Bioequivalence and Bioavailability Forum

dear,
I was included in the model the followling interactions:

Repeated measures analysis of the last three pre-dose concentrations (Ln transformed) will be statistically analyzed using General Linear Model (PROC GLM) of SAS® version 9.1.3 software. The factors included in this model will be the sequence, subject nested within the sequence, period, treatment subject by treatment nested within the sequence, time, time by sequence, time by subject nested within the sequence, time by period, time by treatment and time by treatment by subject nested within the sequence. The error term to test time is the time by subject nested within the sequence. The error term to test time by treatment is time by treatment by subject nested within the sequence.
without interaction term "time by treatment by subject nested within the sequence". i am getting the prob. values for all the above interactions except time*trt*sub(seq).
if included it in the model then all the interactions p values are equal to zero
Note: here the model degrees of freedom for error is zero because of model degrees of freedom is half to the total degrees of freedom.
Please give me clarification why here the model degrees of freedom is half to the total degrees of freedom.

Best Regards
Rams.stat
Please Reply to this mail […]

---

Edit: Category changed (method not required in any guideline). E-mail address removed and personal messaging activated. Users may contact you via contact form. [Helmut]