TriMatch: An R Package for Propensity Score Matching of Non-binary Treatments

The use of propensity score methods (Rosenbaum and Rubin, 1983) have become popular for estimating causal inferences in observational studies in medical research (Austin, 2008) and in the social sciences (Thoemmes and Kim, 2011). In most cases however, the use of propensity score methods have been confined to a single treatment. Several researchers have suggested using propensity score methods with multiple control groups, or to simply perform two separate analyses, one between treatment one and the control and another between treatment two and control. This paper introduces the TriMatch package for R that provides a method for determining matched triplets. Examples from educational and medical contexts will be discussed.