[Limdep Nlogit List] Multivariate mixed model with correlated errors

Alok R. Saboo arsaboo at gmail.com
Mon Dec 20 13:33:51 EST 2010


I am trying to estimate a bi-variate model where one of the dependent
variables (DV) is binary and the other is continuous. Further, I want to
have the errors across the two DVs correlated. Basically, the model looks

y1 = X1*B1 + e1
y2 = X2*B2 + e2

I want to specify the two error terms (e1 and e2) to be bivariate normal. I
am not able to figure out how to estimate this model. The closest that I
could find is a binary probit, but that is not too helpful as I have both
continuous and discrete variables. Any suggestions would be highly


*Alok R. Saboo*
Doctoral Candidate in Marketing
Smeal College of Business,
Pennsylvania State University,
University Park, PA 16802.
Tel: +1 814 865 4091

More information about the Limdep mailing list