System Support for Hypothesis Generation

Abstract

An experiment was conducted on the usefulness of a statistical summary in judging enemy intentions, in the context of a nuclear attack. The statistical summary was based on applications of correlation, factor groupings, and variance ratios. Thirty ROTC students served as Ss, playing the role of intelligence analysts in a simulated command post. Analyses of variance of scores based on analysts' judgments support two major inferences: use of the summary tends to increase the over-all goodness of judgment; and, use of the summary tends to enhance the assimilation of experience as measured by improvement in judgment scores from early to late trials.