 |
 |
> I wondered if anyone on allstat had had problems with the standard errors > for survival predictions produced by SPSS? The standard errors in SPSS are > always noticeably lower than those computed by SAS (BASELINE statement) > and the decrease is often three-fold or more even with large sample sizes. > So although the packages use slightly different algorithms this is > unlikely account for such large differences. The SAS standard errors > appear to be more appropriate when compared to the model parameter > estimates. I'd be interested in hearing from anyone who can help shed any > light on this discrepancy. > > Many thanks, > > Helen Brown > Scottish Cancer Intelligence Unit %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%