| Abstract: | Two experiments were conducted to examine the strategies people use to determine the cue weights in a two-cue prediction task. In order to examine the subjects' strategies, half of the subjects in Experiment 1 were asked to state the weights explicitly for the two cues on each trial rather than directly estimating the criterion value. The subjects' estimates of the criterion were then calculated on the basis of the estimated weights. The results of Experiment 1 suggested that most subjects use a hypothesis-testing approach to determine cue weights. In addition, it appeared that most of the subjects randomly tested hypotheses about the relative importance of the cues, which did not prove to be an effective strategy. The results of Experiment 2 indicated that failure to learn the relative importance of the cues was not due to failure to test the appropriate cue weights at some point during the task, but to failure to test the hypothesized cue weights on enough trials to evaluate their adequacy sufficiently. |
