This paper uses the Brunwik theory of probabilistic functionalism as the backdrop for discussion of interrelationships between the individual learner and the design of computer-assisted instruction (CAI). This Brunswik learning model depicts the acquisition of knowledge as governed by the proximity of various cues and distractions. It is also closely related to research on individual cognitive styles. A study was conducted which investigated the effect of different types of delivery of computerized instruction on 102 Virginia Tech graduate students with differing cognitive abilities. Types of presentations used included text only, text with static graphics, and text with animated graphics. Twenty-four questions were given as a pretest, a posttest, and a recall test. Results indicated that individuals did have different results when they were assigned to different presentation methods, although neither cognitive measure nor presentation method could reliably predict results for all individuals or create an artificially intelligent computer program. Further research should study a full model supporting a semester-long class using resources from the entire department. (Contains 7 figures and 43 references.) (BEW)