I have a pretty good grasp of the pros and cons of different methods of variable selection: LASSO and LARS, AIC, stepwise procedures, etc. But my question is: when modeling, should you conduct variable selection differently if you are using the model for the estimation of fixed effects compared with finding a model that has the best prediction?

I've never used the change-in-estimate variable selection. Is that the best way to do variable selection when you're just looking to estimate fixed effects?

  • 2
    $\begingroup$ To put it shortly, yes, the model building process -- including variable selection -- differs depending on the purpose (explanatory versus predictive modelling). See this thread for some general information. $\endgroup$ – Richard Hardy Nov 3 '15 at 21:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.