Embedding as a pitfall for survey-based welfare indicators : evidence from an experiment

How can we assess the welfare of a society, its evolution over time and predict its change due to particular policy interventions? One way is to use survey-based welfare indicators such as the OECD Better Life Index. It invites people to weight a variety of quality of life indicators according to their individual preferences. 11 broad dimensions aggregate these indicators. Our experiment shows that people do not provide consistent ratings across differently labelled dimensions that embed the same indicators. They also do not adjust the rating of equally named dimensions changing sets of indicators. These results show that survey-based measures might suffer from strong embedding effects and, as a result, may fail to measure citizens' true preferences for the indicators.