PROC GENMOD: allow STRATA statement without EXACT for fixed effect poisson modeling
The strata-statement is also something I miss now and then. Thumbs up!
It is misleading that variables can be specified in strata, but the procedure only account for them in the exact analysis!
A macro doing this analysis can be found as supplementary material in the excellent paper by Richardson & Langholz:
Background stratified Poisson regression analysis of cohort data, Radiation and Environmental Biophysics, Springer-Verlag, 2012, 51, 15-22,
or alternative, one can make a simple transformation of data and make the stratified Poisson regressin with PHREG.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.