Conducted a 6-year followup of almost 4,000 students in Project STAR in Tennessee, a 4-year, large-scale randomized experiment on the effects of class size. Though follow-up data could not be obtained on more than one-half of the students, class size effects persisted for at least 6 years and remained large enough to be of importance for educational policy. Results show the lasting benefits of class size reduction in early grades. (SLD)