st: Multiple failures and discrete time

["Kallimanis, Bellinda" <bkallimanis@fmhi.usf.edu>]

Hi!

I posted a message last week on this topic and Marteen kindly directed
to me to Stephen Jenkins site with lots of really interesting reading on
survival analysis. After reading through much of the material I am still
not sure how to proceed with the analysis of data with multiple failures
and discrete time. 

I went through the examples on discrete analysis using the -logit-
command and now I'm unsure how to build on this syntax to allow for
multiple failures and time varying variables. 

So say I have

id     t        lnt   wander   cps_cat  
8372   1          0        1         1  
8372   2   .6931472        1         1  
8372   3   1.098612        1         1  
8372   4   1.386294        0         2  
8372   5   1.609438        0         2  
8611   1          0        1         1  
8611   2   .6931472        0         1  
8611   3   1.098612        0         1  
8611   4   1.386294        0         1  
8611   5   1.609438        0         1  
8611   6   1.791759        0         1  
8611   7    1.94591        0         1  
8611   8   2.079442        0         1  
8844   1          0        1         0  
8844   2   .6931472        1         1

If I was just running a single failure model I would type

logit wander lnt

however I want to tell the model that the variable cps_cat varies over
time, does this mean I need -xtlogit-? Also, do I need to add a strata
variable to control for the multiple failures like one would if they had
continuous data? And if so how would I go about creating that variable
based on the example above? I am using Stata 9.2.

Any thoughts on this would be really appreciated. 

Kind Regards,
Bellinda


 


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/