This study of the effects of computer-assisted instruction (CAI) on the academic achievement of educationally disadvantaged students in grades 2 through 8 in four urban schools tested three hypotheses: (1) supplementary CAI programs are significantly more effective than non-CAI supplementary instruction approaches for disadvantaged students; (2) school-based supplementary CAI programs are significantly more effective than district-based programs for this group; and (3) vendor-based programs are less effective than programs developed within the system by school personnel. The effects of three types of CAI programs are compared--a vendor supplied microcomputer system, a system-wide CAI program, and a school-based system--with both experimental and control groups for each program. The areas of study included reading comprehension; vocabulary; and mathematical concepts, problem-solving, computation, and composite skills. Analyses of the data failed to produce sufficient evidence to support the contention that disadvantaged students learned more when exposed to CAI, nor was there strong evidence that the impact of CAI varied with the system or the approach used. This report includes the purpose and objectives of the study; information on sample selection and procedures for data analysis and interpretation; study findings; and recommendations for future studies. Tables display data on the study sample by school, program, sex, and grade; pre- and posttest scores; and estimated posttest mean scores. Appendices provide a two-page reference list and analyses of the data on the six areas of study. (DJR)