Thamana Lekprichakul <thamana@hawaii.edu> asks about combining survey data
that is collected over time, and why the -svy- commands are slower in Stata 9
than in Stata 8:
> The first question is about calculating a mean and its variance of a
> series of complexed survey over time. The data are not panel. the
> goal is to calculate average values of variables of interest and their
> confidence interval. What is the best to go about doing this?
Unfortunately, I do not have a reference for this one. Combining survey data
is a nontrivial endeavor, even if the same design is used to produce in both
datasets.
However usually the data centers that produce public use data often provide
updates to the sampling weights in cases like this. For example, the
providers of the NHANES datasets have sampling weight variables for 1999-2000,
2001-2002, and 1999-2002. They collect the data over a 2 year period, so the
first two sampling weight variables are for their respective data collection
periods; whereas the last sampling weight variable is for the accumulation of
the two periods.
> the second question is more for the Stata staff. i noticed that the
> svy: tab command in version 9.0 takes significantly much longer time
> to present results. why does Stata 9.0 regress in term of speed in
> this particular routine? how can i speed up estimation process in
> svy: tab.
Richard Williams <Richard.A.Williams.5@ND.edu> pointed out that I answered a
similar question on Aug. 9. Since then we have taken steps to significantly
speed up the -svy- commands. This "speed-up" will be in the next executable
update. This update should be available within the next few weeks or so.
--Jeff
jpitblado@stata.com
*
* 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/