Advancing Propensity Score Matching Methodology for Quasi-Experimental Field Research: Introducing the Many-Group Matching (MAGMA) Algorithm (Poster 1)

• In Event: AERA Poster Session Five
  In Poster Session: Division D Section 2 Poster Session

Abstract

Balancing quasi-experimental field research for effects of covariates is fundamental for drawing causal inference. Propensity Score Matching deals with this issue but current techniques are restricted to binary treatment variables. Moreover, they provide several solutions without providing a comprehensive framework on choosing the best model. With this paper, we want to address these restrictions. We developed an algorithm that systematically produces the best possible nearest neighbour solution and is applicable for multiple groups. We propose different criteria for optimizing sample size and balance that are independent of the number of groups. Conducting real data examples, we show that our algorithm outperforms an established one in a two-group example and is capable to produce balanced solutions for a three-group example.

Authors

• Markus D. Feuchter, University of Trier
  Presenting Author

• Julian Urban, GESIS - Leibniz-Institute for the Social Sciences
  Presenting Author

• Franzis Preckel, University of Trier
  Non-Presenting Author