Anna Alexandrova on value judgements and the measurement of well-being

Let me say first what I mean by ‘value judgement’. Take the central tasks of science to be building theories—general claims about the world—and explanations—illuminating applications of theories to specific phenomena. These two activities inevitably push scientists beyond the facts they, or those before them, have observed. This is because facts do not uniquely compel a theory, and phenomena do not uniquely compel which theory applies to them. If there is a gap between facts on the one hand and theory and explanations on the other, something closes that gap when a theory or an explanation is adopted. Philosophers have long observed that this role can be played by judgements about what is the simplest, the most compelling, the most elegant, or the most general theory or explanation. In a sense, these are value judgements, but I mean something different: judgements about what is good, virtuous, wholesome, and fair—judgements that are moral, political, or prudential. So should we trust a science that makes value judgements?

If we didn’t allow any value judgement in science, we’d be tiresome sceptics; theorizing and explanation-building inevitably come with the baggage that plugs the gap between facts and theories. Without what philosophers call ‘epistemic values’—simplicity, elegance, generality, and so on—science would be just a collection of facts (if even that). However, values about how we should live seem somehow more easily avoided and less essential to science. Though many people are willing to allow that morals and politics can regulate the pursuit of science—which problems get studied, at what cost, and with what constraints—it does not seem like scientific knowledge itself presupposes value judgements. A government-funded genetics project is not necessarily a statement about how we should live; or, at most, it is so only indirectly. But government-funded research about happy cities, or mental health among soldiers, or the dynamics of poverty, or the suffering of refugees, I hope you’ll agree, sound different. Why? Because in those cases, the very object of research—well-being, health, happiness, suffering—requires as part of its definition a value judgement about what is good, fair, or right. We aren’t only saying that these are the things worth studying, but also taking a stand on what is to be well, to suffer, to be healthy, to be free, and so on. Pursuit of that kind of science engages questions that are normally within the scope of ethics, politics, and theories of good life.

Nothing seems to me as tantalizing as philosophy of social sciences: the study of whether and how it is possible to turn the scientific eye onto ourselves as individuals or communities. But this particular puzzle caught my attention when I came across scientific claims such as the following:

Happiness is not always conducive to well-being

Long commutes are associated with lower well-being

Early learning difficulties have a disproportionate impact on life well-being

I discovered a field within today’s social sciences and medicine that calls itself the science of well-being. As I got to know this field, it struck me that these scientists are, for the most part, regular, bona fide empiricists, doing everything normal scientists do: testing hypotheses, building models, measuring phenomena, respecting facts. But in being such normal scientists, they also make what seem like profoundly value-laden (or ‘normative’) judgements:

If positive emotions get in the way of seeing reality, they take away from well-being

Even if people willingly choose to have longer commutes, they hurt themselves

A child that does not develop her capacities is not flourishing

When I tried to fit the science of well-being within the philosophy of science that I knew, I encountered problems. By the time I started thinking about this ten years ago, philosophers of science were more accepting of the role of values in science, at least as compared to earlier in the field’s history. They were discussing values as arbiters of theories when facts aren’t enough (this is called ‘under-determination’), they discovered ways in which values determine the level of evidence required to accept or reject a hypothesis (this is called ‘inductive risk’), they wrote compellingly about the responsibility of scientists to research topics that affect the most people, rather than just the richest (there isn’t a name for this duty).

But none of these phenomena fit my case. The science of well-being presents a new kind of value-ladenness. I started by defining the phenomenon and giving it a somewhat prosaic name: ‘mixed claims’.

A mixed claim is an empirical claim in which at least one variable is defined in a way that presupposes a moral, prudential, or political value judgement about the nature of this variable.

I called them ‘mixed’ because such claims mix the factual and the normative.

Now I wanted to know whether mixed claims are legitimate in science. I found that when other philosophers noted ‘mixedness’, they seemed embarrassed by it. Ernst Nagel wrote that when faced with apparent mixedness (though he didn’t call it that), we should pry apart the normative content from the factual and only keep the factual in science. The normative is none of scientists’ business. Take my previous example, ‘Happiness is not always conducive to well-being’. Nagel would argue that the proper concern of science is only the following claim:

If well-being is understood as good functioning across many domains and over the course of our lives, then happiness can impede well-being.

This conditional does not take a stand on what well-being is; it only pursues the empirical consequences of defining well-being in terms of lifetime functioning.

