Re: st: compare two nested regression model

[Richard Williams <Richard.A.Williams.5@ND.edu>]

At 09:43 AM 10/23/2006, liu nizi wrote:
> dear all
>  I have develped two nested regression model based on the same data.
> How can I campare these two models by stata command?
> thanks
>                   fengqin
The -nestreg- command provides one easy to use approach.  You could 
also install the -spost9- routines of Long & Freese and use their 
-fitstat- command.  fitstat is quite handy for comparing various 
measures of fit for two models.  Example:

. webuse auto
(1978 Automobile Data)

. quietly regress  mpg  trunk weight

. quietly fitstat, s(m1)

. quietly regress  mpg  trunk weight length turn

. fitstat, using(m1)

Measures of Fit for regress of mpg

                               Current             Saved        Difference
Model:                         regress           regress
N:                                  74                74                 0
Log-Lik Intercept Only        -234.394          -234.394             0.000
Log-Lik Full Model            -194.238          -195.093             0.855
D                              388.477(69)       390.186(71)         1.709(2)
LR                              80.312(4)         78.603(2)          1.709(2)
Prob > LR                        0.000             0.000             0.425
R2                               0.662             0.654             0.008
Adjusted R2                      0.643             0.645            -0.002
AIC                              5.385             5.354             0.031
AIC*n                          398.477           396.186             2.291
BIC                             91.496            84.597             6.899
BIC'                           -63.096           -69.995             6.899
BIC used by Stata              409.997           403.098             6.899
AIC used by Stata              398.477           396.186             2.291

Difference of    6.899 in BIC' provides strong support for saved model.

Note: p-value for difference in LR is only valid if models are nested.



-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
FAX:    (574)288-4373
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu
WWW (personal):    http://www.nd.edu/~rwilliam
WWW (department):    http://www.nd.edu/~soc 
*
*   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/