This paper uses the interval data model to explore the determinants of uncertainty in two-way payment ladder and in multiple-bounded uncertainty choice surveys. It estimates the uncertainty function that relates the size of the willingness-to-pay range to explanatory variables, where one of them is a proxy of the actual willingness-to-pay. The combination of the interval data model and the inclusion of the proxy variable present some advantages over the ordinary least square estimations currently used in the literature. In particular, it reduces the risk of the omitted variable bias and it takes into account that the dependent variable is not fully observed.