Cause typicality and the continued influence effect.

Abstract

A large body of research has focused on whether and how readers update their knowledge of events when an initial piece of causal information is corrected. These studies have indicated that corrections can reduce, but do not eliminate, readers' reliance on the initial cause when drawing inferences or making decisions about the events (i.e., the continued influence effect). Additional studies suggest that supplementing a correction with an alternative cause can further reduce reliance on discredited initial causes. In three experiments, we interrogated the importance of cause typicality for the generalizable utility of these correction strategies by manipulating the typicality of initial and alternative causes. We found evidence that participants showed greater reliance on a typical than an atypical initial cause both before and after correction, but no consistent evidence that this typicality impacted the effectiveness of the correction. Furthermore, the typicality of the alternative causes used to supplement a correction did not seem to matter with respect to updating. These results highlight the importance that characteristics of an initial cause can have for event encodings and corrections, identifying critical boundary conditions for understanding the effects of corrections on knowledge revision and the continued influence effect. Public Significance StatementNews reports often provide detailed information that changes as information is gathered. This study demonstrates that people more strongly rely on event causes that are typical of their expectations as compared to atypical. However, this increased reliance may not interfere with people updating their event understandings after the causes are refuted. In addition, refutations that included atypical or typical alternative causes were similarly effective. These findings highlight the utility of corrections for informing people's understandings of events.