We test the hypothesis that respondents stating divergent reasons for choice uncertainty differ in their probability to vote yes in a CV setting. We introduce the use of a follow-up question used to classify uncertain respondents according to reason for uncertainty. Results indicate that...