Science About What Scientists Do: Distinguishing Between Explanations for Causal Events

The causal reasoning literature suggests that hypothesis testing will only include tests that support a particular hypothesis (confirmation bias), rather than distinguish between possible hypotheses (Wason, 1960; Mynatt, Daugherty, & Tweney, 1977; Klayman & Ha, 1987). Self-generated hypotheses should elicit a stronger confirmation bias than other-generated hypotheses, possibly because the generation of a hypothesis requires an initial assessment of plausibility (Schunn & Klahr, 1993). Plausibility is determined by considering a possible cause within a network of prior knowledge about the world (Koslowski, 1996). Our study examines the testing of genuine explanations in non-emotion-laden, complex causal reasoning situations. We predict that the source of the explanation (self or other) and presence of alternative explanations will influence ability to distinguish between two hypotheses. Furthermore, we predict that the incorporation of prior knowledge into the hypothesis test will allow individuals to distinguish more successfully. Sixty subjects (F = 32, aged 18-22) completed a structured interview evaluating explanations that varied on number of explanations present and source of explanation. Ability to distinguish between a target explanation and its complement (but not a genuine alternative) was shown to differ based on the source (self or other) of the target and whether an alternative was provided. Prior knowledge was only used when distinguishing between the target and a genuine alternative. In general, the use of a contrast or covariation test is the best predictor of ability to distinguish between two explanations