Fig. 1From: Advantages of Bayesian monitoring methods in deciding whether and when to stop a clinical trial: an example of a neonatal cooling trialProbabilities of treatment benefit (log RR) for marginal comparisons of cooling on predischarge mortality. Negative values favor the experimental group. Left panel shows the marginal duration comparison (β 2 + β 3/2) and the right panel the marginal depth comparison (β 1 + β 3/2). Top (bottom) panel shows the neutral (enthusiastic) prior and corresponding posterior for the two-factor marginal comparisonsBack to article page