Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: areg, regress, and cluster()


From   "Mark Schaffer" <[email protected]>
To   [email protected]
Subject   Re: st: areg, regress, and cluster()
Date   Thu, 22 Jul 2004 10:06:18 +0100

Garrett,

Kit answered your question, but I think it's also worth pointing out 
that the standard errors probably can't be used either way.  The 
reason is that you have only 8 clusters.  For the asymptotics of 
cluster-robust covariance matrix estimation to work, the number of 
cluster has to go off to infinity, and 8 is a long way from infinity. 
It's like running a regression with 8 observations (and, in your 
case, 7 unabosorbed regressors).  If you click on the hyperlink of 
the missing F-stat of your regression using -regress-, you'll get a 
more detailed explanation.

--Mark

Date sent:      	Wed, 21 Jul 2004 17:43:14 -0700
From:           	Garrett Glasgow <[email protected]>
To:             	[email protected]
Subject:        	st: areg, regress, and cluster()
Send reply to:  	[email protected]

> Hi,
>   I'm estimating a regression on how changing political party platforms affect
> vote shares.  I included country-specific dummy variables, and I'm also using
> robust clustered standard errors (clustering on countries) as there's likely to
> be (negative) correlation between parties in vote share.
>   I first estimated the model without clustering, first with areg, and then with
> regress and a set of dummy variables.  As expected, the results were identical.
>   However, when I add the cluster option it looks like Stata is making different
> corrections to the degrees of freedom in the t-test for statistical
> significance in these models, as well as doing some other things differently. 
> The output from both models is below:
> 
> . regress vgain vgainone ingovnow dirvshift pshift2a idparty idpshift
> Italy Britain Greece Luxembourg Denmark Netherlands Spain,
> cluster(ctrynum)
> 
> Regression with robust standard errors          Number of obs =     158
>                                                 F(  5,     7) =       .
>                                                 Prob > F      =       .
>                                                 R-squared     =  0.1477
> Number of clusters (ctrynum) = 8                Root MSE      =  4.5572
> 
> -----------------------------------------------------------------------
>              |               Robust
>        vgain |      Coef.   Std. Err.   t   P>|t|  [95% Conf. Interval]
> -------------+---------------------------------------------------------
>     vgainone |  -.1540838   .0951223 -1.62  0.149  -.3790123   .0708447
>     ingovnow |  -2.295443   .9807013 -2.34  0.052  -4.614434   .0235466
>    dirvshift |   3.332989   1.629255  2.05  0.080  -.5195878   7.185565
>     pshift2a |    .808739   .8450504  0.96  0.370  -1.189488   2.806966
>      idparty |   .2054125   1.252439  0.16  0.874  -2.756136   3.166961
>     idpshift |   -3.21951   1.211969 -2.66  0.033  -6.085362  -.3536578
>        _cons |  -.1867649   .8090965 -0.23  0.824  -2.099974   1.726444
> (7 dummy varibles omitted)
> -----------------------------------------------------------------------
> 
> . areg vgain vgainone ingovnow dirvshift pshift2a idparty idpshift,
> absorb(ctrynum) cluster(ctrynum)
> 
> Regression with robust standard errors          Number of obs =     158
>                                                 F(  5,   144) =   12.84
>                                                 Prob > F      =  0.0000
>                                                 R-squared     =  0.1477
>                                                 Adj R-squared =  0.0707
>                                                 Root MSE      =  4.5572
>                    (standard errors adjusted for clustering on ctrynum)
> -----------------------------------------------------------------------
>              |               Robust
>        vgain |      Coef.   Std. Err.   t   P>|t|  [95% Conf. Interval]
> -------------+---------------------------------------------------------
>     vgainone |  -.1540838   .0951223 -1.62   0.107  -.3421002   .0339326
>     ingovnow |  -2.295443   .9807013 -2.34   0.021  -4.233873  -.3570138
>    dirvshift |   3.332989   1.629255  2.05   0.043   .1126434   6.553334
>     pshift2a |    .808739   .8450504  0.96   0.340  -.8615666   2.479044
>      idparty |   .2054125   1.252439  0.16   0.870  -2.270128   2.680953
>     idpshift |   -3.21951   1.211969 -2.66   0.009  -5.615058  -.8239616
>        _cons |   .4581699   .8376469  0.55   0.585  -1.197502   2.113842
> -------------+----------------------------------------------------------
>      ctrynum |   absorbed
> 
> As you can see, the coefficients, standard errors, and t-ratios are identical. 
> However, the p-values associated with those t-ratios differs.
> 
> What accounts for these differences?
> 
> Thanks,
> Garrett
> 
> *
> *   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/

Prof. Mark E. Schaffer
Director
Centre for Economic Reform and Transformation
Department of Economics
School of Management & Languages
Heriot-Watt University, Edinburgh EH14 4AS  UK
44-131-451-3494 direct
44-131-451-3008 fax
44-131-451-3485 CERT administrator
http://www.som.hw.ac.uk/cert

*
*   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/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index