Fig. 1From: Bayesian Mendelian randomization with an interval causal null hypothesis: ternary decision rules and loss function calibrationDirected acyclic graph (DAG) representation of the Mendelian randomisation model we consider throughout the paper. The strength of the \(X \rightarrow Y\) arrow, corresponding to the unknown magnitude of the causal effect of interest, is represented in the model equations by the unknown parameter \(\beta\)Back to article page