Error using MI ICE - Variable Not Found r(111) But Var Was Not Listed in Syntax

[jaykaegee]

Hi everyone,

I am encountering a problem with my syntax for multiple imputation using mi ice. It returns an error stating that a variable (pbc_ctisc_) is not found but I never listed that variable anywhere in my syntax so I do not know why it is telling me that it is an error that it is not found. I did create a variable with the same stem earlier when working with the dataset but it is not referenced in this syntax. Does anyone have any thoughts?



Here is my syntax:


cd "Q:\Data Folder\Data Analysis\For X\Working Directory"

*MI ICE

generate regorgmtg = observation_eventtype_regorgmtg
tab regorgmtg

mi set mlong

mi register regular observation_eventtype_protest regorgmtg ///
observation_eventtype_largeorgmt observation_eventtype_wkshpconf ///
observation_eventtype_commmtg observation_eventtype_panel /// observation_eventtype_hearing
observation_eventtype_lobby ///
f_approvalnetworkcl f_approvalnetworkclcon f_participatorynetworkcl f_participatorynetworkclcon ///
f_targetparticipatorynetworkcl f_targetparticipatorynetworkclco


mi register imputed ///
closef a_organizer_total g_eventtraveltime_minutes ///
disc heard ///
prevev survey_numeventsall ppl poor ///
inc educ g_work_working g_work_hoursperweekall pb b_events_totalnever b_events_totalany ///
g_difficultyreadingenglish g_difficultyspeakingenglish d_prevpart_avg g_usborn g_yearsinusall ///
g_homelanguage_spanish g_benefits_wic g_benefits_foodstamps g_benefits_welfare g_benefits_publichousing ///
organizer_agediff observation_face // move these in real analyses to regular


set more off

mi ice observation_face observation_eventtype_protest regorgmtg ///
observation_eventtype_largeorgmt observation_eventtype_wkshpconf ///
observation_eventtype_commmtg observation_eventtype_panel /// observation_eventtype_hearing
observation_eventtype_lobby organizer_agediff ///
f_approvalnetworkcl f_approvalnetworkclcon f_participatorynetworkcl f_participatorynetworkclcon ///
f_targetparticipatorynetworkcl f_targetparticipatorynetworkclco ///
closef a_organizer_total g_eventtraveltime_minutes /// ADD IN OTHER ORGANIZER VAR?
b_childdiscipline_avg heard ///
o.survey_numeventsclosed survey_numeventsall ppl poor ///
educ inc g_benefits_wic g_benefits_foodstamps g_benefits_welfare g_benefits_publichousing ///
g_work_working g_work_hoursperweekall pb g_usborn g_yearsinusall ///
b_events_totalnever b_events_totalany ///
g_difficultyreadingenglish eng g_homelanguage_spanish ///
d_prevpart_avg if survey_complete==1 , ///
conditional(f_approvalnetworkcl: if survey_long==1 \ f_approvalnetworkclcon: if survey_long==1 \ f_participatorynetworkcl: if survey_long==1 \ ///
f_participatorynetworkclcon: if survey_long==1 \ f_targetparticipatorynetworkcl: if survey_long==1 \ f_targetparticipatorynetworkclco: if survey_long==1 ) ///
add(1)


After the output lists all of the regression equations, this is the output:


> -

Imputing [Note: in regression for organizer_agediff, permanently removing obs
> ervation_eventtype_protest due to collinearity]

[Perfect prediction detected: using augologit to impute survey_numeventsclose
> d]
..........1
file C:\Users\jkg273\AppData\Local\Temp\ST_0f000002.tmp saved
pbc_ctisc_ not found
r(111);

end of do-file

r(111);

[jaykaegee]

Hi everyone,

I am bumping this - hopefully someone can help me out with any advice on how to solve this. I've looked up similar issues but nothing is solving my particular issue.

Any thoughts?

Thank you!

[Federico Tedeschi]

I had the same issue some days ago. The only solution I found was to remove the "offending" variables from the dataset. It looks like Stata somehow "remembers" that in previous codes I was using those variables.