This discussion of questions related to the most appropriate role for the computer in undergraduate postsecondary education begins by describing the Educational Testing Service's evaluations of two large-scale funded projects intended to stimulate the incorporation of computer assisted instruction (CAI) into undergraduate instruction--the network-oriented PLATO (Programmed Logic for Automated Teaching Operations), and TICCIT (Time-Shared Interactive, Computer-Controlled Information Television), which used a stimulus/response learning model. A summary of the evaluation results indicates that CAI should be integrated into classroom instruction rather than serve as a substitute for it; neither project saved instructor time or proved useful for poor students; and students who were initially stronger performed better. A learning model for higher education drawn from a report prepared for the Carnegie Commission by Rockart and Scott Morton is then discussed. This model provides a framework for the integration of new technologies, and is based on the stages of the learning process and the characteristics of the material to be learned. Related studies, including one by James and Chen-Lin Kulik on research on college teaching, are examined to support the model. The concluding statement suggests that extensive use of CAI will become a reality in the next decade. Sixteen references are listed. (LMM)