Peter, I have used the coxph function in S-Plus to fit the logistic model for case control data. A call of the form
coxph(Surv(time,status)~ [variables] + strata(pair),method="exact") In the above call, status is 1 or 0 for case and control, time can be any value, but make them all the same and pair is the variable used to group the cases and controls (in your case patient). Kevin Peter Berg wrote: > Howdy, > > I'm hoping someone might know of an equivalent function in Splus which > allows me to do the same as the PHREG procedure in SAS. Specifically, I'm > interested in Splus functions or procedures which will let me do a > matched-pair conditional logistic regression. > > I need to do bootstrap, jackknife and cross-validation analysis of the > regression, and it's much easier in Splus than in SAS. > > I am using each person in my data set as his/her own case and control, hence > the matched pair requirement, and it's a logistic regression model. I can > provide the SAS code that I used with the PHREG model if that will help. > > This question is time sensitive, so your help is very much appreciated. > > Thanks in advance!! > > -Peter > > > . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
