Synthesis of Research on Computer-Based Instruction

P rograms for computcr-based instruction (CBI) have come a long way in the last two decades. They are available today in a variety of sophisticated shapes and sizes, and show few traces of their origins in B. F. Skinner's modest, fill-in-the-blanks teaching machines. The programs tutor and drill students, diagnose learning difficulties, prescribe remedies for problems, keep records of student progress, and present material in print and diagram form. Pioneers in CBI believed from the start that the computer would bring students great benefits, such as better, more comfortable, and faster learning; opportunities to work with vastly richer materials and more sophisticated problems; personalized tutoring; and automatic measurement of progress. Benefits for teachers were to include less drudgery and repetition, greater ease in updating instructional materials, more accurate appraisal and documentation of student progress, and more time for meaningful contact with learners. Soon after its introduction, educational researchers started to design evaluation studies to determine whether CBI actually produced such benefits. Although these evaluation studies yielded potentially valuable information about the effects of computer-based teaching, the message from the studies was not immediately clear. One problem was that each evaluation report was published separately, making the total picture difficult to see. Another more serious problem was that studies were never exact replications of one another. They differed in experimental designs, settings, and in the types of computer applications they investigated. The first systematic reviews designed to integrate findings from the various evaluation studies concluded that CBI is effective in raising student achievement. especially when it is used to supplement regular instruction in elemenrtarschools. Vinsonhaler and Bass's res-ies (1972), for example, reported that results from ten independent studies showed a substantial advantage for computer-augmented instruction. Elementarv school children who received computer-supported drill and practice generally showed performance gains of one to eight months over children who received only traditional instruction. Edwards and others (1975) reviewed studies of CBI at various educational levels and in various subjects, and thev also reached positive conclusions about its effectiveness in raising achievement test scores. These reviewers noted that CBI reduced the time it took students to learn. Later reviewers have used meta-analysis to integrate evalution findings on CBI (Glass, McGaw, and Smith, 1981). Reviewers taking this approach apply