The uncertainty about the accuracy of estimating the
uncertainty of a class of events. In
information theory, meta-uncertainty is calculated on the
probability distribution of possible samples drawn at
random from a hypothetical universe having the same uncertainty or
entropy as the sample actually observed. E.g., statistical tests of the significance of a null-hypothesis involve a distribution of test results and if the latter is an uncertainty, the uncertainty associated with an uncertainty measure is meta-uncertainty. By the
law of large numbers meta-uncertainty decreases with increasing number of observations. (
Krippendorff )
