0

I came across this post, which was largely nonsensical, but a respondent suggested the original poster follow up with two articles:

Ho, D. E., Imai, K., King, G., & Stuart, E. A. (2007). Matching as nonparametric preprocessing for reducing model dependence in parametric causal inference. Political analysis, 15(3), 199-236.

King, G., & Nielsen, R. (2016). Why propensity scores should not be used for matching. Political Analysis, 1-20.

The first article argues for using PSM, the second basically says it's not good practice.

I have a dataset where I have individuals who moved with assistance from Program A, and individuals who moved without any assistance, and I want to compare outcomes. Has PSM fallen out of favor, statistically, for such analyses?

Richard Hardy
  • 67,272
tchoup
  • 399
  • 3
    Does this answer your question? Propensity score matching - What is the problem? The answer from @Noah on that page is a superb overview of the tradeoffs involved in propensity score matching (PSM), written by an expert in the field. – EdM Apr 27 '22 at 17:33
  • That is a great response, thank you. Follow up question: where should a person start in finding the right methods for comparing two groups, one of which received a non-random treatment - although maybe this is too broad to ask? I'm just not really sure how to approach figuring out what to do with the data at this point. – tchoup Apr 27 '22 at 18:15
  • 3
    Matthay et al. (2020) might be a nice place to start. I also suggest my answer here. As far as choosing the best method, you need to work with someone who has a PhD in statistical methodology for causal inference to determine which statistical method best suites your data, question, and assumptions best. If you're problem is interesting, I'm sure a statistician would love to collaborate. – Noah Apr 27 '22 at 18:39

0 Answers0