High-Dimensional Regression Adjustment Estimation for Average Treatment Effect with Highly Correlated Covariates

Regression adjustment is often used to estimate average treatment effect (ATE) in randomized experiments. Recently, some penalty-based regression methods have been proposed handle the high-dimensional problem. However, these existing may fail achieve satisfactory performance when covariates are highly correlated. In this paper, we propose a novel estimation method for ATE by combining semi-standard partial covariance (SPAC) and methods. Under regularity conditions, asymptotic normality of our SPAC estimator shown. Some simulation studies an analysis HER2 breast cancer data carried out illustrate advantage addressing correlated problem Rubin causal model.