Project activities involved: (1) examination of exploratory measures of social incompetency of junior high school adolescents, (2) development and trial of educational units designed in response to problems in black urban schools, and (3) a compilation and review of cross-sectional Test of Social Inference (TSI) data. Extensive test data was obtained from junior high school students and teachers in two white semi-rural and two predominantly black urban schools. Examination of teacher "labelling" of students as socially incompetent (SI) using behavior descriptions and checklist responses revealed considerable individual arbiter idiosyncrasy. In the semi-urban schools, means for students consensually identified as SI were lower on all experimental tests, outside school activity reports, and home interview ratings and also lower with respect to academic and socio-economic measures. In the urban schools, behavior problem students scored lower on nearly all measures of social awareness, school attitudes, self-concepts, social relations with respect to various groups, classroom teacher ratings and interview reports, and school record comments and file data. Statistical analysis revealed a general nonrelatedness among sets of variables and lack of simplified factor structure. (Author/BJG)