Dear All,
I have 2 questions related to MNL model:
1) IIA: are the results of a Small-Hsiao test influenced by feasible
correlation between covariates in MNL? More precisely, if I drop the
independent variables strongly corr (and therefore I reduce the bias
in my results) with other var on the rhs I get the desired IIA. But if
I leave the model as it was at the beginning (for descriptive purposes
only) I don't get IIA for 1 on category. In other words, does somebody
know if correlation betw covariates induce IIA betw categories?
2) fitstat: this gives some measures of Goodness of Fit. But I read
that there is another parameter, called lamda (which is similar to
R2), which gives a reliable measure for the overall fit. Is there any
routine in Stata to allow that computation?
Thanks a lot
--
Chiara Mussida
PhD candidate
Doctoral school of Economic Policy
Catholic University, Piacenza (Italy)
*
* 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/
