Treating Time as a Repeated Effect

Since you have two measurements for each patient, it makes sense to specify time as a repeated effect as well as a fixed effect. Moreover, from the results of the previous model, you suspect that the two time periods are correlated, and so will choose a non-diagonal covariance structure for the repeated effects. The test of the fixed effect [TIME=1] in such a model is equivalent to a paired t-test.

Next