Modeling Discrete Choice with Uncertain Data: An Augmented Mnl Estimator
8 Pages Posted: 13 Dec 2005 Last revised: 23 Feb 2016
Date Written: March 2, 2004
This article introduces a multinomial logit model that uses ancillary information to control for uncertainty in both the observed choices made by respondents, and in the attributes of a respondent's choice set. Simulated data are used to compare the performance of this estimator versus simpler models, under several different kinds of uncertainty.
Keywords: multinomial logit, simulated estimator, UIV, UDV, uncertain choices, uncertain variables.
JEL Classification: C25
Suggested Citation: Suggested Citation