I have data from a longitudinal parallel groups study where there are 46 subjects randomized to 1 of 5 treatment groups, each subject with roughly 13 observations over time on a given outcome measure. The data trends in a U-shape over time, thus the quadratic term in the model. 4 active treatments are tested against placebo.
I am trying to decide whether it's more appropriate to use a repeated statement
or a random statement
random int time /subject = subjid ;
Can you help me understand the differences in the model implications based on whether the repeated or random statement is used in SAS?