| Abstract: | Four experiments examined change in confidence after hearing two sides of a dispute. The results showed that a case independently judged to weakly support one side often increased confidence that the opposing side was correct. Furthermore, the stronger the first case, the more likely a subsequent weak case had a reverse impact. Traditional belief‐updating models, which tend to focus on change in belief after individual pieces of evidence rather than entire cases, cannot account for these results, and a model that can account for them is introduced. In the new model, case strength is evaluated with respect to a relatively demanding (and malleable) reference point. A weak case can fall below this demanding reference point, resulting in a reverse impact on confidence. Cases must exceed relatively high strength thresholds in order to have their intended impact because they are expected to be biased summaries of evidence. When it is clear that a weak case is unbiased, it affects confidence in the intended direction. Copyright © 2002 John Wiley & Sons, Ltd. |
