Fig. 1.

Panel A illustrates the assumptions underpinning a Mendelian randomization analysis of the association between an exposure (e.g., education) and an outcome (e.g., ADHD). SNPs indicate single-nucleotide polymorphisms. The arrows represent causal pathways. The dashed arrows represent potential causal associations between variables that would violate the Mendelian assumptions. Panel B displays one such possible violation with inclusion of an exposure B (independent exposure, mediator, or confounding factor), in our example proposed to be intelligence. One method to examine the influence of this possible violation is multivariable Mendelian randomization analysis (MVMR) and the inverse-variance weighted method [32] with markers for education adjusted for intelligence. The remaining direct causal effect of education on ADHD is illustrated by the bold arrow