st: Different ways of finding the mean of survey data

[<I.Roland@lse.ac.uk>]

Hi Stata users,

I am working with survey data and at this stage I just would like to produce mean and median stats.

I have noticed a difference in means for variable x when using either of the following command sequences:

svyset id [pweight=pweight], strata(strata)

followed by (1) or (2):

(1) mean x [pweight=pweight]
(2) svy: mean x

I initially used (2) but then I found the following advice for computing the median in the FAQs:

_pctile x [pweight=pweight], p(50)

gen median_x=r(r1)

This led me to try (1) to verify that I get the same result for the mean as in (2) - and I do not. Then: if "svy: median x" existed, it would give a different result. Why would this be a valid method for computing the median if (1) is not a valid method for computing the mean?

Any help in clarifying this is very much appreciated!
Isabelle

Please access the attached hyperlink for an important electronic communications disclaimer: http://lse.ac.uk/emailDisclaimer

*
*   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/