1988 Volume 12 Issue 3 Pages 77-89
The purpose of this study was to interpret the problems of educational evaluation from the perspective of the Bayesian approach, one which necessarily implies maximization of the subjective expectation of utility. Using this premise, the authors suggest several techniques which should be useful in improving the decision process in the following three educational evaluation situations : (1) In making career decisions, the drawing of decision trees should help in more clearly understanding the situation being faced. It is also suggested that employing a total evidence design will improve precision in the evaluation of uncertainties and in calculating expected utility. (2) In comparing teaching methods, it appears that traditional statistical methods such as the t-test and ANOVA are not really as well suited as are the more general Bayesian methods of calculating predictive distributions and expected utilities. (3) Finally, the authors advocate the use of expected values for sample information (EVSI) in evaluating the quality of the data collected as part of the decision-making process. This measure is extremely helpful in evaluating the appropriateness of educational evaluation itself under a great many different situations.
