Hi, I'm working with nonemployment spell data and want to estimate a
piecewise constant proportional hazard rate model. Most persons in my data
set experience more than one episode of unemployment. Moreover I would like
to control for unobserved heterogeneity regarding persons or groups of
persons. Is there a possibility in stata to implement a frailty model with
multiple-failure data? Right now, I end up with the following, because I can
not specify strata in combination with shared
xi: streg i.DUR3 i.CURRAGE3, d(e) iterate(20) strata(UENR2) cluster(PERSNR).
But as I understood, only standard deviations of coefficients are corrected
with this method. Actually I wanted to estimate some kind of random effects
model. Is there any solution to this problem? In case a combination of
multiple events and frailty is not possible, how do I find out, which aspect
is more important to consider? With the Akaike Information Criteria or
LR-Tests?
Thanks
Jana Bruder
*
* 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/
