st: Re: R: How to take into account correlation in Logistic regression using Stata

["Tiago V. Pereira" <tiago.pereira@mbe.bio.br>]

Dear Carlo,

I will check that example!

Once again thank you so much for your time and kind note.

All the best,

Tiago


--------------



Dear Tiago,
if I understand your question right, you may want to add a family_id in the
right hand side of the logistic regression.
Otherwise, there is an example on -vce(cluster hospid)-in Stata 12 .pdf
manual (pages: 938-939) (I suppose this is the one you refer to).

All the best for yor research project.
Kindest regards,
Carlo
-----Messaggio originale-----
Da: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] Per conto di Tiago V. Pereira
Inviato: giovedì 23 febbraio 2012 00:32
A: statalist@hsphsun2.harvard.edu
Oggetto: st: How to take into account correlation in Logistic regression
using Stata

Dear statalisters,

I have data for 500 subjects.

Subjects were sampled from a very small community. Hence, there are
participants from the same family (father, mother, son, etc.).

The dependent variable is a binary one (1=yes, 0=no).

I would like to fit a logistic regression model to test the influence of a
categorical variable on the odds for the binary outcome - allowing for
covariates.

In the -logistic- command, the cluster/robust options account for the lack
of independence among subjects.

However, I am unsure if those options fully suffice  (i.e. if one needs
fancier strategies in Stata).

Would it be possible to you to drop some lines on that issue, please?

All the best,

Tiago



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/