A two-part follow-up study was conducted of students in grade 4 who had participated in the reduced class size experiment during the preceding years. Outcome measures included a range of norm referenced and criterion referenced achievement tests and teachers' ratings of the youngsters' effort, initiative taking, and disruptive behavior in the fourth grade classroom. Statistically significant carry-over effects of small classes were found on every achievement measure. Significant differences were also found in participation levels between students who had attended small classes and those in regular classes, but these were not as consistent as the impact on cognitive performance. The confirmed consequences of placing youngsters in classes with 12 to 17 pupils need to be supplemented by research on other outcomes in order to compare the complete range of benefits with the relatively large monetary costs that could be incurred. There are few organizational features of the classroom that can be easily (not necessarily inexpensively) rearranged with the assurance that student performance will be enhanced. Reducing class size is one of these. The results of Project STAR show clearly that average pupil performance in the primary years can be increased by approximately one fourth to one third of a standard deviation without the introduction of new materials or curricula and without retraining the teachers. Other organizational approaches for facilitating learning include the use of within-class ability grouping, individualized instruction, cooperative learning groups, and peer tutors. Peer tutoring can be highly effective, and both ability grouping and individualized instruction have somewhat larger effects than reduced-size classes in the elementary grades (Slavin, 1989). But these alternatives also require at least some reogranization of course content, teaching strategies, and/or class scheduling; that is, they are not quite as "easy" to implement, at least within the classroom. Further the effects of these strategies are usually limited to one content area at a time. The ability groups or individuals who receive tutoring in reading may not be the same individuals as for mathematics, or the approach might not be used for mathematics, social studies, or science at all. In contrast, the effects of reduced-size classes were found on every achievement measure administered in Project STAR. Had a wider range of tests been used in the evaluation, these would have most likely shown increments of roughly the same magnitude as well. To realize performance gains as extensive as this through any combination of student grouping, individualized instruction, or tutoring would be both difficult and expensive, if it were even possible to implement or maintain such an approach. The pervasiveness of the effects found in kindergarden through grade 3 may well have impacted on the pupils' behavior when they returned to regular size classes in grade 4. This issue was addressed through a two-part study of the Project STAR youngsters during fourth grade. Part A examined the question: "Is the small-class advantage in achievement maintained when pupils return to regular size classes?" Part B investigated the classroom behavior of pupils in grade 4 who were in small or regular classes during the preceding years. The samples for the two studies had a high degree of overlap but the data collection procedures for the two were somewhat different.
