THE SOCIAL CONSTRUCTION OF STORIES
There are a couple of questions that I'm asking myself. What we think we know about happiness nearly always comes from people telling us answers to questions that we ask them about how they're feeling, their moods, and so on. I was struck recently by a couple of papers that show that self-report data are often inconsistent with ways in which we can now start measuring happiness, through looking at facial expressions and conversational analysis.
In the US, for example, self-report data have always shown that conservatives are happier than liberals; that's been taken as a statement of fact. Yet, when you look at people's facial expressions, when you look at the conversations they have, and the tone and mood of the conversations, it's the other way around. I'm interested in trying to find out how happy people are without asking them. Big data and other forms of analysis that look at people's actions and behaviors give us that opportunity. That's one question that I'm asking myself.
This next question is about the role of stories and narratives in people's lives. I'll tell a story to frame this. I went for dinner with a friend who spent the whole of the evening complaining about her job, her boss, her colleagues, and her commute. Everything about her day-to-day experiences was miserable. Then, at the end of dinner, she said, "I love where I work." That's quite common. She was working for an organization where she'd always wanted to work, her parents were proud, her friends were jealous. How could she not be happy when she thought about the story of how happy she was where she was working? Her experiences—day-to-day and moment-to-moment—were telling her something quite different.
I'm interested in where these narratives come from, particularly those narratives that sometimes get in the way of us being happier. There's been a lot of psychological research on how stories are helpful for us; for example, in the case of experiencing adversity or trauma. If we look for explanation and reason through narrative, it helps us cope with the adverse consequences. There's been a lot of work on that. I'm interested more in the social constructions of the stories, in the things that evolution, society, our parents, or historical accident tell us about the lives that we ought to be leading, and in particular, how they might sometimes get in the way of us experiencing better lives.
Relationships—who's not interested in that? We have personal relationships with our parents and our children, but the intimate relationships are particularly interesting. There are so many stories about the types of relationships that we ought to be having. For example, there are massive stories that may have good explanations and reasons why, for some people some of the time, marriage and monogamy are good for them. The idea that romantic love lasts is complete nonsense, from the science. We know from the science; that's very clear. Passionate love—you're probably lucky if you get twelve months out of it, maybe eighteen months if you're really lucky. Then it turns to companionate love. But a lot of people still expect the passionate love to continue, and they think their marriages or their relationships are failures because they're not feeling like they did when they first met. It's just an absolute mess. We get ourselves into the mess by telling a one-size-fits-all story about the kinds of relationships that we ought to be having.
One of the interesting questions has to do with where the stories come from. What is the source of the stories? People's perceptions and views about the relevance and resonance of a story will be contingent upon the sources. Often, we'll give evolutionary reasons, and that's just a story in itself because you can never falsify them. Stories will come from a whole range of different sources. Part of my future work is to look at just how important some of these different causes and sources of stories are.
As an academic, I'm interested mostly in quantitative data, large datasets from ideally many observations of many people over many years. And in the behavioral science space—randomized controlled trials, where we observe the actions of real people in the real world when we randomize them to different treatments and interventions, rather than telling my own stories about what the world should look like.
Here's an interesting story about relationships. There are some data—and most of this is not causal because it's not been done under randomized controlled conditions—that indicates that couples who sleep in separate beds are happier and have longer-lasting marriages than couples who sleep in the same bed. I'm convinced, as a story of how we ought to be living in the modern age, that there's something wrong with marriages and relationships where the couples sleep apart. That's not what you're meant to do, according to notions of romantic and passionate love that feed into the narratives that we tell about our lives.
What I want to explore further, and this is all new research territory for me to some large degree, is to see where the data, the science, the evidence, conflicts with some of the narratives and stories that people tell about relationships and about a whole range of other things, too. We have a good story of recognition, status, and achievement. There may be good evolutionary reasons for why successful people are clearly more likely to have a greater range of partners. But that narrative many times will get in the way of people being happier. We do see that in the data. Again, it is often correlational and not causal. People in high-status jobs, who are sacrificing considerable amounts of time, which is the scarcest resource of all, with people they would otherwise enjoy being with—their family, their friends, and their leisure time—in the quest and pursuit of achievement and success, are just making themselves miserable in order to achieve something that has been constructed for them as being what they should strive or aim for.
