No, I didn't mean to say that. I just included the -regress-
specification to imply that the number of observations used in the
-areg- regressions should be smaller than the number from the original
-regress- results. Sorry if that was confusing. My question is: with
fixed effects using the variable "mother_rc", there are
obervations(children) with only one count per mother_rc, that is
mothers with only one child. Once I include the mother fixed effects,
those with only one child should be dropped from the regression since
with only one child, there is no variation within the mother. I'm
wondering why the number of observations reported is still the same
though. Thanks.
On Thu, Sep 17, 2009 at 4:57 PM, Martin Weiss <[email protected]> wrote:
>
> <>
>
> Are you implying that -areg- and -regress- results must differ? -areg-, as I
> understand it, is meant to replicate -regress- output when there are too
> many fixed effects for -regress- to handle. Example 1 in [R], p. 80, shows
> how the two commands return the same output:
>
>
> *************
> sysuse auto, clear
> regress mpg weight gear_ratio b5.rep78
> areg mpg weight gear_ratio, absorb(rep78)
> *************
>
>
>
> HTH
> Martin
>
> -----Ursprüngliche Nachricht-----
> Von: [email protected]
> [mailto:[email protected]] Im Auftrag von Sue
> Gesendet: Donnerstag, 17. September 2009 19:28
> An: [email protected]
> Betreff: st: -areg- question
>
> Dear Statalisters,
>
> I have a question on the number of observations used in -areg-. I run
> the following regressions:
>
> reg `var'_mort ``var'_controls' _Irel* _Ibirth* _Icountry* _Iethn*
> if touse`var' [pw = weight],
> cluster(clus_rc) robust
>
> areg `var'_mort ``var'_controls' _Irel* _Ibirth* _Icountry*
> _Iethn* if touse`var' [pw = weight],
> absorb(mother_rc) cluster(clus_rc) robust
>
> The only difference between the two is that the second is a regression
> with additional fixed effects using the variable "mother_rc". I'm
> wondering why the number of observations is the same in both
> regressions, even though there are values of mother_rc(the group id)
> that have only one observation and should therefore be dropped and not
> used in the regression. I'm wondering if Stata drops those
> observations but still reports the total sample size, or if there's
> anything wrong. I'd really appreciate your input.
> *
> * 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/
>
>
> *
> * 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/
>
*
* 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/
