The following fit does not make sense to me, please, correct me if I have a logical error.
> moddowsn Call: coxph(formula = Surv(start, stop, resp) ~ sn + matfac2, data = coxsn1, method = "efron") coef exp(coef) se(coef) z p sn2 0.0497 1.051 0.02030 2.450 1.4e-02 sn3 -0.0532 0.948 0.02038 -2.610 9.0e-03 sn4 -0.0410 0.960 0.01979 -2.073 3.8e-02 sn5 -0.0776 0.925 0.01954 -3.973 7.1e-05 sn6 -0.1133 0.893 0.01839 -6.161 7.2e-10 sn7 -0.1252 0.882 0.01846 -6.781 1.2e-11 sn8 -0.1222 0.885 0.01994 -6.130 8.8e-10 sn9 -0.0507 0.951 0.02047 -2.478 1.3e-02 sn10 -0.0444 0.957 0.02056 -2.159 3.1e-02 sn11 -0.0433 0.958 0.02157 -2.008 4.5e-02 sn12 -0.0114 0.989 0.02037 -0.557 5.8e-01 matfac22 -0.2599 0.771 0.01727 -15.048 0.0e+00 matfac25 -0.1804 0.835 0.00924 -19.512 0.0e+00 Likelihood ratio test=651 on 13 df, p=0 n= 253802 This would indicate that in sn6 to sn8 there is less of a chance of an event. ?? do the relative frequencies implied by the following table make any sense?? > table(coxsn1$matfac2,coxsn1$sn,coxsn1$resp) , , = 0 1 2 3 4 5 6 7 8 9 10 11 12 1 3407 3177 3425 3348 3975 3564 3181 3077 2894 2610 3441 3443 2 920 1005 1142 1327 1645 1530 1330 1184 964 864 888 860 5 9036 9507 10258 11888 16826 15575 13394 12346 9938 9001 8970 8599 , , = 1 1 2 3 4 5 6 7 8 9 10 11 12 1 1453 1459 1186 1496 1295 1754 1429 1153 1106 1234 965 1532 2 312 290 330 390 454 539 479 367 295 276 256 267 5 2994 3207 3371 3629 4095 5581 5837 3844 3400 3199 2705 3084 Apparently the frequency of an event is higher in the summer months. I apologize for not being able to disclose the dataset, but think that the table provides enough to address the question. Thanks everybody. Stephen B [[alternative HTML version deleted]] ______________________________________________ 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.