Re: st: ldecomp with weights and mi estimate

[Maarten Buis <maartenlbuis@gmail.com>]

On Thu, Mar 15, 2012 at 4:08 PM, Lauren Beresford wrote:
> I am trying to use ldecomp to decompose direct and indirect effects on a mi svyset dataset. The syntax for ldecomp indicates I can use pweights (see  http://www.maartenbuis.nl/publications/ldecomp.pdf page 11), but I get error messages when I try to run the command with the weight options.  The command runs without the weight options.
>
> mi estimate, cmdok: ldecomp mgr [pweight=wtsurvy], direct(female) indirect(hrswk) nobootstrap   macro e(cmd) is not setmatrix e(V) is not set
>
> Am I doing something wrong, or is just not possible to use both weights and mi estimate with ldecomp?

-ldecomp- is user written (by me). Remember that the Statalist FAQ
askes you to specify where you got user written software from. In case
of -ldecomp- there are multple versions floating around in cyber
space: two on the Stata Journal archive and one on SSC. The most
recent on the Stata Journal archive corresponds with the one on SSC,
but this is not always the case. To avoid confusion you must say which
version you are using.

It is not possible to use -ldecomp- with weights and get standard
errors. This is implicit in the help-file of -ldecomp- (type in Stata
-help ldecomp-), which states: "fweights, pweights, and iweights are
allowed when the nobootstrap option is specified." The absense of
standard errors is what -mi- is complaining about. So, the answer is
that it is not possible to use -ldecomp- with weights and multiple
imputation.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------

*
*   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/