A comparison of approaches to estimating confidence intervals for willingness to pay measures

This paper describes four approaches to estimating confidence intervals for willingness to pay measures: the delta, Fieller, Krinsky Robb and bootstrap methods. The accuracy of the various methods is compared using a number of simulated datasets. In the majority of the scenarios considered all four methods are found to be reasonably accurate as well as yielding similar results. The delta method is the most accurate when the data is well-conditioned, while the bootstrap is more robust to noisy data and misspecification of the model. These conclusions are illustrated by empirical data from a study of willingness to pay for a reduction in waiting time for a general practitioner appointment in which all the methods produce fairly similar confidence intervals.