Characterize hypotheses susceptible to false confidence

Characterize the class of statistical hypotheses for which probabilistic uncertainty quantification assigns high probability to false assertions, and quantify the severity of false confidence for those hypotheses.

Background

The paper argues that probabilistic uncertainty quantification has inherent unreliability due to false confidence. Although nonlinearity is conjectured to play a role, a comprehensive delineation of affected hypotheses is lacking.

In the concluding open problems, the author calls for a precise classification of hypotheses to which false confidence applies, addressing both theory and practice.