Disagreeing Perspectives Enhance Inner-Crowd Wisdom for Difficult (but Not Easy) Questions

Recently, it has been demonstrated that taking a disagreeing perspective increases the accuracy of inner crowds by enhancing estimation diversity. An insightful commentary reanalyzed the data using maximal random structure models and found no increase in accuracy when taking a disagreeing perspective. These findings present a curious challenge for inner-crowd research and hint at the importance of question variability. Here, we present the results of three preregistered experiments (total N = 2,884, with online adult participants from the United States and the United Kingdom) that reconcile these findings by discerning between the ease and difficulty of questions. The results support the notion that taking a disagreeing perspective is beneficial for difficult questions, yet harmful for easier questions. We emphasize that question difficulty is a key factor to consider when evaluating the effectiveness of any intervention designed to improve the accuracy of aggregate estimates through the enhancement of diversity.