Causal Inference in Statistics: A GentleIntroduction

摘要

This paper provides a conceptual introduction to causal inference, aimed tornassist researchers beneft from recent advances in this area. The paper stressesrnthe paradigmatic shifts that must be undertaken in moving from traditionalrnstatistical analysis to causal analysis of multivariate data. Special emphasisrnis placed on the assumptions that underly all causal inferences, the languagesrnused in formulating those assumptions, and the conditional nature of causalrnclaims inferred from nonexperimental studies. These emphases are illustratedrnthrough a brief survey of recent results, including the control of confounding,rnand a symbiosis between counterfactual and graphical methods of analysis.