st: mlogit and outreg after mlogit

[Dilek Cetin <dcetin@metu.edu.tr>]

Dear statalisters

I have two problems with multinomial logit estimation.

1. I run the first regression, mlogit y x1 x2 x3 x4 x5 x6 x7 .... x20
then I run the second regressin which is mlogit y x1 x2 x3 x7 ... x20  
x4 x5 zx6
y=1, ...,5

In the second regression the coefficents, standart errors and  
significance levels are totally different.

Even if I have tried the robust estimation, the problem still exists.

In the multinomial logit estimation why the order of the independents  
are matter???

2. When I used the outreg command after mlogit in my desktop, I have 4  
categories in the output file (y=1, ...,5 - outcome(1) is the base  
outcome). When I used this command in my laptop, I have only one  
category in the output file. How can I solve this problem?

All the best
Thanks

Dilek


*
*   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/