Model selection in iterative valuation questions

In this article, we propose a unified framework that accomodates many of the existing models for dichotomous choice contingent valuation with follow-up and allows to discriminate between them by simple parametric tests of hypothese. Our empirical results show that the Range model, developped in Flachaire and Hollard [2007], out-performs other standard models and confirms that, when uncertain, respondents tend to accept proposed bids.