This question already has an answer here:

Is it formally correct to compare two separate models with different independent variables?

For instance in R a pseudo code would be something like:

AIC( lm( y ~ X ) )


AIC( lm( y ~ W ) )

(X and W cannot feature in the same model as for instance lm( y ~ X*W ) )


marked as duplicate by usεr11852, gung, Community Dec 12 '16 at 14:02

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • 1
    $\begingroup$ See stats.stackexchange.com/questions/116935/… $\endgroup$ – Tim Nov 30 '16 at 10:53
  • $\begingroup$ That question is about using the AIC with different Y variables, this question is about using the AIC to compare models w/ different X variables. I'm voting to leave open. $\endgroup$ – gung Nov 30 '16 at 22:36
  • $\begingroup$ That seems like a winner, @RichardHardy. $\endgroup$ – gung Dec 12 '16 at 13:39