1

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?

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
LindsayL
  • 616
  • 5
  • 9
  • 2
    To put it shortly, yes, the model building process -- including variable selection -- differs depending on the purpose (explanatory versus predictive modelling). See [this](http://stats.stackexchange.com/questions/1194/practical-thoughts-on-explanatory-vs-predictive-modeling) thread for some general information. – Richard Hardy Nov 03 '15 at 21:02

0 Answers0