Mixed effects logistic regression - stepwise procedure?

[Claire Boulange]

Hi,

I am trying to run a multilevel mixed-effects logistic regression(melogit) model and to do a stepwise procedure so that I can get the 'best' predictors for my model. I read that melogit is not supported by stepwise / how should I proceed? how can I figure out which variables to include or not?

thanks

[Friedrich Huebler]

This forum is for questions about Statalist, not Stata. Please ask in the General forum. You may want to review the FAQ for advice on how to ask questions on Statalist.

[Carlo Lazzaro]

Claire:
welcome to the list.
I second in full Friedrich's recommendation and just add that step-wise procedure is, in general, a bad way to go. See for instance: http://www.statalist.org/forums/foru...s-e-adjustment.