The Program for Quantitative and Analytical Political Science (Q-APS) at Princeton University hosts numerous conferences throughout the year. The links below provide detailed information on upcoming conferences.
Wednesday, Aug 29, 2012, 1:00 PM-4:00 PM APSA 2012 Annual Meeting, New Orleans, Louisiana
Sponsored by Kosuke Imai, Luke Keele, Dustin Tingley, Teppei Yamamoto
Understanding causal mechanisms is a fundamental goal of social science. Researchers seek to study not only whether one variable affects another but also how such a causal relationship arises. This three-hour workshop will introduce the latest developments of causal mediation analysis to applied researchers who wish to learn about causal mechanisms from experimental and observational studies. The workshop will be in part based on the recently published American Political Science Review article (Vol. 105, No. 4, pp. 765-789).
The materials of the workshop consist of five distinct components. First, participants will learn how to define and interpret the key quantities of interest that operationalize causal mechanisms within the modern statistical framework of causal inference. Under this framework, we introduce a minimal set of assumptions required for identifying causal mechanisms when applying commonly used mediation analyses. Second, participants will learn how to conduct causal mediation analysis under a variety of modeling assumptions. Going beyond the traditional linear regression modeling framework, we show how to estimate the quantities of interest under nonlinear and even nonparametric models. Third, given that the standard mediation analysis rests upon a strong assumption, sensitivity analysis plays an important role. We show how to examine the robustness of empirical findings to the potential violation of the required assumption. Fourth, to further improve the credibility of causal mediation analysis, we introduce alternative strategies for designing experiments and observational studies that allow researchers to identify causal mechanisms with weaker assumptions. Finally, the workshop will include a demonstration of easy-to-use software, "mediation." This package implements all of the suggested methods and is available in both R and Stata.
The workshop is free of charge but the registration is required. Those who wish to attend the workshop should email Esther Kim at firstname.lastname@example.org