Respondent Uncertainty and Ordering Effect on Willingness to Pay for Salt Marsh Conservation in the Brest Roadstead (France)

- We test for sensitivity of willingness to pay to the order of presenting bid amounts.
- The test is carried out for fully sure respondents and unsure respondents separately.
- An order effect appears to be not present for fully sure respondents.
- We find strong evidence of order effect for unsure respondents.

This paper explores the potential link between the sensitivity of willingness to pay (WTP) to the order of presenting bid amounts in contingent valuation questions (ordering effect) and respondent uncertainty. The resource being valued is a public project to protect salt marshes against the spread of an invasive aquatic plant in the Brest roadstead (France). Valuation uncertainty is captured through a variant of payment card format where respondents are given the opportunity to report their WTP as either a single value (Option A) or an interval of values (Option B). The ordering effect is tested using both parametric models that ignore and control for the potential sample selection bias related to the choice between Option A and Option B, as well as non-parametric models. The results suggest that (1) respondents place substantial WTP values on salt marsh conservation, and (2) the ordering effect is linked to respondent uncertainty since only uncertain respondents react differently to changes in the order of presenting bid amounts. Specifically, for uncertain respondents, putting bid amounts in ascending order yields lower welfare estimates than putting bid amounts in descending order or random order. Policy recommendations and options to deal with ordering effect are discussed.