Ambiguity aversion has shown to be economically relevant and has been proposed as an explanation for many phenomena in economics and fi nance. While the literature has suggested a large variety of elicitation methods to measure ambiguity preferences, their consistency and reliability it is rarely evaluated. This is the fi rst study that systematically analyses the consistency of individual ambiguity preferences elicited using a variety of incentivized tasks, non-incentivized thought experiments and survey questions. We fi nd a high degree of aggregate consistency across elicitation methods, but large discrepancies in degrees of individual consistency in pair-wise tasks comparisons. Finally, the study identi es a set of non-incentivized tasks that predict ambiguity attitudes elicited experimentally which may serve as a viable alternative when running laboratory experiments is unfeasible.