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

Re: st: Difference in models

From   Maarten buis <>
Subject   Re: st: Difference in models
Date   Mon, 29 Oct 2007 16:40:44 +0000 (GMT)

--- "Herrnstadt, Evan" <> wrote:
> I have run two regressions, identical except for the main independent
> variable.  I need to figure out if the R-square is significantly
> larger in the first compared to the second.  Is there a simple test
> statistic?


> reg pct_error gdp_real_error lagdum*, cluster(trend)


> reg pct_error cbo_real_error lagdum*, cluster(trend)

The problem is that these two models are not nested, so there is not
test I know of (but I don't know everything) that can help you here.
What you can look at here are Information Criteria (BIC or AIC) that
you can recover using -est stats-. 

The real question is, do you really want to use a test here? Do you
have a hypothesis you substantively care about? Often when people
formulate there test in terms of R2 they are in a fase prior to
testing: selecting the model. If that is your situation, use the
formidable aresenal of model diagnostic tools discussed in 
-help regress postestimation- to get a feel for which variable performs

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

Want ideas for reducing your carbon footprint? Visit Yahoo! For Good
*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index