Published online by Cambridge University Press: 14 March 2022
Several authors, e.g. Patrick Suppes and I. J. Good, have recently argued that the paradox of confirmation can be resolved within the developing subjective Bayesian account of inductive reasoning. The aim of this paper is to show that the paradox can also be resolved by the rival orthodox account of hypothesis testing currently employed by most statisticians and scientists. The key to the orthodox statistical resolution is the rejection of a generalized version of Hempel's instantiation condition, namely, the condition that a PQ is inductively relevant to the hypothesis (x)(Px ⊃ Qx) even in the absence of all further information. Though their reasons differ, it turns out that Bayesian and orthodox statisticians agree that this condition lies at the heart of the paradox.
The author's work on the foundations of statistics has been supported in part by the National Science Foundation under grant GS 2525.