Randomization-Based Inference Within Principal Strata

Randomization-Based Inference within Principal Strata.

In randomized studies, treatment comparisons conditional on intermediate post-randomization outcomes using standard analytic methods do not have a causal interpretation. An alternate approach entails treatment comparisons within principal strata defined by the potential outcomes for the intermediate outcome that would be observed under each treatment assignment. In this paper, we develop method...

A Paradox From Randomization-Based Causal Inference

Under the potential outcomes framework, causal effects are defined as comparisons between potential outcomes under treatment and control. To infer causal effects from randomized experiments, Neyman proposed to test the null hypothesis of zero average causal effect (Neyman’s null), and Fisher proposed to test the null hypothesis of zero individual causal effect (Fisher’s null). Although the subt...

Randomization-based inference using the mosaic package

The mosaic package is intended to support teaching statistics and modeling (as well as calculus) in a way that embraces the possibilities offered by modern computational techniques. Our goal is to make effective computation accessible to university-level students at an introductory level. With the broad availability of inexpensive, fast computing and powerful, free software such as R, the rate-...

Causal Inference with Intermediates: Simple Methods for Principal Strata Effects and Natural Direct Effects

Political Science 236 Randomization inference

Most of this course will be devoted to the study of treatment effects in the absence of random assignment of subjects to treatments. As we will see, performing causal inference in the absence of random treatment assignment requires that we make fairly strong assumptions. In contrast, when treatment is assigned randomly, treatment effects can be estimated with very mild assumptions and, very imp...