Plattner, Dankwart
>
> Dear Nick, dear H�kon,
> thank you very much for your helpful replies.
> The method H�kon suggested works, and it might also work to
> define instead
> another category of the variable as omitted.
> However, Nick, when I have another base category then
> r2==15 the problem
> with the dropped variables disappears: they are not any
> more dropped and
> have good explanatory power, as expected. Though I see your
> point I feel
> that the division of labour, as you call it so nicely,
> between xi and the
> reg command is not really helpful. At least it caused some
> headaches to me.
Normally it should not matter to a user how a command is
implemented. However, it seems that to understand
what you got, which was your question, it is
necessary to know that -xi- is a wrapper program
which does some preliminary processing of your
command before passing over to another command, in
your case -regress-.
I can't diagnose precisely why you got different results
in different circumstances.
Nick
[email protected]
*
* 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/
