# st: RE: calculrating confidence Intervals in svyprop statements

 From "Nichols, Austin" To "'statalist@hsphsun2.harvard.edu'" Subject st: RE: calculrating confidence Intervals in svyprop statements Date Tue, 24 Aug 2004 16:15:29 -0400

```Hmm. The syntax diagram for -svyprop- in online help shows no
[common_options], so you can't specify the [,ci] option.  You might try

svylogit x
di "(" exp(lcb)/(1+ exp(lcb)) "," exp(ucb)/(1+ exp(ucb)) ")"

where lcb and ucb are the lower and upper confidence bounds reported by
-svylogit- so your interval is at least guaranteed to lie in [0,1].  Note
that if your interval lies absurdly close to 0 or 1, you should prefer to
lie close to zero, since Stata's precision will be better on that end.  In
other words, define your proportion so there are fewer ones than zeros,
since the usable logit coefficients are in (-706,39), approximately.

-----Original Message-----
From: hk489@nyu.edu [mailto:hk489@nyu.edu]
Sent: Tuesday, August 24, 2004 2:15 PM
To: statalist@hsphsun2.harvard.edu
Subject: st: calculrating confidence Intervals in svyprop statements

Hello,

I did survey proportion estimation by combining three variables, cat1Rr
cat2Rr  cat3Rr. Below is the simple statement I used:

svyprop  cat1Rr cat2Rr  cat3Rr

As a result, I got 8 mutually exclusive groups. The STATA output includes as
follows: Obs. / Est. Prop. / Std. Err. It doesn't show confidence intervals.

This is my question: How could I calculate the confidence interval for the
each of the eight groups separately?

All best,
Hongsoo

*
*   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/
```