We investigate the relationship between visual attention and attendence with the rate of "preference reversals" in a discrete choice experiment (DCE) that employed eyetracking. We nd that visual attention and attendance, counter to our initial expectations, is positively related to the rate of preference reversal. Our results indicate that moderately low levels of visual attention should not be used as a way of identifying individuals with low levels of engagement, nor should preference reversals necessarily be assumed to indicate low levels of participant engagement. We nd that those reversing preferences do not substantively di¤er from the rest of the population in terms of their underlying preferences. Rather, these respondents spend longer looking at tasks that are similar in terms of utility, so more complex, and as a result these respondents are more uncertain of the choice to make.