Choice experiments have been extensively used in many stated preference studies and disciplines. In order to give external validity to stated preferences, participants must behave the same in an experiment setting as they do in reality. The difference between a decision made in a real experiment and a hypothetical experiment is commonly attributed to "hypothetical bias." In this paper, we investigate whether virtual reality can reduce hypothetical bias in choice experiments. We conducted a set of experiments using the Food and Drug Administration's new Nutrition Facts Label for yogurts. Participants were randomly assigned to non-hypothetical and hypothetical experiments, which differed in terms of visual presentations of the product in question: text, picture, and a virtual grocery store. Given the absence of real markets for the new label at the time of the research, we approximate the real market with a binding choice in our non-hypothetical experiments. Our results suggest that although the hypothetical bias with virtual reality is not statistically significantly different from those using text or picture, it is the smallest. Among participants who did not exhibit high simulator discomfort, results suggest that virtual reality can significantly reduce hypothetical bias in choice experiments.