st: Why am I getting this error: "V1: string variables may not be used as factor variables"

[Peter Wielhouwer <peter.wielhouwer@wmich.edu>]

I'm running a multivariate mlogit on a subset of a dataset (the subset
N=945, and there are not very much missing data), and generating dy/dx
and margins for all of my covariates, many of which I've included as
indicator variables. When I run this from my do-file, after the first
set of dy/dx output I get the red error message:
"V1:  string variables may not be used as factor variables r(109);"
None of my covariates are string variables, however. Any thoughts?

From my do-file:
mlogit mvote i.pid7 b4.ideolresp_1 b1.vote12full b8.educ male age
b4.newsfollow  b5.marital bornagain2 anychildren childschoo2 c.income
b1.grpidmet3 i.bornus i.conditions if pid3B==1, baseoutcome(5)
    forvalues i = 1/5 {
        margins, dydx(*) predict(outcome(`i'))
        margins (*), predict(outcome(`i'))
    }

Thanks for your help!
Peter

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Peter W. Wielhouwer, Ph.D.
Western Michigan University
Associate Professor of Political Science
Director, Institute for Government and Politics
Email: peter.wielhouwer@wmich.edu
Phone (office): 269.387.5685
Phone (cell): 269.599.3544



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/