For "better" or "worse": a direct approach to elicit preference rankings from life-satisfaction data

We provide empirical evidence that intra-personal changes in life-satisfaction are a biased measure of preference rankings and, as a remedy, propose the ranking measure where subjects state whether their life has become "better" or "worse". Three representative datasets reveal: intra-personal changes in satisfaction levels are dominated by noise and less well explained by socio-economic variables than the ranking measure; the deviation between the two measures is systematic and adaptation (e.g. to income and unemployment) is only observed for changes in satisfaction levels but not for rankings, indicating that adaptation is driven by the elicitation method rather than changes in preferences.