[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 10:10:22 +0000 (GMT)

--- Chiara Mussida <> wrote:
> my question today is: if the null of testparm is rejected by the data
> (my 4 categories related to geographical regions of residence) can I
> quietly proceed with the estimation of (4 in my case) separated
> models for the duration of unemployment? or are there other
> procedures/test to perform?

Doing analyses seperately is in effect not only including the main
effects of region to your model, but also interaction effects between
region and the other explanatory variables. With testparm you are only
testing the main effects and not the interactions.

You don't need a test to see if you can add a variable. If you add a
variable or an interaction that is insignificant, than worst case
scenario is that it is inefficient. In the case of adding variables the
real issue is interpretation, you want to add confounding variables,
but often you do not want to add intervening variables to your model,
and you definately do not want to add two variables that are two
imperfect measures of the same concept.

However, if you do these analyses separately for different regions, the
differences in parameters across regions do not have a causal
interpretation in non-linear models (like logit and probit, but also
duration models). There has been a long and pretty exhaustive thread
with working papers and lecture notes you can download, literature
references, and simulations on this issue recently on the statalist
starting with:

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

Sent from Yahoo! Mail.
A Smarter Inbox.
*   For searches and help try:

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