Hello all, I am currently running a linear mixed model using Proc Mixed for a longitudinal and clustered data set. I have two questions relating to the residuals: (1) As each participant has two data points (longitudinal data), should I indicate (effect=participant ID) in the model line to obtain influence diagnostics for each observation for each participant? (2) When I use the code "residual outp=lib.XXX", the residuals in the output file "XXX" are different from the ones indicated in the figures when I use the "influence". Does anyone know why these residuals are different, and importantly, which one I should use? Thank you in advance for your help. Best, Tamara
... View more