0
$\begingroup$

If the Independent variables are in percentages, then will it make sense to calculate interaction effects using the original figures? or do I need to make any changes?

$\endgroup$
  • $\begingroup$ on the face of it, looks unproblematic, but for an informed answer we want to have some context, please! $\endgroup$ – kjetil b halvorsen Feb 25 at 8:53
  • 1
    $\begingroup$ I'd say that it's potentially problematic -- because near the bounds the response can't be additive, even if it's additive near the middle. This nonlinearity makes interpreting interactions on the original scale confounded with interpreting nonlinear "main effects". More details would help, but typically I would not be working with a model on the original scale; I'd expect to use a nonlinear link function. $\endgroup$ – Glen_b Feb 25 at 11:47

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.