On 7/29/08, Tiffany Davenport <[email protected]> wrote:
> I am estimating the following three equations:
>
> 1. Outcome= &agr;1 + &bgr;1TreatmentAContact + &bgr;2TreatmentBContact + &bgr;3Stratum1 +
> &bgr;4Covariate1 +.+ &bgr;9Covariate6 + &egr;1
>
> 2. TreatmentAContact= &ggr; + &dgr;1Treatment A + &dgr;2Stratum1 + &dgr;3Covariate1 +.+
> &dgr;8Covariate6 + &egr;2
>
> 3. TreatmentBContact = &pgr; + &thgr; 1TreatmentB + &thgr;2Stratum1 + &thgr;3Covariate1 +.+
> &thgr;8Covariate6 + &egr;3
>
> Where
>
> - All variables are binary including the "Outcome" dependent variable and
> all explanatory variables.
If all your variables are binary, then some sort of loglinear analysis
should be applicable.You are essentially manipulating a contingency
table. I believe that is a pretty standard biometric technique for
binary/count variables, and should be covered in Agresti's Categorical
Data Analysis and/or Hosmer and Lemeshow's Applied Logistic
Regression. I've never understood it well though. I don't know whether
the multivariate extensions are feasible, either.
--
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.
*
* 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/
