Realism in metacognitive judgments: Effects of social factors

Abstract: Theoretical accounts and models of calibration typically explain the degree of realism (validity) in confidence and frequency judgments in terms of different cognitive and methodological factors. The present thesis demonstrates that also social factors are of importance for the degree of realism in these types of metacognitive judgments. The results of Study I showed improved realism in pairs’ confidence and frequency judgments of episodic memory reports to be contingent on how the pair collaboration was organized relative to individual effort. Confidence and frequency judgments (Study II and III) as well as ratings of knowledge extent (Study III) were generally lower when they concerned one’s own knowledge, compared with that of another person’s. Concerning the difference in realism between confidence judging one’s own and another’s knowledge assertions, the results of Study II did not support those reported in previous research and suggested that the results of these studies may have been due to methodological factors. Study II also showed that higher confidence and better discrimination ability (resolution) resulted when additional information about another’s answers was provided. Moreover, Study III showed that ratings of one’s own or another’s extent of knowledge were inconsequential for the realism in confidence judgments of the other person’s knowledge statements. This study also showed that the difference in frequency judgments between one’s own and another’s performance could be eliminated within subject. Lastly, partial support was found for the hypothesis that the influence of information about another person’s memory reports on the realism in confidence judgments can be increased by divided attention during study (Study IV). In discussing the implications of the results, it is argued that their full complexity cannot be accommodated by the extant calibration accounts.