This paper describes an experimental method to calibrate an individuals ambiguity attitude and beliefs under a general decision model in risk and ambiguity. In the experiment, subjects are asked to state separately their desired level of investment at different return rates in situations of ambiguity and risk. This data allows us to calculate bounds for ambiguity aversion on a subject by subject basis, and then calibrate ambiguity attitudes and beliefs for each subject. The measures obtained are orthogonal to existing measures of risk attitude: there is no signi cant correlation between the Holt-Laury risk attitudes and our measured ambiguity attitudes and beliefs. Furthermore, these measures allow us to predict the subjects decisions in other ambiguous and uncertain situations. The experiment also uncovers differences in attitudes across individuals and environments: among the U.S. subjects, while almost all consistently behaved subjects are ambiguity averse or ambiguity neutral, female subjects are more ambiguity averse than males and Asian subjects are more ambiguity averse than Caucasian subjects. Similar experiments are also conducted using Chinese subjects but the gender difference is less significant.

We conducted quasi-field experiments on the stock trading floors in China to investigate how investors react to ambiguity relative to quantifiable risks and the degree of heterogeneity in these reactions. Our experiment consists of three sections; a background survey; individual self-reports of emotional states; and a series of individual portfolio choice questions involving ambiguous assets and assets with a known probability of success. By conducting these experiments in China, we were able to measure the degree of ambiguity aversion among a sample of experienced and accessible investors who face ambiguous decisions on a daily basis. We propose an index of ambiguity aversion that controls for risk aversion, show how to calculate it through a series of simple choices and demonstrate its outside validity. We find a significant degree of heterogeneity in ambiguity attitudes. We also discuss correlation between ambiguity attitudes and risk attitudes, and some demographic or emotional factors that might contribute to the heterogeneity of ambiguity attitudes.