Abstract
In the present study, we experimentally clarified whether assessing other-generated hypotheses could facilitate hypothesis revision using a rule-discovery task. In the first phase, participants were asked either to generate their own hypotheses and assess them or to evaluate the hypotheses presented as other-generated hypotheses. In the second phase, all participants were required to generate their own hypotheses and evaluate their plausibility. The results showed that participants who assessed the other-generated hypotheses prior to generating and assessing their own hypotheses produced better performance than those who generated their own hypotheses and assess them thoroughly. In addition, while confidence on their own hypotheses became slightly higher as the participants saw instances in the Self-generated hypothesis condition, confidence on the hypotheses became lower in the Other-generated hypothesis condition. It was concluded that the source of hypotheses affected the confidence on the hypothesis, especially after the participants faced the counterevidence. In other words, other-generated hypotheses were less confident than self-generated ones after the data inconsistent with the hypotheses was presented, which enabled the participants to revise hypotheses more easily in the other-generated hypothesis condition than in the self-generated hypothesis condition.
