Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 576 - Matching Methods for Causal Inference with Emerging Data and Statistical Challenges
Type: Topic Contributed
Date/Time: Thursday, August 6, 2020 : 3:00 PM to 4:50 PM
Sponsor: Biometrics Section
Abstract #311065
Title: Individualized Matched Design of Observational Studies
Author(s): Jose Zubizarreta*
Keywords: Causal inference; Observational studies; Multivariate matching; Individualized treatments; Integer programming

We present a new matching approach for building balanced observational studies with respect to a particular individual. This individual can be a patient receiving treatment in a hospital, a student pursuing a given school degree, an ex-convict in a program of social re-integration. The approach first profiles the individual in terms of observed covariates; then it solves an integer programming problem in order to find the largest balanced sample that contains the individual close to the centroid of its convex hull. The integer program problem can be solved in large data sets quickly. The approach can be augmented with flexible machine learning regression methods, in the spirit double robust estimation. We discuss real and simulated data examples.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program