Closed samyogita-hardikar closed 1 year ago
What you see is a consequence of the model you fitted. In that model, gender
does not interact with the other factors; in so doing, you have specified that the effects of phase
, age
, and phase:age
are the same for both gender
s.
Aah, I see. Thanks for the quick help!
Hello! I was wondering how to get appropriate emtrends in a model that has 2 interacting categorical predictors and a numeric predictor.
Reproduced below using the obk.long dataset
My data and model are comparable to the following where "phase" is a within subject variable, and "age" and "gender" are between.
I am interested in the effect of age at every level of phase but also for both genders. The first part is achieved by doing
but after adding gender as an interaction to emtrends ...
even though 6 "age.trend" rows are generated, only the pvalues are adjusted to reflect 6 tests instead of 3. But the actual age.trend is still only given for each level of phase, while averaging over two levels of gender.
What would be the solution here?