Regency B
Counterfactual Analysis of Cross-Sectional Data Using Quantile Process Regression (303957)
*Yonggang Yao, SAS Institute, Inc.
Keywords: counterfactual inference, distribution prediction, weighted bootstrap, selection bias
This presentation introduces the basic concepts of quantile process regression (QPR) and demonstrates a standard workflow for using the QPR method to analyze cross-sectional data. The workflow predicts counterfactual distribution of the treatment-group response by applying the control-group model; estimates selection bias and treatment effects; and performs a variety of statistical inferences for evaluating means, quantiles, probabilities, and distributions. In comparison to the popular propensity score and doubly robust methods, the QPR method is distribution-agnostic and more straightforward, can output richer information, and can incorporate other nonparametric statistical methods such as Mann-Whitney-Wilcoxon U test. This presentation uses two real-data applications to illustrate the usage of QPR method. The first example analyzes the impact of smoking on newborn body weights, and the second example evaluates the employment-and-salary effect of the 1975– 1979 National Supported Work Demonstration project.