Causal diagrams for empirical research

The primary aim of this paper is to show how graphical modelscan be used as a mathematical language for integrating statisticaland subject-matter information. In particular, the paper developsa principled, nonparametric framework for causal inference,in which diagrams are queried to determine if the assumptionsavailable are sufficient for identifying causal effects fromnonexperimental data. If so the diagrams can be queried to producemathematical expressions for causal effects in terms of observeddistributions; otherwise, the diagrams can be queried to suggestadditional observations or auxiliary experiments from whichthe desired inferences can be obtained.