|
|
Chi-square difference test |
|
Message/Author |
|
|
Good morning, I want to test if the quadratic model is superior to the linear model for repeated measures of an ordered categorical variable. I'm using the difftest option but am unable to figure out how to create a nested model. I tried running the quadratic model and then setting the quadratic mean to zero. Isn't that model nested within the model with the quadratic mean freely estimated? |
|
|
Those models should be nested. But looking at the z-score of the quadratic mean should be enough. If it is significant, the quadratic model is better. |
|
|
Thank you! |
|
Back to top |
|
|