The second of two reports on a national evaluation of the immediate effects of Project Head Start describes the characteristics of children, families and programs in samples of full-year classes operating in 1966-67 and 1967-68. No control groups were used; comparisons were made within the Head Start sample to see what kinds of classroom experiences "work best" for what kinds of children. Where appropriate in this report, comparisons are made across years concerning the patterns of child-program interaction effects. Chapters 1 through 10 concern the 1967-68 data: (1) design; (2) measures for analysis; (3) the children: entering characteristics; (4) their families: entering characteristics; (5) programs--class and site characteristics, observed classroom use of materials, observed classroom activities, teachers' education and experience; (6) gains associated with Head Start--cognitive and social-emotional measures, parent attitudes; (7) differences in gains for different types of children; (8) differences in gains associated with different program approaches; (9) relationships between program approaches and performance for different subgroups of children; and (10) summary. Chapter 11 presents similar data for the 1966-67 evaluation. Appendices provide frequency distributions for scores on dependent variables for both years. References are provided. (KM)