This study investigates whether choice experiments (CEs) are incentive compatible by comparing hypothetical and nonhypothetical responses to questions about beef ribeye steaks with different quality attributes. The results show that hypothetical responses generally predict higher probabilities of purchasing beef steaks than nonhypothetical responses, indicating that hypothetical choices overestimate total willingness-to-pay (WTP) for beef steaks. However, marginal WTP for changes in steak quality is not statistically different between hypothetical and actual payment settings. The study involved a random sample of consumers who evaluated quality-differentiated beef steaks. Participants were asked to choose between five steaks and a "none of these" option in seventeen repeated CE questions. Two treatments were used: hypothetical and nonhypothetical. In the hypothetical treatment, participants were told that payment would not actually occur, while in the nonhypothetical treatment, they were informed that one of the questions would be binding and they would purchase the steak they chose. The study used various econometric models, including multinomial logit (MNL), universal logit, heteroskedastic extreme value (HEV), multinomial probit (MNP), and random parameters logit (RPL), to analyze the data. Results from these models consistently showed that hypothetical and nonhypothetical responses generated different taste parameters, indicating that hypothetical bias exists in CE responses. However, marginal WTP for steak attributes was similar in both settings. The study found that total WTP to obtain beef steaks was greater in the hypothetical setting than in the nonhypothetical setting, but marginal WTP for changes in steak quality was generally equivalent. The results suggest that while hypothetical bias exists in CE responses, it is relatively small compared to other studies. The study also found that individuals in both treatments behaved consistently with the weak axiom of revealed preferences across the seventeen CE scenarios. Overall, the study concludes that CE responses are not fully incentive compatible, and that nonhypothetical valuations are necessary for accurate valuation of private goods.This study investigates whether choice experiments (CEs) are incentive compatible by comparing hypothetical and nonhypothetical responses to questions about beef ribeye steaks with different quality attributes. The results show that hypothetical responses generally predict higher probabilities of purchasing beef steaks than nonhypothetical responses, indicating that hypothetical choices overestimate total willingness-to-pay (WTP) for beef steaks. However, marginal WTP for changes in steak quality is not statistically different between hypothetical and actual payment settings. The study involved a random sample of consumers who evaluated quality-differentiated beef steaks. Participants were asked to choose between five steaks and a "none of these" option in seventeen repeated CE questions. Two treatments were used: hypothetical and nonhypothetical. In the hypothetical treatment, participants were told that payment would not actually occur, while in the nonhypothetical treatment, they were informed that one of the questions would be binding and they would purchase the steak they chose. The study used various econometric models, including multinomial logit (MNL), universal logit, heteroskedastic extreme value (HEV), multinomial probit (MNP), and random parameters logit (RPL), to analyze the data. Results from these models consistently showed that hypothetical and nonhypothetical responses generated different taste parameters, indicating that hypothetical bias exists in CE responses. However, marginal WTP for steak attributes was similar in both settings. The study found that total WTP to obtain beef steaks was greater in the hypothetical setting than in the nonhypothetical setting, but marginal WTP for changes in steak quality was generally equivalent. The results suggest that while hypothetical bias exists in CE responses, it is relatively small compared to other studies. The study also found that individuals in both treatments behaved consistently with the weak axiom of revealed preferences across the seventeen CE scenarios. Overall, the study concludes that CE responses are not fully incentive compatible, and that nonhypothetical valuations are necessary for accurate valuation of private goods.