Utilization of Temporally Ordered Evidence in Evaluative and Predictive Judgments

Abstract

90 subjects were presented with 3, 4, or 5 previous semesters of students' academic performance and asked to rate how likely a student is an “above average” student (evaluation group) or rate how likely the student would perform “above average” the following semester (prediction group). Use of evidence at each temporal position was estimated with a linear model, Y = X₁b₁ + X₂b₂ + X₃b₃, + X₄b₄ + X₅b₅, where Y refers to the subject's judgment, X₁ refers to the i_th academic performance, and b_i to the weight, or contribution of the performance in the i_th temporal position. Subjects in the prediction groups used the temporal position of the displayed evidence more than subjects in the evaluation groups for both the 3 and 4 previous semesters' displays but not for the 5-semester display. The results were interpreted as supporting an hypothesis that temporal features of the evidence are minimally relevant unless the task requires involvement of the temporal dimension such as making a prediction to a future point in time.