Fig. 5From: Frameworks for estimating causal effects in observational settings: comparing confounder adjustment and instrumental variablesDAG with more IVs and observed confounders as well as stochastic errorsBack to article page