Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Re: survival analysis with unknown censoring

From   Steven Samuels <>
Subject   Re: st: Re: survival analysis with unknown censoring
Date   Wed, 13 Apr 2011 11:35:39 -0400

Dear Leo:

I believe you are referring to a post by Stefan Wagner at I responded to that post 

In Stefan's case there were two types of failure, but he observed only one. I can't tell for sure from your question whether you have the same problem, but I don't think so.  I believe that you have four types of firms; please correct me if I am wrong.  All enter in 1988 and observation ends in 2008.

1. Those that have no failures in the study period 
2. Those that have one failure in the study period but do not re-enter.
3.  Those that have one failure, re-enter at a later date,  then  fail a second time  in the study period
4. Those that have one failure, re-enter at a later date,  but do not fail by the end of the study period

We would nee to see an example of the real data for one firm of each type, including whatever dates you had, with your actual variable names (not made-up names like "dep.var") 



Steven J. Samuels
Consultant in Statistics
18 Cantine's Island
Saugerties, NY 12477 USA
Voice: 845-246-0774
Fax:   206-202-4783

On Apr 5, 2011, at 11:25 AM, lquadros wrote:

Dear Mr. Wagner,

All dates in my data is known (1988-2008).  However, there are no failures
in the period of analysis for some firms, while other firms have just one
observation and fail.
My doubt is the time duration of the observations that do not fail. In
Stata, firms are without any value in the cases there are no failure. 
However when I run stset time, failure(dep.var) id(idfirm), the message
As my goal is to study the second entry in the market, for firms that have
just one observation and did not fail NO value is related to them, while
firms that have a failure in the first event has the value "zero". And if
the second event is a failure it has value "1".

Please, should I run other commands to avoid probable errors? Am I using the
right values for the circumstances that I mentioned?
Thank you very much.
Leo Quadros
Universidad Autonoma de Barcelona

View this message in context:
Sent from the Statalist mailing list archive at
*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index