Linking Structures and Sensitivity to Judgment-Relevant Information in Statistical and Logical Reasoning Tasks

It is proposed that in solving statistical and logical reasoning tasks, reasoners form a bilateral linking structure connecting the 2 problem focal categories with each other. This structure includes two links. Each link may be conceived as full or partial. A full versus partial link, relevant to the inference, was predicted to promote unqualified and confident conclusions and to decrease sensitivity to incoming judgment-relevant information. Similar, though weaker, effects were predicted for the converse (irrelevant) link. Undergraduates (N = 248) were exposed to 1 of the 4 possible linking structures and then performed a series of judgments related to pseudodiagnosticity, insensitivity to sample size, judgmental overconfidence, and so on. Strong evidence for the relevant link hypothesis and some evidence for the irrelevant link hypothesis were found.