Re: st: "Separation" issue in clustered/Longitudinal binary data.

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Wed, 22/12/10, sigontw@uchicago.edu wrote:
> I am now working on a longitudinal dataset. The outcome
> variable is a binary variable (a patient-reported drug's
> side effect) with repeated measures for three waves. Now I
> have an intervention (whether the participant received the
> drug), and I have used xtgee, xtlogit and xtmelogit to model
> the effects of this intervention on the outcome in a few
> different ways. However, no matter which method I used, I
> always encountered the separation issue. 

I may be missing something obvious, but don't you need to use 
the drug in order to experience its side-effects. This is in 
part a substantive/medical issue, but also a matter of how the
data were collected. Even if you could experience the same 
symptoms without using the drug, I can easily imagine situations 
where questionnaires redirected respondents who do not use the 
drug to the next question, so they trivially cannot have 
reported the side-effects/symptoms, or in register data where 
these symptoms are defined as side-effects only when the drug is 
used, etc. If something like that is happening in your data, 
then it is hard to see how an "effect" of your treatment could
have a meaningful substantive interpretation. In that case the
problem is no longer "what kind of technique can I use to 
estimate my effect?" but "what effect do I want to estimate?"

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   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/