From the Miscellany Section of Economic Inquiry:
We collect contingent valuation data from 524 student survey respondents over a 3-day, 72-hour period. Data analysis of a hypothetical campus referendum focuses on time-of-day effects on willingness to pay for a renewable energy project. We find that subjects responding to the survey during the night-time hours (i.e., between 12 a.m. and 6 a.m.) do not display the law of demand, offering theoretically invalid responses to questions with important policy implications. Results from this research may have serious implications for the contingent valuation method (CVM). In short, just like your father said, nothing good happens after midnight when using the CVM. (JEL Q51)
The last line of the introduction says it all:
A guiding principle, consistent with the work of the CVM critics, is that a hypothetical valuation question, especially one used for major policy issues and natural resource damage assessments, should pass the most important theoretical validity test no matter what time of day the survey is administered.
I'll point out the funny parts if, like the folks we passed it around to for a quick read and the journal referees, you can't spot the humor (here's one: footnote 1 drops "cromulent" for the first time, I think, in the economics literature). And thanks to the editor for accepting the paper even when a referee said it should be rewritten as a real paper (and submitted where?). That was the original intention but we could never get motivated to do so. It was only during the hopeless to curious phase of my life that inspiration stuck.
*Dickinson, D. L. and Whitehead, J. C. (2014), DUBIOUS AND DUBIOUSER: CONTINGENT VALUATION AND THE TIME OF DAY. Economic Inquiry. doi: 10.1111/ecin.12161