[Limdep Nlogit List] Cox proportional hazard model

Shao-Hsun Keng SHKeng at nuk.edu.tw
Thu May 15 14:39:25 EST 2008


 I want to use Cox proportional hazard model with time varying covariates to estimate  mortality hazards.  In the program, I need to specify a variable, t0 indicating the time of the beginning of the interval.  The data I have is a longitudinal data with 5 panels.  Each respondent should have 5 lines of data. What should the value of t0 in the first line of the data for respondent i be given?  There is no clear explanation in the manual.  My hunch is that t0 should be the first year that we observed the respondent in the survey.  For example, Respondent i was first observed in 1989 and his age was 60 at that time.  Then t0 in the first line of data for respondent i should be 60.  Am I right?

Best

Shao-Hsun Keng


More information about the Limdep mailing list