"Invalid psu variable name" error with mi svyset prefix

[Andrew Steward]

Hi, I am getting an error when attempted to svyset data and run a logistic regression after multiple imputation. I ran 20 multiple imputations. Then, I svyset the data as a whole. Then, when trying to run a logistic regression with the prefix "mi svyset," the error arises "invalid psu variable name r(111);". Output is below. I am confused because I was not aware I need to enter the psu when running the logistic regression command; I was under the impression I only needed to enter "mi svyset" as a prefix. I tried entering the psu as part of the prefix, but that also gave me the same error message.
Thank you in advance!


mi svyset c5varunit [pweight=w5cgfinwgt0], strata(c5varstrat)

pweight: w5cgfinwgt0
VCE: linearized
Single unit: missing
Strata 1: c5varstrat
SU 1: c5varunit
FPC 1: <zero>

mi svyset logistic cac5diffinc chi5medicare edudummy martstatdummy che5health c5dgender chd5age cdc5hlphrsdy hispdummy crl5dyourrace5 crl5dyourrace4 crl5dyourrace3 crl5dyourrace2 crl5dyourrace1
logistic cac5diffinc chi5medicare edudummy martstatdummy che5health c5dgender chd5age cdc5hlphrsdy
hispdummy crl5dyourrace5 crl5dyourrace4 crl5dyourrace3 crl5dyourrace2 crl5dyourrace1 invalid psu
variable name
r(111);

[daniel klein]

Then, when trying to run a logistic regression with the prefix "mi svyset," [...]

You only svyset your dataset once. The prefix(es) for estimation should then be

mi estimate : svy : ...

[Andrew Steward]

Thank you so much!
I believe this worked.

Now my results are non-significant, but that's the way it goes in doing science...
Thank you again