About 180 Negro Head Start children in Dade County, Florida, were tested (1) to discover if the county's program contributed significantly to language skills, social skills, and self-concept development and (2) to determine if an efficient instrument could be developed to measure self-concept in the disadvantaged child. Pretests and posttests used were the Children's Projective Pictures of Self-Concept, the Preschool Attainment Record, the Self-Concept Rating Scale, and an anxiety scale. In the posttest phase, 20 subjects in a proportional random selection were compared with 20 control children (from the same districts) with no preschool experience. The Head Start sample performed significantly better on tests measuring social skills, language skills, and self-concept. The meaning of the significance was unclear because of the possibility of an interaction effect between the pretest and the experimental variable; namely, exposure to the Head Start program. It was suggested that Head Starters entering the first grade the following fall be tested and that their performance be compared with the performance of a control group with no preschool experience. (DO)