Judging My-side versus Thy-side: Meta-Forecasting Accuracy Predicts Forecasting Accuracy in the Political Domain

Forecasting research has capitalized on performance-related individual difference measures to improve forecasting accuracy by indicating who is likely a better forecaster, but few studies have examined how socially-relevant factors influence forecasting accuracy. In two experiments (N = 2,921), we examined how accurately participants estimated the average forecasts of political ingroups and outgroups (meta-forecasts). We tested whether these meta-forecasts predicted participants’ forecasts of the actual outcomes. We found that participants were better at predicting the forecasts of ingroups than of outgroups on group-relevant topics (politics in our study), but not on group-adjacent (i.e, non-political) topics. While meta-forecasting accuracy was predictive of forecasting accuracy, in general, ingroup meta-forecasts on group-relevant topics were particularly good at predicting forecasting accuracy on the topics themselves. Moreover, this did not depend on participants’ political affiliation. Meta-forecasts added indicative value regarding forecasting accuracy above and beyond numeracy and an actively open-minded thinking style.