This quote from Michael Moss's book Salt, Sugar, Fat accurately sums up one of the main reasons I see economic analysis as preferable to psychological explanations (and it is one of the main reasons I often prefer non-hypothetical economic experiments to hypothetical surveys).
Pg. 150: “There is not a lot to be gained from asking people why they like something because they don’t bloody know.” - Fancis McGlone, former Unilever scientist