Research reports issued during the last decade and a half on the effectiveness of computer-based education (CBE) are examined. Surveys of research on CBE over this time period are considered rather than individual studies. Thus, this report represents a synthesis of a number of review articles which provide snapshots across two decades of the kind of CBE work that was being done and the impact that CBE was having on learning. Specific questions addressed are: (1) How effective is CBE in promoting learning? (2) Under what conditions does CBE have differential effects? (3) Are some forms of CBE more effective than others? (4) What trends in the effectiveness of CBE are evident over the years? Among the answers are these: CBE is effective in promoting learning. Most studies show CBE to be better than conventional instruction. Young children seem to be affected more powerfully than older learners and low ability more strongly than high ability. Computer-assisted instruction shows a larger impact on achievement than computer-managed instruction. Supplementing instruction with computers may be more effective than providing a total computer environment. Microcomputers are now becoming common and instruction with them will become the dominant form of the future. (JN)