RE: st: Understanding Factor variables - is order significant ?

[Richard Williams <richardwilliams.ndu@gmail.com>]

At 03:17 AM 5/26/2010, Jesper Lindhardsen wrote:
> Hi All,
>
> Thank you for taking such an interest in my (hairpulling) problem.
> I also think it is a bug in -poisson- as there is no other reasonable
> explanation for this behaviour in this specific regression command.
> I am not able to update Stata/MP to the April edition as it is server
> based, so I can't check if that does the trick.
> I guess I'll have to hard code the dummy variables until further
> clarification .....

If by "hard code" the dummy variables you mean code them so that 0 is 
the desired reference category, yes, that would work, and might be a 
good idea even if there was no problem (all these complicated 
notations may get a little confusing).  I wouldn't hard-code the 
interactions though, as that could cause problems with 
post-estimation commands like -margins-.  You could also just enter 
the terms in the order that does not cause problems.  Or, use ## 
instead of #.  I prefer ## anyway, as it separates the main effects 
from the interaction effect, and hence you may find that you don't 
need an interaction term anyway.


-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu
WWW:    http://www.nd.edu/~rwilliam

*
*   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/