Weighted database regression analyses

[William Mitchell]

Dear Stata List,

I'm having trouble completing a weighted linear regression model using NHANES data, that's previously not been problematic.

My code (which usually works) is

svyset psu [pweight=exam_wts], strata(stratum) vce(linearized)
svy:regress vis_difficulty_score_7 i.MD_3level

, however, I'm getting the following error message:

no; data are mi set
    Use mi svyset to set or query these data; mi svyset has the same syntax as svyset.

    Perhaps you did not type svyset.  Some commands call svyset to obtain information
    about the settings.  In that case, that command is not appropriate for running
    directly on mi data.  Use mi extract to select the data on which you want to run
    the command, which is probably m=0.
r(119);

I have previously run correlation matrices, factor analysis and IRT (with imputed data, and _mi_m, _mi_id, and _mi_miss variables - which seem to be causing the problem)..

If anyone had any experience using weighted national database data, and how to troubleshoot this problem, I would really appreciate any advice. Thanks very much for your consideration.

Will

[William Mitchell]

Dear Stata List,

If it helps at all: I have found a document by Yulia Marchenko from 2010, which outlines using the mi estimate commands for multiple imputations (which I may be running in to?); which I have tried -

mi svyset psu [pweight=exam_wts], strata(stratum) vce(linearized)

mi estimate: svy: regress vis_difficulty_score_7 i.MD_3level

But unfortunately with this initial output:

     
pweight: exam_wts
VCE: linearized
Single unit: missing
Strata 1: stratum
SU 1: psu
FPC 1: <zero>

I get this subsequent error message:

no imputations
r(2000);

If anyone has any advice, it would be really great. Thanks again for looking.
William