Re: st: GLS interpretation

[Maarten Buis <maartenlbuis@gmail.com>]

On Thu, Sep 15, 2011 at 3:15 PM, bucur sorana wrote:
> Can you please tell me anything about the Wald chi2(5) value and the z for income variable, because the income z is very high in comparison to the other variables?
> . xtgls growth income trade population2 school_sec2 kaopen2, panels(correlated)corr(ar1)rhotype(dw)
>
> Cross-sectional time-series FGLS regression
>
> Coefficients:  generalized least squares
> Panels:        heteroskedastic with cross-sectional correlation
> Correlation:   common AR(1) coefficient for all panels  (0.5519)
>
> Estimated covariances      =       190          Number of obs      =       304
> Estimated autocorrelations =         1          Number of groups   =        19
> Estimated coefficients     =         6          Time periods       =        16
> Wald chi2(5)       =  3.54e+07
> Prob > chi2        =    0.0000
>
>
> growth            Coef.         Std. Err.          z         P>z         [95% Conf. Interval]
>
> income          .984125     .0001666     5907.06  0.000     .9837985    .9844515
> trade             .0010546   .0000313         33.68  0.000     .0009933     .001116
> population2    -1.025569  .0011646      -880.65   0.000    -1.027852   -1.023287
> school_sec2  -.0004798  .0000128        -37.55  0.000    -.0005048   -.0004547
> kaopen2        .003106     .0004466          6.96   0.000     .0022308    .0039813
> _cons           .0189592   .0024439           7.76  0.000     .0141693    .0237491

"Wald chi2(5)       =  3.54e+07" means 3.54*10^7, i.e. 35.4 million.
That is a lot, under the null hypothesis you expected a draw from a
chi square distribution with 5 degrees of freedom, i.e. an average
value of 5. So this is a huge deviation form what you would expect
under the null hypothesis. I don't think the z-value of income is on
its own a problem, it looks to me consistent with all your other
z-values, all of which I would normally consider suspiciously large.
However, the kind of z-values you can reasonably expect differ greatly
from problem to problem. To see what is normal for your problem I
would look for articles on similar problems and see what they find.

Hope this helps,
Maarten
--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/