Meeting US ethanol blending mandates proposed by the Environmental Protection Agency will require a substantial number of motorists with flex-fuel vehicles to switch from low ethanol-gasoline blends to high ethanol-gasoline blends. The lower the willingness to pay for high-ethanol blends, the greater the cost of complying with the proposed mandates. Existing estimates of the willingness to pay for high-ethanol blends use data from Brazil (where consumers have knowledge of and experience with high-ethanol blends), data generated when retail prices greatly favored low-ethanol blends, or stated data collected from mail and online surveys. To obtain more accurate estimates of US willingness to pay, we conducted an intercept survey in five US states of motorists with flex-fuel vehicles as they were refueling. We address a sample-selection problem caused by the lack of stations that sell high-ethanol blends; consumers who have a high willingness to pay are more likely to seek out the stations and hence to show up in our sample. We attempt to overcome the problem caused by prices favoring low-ethanol blends by augmenting revealed preference data with stated preference data generated by hypothetical prices that tended to favor high-ethanol blends. Our estimates of mean willingness to pay shows that the price at which the average US consumer will switch fuels is substantially below the price that equates the cost per mile of driving. The large discount that the average US consumer requires to switch suggests that the cost of proposed ethanol mandates will be higher than previously estimated.