Ashraf Chaudhary
2006-Jul-29 02:02 UTC
[R] Help with clogit in survival - conditional logistic regression
Dear All: I have been struggling to run the conditional logistic regression on a dataset. I am using clogit function in survival package. Here is how the data is generated. I tried to shorted the data by using the second count variable but still the clogit did not run. Why I am not able to run clogit on my dataset. I appreciate any input. Regards, Ashraf age <- c(rep(0,8),rep(1,8)) scc <- c(1,1,1,0,0,1,0,0,1,1,1,0,0,1,0,0) case <- rep(c(1,0),8) count <- c(5,5,216,216,110,110,40,40,5,5,308,308,212,212,21,21) #count <- c(2,5,3,5,4,6,2,4,2,5,1,3,1,4,2,5) agescc <- age*scc # Grouped data clmatch <- data.frame(age,case,scc,agescc,count) # Ungrouping the data as clogit as I can not use weights=count clmatch1 <- data.frame(lapply(clmatch, function(x) rep(x,clmatch$count)))[,1:4] # Sorting by age and case for clogit - I think it is not necessary ii <- order(clmatch1$age,clmatch1$case) clmatch2 <- clmatch1[ii,] clmatch2 clogit(case~scc+agescc+strata(age),data=clmatch2) -----------------______________________________________ M. Ashraf Chaudhary, Ph.D. Associate Scientist/Biostatistician Department of International Health Johns Hopkins Bloomberg School of Public Health 615 N. Wolfe St. Room W5506 Baltimore MD 21205-2179 (410) 502-0741/fax: (410) 502-6733 mchaudha at jhsph.edu Web:http://faculty.jhsph.edu/?F=Mohammad&L=Chaudhary