MR Dictionary

Non-parametric methods with bounds of causal effect

A non-parametric method that estimates upper and lower bounds for an MR causal effect. The bounds are not the same as confidence intervals they reflect the upper and lower limits of an estimated distribution of causal effects. 

Genetic variants must satisfy the IV assumptions. It does not require any of the additional IV assumptions. This method can only be used in the situation of binary (or categorical with few levels) risk factor, IV and outcome. A method for implementing these in Stata has been developed. 


Other terms in 'One-sample MR methods':