MASSIVE: Model Assessment and Stochastic Search for Instrumental Variable Estimation

One popular approach is to utilize these many-dimensional measurements as instrumental variables (instruments) for improving the causal effect estimate between other pairs of variables. Unfortunately, searching for proper instruments in a many-dimensional set of candidates is a daunting task due to the intractable model space and the fact that we cannot directly test which of these candidates are valid.

We propose a general and efficient causal inference algorithm (MASSIVE) consisting of Model Assessment and Stochastic Search for Instrumental Variable Estimation. The MASSIVE algorithm accounts for model uncertainty by performing Bayesian model averaging over the most promising many dimensional instrumental variable models, while at the same time employing weaker assumptions regarding the data generating process compared to similar methods.