Regressive Judgments: How People Fail to Account for Regression in Other People’s Judgments

Regression to the mean is a well-established statistical phenomenon. In probability judgments, for example, people tend to overestimate low probabilities and underestimate high probabilities. Although the consistency of this regressiveness has been widely demonstrated, it remains unclear whether people take it into account when interpreting probability estimates made by others. Across four preregistered experiments (total N = 601), we demonstrate that, rather than adjusting estimates towards the extremes to correct for the regressiveness of the original judgments, participants consistently shifted their estimates of the true underlying probabilities towards the mean, thereby reinforcing regression effects. These findings suggest that people fail to account for regression when interpreting others’ probability estimates.