07-20-2017 02:50 PM - edited 07-20-2017 02:51 PM
I have run both a proc mixed and a proc glm on my data the Type 3 table keeps coming upt empty. on the Type 1 table items are significant but Type 3 is blank. What does this mean?
Source DF Type III SS Mean Square F Value Pr>F
Location 0 0
Age 0 0
Meds 0 0
07-20-2017 03:07 PM
Are you fitting too many interaction terms (a saturated model)? Or are one of these variables categorical with a large number of parameters? (For example, almost as many parameters as valid observations?) It looks like you are including some term in the model that sucks up all the variation among the means, thus leaving no variation for the other effects.