Hi Maarten,
As I grapple with the survival analysis, I have many
questions but one I would like to ask you first is: why
does the data of Hosmer and Lemeshow (that you use in your
note as well) mainly consist of categorical covariates and
not continuous variables? Do continuous variables not lend
easily to survival analysis?
Except very few variables (like age), both the UIS data and
WHAS data (in tables 1.3 and 1.4 of their book - 1999
edition) mainly consist of categorical variables only. I
intend to analyze firms and would like to "control" for
firm-characteristics, which are all "continuous" variables.
(I say "control" because I don't see them as "predictors"
of hazard.) So my question is - can these variables be
used, given that they are continuous and only included as
"controls"?
I would sincerely appreciate your help on this.
Thank you,
Nishant
--- Maarten Buis <[email protected]> wrote:
> Nishant
> In your case I'd look at AICs and BICs since you want to
> compare non-nested
> models. -estout- will allows the options aic and bic,
> even if these statistics
> aren't "ereturned". However I wouldn't rely only on one
> statistic (or two pretty
> similar statistics) to choose between models. For
> instance for -stcox- and
> some of the -streg- models you'll want to check the
> proportional hazard
> assumption. Also for -streg- you will probably want to
> check whether you have
> used the right model. A very short overview of some
> methods that could help
> you choose a model in -streg- can be found on page 13 of
> my introduction to
> survival analysis text.
> http://home.fsw.vu.nl/m.buis/wp/survival.html . Also see
> the excellent book by Mario Cleves, William W. Gould, and
> Roberto Gutierrez
> "An Introduction to Survival Analysis Using Stata,
> Revised Edition" from
> Stata Press 2004. You can get this book from the Stata
> bookstore
> http://www.stata.com/bookstore/saus.html
>
> HTH,
> Maarten
>
> -----------------------------------------
> Maarten L. Buis
> Department of Social Research Methodology
> Vrije Universiteit Amsterdam
> Boelelaan 1081
> 1081 HV Amsterdam
> The Netherlands
>
> visiting adress:
> Buitenveldertselaan 3 (Metropolitan), room Z434
>
> +31 20 5986715
>
> http://home.fsw.vu.nl/m.buis/
> -----------------------------------------
>
> --- Nishant Dass wrote:
> > I want to estimate a hazard-rate model and I would
> like to
> > compare estimates from -streg- with those from -stcox-,
> > however, I am puzzled as to why -streg- and -stcox-
> report
> > different statistics for the overall fit of the model
> (I
> > used -ereturn list to check this).
>
>
>
> *
> * 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/
>
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
*
* 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/
