Empirical implications of response acquiescence in discrete‐choice contingent valuation

The use of discrete-choice contingent valuation (CV) to elicit individuals' preference, expressed as maximum willingness-to-pay (WTP), although primarily developed in environmental economics, has been popular in the economic evaluation of health and healthcare. However, a concern with this method is the potential for 'over-estimating' WTP values due to the presence of response acquiescence, or 'yea-saying' bias. Based on a CV survey conducted to estimate physicians' valuation of clinic computerization, the extent of such bias was estimated from a within-sample open-ended valuation question following the respondents' discrete choice response. Analysis of this data suggests that not only was response acquiescence an issue, but also that the parametric estimation of mean and median WTP, the most common approach to estimating WTP from discrete-choice data, would potentially magnify such bias (to various degrees depending on the distributional assumptions applied). The possible extent of CV design versus analysis in discrete-choice methods therefore warrants further exploration.