The existing literature has addressed choice set complexity, as well as other choice set characteristics, as ancillary conditions (i.e. context effects) that can affect performance in stated choice tasks. Choice set characteristics can contribute to the perceived difficulty of a choice task and can lead respondents to adopt choice heuristics which may suggest that they are not choosing rationally based upon the full complement of information provided in the choice scenario. However, “choice difficulty” is not usually observable by the researcher. Objective measures of choice set complexity have instead been assumed to proxy for choice difficulty and these measures have been used empirically to shift the scale of the error term or the slope coefficients in choice models. In our stated preference survey, respondents are asked directly to rate the subjective difficulty of each of their choices. We use this unique opportunity to explore the determinants of subjective choice difficulty to assess how well the customary reduced-form proxies are likely to capture this behavioral aspect of subjects’ interactions with choice tasks. Common measures do not fully explain subjective choice difficulty, which also depends on the interplay among objective attribute-space complexity, the similarity of alternatives in utility space, and cognitive resource constraints.