USING DATA OBTAINED BY THE QUALITY MEASUREMENT PROJECT ON SOME 70,000 PUPILS, THIS STUDY ATTEMPTED TO PRODUCE EVIDENCE REGARDING THE INFLUENCE OF CLASS SIZE AND CLASS HOMOGENEITY ON ACHIEVEMENT GAINS IN GRADES 7 AND 8. A TOTAL OF 130 ENGLISH AND 135 MATHEMATICS CLASSES CLASSIFIED ACCORDING TO SIZE AND HOMOGENEITY WERE EXAMINED. THE READING COMPREHENSION AND ARITHMETIC TEST SCORES ON THE IOWA TEST OF BASIC SKILLS WERE USED AS THE MEASURES OF ACHIEVEMENT. RESULTS INDICATED THAT GAIN DIFFERENCES IN RESPECT TO CLASS SIZE AND CLASS VARIABILITY WERE GENERALLY VERY SMALL AND INCONSISTENT. BECAUSE TWO-THIRDS OF THE CLASSES STUDIED CONSISTED OF FROM 23 TO 32 PUPILS, THE LARGEST AND SMALLEST CLASSES (LARGER THAN 34 AND SMALLER THAN 24 STUDENTS) WERE ISOLATED FOR SEPARATE COMPARISON. RESULTS CONFIRMED THAT THERE WAS NO INSTANCE OF A SIGNIFICANT DIFFERENCE IN ACHIEVEMENT GAIN EVEN BETWEEN THESE EXTREME GROUPS. ALTHOUGH THESE TESTS DO NOT MEASURE ALL TYPES OF ACHIEVEMENT, THEY DO SUGGEST THAT ATTENTION MIGHT MORE PROFITABLY BE DIRECTED TOWARD REDUCING THE NUMBER OF CLASSES ASSIGNED TO ONE TEACHER THAN TOWARD REDUCING THE SIZE OF THE CLASSES THEMSELVES. (THIS ARTICLE APPEARED IN "THE SCHOOL REVIEW," VOL. 75, NO. 3, AUTUMN 1967.) (DL)