I am using gllamm and looking for some advice on model construction. If there is an obvious source for the info, my apologies.
I am a frequent user of clogit in stata, and have used the gllamm framework to estimate a 'standard clogit' model. I have also got myself familiar with estimating a logit model with covariate measurement error using gllamm (following section 6 in the manual).
What I would like to do is estimate a clogit model with covariate measurement error. In particular, I would like to specify the utility function to be:
U = (a+bz)X +....
Where X is the attribute , z an unobserved latent variable, and a and b parameters. I then have observed repeated measures associated with z for each individual. The intuition is that the marginal utility associated with an attribute in the choice model is affected by a factor such as 'risk aversion' and I have a number of measures of risk for each individual.
My intuition says that I need to 'stack' both clogit and measurement model variables, in a similar manner to the logit example. What is unclear to me is how to indicate that z appears as an interaction with one of the variables in the utility function, rather than a variable itself.
Any advice would be very gratefully received.
Michael
NOTE NEW PHONE NUMBERS
__________________________________________
Dr Michael Burton
Associate Professor
School of Agricultural & Resource Economics (MO89)
Faculty of Natural & Agricultural Sciences
University of Western Australia
35, Stirling Highway
Crawley (Perth), WA 6009 Australia
tel +61 8�6488 2531
fax +61 8�6488 1098
The University of Western Australia: CRICOS Provider No. 00126G
�
*
* 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/
