0
$\begingroup$

I tested an interaction in a correlative study (N = 220). The interaction was non-significant but one of the simple effects was. Is it somehow possible to calculate power based on this simple effect? Because both variables were continuous I don't have an exact number of participant in the "cell" in which I found the simple effect.

$\endgroup$

1 Answer 1

1
$\begingroup$

Post hoc power is a misleading and fallacious concept. Don't do it.

For an in-depth argument, see Hoenig & Heisey, "The Abuse of Power - The Pervasive Fallacy of Power Calculations for Data Analysis", The American Statistician, 2012.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.