This study shows the usefulness of a classroom observation measure of social and mastery behaviors in a small longitudinal study. A collaborative day care and public school intervention program for children in public housing was evaluated with the Bronson Social and Task Skill Profile. The intervention group and comparison children from the same classrooms, were observed four times over a two year period in their prekindergarten and kindergarten years. Comparison children came from the same neighborhood, but were economically more advantaged had few minority group members. The results demonstrate a progressive $quot;catch-up$quot; effect in most observed behavior categories for the intervention group in relation to the comparison sample. Very poor performance in a few children, eligible for the intervention but not served, underlines the effectiveness of the intervention. Observational assessment of both social and mastery variables provided valuable information on areas of skill and weakness in the samples.