Hi. When I calculate proportions of a binary outcome (0, 1) using
svy , subpop (if region == 1) : tab outcome , ci
svy , subpop (if region == 2) : tab outcome , ci
svy , subpop (if region == 3) : tab outcome , ci
svy , subpop (if region == 4) : tab outcome , ci
(where region takes the values 1-4 for four different geographical regions)
I get different CIs from when I cross-tabulate outcome and region:
svy : tab outcome region , col ci
although I assumed that the CI should be for the column, i.e. the
region. The proportions are obviously the same, as is the CI for the
total population.
(I use "tab" rather than "proportion" in order to assure my CI stays
within the limit 0,1.)
Could anybody explain me the reason for the different CIs and which one
would be the appropriate estimation?
Thank you,
Manuel
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
