Fig. 2From: Bayesian Mendelian randomization with an interval causal null hypothesis: ternary decision rules and loss function calibrationLoss of the frequentist (grey) and our Bayesian MR with binary Y when \(\beta = 0\) for different combinations of missing rate of X (80%, 40% and 0%) and IV strength (0.3, 0.1 and 0.05), based on 200 simulated datasets per combination. Loss is pictured for different values of the threshold, T, and of the score aBack to article page