Can you test assumptions for a categorical variable? This doesn't seem to make sense, because looking at residuals for a contrast-coded variable doesn't seem to provide useful information. So maybe rather than can you test the assumptions, my question is am I supposed to?
Also, for a moderated model, do you check assumptions for the moderator variable? I know there was an earlier post that said you don't have to check assumptions of the predictor for each level of the moderator, but I'm unclear about checking the moderator as its own variable.
Subscribe to:
Post Comments (Atom)
2 comments:
Yes, you can (and should) test assumptions with categorical predictors. The scatterplots will not look the same, and you don't look for the same things exactly, but you do want to test.
Linearity is the hardest, because a smoothing line with only two points to work with may not be so smooth, so don't worry too much about strange graphs there.
Homoskedascity is actually easier to assess, because if the variance at one level of the categorical variable is 10 times that of the other, you will have a graph that looks something like this:
o
o
o
o
o
o o
o
o
o
o
only with hundreds of circles, not just 11.
Normality is assessed in the usual ways.
Well, that pictograph didn't work too well. I was trying to indicate that there would be two columns of circles, so imagine that one circle about an inch further to the right, all by itself, and you'll see what I meant.
Post a Comment