search

mhamjediers / nopo_decomposition

Stata ado files to run matching-based decompositions
MIT License
3 stars 0 forks source link

Strata #15

Closed maximilian-sprengholz closed 1 year ago

maximilian-sprengholz commented 1 year ago

There was a mistake in the code to the effect that the _nopo_strata captured the number of unique values of the propensity score (which we knew). After correction (which is pushed), the no. of strata for ps will always be 1 (the same is true for md, where strata anyhow make no sense in the distance metric).

What should we do? Show strata only for em? We could also show the unique values for the ps, but then _nopo_strata would be different from _KM_strata.

mhamjediers commented 1 year ago

do we need the _KM_stratavariable after ps/md-matching?

I would generally think, that we should report the number of strata only for em, and also just report a variable of the propensity score instead of the strata

maximilian-sprengholz commented 1 year ago

Closing.