Reading too much between the lines: illusory correlation and the word association implications test

We examined the illusory correlation phenomenon with the Word Association Implications Test (WAIT), a task where diagnoses and signs are causally connected due to priming effects. The WAIT is an analogue to clinical assessments in which subjects “read between the lines” of target persons’ word associations which have been primed by fantasized scripts. 164 undergraduates were randomly assigned to study WAIT protocols with either 0, 30, 70, or 100% of the targets veridically identified. Following subjects’ examination of WAIT protocols, we assessed their incidental learning of valid diagnostic clues (i.e., their clue schemata). Subjects given no veridical diagnoses showed minimal incidental learning. However, those given 30, 70, and 100% veridical diagnoses showed equivalent incidental learning of diagnostic clues and all exceeded an intuition (no experience) comparison group. The results suggested that an illusory correlation operated even when clues and diagnoses have causal, not just contingent, connection. Successful judges must contend not only with others’ tools for avoiding prediction but with their own tendencies to read too much between the lines.