Learning theory and knowledge structures in computer-aided instruction

The development of computer-aided instructional (CAI) systems suffers from a lack of a cohesive theory of learning--how do students acquire and store knowledge? From studies of computer systems that learn and tutor, we can infer generic activities that appear to be integral parts of the learning process, such as aggregation, clustering, characterization, and storage for later retrieval. Learning is faster and more efficient if the goal of a task is made explicit. Hints should be given with the correct timing in relation to an objective so that students can advance in their own problem-solving strategies with the prerequisites in mind. The general form of a rule should usually be taught first, followed by exceptions and special instances. We review theories of learning associated with CAI that illustrate the classification of different types of knowledge. Rule-based (if-then) knowledge forms are represented in these theories, as are declarative and causal knowledge structures. Extracting the common themes from different classifications of knowledge may help us create better CAI.