The primary objective was to estimate θ 1, with the causal effects set to: θ 1=0(null causal effect) or θ 1=0.3(positive causal effect); θ 2=0.1; and θ 3=−0.3. The data were simulated to consider the following four scenarios: No pleiotropy ( αj′=0 for all j), InSIDE assumption automatically satisfied;Balanced pleiotropy (μ=0), InSIDE assumption satisfied;Directional pleiotropy (μ=0.01,0.05 or 0.1), InSIDE assumption satisfied;Directional pleiotropy (μ=0.01,0.05 or 0.1), InSIDE assumption violated.
