Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: discrete survival time analysis with pgmhaz8
From
<[email protected]>
To
<[email protected]>
Subject
st: discrete survival time analysis with pgmhaz8
Date
Thu, 13 Feb 2014 09:28:28 +0000
Why would you expect to get the "correct" results, whatever "correct" is?
Before proceeding, you should check the background literature to appreciate how a discrete time proportional hazard with Gamma frailty can be fitted. Look in particular at Bruce Meyer's Econometrica article (cited in full in the -pgmhaz8- help file), or the discussion in my Survival Analysis MS at my website (see below signature).
It happens that in the case of single-spell discrete time survival data, the frailty can be 'integrated out', giving a closed-form expression for the likelihood. -pgmhaz8- fits that model. There is no corresponding expression if one uses a proportional odds discrete time hazard model (one with a logit link) and Gamma frailty, or indeed Gaussian frailty. The integrating out has to be done numerically. Hence, a proportional odds discrete time hazard model with Gaussian frailty can be fit with Stata programs such as -xtlogit- and their Stata 13 equivalents. (The frailty is a random intercept assumed normally distributed.) More about all this in the web materials (and accompanying Lessons).
Please note Statalist protocols: cite the provenance of user-written programs.
Stephen
------------------
Stephen P. Jenkins <[email protected]>
Survival Analysis Using Stata: http://www.iser.essex.ac.uk/survival-analysis
Stephen P. Jenkins <[email protected]>
------------------------------
Date: Wed, 12 Feb 2014 11:25:55 +0000
From: <[email protected]>
Subject: st: discrete survival time analysis with pgmhaz8
Dear all,
I'm performing a discrete survival time analysis use pgmhaz8. As the pgmhaz8 command uses the cloglog link, I'm not sure if I could use the same command with the logit link to perform a discrete survival time analysis with gamma fraility. So, do I still get the correct results when I change the cloglog link to a logit link?
Thanks,
Andri
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
Please access the attached hyperlink for an important electronic communications disclaimer: http://lse.ac.uk/emailDisclaimer
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/