Adding main effects when main effects model is insignificant

#1
Hi,

I have a linear model where both main effects are insignificant, and an overall insignificant model. Should I stop just stop there, or can I add an interaction term and check the model again?

Thanks!

-Bryan

*EDIT* Title should be "Adding interactions when main effects model is insignificant"
 
Last edited:

hlsmith

Omega Contributor
#2
Interaction term should be examined if it make sense given the context. And you should keep the main effects in the model.