[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: Does sample size and survival time pattern matter baseline survival calculation?
I am now using STATA ver7.0 to analyze banking failure
in Thailand by cox proportional hazards model. The
problem is that I can not obtain accurate baseline
survival probability for each observation. (most
observations have baseline survival prob. of 1!! and
almost final observations got the value of nearly 1!!
That is much different from the result i obtained from
estimating the sample cancer dataset that the
observations' baseline survival decrease
proportionally) Consequently, some observations have
zero hazard ratio!!! That means the prediction
accuracy result is very bad!! Is this possible?
I think there must be something wrong but wonder what
exactly cause this problem. My sample size 279
observations (15 subjects , 9 non-failed subjects and
6 failed). For 6 failed subjects, they have almost the
same survival time pattern since they became default
during the same period(crisis). Do different sets of
covariates affect these baseline survival prob?
I do hope that someone could help me on this..
Thanks in advance...
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
* For searches and help try: