Class Size Reduction: Been There, Done That

Published May 1, 1998

A very consistent picture emerges from a new and extensive investigation into research on the effects of class size on student performance: Little systematic gain in performance is achieved by reducing class size. Even the oft-cited Tennessee Project STAR study provides no support for widespread class size reductions, but merely offers hope that kindergarten classes of 15 or less may boost achievement.

“Existing evidence indicates that achievement for the typical student will be unaffected by instituting the types of class size reductions that have been recently proposed or undertaken,” concludes Eric A. Hanushek, professor of economics and of public policy at the University of Rochester. “The most noticeable feature of policies to reduce overall class sizes will be a dramatic increase in the costs of schooling,” he adds, noting that this increase would be “unaccompanied by achievement gains.”

Although reducing class sizes would certainly add more resources to the public schools, incentives and organizational factors within the existing school structure have proven very ineffective at translating resources into improved student performance. Despite a dramatic reduction in pupil-teacher ratios over the past three decades, student performance has remained virtually unchanged.

“Much of the past expenditure growth comes from pursuing exactly the policies being proposed today,” writes Hanushek in his new monograph, The Evidence of Class Size. “If such policies failed in the past,” he asks, “why should we believe that the next round will be any different?”

As well as finding no evidence to support lower class sizes in the U.S., Hanushek finds none in international comparisons, where “extraordinarily large” differences in class sizes occur without commensurate differences in student performance. And in hundreds of school-level analyses and experimental studies, positive and negative effects almost evenly balance each other out, “underscoring the ineffectiveness of overall class size policies such as those being currently advocated.”

Although the 1980s Project STAR is frequently used by policy-makers to support calls for a generalized reduction in class sizes, the actual results of Project STAR provide no such support. The experiment, which involved dramatic reductions in class sizes to below 15 students, provides no evidence on intermediate class sizes of 18-20. Also, the improvement in student achievement produced by small classes was found to occur only in kindergarten as a one-time increase that is unaffected by smaller classes in later grades.

“None of the STAR data support a wholesale reduction of class sizes across grades in schools,” explains Hanushek, who also points out that many questions about STAR’s design, procedures, and methodology remain unanswered because the data from the experiment have not been made generally available to other researchers for their analysis and interpretation.

George A. Clowes is managing editor of School Reform News. His email address is [email protected].