摘要
Subjects were asked to judge interevent relations exhibited in 2 × 2 contingency tables such as the following: number of plants which bloomed when given a fertilizer (Cell A), number of plants which did not bloom when given the fertilizer (Cell B), number of plants which bloomed when not given the fertilizer (Cell C), and number of plants which did not bloom when not given the fertilizer (Cell D). Subjects rated the extent to which the fertilizer had a positive or a negative effect on the plant′s blooming. Optimally, equal weights should be given to each cell. Several analytic techniques were developed in Experiment 1 to determine the subjects′ actual weightings: pairwise comparison of problems differing in only one cell; group- and individual-level analyses of variance; and regression analysis. Converging evidence across problem contexts and analytic techniques revealed that most subjects′ judgments were unequally affected by evidence in the four cells: Cell A > Cell B > Cell C > Cell D. Experiment 2 examined the processes underlying this biased weighting pattern. Subjects′ tendency to overweight the information in Cell A was related to their a priori hypothesis about the direction of the interevent relation. The weight given to Cell A was significantly greater for subjects who expected a positive relation than for subjects who expected a negative relation. © 1993 Academic Press, Inc.