[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: weibull with frailty and ancillary parameters

From   Maarten buis <>
Subject   Re: st: weibull with frailty and ancillary parameters
Date   Wed, 15 Aug 2007 12:24:48 +0100 (BST)

--- Aysegul Aydin <Aysegul.Aydin@Colorado.EDU> wrote:
> I have a question about a Stata command that Blossfeld, Golsch and
> Rohwer use in their book "Event History Analysis with Stata" (p.
> The command is;
> streg edu coho2 coho3 lfx pnoj pres, dist(w) time///
> ancillary (edu coho2 coho3 lfx pnoj pres) 
> where the ancillary option links the covariates to what the authors
> call b parameters.
> My question is;
> 1) What purpose do the ancillary parameters serve? It is my
> understanding that the ancillary parameters directly estimate the
> individual effect of a covariate on duration dependence whereas in
> standard weibull regression, the model produces a duration dependence
> parameter p but does not allow the researcher to analyze such
> individual impacts. Is this right? 

No, the way I think about this is that there is a baseline hazard
function with a fixed shape (typically first increasing than
decreasing) and the normal paramters tell by how much this baseline
hazard function should be sqeezed or streched for every unit change in
the explanatory variable. What the ancilliary parameters allow is to
also change the shape of that baseline hazard function beyond just
squeezing or streching it.

> 2) Is it possible to use this command with the frailty option? Right
> now, Stata refuses to allow the frailty option with 'ancillary'. Are
> you aware of other ways to integrate frailty into the model? 

Quoting from the helpfile of -streg-: "ancillary(varlist) <snip> This
option may not be used with frailty(distname)."

Hope this helps,

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Want ideas for reducing your carbon footprint? Visit Yahoo! For Good
*   For searches and help try:

© Copyright 1996–2022 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index