This three-year study of early learning programs in Washington (District of Columbia) public schools found that the extension of formal educational experiences to preschool and kindergarten classrooms does little to promote academic preparation and can actually hinder future academic achievement and overall child development. The Vineland Adaptive Behavior Scales were used to measure the progress of three cohorts of children from enrollment in prekindergarten or Head Start through grade 1. Preschool students were exposed to the following educational models: (1) Model M, combining child-initiated and academic, teacher-directed approaches; (2) Model AD, an academic, teacher-directed approach; and (3) Model CI, a child-initiated approach. Kindergarten students were exposed to ModAcK/SE, emphasizing socioemotional goals, and ModAcK, emphasizing academic preparation. The following summary conclusions are presented: (1) Model M is ineffective and should be eliminated; (2) Model CI fosters a higher level of social development and basic skills than the other two models; (3) Model AD places children at a disadvantage for later social development; and (4) ModAcK/SE programs in kindergarten emphasizing socioemotional goals are more effective than ModAck programs emphasizing academics in kindergarten. Statistical data are presented in 38 tables and six graphs. The survey questionnaire and one table of statistical data are appended. (FMW)