Stated preference methods assume respondents' preferences are consistent with utility theory, but many empirical studies report evidence of preferences that violate utility theory. This evidence is often derived from quantitative tests that occur naturally within, or are added to, stated preference tasks. In this study, we use qualitative methods to explore three axioms of utility theory: completeness, monotonicity, and continuity. We take a novel approach, adopting a 'think aloud' technique to identify violations of the axioms of utility theory and to consider how well the quantitative tests incorporated within a discrete choice experiment are able to detect these. Results indicate that quantitative tests classify respondents as being 'irrational' when qualitative statements would indicate they are 'rational'. In particular, 'non-monotonic' responses can often be explained by respondents inferring additional information beyond what is presented in the task, and individuals who appear to adopt non-compensatory decision-making strategies do so because they rate particular attributes very highly (they are not attempting to simplify the task). The results also provide evidence of 'cost-based responses': respondents assumed tests with higher costs would be of higher quality. The value of including in-depth qualitative validation techniques in the development of stated preference tasks is shown.
