MR Dictionary

MR using Robust regression (MR Robust)

Robust MR method developed for two-sample MR settings that downweights and excludes instrumental variables (IVs) with heterogeneous causal estimates that are likely invalid using robust regression.

This method specifically MM-estimation, a combination of "maximum-likelihood-type" estimation (i.e., an M-estimator) and "scale-type" estimation (i.e., an S-estimator) to provide a model that is robust to outliers in an MR analysis. See also MR Lasso and MR with penalized weights.

References

Other terms in 'Pleiotropy-robust two-sample MR methods':