I want to know how to do propensity matching in the JMP student version.
How do I perform propensity score matching in JMP Student Edition?
My name is Ono, and I am a senior tester at the JMP Japan division of SAS Institute Japan.
■JMP Student Edition includes the functionality of JMP Pro. The current version of JMP Student Edition is JMP Pro 18. Unfortunately, the current versions of JMP 18 and JMP Pro 18 do not have the functionality to perform propensity score matching.
■JMP and JMP Pro have a language called the JMP Scripting Language, which allows programming. I created an add-in that performs 1:1 nearest neighbor matching, although it is an old one created about 10 years ago. The download location for this add-in is included in the description file below.
https://www.jmp.com/content/dam/jmp/documents/jp/support/propensityscore201412.pdf
This add-in performs 1:1 matching using a greedy method after sorting by random numbers. It cannot perform 1:k matching or optimal matching. It is provided as an as-is sample only.
The next major release, JMP Pro 19, will include a new feature called the Causal Treatment Effects platform, which will enable nearest neighbor matching. The platform also offers three methods for estimating average causal effects (ATE) and average treatment effects (ATT): inverse probability weighting, regression adjustment (standardization), and doubly robust methods.
(This discussion is about the next major release, JMP Pro 19. The current release is JMP Pro 18, and the current version of JMP does not offer propensity score matching functionality.)
■Note:
Although it is unrelated to your question, the following is a famous paper that criticizes propensity score matching.
King, G. and Nielsen, R. (2019)
Why Propensity Scores Should Not Be Used for Matching
Political Analysis, 27(4), 435-454.
https://gking.harvard.edu/publications/why-propensity-scores-should-not-be-used-formatching
Additionally, the PDF file "Nearest Neighbor Matching Script Explanation" that is installed with the add-in mentioned above (which can be accessed by going to [Add-ins] → [Nearest Neighbor Matching] → [Nearest Neighbor Matching Script Explanation]) introduces some cautions (criticisms) about propensity scores by Stephen Senn.
In "What is Causal Inference?" by Rosenbaum, supervised translation by Takada Yuya and Takahashi Koji, Asakura Shoten, pp. 49-53, what is being matched in propensity score matching is explained.
