Why ex-post power using estimated effect sizes is bad, but an ex-post MDE is not.

less than 1 minute read

Published:

Why ex-post power using estimated effect sizes is bad, but an ex-post MDE is not. Post by David McKenzie and Owen Ozier on the World Bank Development Impact blog.

With supporting Stata code: File 1 (simulation); File 2 (visualization)

Via the World Bank Development Impact blog.