Experimental Assessment of Delphi Procedures with Group Value Judgments

One of a series of studies using Delphi procedures to aid decisionmakers in dealing with value judgments. Previous studies have not clearly shown that there is an appropriate population of factual questions to compare with value judgments; the variability of performance on factual questions is large, depending on the type of questions asked. With this in mind, some comparisons were made: Two groups of UCLA students were asked to generate and rate lists of value categories that they considered important to higher education and the quality of life. Analyses showed that (1) distributions were generally single-peaked and roughly bell-shaped, (2) the correlations between different groups and different rating methods were high, and (3) the number of changes and degree of convergence for value judgments (reduction in standard deviation) were comparable to similar indices for factual judgments. The experiment supported the conclusion that Delphi procedures are appropriate for processing value material as well as factual material. (See also RM-5888, RM-5957, RM-6115, RM-6118.)