6 June 2017Ever since Ellsberg (1961), the distinction between risk, where agents assign well-defined probabilities to possible outcomes, and ambiguity, where agents do not, has been of particular interest. Using a carefully-designed field experiment, we elicit information about risk and ambiguity preferences among 197 French farmers and structurally estimate these preferences. e use cumulative prospect theory and a multiple-prior model in order to model risk and ambiguity preferences, respectively. We find that farmers are risk, ambiguity, and loss averse, and that probability distortion differs in gains vs. losses, as well as in risk vs. ambiguity. These findings can have important implications for policy design.