Covariate adjustment in randomization-based causal inference for 2K factorial designs
Author:
Preprint
Abstract
We develop the finite-population asymptotic theory for covariate adjustment in randomization-based causal inference for 2K factorial designs. In particular, it is confirmed that both the unadjusted and covariate-adjusted estimators of the factorial effects are asymptotically normal, and the latter is more precise than the former.Keywords
All Related Versions
This publication has 9 references indexed in Scilit:
- On randomization-based and regression-based inferences for factorial designsStatistics & Probability Letters, 2016
- A Bayesian Perspective on the Analysis of Unreplicated Factorial Experiments Using Potential OutcomesTechnometrics, 2016
- Construction of alternative hypotheses for randomization tests with ordinal outcomesStatistics & Probability Letters, 2015
- Causal Inference from 2K Factorial Designs by Using Potential OutcomesJournal of the Royal Statistical Society Series B: Statistical Methodology, 2014
- Agnostic notes on regression adjustments to experimental data: Reexamining Freedman’s critiqueThe Annals of Applied Statistics, 2013
- Adjusting Treatment Effect Estimates by Post-Stratification in Randomized ExperimentsJournal of the Royal Statistical Society Series B: Statistical Methodology, 2012
- For objective causal inference, design trumps analysisThe Annals of Applied Statistics, 2008
- On regression adjustments in experiments with several treatmentsThe Annals of Applied Statistics, 2008
- Some heteroskedasticity-consistent covariance matrix estimators with improved finite sample propertiesJournal of Econometrics, 1985