This paper examines the degree of risk aversion in decision-making through a series of lottery choice experiments. The authors present subjects with paired lottery choices, where the crossover point to the high-risk lottery is used to infer risk aversion. They find that most subjects exhibit risk aversion, with few showing risk-loving behavior. When hypothetical payoffs are scaled up by factors of 20, 50, and 90, the high payoffs have little effect on risk aversion. However, when the high payoffs are actually paid in cash, subjects become significantly more risk averse. The authors propose a hybrid "power/expo" utility function that captures both increasing relative and decreasing absolute risk aversion, which fits the data patterns over a wide range of payoffs from several dollars to several hundred dollars. The results suggest that subjects cannot accurately predict their behavior under high incentive conditions, and that risk aversion is more pronounced than previously thought, even at low stakes.This paper examines the degree of risk aversion in decision-making through a series of lottery choice experiments. The authors present subjects with paired lottery choices, where the crossover point to the high-risk lottery is used to infer risk aversion. They find that most subjects exhibit risk aversion, with few showing risk-loving behavior. When hypothetical payoffs are scaled up by factors of 20, 50, and 90, the high payoffs have little effect on risk aversion. However, when the high payoffs are actually paid in cash, subjects become significantly more risk averse. The authors propose a hybrid "power/expo" utility function that captures both increasing relative and decreasing absolute risk aversion, which fits the data patterns over a wide range of payoffs from several dollars to several hundred dollars. The results suggest that subjects cannot accurately predict their behavior under high incentive conditions, and that risk aversion is more pronounced than previously thought, even at low stakes.