We generated 10,000 simulated datasets for each scenario in a two‐sample setting with two values of the causal effect (β=0, null causal effect; β=0.1, positive causal effect); the simulations are repeated in Supporting Information Appendix 3 in a one‐sample setting. Only the summary data on the genetic associations with the exposure and with the outcome (and their standard errors) are used as data inputs in the analysis methods (individual‐level data are not used). Details of the data‐generating model and the parameters used in the simulation study are given in Supporting Information Appendix 3.
