The application of the Minnesota Adaptive Instructional System (MAIS) decision procedure by R. D. Tennyson et al. (1975, 1977) is examined. The MAIS is a computer-based adaptive instructional system. The problems of determining the optimal number of interrogatory examples in the MAIS can be formalized as a problem of Bayesian decision making. Two features of the MAIS decision procedure can be improved by using other results from this decision-theory approach. The first feature deals with the determination of the loss ratio "R." A lottery method for assessing this ratio empirically is discussed. The second feature concerns the choice of the loss function involved. It is argued that in many situations, the assumed threshold loss function in the MAIS is an unrealistic representation of the loss actually incurred. A linear utility function is proposed to meet the objections to threshold loss. Whether or not these two innovations are really improvements of the present decision component in the MAIS in terms of student performance on posttests, learning time, and amount of instruction must be decided on the basis of experiments. Research projects for these areas have already been planned. One table and one figure illustrate the decision theory approach. A 38-item list of references is included.
|Name||OMD research report|
|Publisher||University of Twente, Faculty of Educational Science and Technology|
- Statistical Analysis
- Academic Achievement
- Decision Making
- Computer Assisted Instruction
- Bayesian Statistics