Displaying 1 result from an estimated 1 matches for "daytomort".
2008 Apr 29
0
Looking for Post-hoc tests (a la TukeyHSD) or interaction-level independent contrasts for survival analysis.
...would be easiest for me if I performed a responsible
multiple comparisons test like TukeyHSD to see how each of the six
Treatment:Group subgroups compared to each other. TukeyHSD does not
appear to be defined for outputs from the function coxph -- (see
survival library).
cph <- coxph(Surv(DayToMort, Censor) ~ Treatment*Group, data=subb)
--> Does anyone know of an implementation of TukeyHSD that would
work, or of another post-hoc multiple comparison test?
I believe that another responsible tack would be to clearly define
the contrasts I'd like to make within the interaction term. Ho...