In this paper, we introduce a methodology for estimating parameters of a nested logit modelwhen decision makers who choose one of the alternatives are systematically excluded from thesample data that is used to estimate the model (i.e., censored). Unlike existing methods forestimating discrete choice model parameters with censored data, which require exogenousinformation beyond the specification of the model to be estimated and the available sampledobservations, the proposed method requires no additional outside information. We demonstrateempirically that this approach is able to recover not only generic model parameters that apply tocommon attributes of all alternatives, but also parameters for alternative specific constants andvariables associated with both observed and censored alternatives. While the standard errors ofthe estimated parameters are larger than those of models estimated with uncensored data,censored data methods still hold great potential for applications where uncensored data isexpensive or impossible to collect.
展开▼
