Good afternoon, We are working on a multiple subject aging study with PET. We have setup the experiment to study longitudinal trends in the PET signal. A condition indicator models each time point (regularly spaced in time) and each subject has an independent mean. We include age at the baseline scan as a covariate. Our design matrix is located here: https://jshare.johnshopkins.edu/xythoswfs/webui/_xy-2399293_1 My question is: How does covariate centering work? When I examine linear contrasts on the conditions (significance of trends), I see no differences between centering age and not centering age? Is centering done automatically? Regards, Bennett ------------------------------------------- Bennett Landman, <mailto:[log in to unmask]> [log in to unmask] Ph.D. Candidate, Biomedical Engineering Johns Hopkins University School of Medicine