Group Judgment Accuracy: Reliability and Validity of Postdiscussion Confidence Judgments

Abstract The present study investigated the group judgment process, specifically the issue of whether individuals in a group setting are able to identify the most accurate group member under conditions of uncertainty. Individuals first estimated the answers to nine factual questions, then discussed the questions in a group until consensus was reached. Following the group discussion, individuals were asked to indicate which individual judgments (shared during the group discussion) they believed to be the most accurate for each of the nine judgment items. Despite the lack of forewarning that such assessments would be requested, these postdiscussion judgments of judgment accuracy (i.e., confidence judgments) had a high degree of reliability (i.e., consistency across group members) as well as validity (i.e., correct identification of the most accurate group member). These results, in conjunction with analyses of the group judgment data, provide evidence that individuals are able to identify the most accurate group member at a rate that far exceeds chance expectation. There is also indirect evidence indicating that group members may make these judgments as a normal part of the group judgment process.