Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | Wameq Raza <wameq.r@gmail.com> |
| To | statalist@hsphsun2.harvard.edu |
| Subject | st: Alternatives to clogit? |
| Date | Wed, 10 Oct 2012 01:43:56 +0200 |
Hello members, At this I point I don't have a question pertaining to STATA rather about potential modeling choices. The point of my paper is to see the determinants of health seeking behaviour and I have a list of 5 providers with "no care" as a base alternative. Upon running the Hausman test I discovered that my alternatives do in fact violate the IIA properties and I was forced to deviate from the mixed logit model. As for the case specific variables, I'm using the MNP model, but I'm somewhat confused about how to proceed in terms of the alternative specific variables such as cost and distance. One potential alternative I've been reading about is the conditional probit model, but the predominant impression I get is that beyond 4-5 alternatives, the results are quite dubious. Can any one suggest any potential solutions? My apologies for the non-STATA related question, but I'd very much appreciate any help I could get! Best Wameq -- W A M E Q R A Z A * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/