Multivariate Matching Methods for Causal Inference

• Activity Number: 177 - Section on Statistics in Epidemiology P.M. Roundtable Discussion (Added Fee)
• Type: Roundtables
• Date/Time: Tuesday, August 10, 2021 : 12:00 PM to 1:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Jose Zubizarreta*
• Companies: Harvard University
• Keywords: Causal inference; Observational studies; Matching methods

Abstract

In this roundtable, we will discuss the role of matching methods for causal inference in observational studies. We will establish a sharp distinction with related weighting and regression modeling approaches. We will deliberate on the use of matching methods with multi-valued and time-varying treatments. We will explain how these methods can be used to build matched samples that are balanced relative to target populations and specific individuals. As time permits, we will talk about sensitivity analyses to hidden biases.