Dear Statalist,
For my dissertation, I am using data with a complex survey design and
am estimating a user written maximum likelihood program using the svy
option to obtain the correct standard errors. It seems to be working
the way I anticipate (i.e. the coefficients are not altered and the
standard errors are altered).
For background: I originally intended to use "method 2" outlined in
Rabe-Hesketh and Skrondal (2006)
"Multilevel modelling of complex survey data",
J.R. Statistic. Soc. A, 169, Part4, 805-827
to obtain the correct standard errors post-estimation, but realized
that Stata's ml, svy option should work for me and save me some time.
My problem is that I do not seem to be able to find detailed
documentation on the procedure that the "ml ..., svy " command
utilizes to obtain its standard errors. I will obviously need this to
confirm whether I am obtaining these standard errors correctly and
document my methodology within my dissertation.
Could someone please point me in the right direction or refer me to
the relevent documentation or articles reviewing Stata's methods
and/or performance in this regard?
Thank you ahead of time,
Ian Breunig
Ph.D. candidate
Colorado State Univeristy, Economics Dept.
*
* 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/
