Brookings-Wharton Papers on Urban Affairs: 20088. Discrete choice models with random parameters are specified to capture unobserved preference heterogeneity.In practice, researchers have often found it necessary to include more than one observation for each individual in the sample to obtain precise estimates of unobserved deviations from mean tastes.In the case of a disaggregate model of the choice of residential location, that could be done by estimating a household's ranking of alternative residential locations.But such data are not publicly available, so it is likely that a researcher would have to conduct a new and expensive survey of residential location choices to obtain satisfactory estimates of preference heterogeneity.
Stored only when both papers are in the corpus — 22 citations exist upstream.