Does Virtual Reality Enhance Validity in Discrete Choice Experiments: A Comparison With Traditional Presentation Modes

Abstract

Stated preference methods such as discrete choice experiments (DCEs) are used to elicit respondent's preferences and willingness to pay (WTP) for a good or service. However, DCEs may deliver biased estimates, as respondents have been found to be unfamiliar and uncertain about what they are supposed to be valuing. There is evidence that visualization techniques such as photos can improve the cognitive ability of the respondents and thus increase their choice certainty. To complement the research on the effect of different visual representations of the information provided in choice sets, on people's choices we developed a split sample experiment. We presented people with corresponding attributes and levels using 3 different presentation formats, being (1) a text only matrix, (2) a video, and (3) an immersive virtual reality (VR) environment, in order to uncover respondent's preferences and WTP for different urban green scenarios. We found differences in the WTP estimates elicited using the VR and the other presentation formats. Moreover, changing the presentation mode even gave rise to preference reversal for some attributes. We also found a significantly higher scale parameter in the VR experiment compared to the other presentation formats, suggesting that the use of VR results in a lower degree of error variance, meaning that lay individuals provide more deterministic responses compared to when using traditional methods. Taken together, our results suggest that (1) the presentation mode in DCEs can have a significant impact on WTP estimates and that (2) uncertainty can be reduced by employing more realistic visualization techniques.