individual measures to students, and not as cooperative group activities. …