I appreciated Nagel’s concerns, namely, that normative claims are too important to be left to science, and mixed science can be ideological, compromising the authority and neutrality of science. But I hated the division of labour that his picture implied: scientists take care of facts, while values are left to… who? Nagel didn’t say, but perhaps the democratic process? This neat division seemed suspicious, because the more I learned about the science of well-being, the more I saw that choices about values permeate the field, through and through. Suppose you want to measure whether people are satisfied with their lives. Do you just ask them? What if they are not truthful, or not sincere, or not serious? You might try to create conditions that would make their judgements more authoritative, such as giving them anonymity, or time and space for serious reflection and authenticity. But what counts as serious reflection? What counts as authenticity? These are value judgements themselves and it seems comical to stop scientists from making them, or to send them running to whoever is the authority on values.

Not only is the Nagelian arrangement messy, it is also dismissive of scientific knowledge. Dismissive because (and here was a moment of epiphany for me) some scientific knowledge is about values! I owe this epiphany to reading the work of Elizabeth Anderson. She observed that when feminist sociologists started studying divorce, they adopted different methods than previous scientists. They allowed for the fact that divorce can be an opportunity for growth as well as a trauma, and thus discovered new empirical facts about the consequences of divorce. Being feminist may change your perspective, but it does not give you license to ignore facts, which these scientists didn’t. Similarly, it seems entirely uncontroversial that a child psychologist knows that institutionalization is bad for children (a normative claim), because she has seen what living in orphanage does to young children (a factual claim). So values and facts can be entangled without fraud or bias. Why didn’t I think of it before, I wondered?

After this breakthrough, I didn’t look back. I knew that mixed claims could be epistemically legitimate. But my job wasn’t over yet. I had seen some dark and problematic trends developing in the science of well-being. Here’s an example: Some psychologists (call them ‘hedonists’) succeed impressively in the awesome task of detecting the positive and negative emotions of their subjects. They called the ratio of one to the other ‘happiness’, and argued that it measured true happiness or well-being, and that policy should aim at improving this ratio. When others objected that people report and behave as if other things besides happiness matter too—meaning, autonomy, achievement, status—the hedonists sometimes replied with dangerous lack of moral awareness that those other things are just biases, and that true well-being is only what’s experienced. This is not a dig at psychologists; economists sometimes behave as if their way of detecting value (by willingness to pay) is correct by definition too.

I saw the debate between hedonists, traditional economists, and psychologists of a non-hedonist persuasion play out in the public sphere in 2010, when the Office of National Statistics was formulating a measure of the UK’s national well-being. Different scholars advocated different questionnaires based in part on their views about the nature of well-being. As far as I could tell, the ONS handled that debate rather well: different parties were heard, and the resulting measure—though clunky and inelegant—does a decent job of reflecting the variety of answers to the big question, namely, what is it for our community to be well?

This example, plus other trends in political theory and philosophy of science, led me to propose that mixed claims are legitimate, and can even be objective, when they are properly vetted. I proposed three rules for such vetting. First, scientists should make explicit the values presupposed by mixed claims. They shouldn’t just collect a bunch of indicators and call them ‘child well-being’, or ‘freedom’, and so on. They must own up to the value judgements that underlie their choice of measurement instruments. Secondly, they must check whether the empirical claim they are pursuing is subject to disagreement on the basis of values. Unemployment, evidence indicates, hurts all aspects of a person’s life and so the hard debate on the precise nature of well-being can perhaps be avoided. But often things are trickier. For example, greater earnings predict greater satisfaction with life but not greater emotional happiness. What matters more for well-being: how you feel or how you think you are doing relative to others? If scientists wish to make a claim that takes a stand on this question, and this is my third rule, the claim will be objective only to the extent that a given measure of well-being passes a deliberative test. I think of this test as a public debate, though sometimes it might only be hypothetical, in which the relative advantages of different measures of well-being are heard and the perspectives of both experts and the public whose lives could be affected by the choice are included.

Some scientists will be averse to my proposal because it forces them to be philosophers, to make and to defend value judgements. To them I reply that they face a choice between staking their identities as scientists on not being philosophers on the one hand, and making their hard-earned knowledge objective and trustworthy on the other. I hope they choose the latter.

Anna Alexandrova is Senior Lecturer in Philosophy of Science at the University of Cambridge. This essay is based on her article ‘Can the Science of Well-Being Be Objective?’, published in the British Journal for the Philosophy of Science. Her research interests lie in the philosophy of economics and the social sciences, and she is currently working on a project entitled ‘Limits of the Numerical.

Image credit: Owen Davey