MacMillan-CSAP Workshop on Quantitative Research Methods: Donald Green, “Double Sampling for Nonignorable Missing Outcome Data in Randomized Experiments”

“Double Sampling for Nonignorable Missing Outcome Data in Randomized Experiments”

Guest Speaker: Donald Green, Professor of Political Science, Columbia University

Abstract: Missing outcome data plague many randomized experiments. Common solutions often rely on ignorability assumptions that may not be credible in a given application. We propose a measurement and estimation strategy for dealing with missing outcome data that makes minimal assumptions but still yields substantively informative bounds on the average treatment effect. Our approach is based on a combination of the double sampling design and non-parametric worst-case bounds. We propose a method for covariate adjustment using post-stratification, and provide analytic expressions for confidence intervals for the average treatment effect. Using our approach, we show that double sampling greatly reduces uncertainty in a placebo-controlled randomized field experiment on the effects of persuasion on social attitudes with survey-based outcomes.

Speaker Bio: Donald P. Green is Professor of Political Science at Columbia University.  The author of four books and more than one hundred scholarly articles, Green has written on a wide array of topics: voting behavior, partisanship, campaign finance, hate crime, and research methods. Much of his current work uses field experimentation to study the ways in which political campaigns mobilize and persuade voters. He looks forward to returning to ISPS, which he directed from 1996 to 2011.