Attitudes towards risk and uncertainty have been indicated to be highly context-dependent, and to be sensitive to the measurement technique employed. We present data collected in controlled experiments with 2,939 subjects in 30 countries measuring risk and uncertainty attitudes through incentivized measures as well as survey questions. Our data show clearly that measures correlate not only within decision contexts or measurement methods, but also across contexts and methods. This points to the existence of one underlying "risk preference", which influences attitudes independently of the measurement method or choice domain. We furthermore find that answers to a general and a financial survey question correlate with incentivized lottery choices in most countries. Incentivized and survey measures also correlate significantly between countries. This opens the possibility to conduct cultural comparisons on risk attitudes using survey instruments. (JEL: C90, D0, D81, J10)The editor in charge of this paper was Stefano DellaVigna. Acknowledgments: Ferdinand Vieider gratefully acknowledges generous financial support from the Excellence Initiative at the University of Munich and the WZB Berlin. The experiments in Costa Rica and Nicaragua were financed by Formas through the COMMONS program. The following people provided help with translations and in organizing and running the experiments: Clara Inés Villegas-Palacio,

No abstract