A longitudinal analysis was conducted of school performance by children enrolled in kindergarten between 1968 and 1978. The total Follow Through Program was studied, as well as each of the seven teaching models, focusing on achievement, school experience type, absence, and prekindergarten data. A base-year and a later year were identified for each comparison. The percentages of children remaining at or above the 50th percentile, nationally, in reading and mathematics achievement were compared for Follow Through versus the non-Follow Through comparison group. Data indicated that only the Educational Development Center Model had a consistent positive, long-term effect on children's post-program achievement (grades 4-9). Two models appeared to have a positive effect on children's achievement during program years: the Behavior Analysis Model and the Philadelphia Process Model. The Follow Through Program, studied as a single unit, appeared to have a positive effect on mathematics achievement during both program and post-program years. Prekindergarten experience or low absence rate appeared to have a long-term positive effect on mathematics achievement participating in four of the seven models. (Author/GDC)