Sampling: Weighted Exogenous Maximum Likelihood Estimation

This lecture covers the Weighted Exogenous Maximum Likelihood Estimator (WESML) in the context of discrete choice models. The WESML is introduced as an estimation method similar to weighted least-squares in linear regression, providing consistency but not efficiency. It is recommended when other methods are not applicable. The lecture also discusses the use of WESML in different scenarios, such as with simple random sampling, endogenous sampling, and mixed endogenous variables. Additionally, the importance of incorporating weights in forecasting is highlighted. References to relevant literature, including the work of Manski and Lerman from 1977, are provided for further details.