Illusory correlation: a function of availability or representativeness heuristics?

The present study sought to investigate the illusory correlation phenomenon by experimentally manipulating the availability of information through the use of the "lag" effect (Madigan, 1969). Seventy-four university students voluntarily participated in this study. Similar to Starr and Katkin's (1969) methodology, subjects were visually presented with each possible combination of four experimental problem descriptions and four sentence completions that were paired and shown twice at each of four lags (i.e., with 0, 2, 8 and 20 intervening variables). Subjects were required to make judgements concerning the frequency with which sentence completions and problem descriptions co-occurred. In agreement with previous research (Starr & Katkin, 1969), the illusory correlation effect was found for specific descriptions and sentence completions. Results also yielded a significant effect of lag for mean ratings between 0 and 2 lags; however, there was no reliable increase in judged co-occurrence at lags 8 and 20. Evidence failed to support the hypothesis that greater availability, through the experimental manipulation of lag, would result in increased frequency of co-occurrence judgements. Findings indicate that, in the present study, the illusory correlation effect is probably due to a situational bias based on the representativeness heuristic.