Thanks - sorry about the missing "no[1,1]" in my original di line
corrected below (but not correct of course see Jeff's reply)!
On 1/25/06, Jeff Pitblado, StataCorp LP <[email protected]> wrote:
> Richard Palmer-Jones <[email protected]> asks
>
> > How do I recover the CIs from svymean?
> > Using (for example)
> >
> > svymean s
> > mat no = e(_N)
> > di ci = _b[s] +/- invttail(no[1,1]-1,0.025)* _se[s]
> >
> > does not give exatly the same figures as reported by the output
>
> Richard has most of the formula correct, except for the degrees of freedom.
> -svymean- (and -svy: mean- in Stata 9) save the degrees of freedom in
> -e(df_r)-. To reproduce the CIs reported by -svymean-, type
>
> . svymean s
> . di "95% lower limit = " _b[s] - invttail(e(df_r), 0.025) * _se[s]
> . di "95% upper limit = " _b[s] + invttail(e(df_r), 0.025) * _se[s]
>
> --Jeff
> [email protected]
> *
> * 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/
