Re: st: Different results for confidence intervals for proportions using "svy: proportion" and "espost svy: tabulate" ??

[Steve Samuels <sjsamuels@gmail.com>]

Aldana:

The description of -estpost- states:

"estpost svy: tabulate posts a one-way or two-way table for complex
survey data computed by svy: tabulate." 

So the problem is the difference between -svy: tabulate- and -svy: prop-. For this, see:

http://www.stata.com/statalist/archive/2006-10/msg01125.html

Regards,

Steve

On Sep 9, 2013, at 8:49 AM, Aldana Rosso wrote:

Hi,

I am running the svy proportion command to obtain the confidence intervals
for a proportion. I realized that I obtained different results using  the
"official commands: svy: propotion"  and the user written "estpost svy:
tabulate race". Does someone know why?

Here is an example with the data obtained in the espost website
http://repec.org/bocode/e/estout/hlp_estpost.html#svy_tabulate

**************************
*Results with estpost
******************************

webuse nhanes2b, clear
svyset psuid [pweight=finalwgt], strata(stratid)

estpost svy: tabulate race
(running tabulate on estimation sample)

Number of strata   =        31                 Number of obs      =
10351
Number of PSUs     =        62                 Population size    =
117157513
                                              Design df          =
31

-----------------------
1=white,  |
2=black,  |
3=other   | proportions
----------+------------
   White |       .8792
   Black |       .0955
   Other |       .0253
         | 
   Total |           1
-----------------------
 Key:  proportions  =  cell proportions

. matrix list e(lb)

e(lb)[1,4]
       White      Black      Other      Total
r1  .84078371  .07248092  .01076856         .z

. matrix list e(ub)

e(ub)[1,4]
       White      Black      Other      Total
r1  .90927575  .12486062  .05846181         .z

**************************
*Results with svy
******************************

svy: proportion race
(running proportion on estimation sample)

Survey: Proportion estimation

Number of strata =      31       Number of obs    =      10351
Number of PSUs   =      62       Population size  =  117157513
                                Design df        =         31

--------------------------------------------------------------
            |             Linearized
            | Proportion   Std. Err.     [95% Conf. Interval]
-------------+------------------------------------------------
race         |
      White |   .8791545    .016689       .845117    .9131919
      Black |   .0955059   .0127491       .069504    .1215079
      Other |   .0253396   .0105423      .0038385    .0468408
--------------------------------------------------------------


Aldana

Aldana Rosso, MSc, PhD
Statistician
RC Syd - National centre for health care quality registers, Klinikgatan 22,
Skane University Hospital, SE-221 85 Lund, Sweden, Tel. +46 46 17 58 12
ar@rcsyd.se


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