Abstract
This paper introduces model specifications that can be used to explain response incentive effects that might occur with discrete response contingent valuation data when follow-up responses are collected. The models allow for possible random response shocks, structural shifts in willingness to pay between payment questions and heteroskedasticity between and within responses. Three well-known contingent valuation survey datasets that include follow-up payment questions are used to empirically test the models.
* University of Colorado, Boulder
** University of Minnesota, Minneapolis