Why Propensity Scores Should Not Be Used for Matching

By Gary King, Richard A. Nielsen. Political Analysis, 2019

1,642 citations recorded by the source corpus.

We show that propensity score matching (PSM), an enormously popular method of preprocessing data for causal inference, often accomplishes the opposite of its intended goal—thus increasing imbalance, inefficiency, model dependence, and bias. The weakness of PSM comes from its attempts to approximate a completely randomized experiment, rather than, as with other matching methods, a more efficient fully blocked randomized experiment. PSM is thus uniquely blind to the often large portion of imbalance that can be eliminated by approximating full blocking with other matching methods. Moreover, in data balanced enough to approximate complete randomization, either to begin with or after pruning some observations, PSM approximates random matching which, we show, increases imbalance even relative to the original data. Although these results suggest researchers replace PSM with one of the other available matching methods, propensity scores have other productive uses.

Subfields: Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/3c99b826b76319e99abfd227846033e78a7cb9bd

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

1. Markku V. J. Maula
2. W. Stam
3. Qi Chen
4. M. Fraser
5. Serena Fatica
6. Roberto Panzica
7. Shenyang Guo
8. J. Rohrer

Citations in the corpus

• The Effect of Cultural Similarity on Mergers and Acquisitions: Evidence from Corporate Social Responsibility (2018, positive): “Indeed, King and Nielsen (2016) suggest that propensity scores should not be used for matching, and that there are many advantages to using CEM instead.”
• Thinking Clearly About Correlations and Causation: Graphical Causal Models for Observational Data (2017, background): “The use of propensity scores for matching is controversial, and critics have indicated that other procedures might be preferable (King & Nielsen, 2016).”
• Propensity Score Analysis: Recent Debate and Discussion (2020, methodology): “For the three methods considered by King and Nielsen (2019), researchers often must choose between inexact matching and incomplete matching (Guo & Fraser, 2015; Parsons, 2001).”
• Adjusting for Confounding with Text Matching (2020, background): “However, propensity score matches are difficult to manually assess because propensity score matching approximates a fully randomized experiment rather than a block-randomized experiment (King and Nielsen 2019); documents with similar treatment probabilities may look very different.”
• Newspaper Closures Polarize Voting Behavior (2018, background): “King and Nielsen (2016) show that the inclusion of propensity scores can increase imbalance on some covariates.”
• Enhancing Rigor in Quantitative Entrepreneurship Research (2019, background): “However, PSM has been found to suffer from important limitations (King & Nielsen, 2019), with coarsened exact matching (CEM) being recommended as a robust alternative (Iacus et al., 2009; 2012; King & Nielsen, 2019).”
• Green bonds as a tool against climate change? (2021, positive): “Compared to other standard matching methods, such as the propensity score, coarsened exact matching has the important advantage of reducing the imbalance in the empirical distribution of the confounders between the treated and control group, thereby minimizing concerns of biased statistical inferenc”