Are Willingness to Pay Estimates Derived from Discrete Choice Experiments Plausible? An Investigation of Overshooting Using a Simple Criterion for Face Validity

We define a simple criterion for assessing the plausibility of Willingness to Pay (WTP) estimates obtained from discrete choice experiments. The criterion compares WTP for the bundle of attributes yielding the highest utility with a threshold value related to the highest level of the cost attribute. Estimated WTP that exceeds the threshold is considered to be implausible and overshoot. Failing this simple face validity test points to issues with i) the design of the choice experiment and/or ii) respondents’ behavior deviating from assumptions underpinning choice experiments. A random sample of choice experiment publications including 305 observations confirms that overshooting of WTP is an issue in the published choice experiment literature, with 66% of observations failing to pass the simple face validity test. We provide an exploratory analysis to identify factors associated with overshooting of WTP across studies, and discuss contributing factors related to study design and respondent behavior. Based on that, we e recommend that researchers dedicate greater effort to cost vector design. Studies that find that WTP exhibits overshooting should acknowledge the problem and scrutinize their design and data for potential contributing factors