[Limdep Nlogit List] introducing socio-economic characteristics in binary choice experiments
miq at wne.uw.edu.pl
Thu Jan 22 21:21:34 EST 2009
Rady Talaat wrote:
> For the binary unlabeled experiments, how can I incorporate the individual characteristics in the model? Although they may be introduced as interactions with the attributes to avoid hessian singularities, this inclusion of the separate attributes along with their interaction is likely to induce multicollinearity.
You could try other models, e.g. Error Component Model with
heterogeneity in the variances of the error components, Covariance
Heterogeneity Nested Logit or Random Parameter Logit. See NLogit manual
and the literature for details.
Warsaw Ecological Economics Center
University of Warsaw
More information about the Limdep