Propensity score matching

General
1

Hi,

I'm trying to do propensity score matching using R, which I've never used before. I mostly use SAS due to big data and I already have a table imported from SAS with

id
x (exposure variable)
logit (value of linear predictor)
ps (estimated probabilty)

It's about 2 million rows. Now, I want to try to do matching 1:2, option to replace a matched individual after matching, caliper about .2*sd.

I have tried some snippets from some code examples but in some cases I get some kind of overflow error and in some cases I just can't get it to work.

Can someone give me a short rundown on a possible solution?

Thank you so much in advance.

2

Its hard to help you with data and code we can't see.
In general when trying example code; its best to run their entire lesson on their data, and try to understand their steps; where your data/issues may differ, so that you can a) be confident that they example code at least works on their data... b) have more knowledge when adapting theirs to yours.

I can tell you that * R: propensity score matching is available as part of the MatchIt,[7][8] optmatch,[9] or other packages.

I would really encourage you to review the following guide, FAQ: Tips for writing R-related questions.
For example, the guide emphasizes asking coding questions with formatted code-chunks and a reprex.

You may have noticed folks here requesting minimal reprexes, that's because asking questions this way saves answerers a lot of time.

Reproducible Examples:

  • help make your question clear and replicable
  • increases the probability folks will reach out and try to help,
  • reduces the number of back-and-forths required to understand the question,
  • and makes your question and suggested solutions more useful to folks in the future researching similar problems.
3

This topic was automatically closed 42 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.