Class Size

The Research

In the 1970s, Gene V. Glass and his associates conducted a series of meta-analyses of previous studies and found that smaller classes did improve student achievement across all grade levels and subject matter. They found an even greater effect on nonachievement outcomes, such as the instructional environment and effects on teachers. But their meta-analyses were criticized, and the conclusions were refuted by many.

Among the critics were Glen E. Robinson and James H. Wittebols, who published a review of 100 class size studies that had been conducted between 1950 and 1985. They concluded that the research did not support an optimum class size for all situations. However, they did find evidence that students in the primary grades benefited from smaller class sizes for reading and mathematics. Also, they concluded that students of lower ability and those of minority ethnicity or from impoverished homes benefited from smaller class sizes. They found that changing instructional approaches and teaching practices were more effective than reducing class size in improving achievement. The research they reviewed did show that class size was a major concern of teachers and of the public and that students in the primary grades behaved better in small classes. Finally, they found that class size had a large impact on school budgets.

Although there have been a lot of studies on the relationship between class size and student achievement, most of them have been nonexperimental in design. Nonexperimental studies inspire less confidence in their results than controlled experiments do. Some of the studies in which class size was an independent variable and student achievement was the dependent variable showed a relationship, while others did not. Studies have varied a great deal in the way they defined small and large classes; a class size of say, 20, might be considered small in one study and large in another. Teaching methods and styles also varied in the different studies; thus, it was difficult to draw conclusions across the studies.

The Tennessee STAR study was a great improvement over previous studies because it was large, with over 6,000 students across the state of Tennessee, and because of its strong research design. The design was a true experimental design over a 4-year period. Students and teachers were randomly assigned to a small class (13 to 17 students), a regular class (22 to 26 students), or a regular class with a teacher aide. Results of standardized tests showed that students in the small classes performed better than students in the other classes in all subjects and across Grades K through 3. Follow-up studies showed that students who had been in small classes in the early grades maintained their advantage in Grades 4 through 7. There were no significant differences between the regular-sized classes with an aide and the regular-sized classes with no aide.

...