Hi All, I am working on a BE 2X2 study and got negative variance(sequence*subject) for one parameter (eg. Cmax). Based on the explanation by Helmut and others, this indicates intersubject CV% < intrasubject CV% and a fix to this is to remove the subject (sequence) term in random effect and add that to the fixed effect as: sequence+period+treatment+subject(sequence) and will resolve this issue. However the final variance parameters only show the Var(residual) which can be used to calculate to intrasubject CV% while the intersubejct CV% are not displayed. Is there still a way to find these intersubject CV% or data to calculate these intersubject CV% from the output? Thanks and best regards, zan