Relationships between first and third grade classroom instructional practices and child outcomes (i.e. test scores, days absent, and observed child behavior) were assessed in seven Project Follow Through educational programs. The programs chosen represented a wide spectrum of innovative educational theories. The range included two models based on positive reinforcement theory, a model based primarily on cognitive developmental theory, an open classroom model, and three other models drawn from Piaget, Dewey, and the English Infant Schools. Non-Follow Through classrooms were observed for comparison. Results showed that time spent in reading and math activities and a high rate of drill, practice, and praise contributed to higher reading and math scores. Children taught by these methods tended to accept responsibility for their failures but not for their successes. Lower absence rates and higher scores on a nonverbal program solving test of reasoning were attributed in part to more flexible instructional approaches in which children were provided with a wide variety of activities and materials and where children engaged independently in activities, selecting their own groups part of the time. (Author/BRT)