# st: Re: loneway - comparing SD's

 From "Steichen" To Subject st: Re: loneway - comparing SD's Date Tue, 11 Nov 2003 10:11:57 -0500

```Rebecca,

That sd is returned in r(sd_b).

Generally, you can list all returned elements by the command -return list-
then pick out the one you want.

Tom

----- Original Message -----
From: "Rebecca Thornton" <rlthornt@fas.harvard.edu>
To: <statalist@hsphsun2.harvard.edu>
Sent: Monday, November 10, 2003 6:49 PM
Subject: st: loneway - comparing SD's

> I'm trying to compare the estimated SD of "between" and "within" variances
> for two groups, a treatment group and a comparison group.  Is there a
> suggested way to recover the standard errors in order to test the
> hypothesis that (estimated SD between effect Treatment group) = (estimated
> SD between effect Comparison group) ?
>
> For example:
> The output from loneway for the treatment group is:
>
>  loneway score00nm schid if std==6 & teso==0 & male==0 & treat==1
>
>                   One-way Analysis of Variance for score00nm:
>
>                                               Number of obs =       345
>                                                   R-squared =    0.5995
>
>     Source                SS         df      MS            F     Prob > F
> -------------------------------------------------------------------------
> Between schid          160.00569     30    5.3335229     15.67     0.0000
> Within schid           106.89158    314    .34041906
> -------------------------------------------------------------------------
> Total                  266.89727    344    .77586416
>
>          Intraclass       Asy.
>          correlation      S.E.       [95% Conf. Interval]
>          ------------------------------------------------
>             0.57069     0.07739       0.41900     0.72238
>
>          Estimated SD of schid effect            .6727015
>          Estimated SD within schid               .5834544
>          Est. reliability of a schid mean         0.93617
>               (evaluated at n=11.03)
>
>
> and for the comparison group is :
>
> loneway score00nm schid if std==6 & teso==0 & male==0 & treat==0
>
>                   One-way Analysis of Variance for score00nm:
>
>                                               Number of obs =       404
>                                                   R-squared =    0.5171
>
>     Source                SS         df      MS            F     Prob > F
> -------------------------------------------------------------------------
> Between schid          169.09683     33    5.1241465     12.01     0.0000
> Within schid           157.89682    370    .42674816
> -------------------------------------------------------------------------
> Total                  326.99365    403    .81139864
>
>          Intraclass       Asy.
>          correlation      S.E.       [95% Conf. Interval]
>          ------------------------------------------------
>             0.48329     0.07764       0.33111     0.63546
>
>          Estimated SD of schid effect            .6317745
>          Estimated SD within schid               .6532596
>          Est. reliability of a schid mean         0.91672
>               (evaluated at n=11.77)
>
>
> I want to test that hypothesis that 0.6727 = 0.63177
>
>
> but don't know how to recover the standard errors.
>
>
> Thanks,
>
>
>
> Rebecca
> *
> *   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/
>

*
*   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/
```