Referendum Contingent Valuation, Anchoring, and Willingness to Pay for Public Goods

This study reports on experiments that examine anchoring in single referendum questions in contingent valuation surveys on willingness to pay for public goods, and on objective estimation. Strong anchoring effects are found that lead to systematically higher estimated mean responses from Yes/No referendum responses than from open-ended responses. This response pattern is similar for contingent valuation questions and for objective estimation questions. The paper concludes that psychometric anchoring effects, rather than incentive effects, are the likely cause of results commonly found in contingent valuation studies, and that the currently popular single referendum elicitation format is highly vulnerable to anchoring.