Richard Slavin, of the Center for Research on Elementary and Middle Schools at Johns Hopkins University, is critical of research that claims that reductions in class size improve academic achievement. Slavin states that even where class size was reduced to an average of 16 or less, the average achievement gain was small.
Therefore, he concludes that if the goal is to raise achievement scores, this research does not provide an adequate basis for a policy on reducing class size. However, reduction in class size does not improve student and teacher morale, so there may be other reasons to keep class size down. Slavin reports that data does indicate that dramatic achievement gains can be obtained with one to one tutoring.
Slavin argues that too often ‘meta analyses’ of research (the analysis of many research studies together) indiscriminately lump together all studies on a subject. Slavin uses a technique called “best-evidence synthesis” in which he chooses studies to analyze based on the quality of research methods and compares only those studies which measure the same factors. In analyzing the research on small class size, he included only those studies which:
1. Compared achievement effects on standardized tests of reading or math between alternative size classes in grades K-6 over a period of at least one year.
2. Compared classes at least 30% smaller than the control group and limited to a maximum of 20 students.
3. Randomly assigned students to control and experimental classes or which demonstrated other evidence of equality between the classes at the outset of the experiment.
Only eight studies were found to meet the “best evidence” criteria. Findings from these studies demonstrate that significantly smaller class size does have a positive effect on student achievement, but it is small and there is no cumulative effect. For example, Slavin points out that the small classes in the Nashville, Tennessee, study (1985), a study in which the greatest achievement gains had been demonstrated during the first year, lost their advantage by the end of the second year.
Just why achievement differences between large and small classes are so small is difficult to determine. Slavin suggests that teacher behavior does not seem to vary much with class size. It is possible that there may be very effective techniques which are usable only with small classes, but, at this time, we have no evidence to prove this.
Slavin concludes that education funds can be put to better use in implementing one-to-one tutoring rather than in reducing class size. Of particular importance, Slavin cites evidence that tutoring in first grade (15-20 minutes per day) can prevent the development of reading deficits in “at-risk” children. Tutoring can be provided by a teacher or by a trained parent volunteer or teacher’s aide. Also, greater achievement gains can be produced by implementing effective instructional strategies than by simply reducing class size.
“Class Size and Student Achievement: Small Effects of Small Classes” Educational Psychologist Volume 24 Number 1, pp. 99.
Published in ERN May/June 1989 Volume 2 Number 3