Stopping rule and Bayesian confirmation theory

Abstract

This article mainly investigates whether common Bayesian confirmation measures are affected by stopping rules. The results indicate that difference measure d, log-ratio measure r, and log-likelihood measure l are not affected by non-informative stopping rules, but affected by informative stopping rules. In contrast, Carnap measure \(\tau \), normalized difference measure n, and Mortimer measure m are affected by (non-)informative stopping rules sometimes but sometimes aren’t. Besides, we use two examples to further illustrate that confirmation measures d, r, and l are better than \(\tau ,n\), and m.