Hi everyone,
I am trying to use the ei program to analyze voting data, by ethnicity,
within African countries. The problem I am encountering is very similar to
one discussed in the Feb 2012 archived messages.
No matter how I specify my variables (as counts or proportions), I get an
error message:
> formula2 <- cbind(turnout04, noturnout04) ~ cbind(Akan, Ga, Ewe, Guan,
Gurma, Mole, Grusi, Mande, Others)
> dbuf2<- ei(formula2, total = "pop18", data=STATA)
[1] "Running eiRxC"
Error in BayesMDei(formula, data, total = total, lambda1 = lambda1, lambda2
= lambda2, :
column marginals are neither counts nor proportions - please
respecify data
I have looked over my proportions and counts and cannot find any problem.
One group did make up a sizeable part of the overall ethnic data, so I
replaced it with disaggregated ethnic totals to replace that aggregated
group (my aim is to use these disaggregated ethnic totals in the long run).
The same error message was returned.
I have also successfully run a 2x2 analysis for this data.
Any clues, similar experiences, etc??
Thank you for any and all help!
Jennifer Boylan
--
PhD Candidate
Department of Political Science
University of Florida
PO Box 117325
Gainesville, FL 32611-7325