The Processing of Conflicting Information in a Simulated Tactical Decision-Making Task

Abstract

In a simulated tactical situation, 40 subjects were required to decide which one of three alternative hypotheses relating to future enemy action was the most likely to occur. The decisions involved combining probabilities from three independent sources to arrive at the overall probability associated with each hypothesis. In addition to specifying the hypothesis which had the highest probability, the subject also had to give his estimate of this probability. The independent variable was the type of feedback given after each problem. Although the proper strategy for selecting the correct hypothesis was to multiply the probabilities, most subjects used an adding strategy. This strategy was particularly ineffective on the more difficult problems. There was a general tendency to underestimate high probabilities and to overestimate low probabilities. The effect of feedback on error rate was contingent upon the particular problem type, and probability feedback improved probability estimation.