Correcting for On-Site Sampling in Random Utility Models

Abstract

This study demonstrates how the joint distribution of a set of conditional trip counts to a system of recreation-sites can be adjusted for on-site sampling. An econometric approach is proposed that addresses both the size-biased distribution of the sampled visits and the weighted distribution of reported visits to ancillary destinations in a multivariate random utility framework. Estimation results indicate that uncorrected models produce biased estimates of trip counts and welfare measures. The empirical application examines jet skiing in the Lake Tahoe region.