Does Willingness to Pay Reflect the Purchase of Moral Satisfaction? A Reconsideration of Kahneman and Knetsch

The first study in a recent paper by Kahneman and Knetsch (1992) (KK) demonstrates 2 different embedding effects in a set of data collected using the contingent valuation method. KK offer the hypothesis that responses to willingness to pay questions express willingness to acquire a sense of moral satisaction by a volunatry contribution to the provision of a public good. They then report the results of a follow-up study testing the hypothesis. KK incorrectly believe that in testing and supporting their 2nd hypothesis they have also tested and supported their first. KK's error is an example of a subtle, fascinating, and sidious methodological problem known as the cross-level inference. In KK's study, a cross-level inference problem occurs because aggregate correlational analyses based on means or medians do not necessarily provide the same results as individual correlational analyses.