This article presents an econometric approach to modeling uncertainty, unwillingness to pay, and protest behavior in contingent valuation studies. For that purpose, a mixture model with sample selection is developed for a multiple‐bounded uncertainty elicitation format. The proposed theoretical framework is applied to evaluate the social welfare impact of implementing a sustainable rural development program. Results show that a “naive” analytical approach that excludes protesters from the analysis would result in significantly higher willingness to pay estimates for those individuals who favor the implementation of the program and agree to reveal their true reservation prices.