In a Good Society Debate article earlier this year, Christian Kroll argued that findings from the science of well-being should be used to inform policy-making. His argument is essentially the same as those presented by Lord Richard Layard (in Happiness: Lessons from a New Science) and Derek Bok (in The Politics of Happiness).
The argument runs along the following lines: Scientists have been investigating well-being and they have discovered some of its causes. Well-being is more important than traditional economic measures of progress. Therefore, policy-making should be informed (at least partially) by scientific findings about well-being.
Unfortunately, the argument suffers from an important oversight; it assumes that the scientists have been investigating the same phenomena – well-being. It is very unlikely that more than a few of these scientists are in fact measuring well-being because they are all measuring different things. Despite this problem, the following discussion does not argue against the use of the science of well-being to inform policy-making. Rather, a solution is suggested for how to get the best out of the science of well-being, now and in the future.
There are a number of different measures currently claimed to assess well-being. These measures include brain scans, daily reports about how participants have been feeling, the opinion of a friend, colleague or expert, the amount the participant smiles and, most commonly, survey questions. The survey questions are often markedly different. They might ask about happiness, satisfaction with life, the amount one would choose to change one’s life and many other distinct notions. It is the findings produced by analyses of these surveys that have been recommended as aids to inform policy-making. But should we trust that these surveys are actually measuring well-being?
Kroll, Layard and Bok all argue that we should because answers to survey questions about well-being are significantly correlated with many other measures of well-being. It is true that most of the diverse measures of well-being mentioned above are significantly correlated for most test participants. However, that certainly does not indicate that the different measures are evaluating the same phenomena. The important point that these researchers ignore is that most of these correlations are not particularly large. The significant, but small, correlations between various measures of well-being tell us two things: that the different measures are very likely to be tracking phenomena that are related in some way, but also that the phenomena being tracked are clearly distinct.
The fact that different questions in well-being surveys are not measuring the same phenomena is no revelation to many researchers. Philosophers have been discussing the merits of various conceptions of well-being for a long time. To a philosopher, the different kinds of questions used in these surveys often endorse one particular conception of well-being. It might be suggested that these philosophical differences are trifling distractions since all of the measures are assessing something that is obviously good. There are two serious problems with this suggestion.
First, not all of the measures are tracking something that is obviously good in a way relevant to public policy. If it were really obvious that a positive balance of pleasant feelings over negative feelings is good for someone, then we should want governments to lace the water supply with a side-effect-free pleasure drug. Not many people would vote for such a policy.
Second, what should policy-makers do when findings based on different measures of well-being imply different policies? This second problem might be easily pushed aside if it were not for the fact that many well-being surveys produce results that contradict the extant findings of the science of well-being. For example, according to Kroll it is well established that increases in income have no affect on the overall self-reported well-being of countries with per capita GDP over 10,000 Euros. However, well-being researchers using a different question have found a positive significant relationship between income and self-reported well-being amongst both poor and rich countries.
Since both of these findings are statistically significant, the ‘apparently’ very similar questions are the most likely cause of the divergent results. In this particular case, the findings that increasing income improves well-being well above 10,000 Euros are based on data from Gallup Polls. The question about well-being in the Gallup Polls is worded in a way that is more likely to elicit comparisons with all other people in the world, not just the respondents’ immediate reference group. It is reasonable to assume that survey respondents will report being more satisfied with their financial situation, and their whole lives, when they are encouraged to compare them with those who are destitute, rather than their with those of their friends and neighbours.
When the science of well-being produces contradictory findings in this way, it creates a problem for policy-makers. To prevent this from occurring, researchers analysing well-being survey data should never generalise findings from different questions unless those questions really are asking about the same phenomena. But what should policy-makers intent on using the science of well-being do when the findings are contradictory?
Policy-makers should always investigate the original surveys to find out if the well-being findings are about one conception of well-being. Such an investigation might lead to the identification of different questions about well-being as the source of conflict. Discovery of such conflicts illuminates the fundamental problems policy-makers intent on using the science of well-being face. Which question about well-being is the most appropriate basis for policy-making? What exactly is well-being anyway?
Despite thousands of years pondering these questions, philosophers have not yet come up with definitive answers. They have, however, identified the main conceptions of well-being, their advantages and their disadvantages. This knowledge should be shared and discussed widely. With a greater awareness of these problems and deeper understanding of what well-being might consist of, citizens can exercise their democratic rights and lobby their government to adopt their preferred conception. When this occurs, governments can encourage the use of appropriate measures of well-being. Only then can well-being researchers be confident that they are producing findings that are really relevant for policy-making. And only then can policy-makers get the most out of the science of well-being.
 For Example, see Angus Deaton, 2010. “Income, Aging, Health and Well-Being around the World: Evidence from the Gallup World Poll,” NBER Chapters, in: Research Findings in the Economics of Aging, pages 235-263 National Bureau of Economic Research, Inc.