When would you be forced to use a mixed model for a repeated measures ANCOVA instead of the General Linear Model?

Similar questions and discussions