Yes, in this case it's comparing to the value '1'. In general with factors it's comparing to whatever level is missing.
Sam On Thu, Dec 10, 2009 at 10:40 AM, Ashta <sewa...@gmail.com> wrote: > David, > > Thank you very much for your response. > > I fitted the model as factor instead of numeric. > > coxfit1 <- coxph(Surv(sdat$time, sdat$cens)~factor(y1)+factor(x2) > coef exp(coef) se(coef) z > Pr(>|z|) > factor(y1)2 0.036161 1.036822 0.083921 0.431 0.6665 > factor(y1)3 -0.510124 0.600421 0.088901 -5.738 9.57e-09 *** > factor(x2)2 -0.510124 0.600421 0.088901 -5.738 9.57e-09 *** > > > > What are those values? Is it comparing in reference to the first > class of each covariate? > > Thanks again. > > > > > > > On Thu, Dec 10, 2009 at 8:33 AM, Ashta <sewa...@gmail.com> wrote: >> Hi all, >> >> I want to calculate hazard ratio within each covariate >> >> >> Example, one covariate has 3 classes (1,2 and 3) and x2 has 2 classes >> >> I want to compare the relative risk ratio within each class of the covariate. >> How do I get this result ? . >> >> >> The other question is that how do I interpret the second column in >> the second panel (i.e., exp(-coef)) >> >> I used the model >> coxfit1 <- coxph(Surv(sdat$time, sdat$cens)~ y1+x2) >> >> coef exp(coef) se(coef) z Pr(>|z|) >> y1 -0.024084 0.976204 0.003077 -7.828 5.00e-15 *** >> x2 0.036161 1.036822 0.083921 0.431 0.6665 >> >> exp(coef) exp(-coef) lower .95 upper .95 >> y1 0.9762 1.0244 0.9703 0.9821 >> x2 1.0368 0.9645 0.8796 1.2222 >> >> >> Thanks in advance >> > > ______________________________________________ > R-help@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
