[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: testparm

From   Maarten buis <>
Subject   Re: st: testparm
Date   Tue, 4 Mar 2008 11:16:17 +0000 (GMT)

--- Chiara Mussida <> wrote:
> my results therefore do not allow me to estimate separate models? 

No, I stated -testparm- is irrelevant for that decision. If you want to
do it separately than by all means do so, the results from -testparm-
will not give you any information about whether that is a good idea or
not. I also made the point that in non-linear models, like the ones you
want to estimated, the differences across regions do not have a causal
interpretation. But this has nothing to do with -testparm-, this is
always the case.

> as regards the interactions, do you mean to estimate a fully
> interactive models w r to the region of residence for every separate
> category?
> in this case I already have models separated by groups of regions.
> This means that their geographical bunary variable is 1 (it is the
> criteria for the separation), individuals in each model are all in
> the same geographical location, and i don't know if it is meaningful
> to estimate fully interactive models...

I stated that doing the models separately for each region is the same
as a full interactive model. This means that you could either have all
regions in one model and add all the interaction terms, or do the
models separately for each region. The result would be the same. 

Hope this helps,

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Rise to the challenge for Sport Relief with Yahoo! For Good
*   For searches and help try:

© Copyright 1996–2019 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index