Statistics and Data Science Seminar

  • Ph.D. Candidate
  • Department of Statistics
  • Harvard University

Is My Matched Dataset As-If Randomized, More, Or Less? Unifying the Design and Analysis of Observational Studies

Causal analyses for observational studies are often complicated by covariate imbalances among treatment groups, and matching methodologies alleviate this complication by finding subsets of treatment groups that exhibit covariate balance. However, standard methods for conducting causal analyses for matched datasets do not condition on the strong covariate balance that modern matching algorithms provide. We find that, as a result, these analyses can be unnecessarily conservative. We develop an alternative approach that involves three contributions. First, we formulate an as-if randomization assumption that---unlike previous works---can incorporate any assignment mechanism of interest, including mechanisms that ensure covariate balance constraints, as in matching algorithms. Second, we develop a valid randomization test for this as-if randomization assumption, which allows researchers to determine the experimental design that their matched dataset best approximates. Thus, this test encapsulates the design stage of the observational study. Third, we provide a treatment effect estimation strategy that utilizes the experimental design determined during the design stage, thereby unifying the design and analysis stages of the observational study. We find that our approach yields more precise causal inferences for observational studies than standard approaches by conditioning on the covariate balance in a given matched dataset. We also discuss several applications of our method, with a focus on a regression discontinuity design in education.

Draft of job market paper

Draft of regression discontinuity applicatiom

Zach Branson is a PhD Candidate in Statistics at Harvard University. His main research interests are in experimental design and causal inference, and his work has explored applications in economics, education, epidemiology, engineering, medicine, and text analysis. Before Harvard, Zach received a B.S. in Economics and Statistics and a B.A. in Professional Writing from Carnegie Mellon University in 2014.

For More Information, Please Contact: