Hello all: I'm having some difficulty interpreting the results of a Cox proportional hazards analysis that I recently performed. I am running the fit with a fixed categorical variable (Treatment) and a nested categorical variable (Chamber) within Treatment. The likelihood ratio tests of both of these effects have a low p (<0.001), but I am unable to discern which specific treatments (or chambers) are statistically different from one another. I was wondering if there are any aposteriori or post-hoc tests that are performed in relation to proportional hazards analysis and can help pinpoint which treatments are responsible for these results. My experience with proportional hazards tests is minimal, so if I need to provide more details, please let me know. Thanks for any help. Justin [[alternative HTML version deleted]]