2

After removing the insignificant variables(p-value >.05), I fitted the OLS model again. I found there are still many variables which had p-value < .05 earlier have p-value > .05 now. Do I need to keep removing them and till when?

AlexK
  • 1,007
  • 4
  • 11
Saurabh Verma
  • 21
  • 1
  • 3
    You are suggesting a stepwise regression approach to arriving at the final model. If that's what you are looking to do (and keep in mind that stepwise regression is viewed negatively by most statisticians today and there are better alternatives), look up methods for backward and forward variable selection. – AlexK Apr 04 '19 at 07:09
  • 1
    In particular see [this post](https://stats.stackexchange.com/questions/13686/what-are-modern-easily-used-alternatives-to-stepwise-regression) – Peter Flom Apr 04 '19 at 10:36

1 Answers1

0

Welcome to CV. This may get closed as a duplicate question but maybe you will see this before that happens: You don't even necessarily need to remove the nonsig. variables, which you say you did in your first sentence.

Model building is complex and should not be automated unless it absolutely has to be.

Peter Flom
  • 94,055
  • 35
  • 143
  • 276