AN EXPERIMENTAL COMPARISON OF DISCOVERY AND DIDACTIC COMPUTERIZED INSTRUCTIONAL STRATEGIES IN THE LEARNING OF COMPUTER PROGRAMMING

The psychological theories of Piaget and Bruner suggest the value of exploratory discovery-oriented instructional experience. The theories of Skinner and Ausubel emphasize more systematically controlled expository instruction. Experimental studies comparing discovery and didactic instruction have been hampered by two major methodological weaknesses: (1)artificial laboratory implementation of discovery learning has unreasonably constricted the operationalization of discovery processes; and (2)instructional effectiveness has typically been measured by a single achievement posttest rather than a range of outcomes.;Outcome differences among discovery treatments were minor but differences between discovery and didactic instruction were substantial. Didactic programmed learning resulted in significantly better reception and recall of programming facts and rules on a multiple-choice posttest. Discovery learners, however, were significantly better able to write original programs and tended to score higher on posttest transfer problems requiring interpretation or extrapolation. Discovery instruction also led to significantly higher self-confidence ratings on correctly answered posttest items and to significantly higher levels of expressed interest in continuing instruction in additional educational settings.;There were no ability by treatment interactions on cognitive outcomes, but low ability learners expressed significantly more positive affect about instruction after discovery treatments than after programmed learning. Learners with an "internal" locus of control did better under discovery instruction while "externals" did better under didactic instruction. The pattern of this interaction was remarkably consistent across all eleven outcome measures and reached statistical significance for measures of risk taking, continuing motivation, posttest effort, and positive affect.;The findings of the present study lend empirical support to important expectations of discovery learning theory and demonstrate the usefulness of computer technology in comparative instructional research.;The present study used a computer to teach introductory computer programming to 81 college undergraduates using either a discovery or didactic instructional strategy. The didactic strategy used a programmed learning sequence. The expository question-answer format of this approach is representative of many current instructional applications of Computer Assisted Instruction (CAI). The discovery strategy used a computerized discovery learning environment based on instructional heuristics and epistemological assumptions of the discovery approach. The discovery environment consisted of short model programs which could be learner-modified and executed. During program execution, the computer screen displayed simulated dynamic computer transactions. The amount of personal initiative required of discovery learners was varied by prohibiting or encouraging the learner to initiate their own changes in the model programs. Instructional treatments (approximately one hour long) were compared for effectiveness on a range of cognitive, affective and motivational outcomes. Two individual difference variables, I-E locus of control and cognitive ability were examined for possible aptitude by treatment interactions (ATI).