This is not going to be a view that I can adopt and hold in every instance for everything, of course, context matters. I'd like to try to push the case that we need to get rid of stories and narratives and live life in experiences that are essentially about pleasure and purpose, rather than evaluations and constructions about things we think we should be doing and the lives we should be leading.
When I was a health economist for the best part of a decade of my academic life, I was pretty successful at that. I did a good job of writing papers and having a policy impact. The work I did asked people to imagine what life would be like to experience different health conditions for the purposes of then informing the use of scarce public resources. For example, you would be asked to imagine what it would be like to be in a health state where you had some physical functioning problems, or you had anxiety or depression. People were projecting and imagining what it would be like to be in those states. Those numbers and those answers are still being used by the regulatory authorities to inform allocation decisions.
It struck me quite quickly that we are pretty poor, as a lot of psychologists have shown, at being able to predict the impact of changes in life circumstances—health and other things—on our well-being. I picked physical functioning and anxiety and depression because when you ask people to imagine how bad those conditions are, they think they would be about as bad for them if they were to experience them in the future.
The evidence from happiness research, as I then started to find out, tells us something quite different. It tells us that living with, over long periods, mental health problems is much worse than living with, over long periods, physical functioning problems. The reason is because of attention. Mental health problems continue to draw attention to themselves in the experience of our lives over time. You don't wake up a year after being depressed less so because it's been a year. But with physical functioning limitations, insofar as there's no other attention‑seeking attributes associated with it, like pain, you get used to it. A year after you first started limping is not as bad as the first day.
Our imaginations are not particularly good at being able to predict that. There's obviously quite a lot of research evidence on that. What it led me to argue was that we're misallocating resources. We're spending much more on physical health problems than mental health problems, given the impact that those conditions have on people's lives. One of the great advantages of happiness data are that they enable us not only to find out what affects people as they experience the conditions of their lives, but to do so without having to ask them how much they think those conditions affect their lives.
All of the preference data that we use in health economics is to say to someone, "Imagine how much this would be a problem for you when you're imagining how much it's a problem." The happiness data allow us to find out how happy someone is, find out all sorts of other stuff about their lives, including their health status—insofar as that's what we're interested in—and then the regression models tell us the impact of those events, and circumstances, and health states on someone's well-being. That's a huge advance. It's taken us a while to wake up to that, in economics particularly, and also public policy, but we're finally getting there.
We have essentially two fields of activity. We've got happiness research on the one hand, where you've got people looking at survey data, self-reports of the kind that I alluded to earlier—how satisfied are you with your life overall?—and then trying to infer cause and effect from sophisticated regression techniques. On the other hand, you've got behavioral science doing interventions, nudging people to behave differently, using financial incentives or social norms or whatever other effects to get people to behave differently by conducting randomized controlled trials in natural settings.
You've got two different research methods being used—survey research and randomized controlled trials—but you've got no one thinking about what the well-being and happiness consequences of the nudges are. They're looking at the change in behavior. Say I send you a letter that says you're a high‑end energy user compared to other energy users, and I show that that might have some effect on population, on average, on your energy use. What if you're the kind of person who hates getting this letter about what a terrible person you are for using more energy than the average person? What if you might be willing to pay money not to get the letter that tells you what a bad person you are, relative to other people? What if it makes you feel worse by being told you're a bad person, relative to other people?
These are all questions that are not being answered by keeping those two research areas distinct. My own research tries to do both things. To use the research methods of behavioral science, we should be doing more randomized controlled trials that look at the cause and effects on and of happiness, because most of it is correlational. It's largely nonsense about whether children make people happier. We've got no randomized controlled trials where people get allocated children or not to show a treatment effect. There's a huge selection effect problem there. With other behaviors and other conditions, we use randomized trials to assess the cause and effects on and of happiness, because there's a whole range of consequences that follow from people being happy or otherwise.
Most significantly, looking at the happiness consequences of the nudges that we make. I'm interested in showing and knowing whether, when we nudge people to be a little bit healthier, or save a bit more for their retirement, or to use less energy, they're better off as a result of doing that. It's not obvious that we always are.
~ ~ ~
As a health economist, I thought like an economist for a decade. This is the great thing about stories. We can always have a nice narrative that makes our lives cohere and makes sense of things. I serendipitously met Danny Kahneman. That's what happened. I went to an Economics of Happiness conference in Milan. I won a small prize in the UK to buy out some teaching and give me some time to go and do other things. I saw this conference and thought it sounded all right. Danny was there. This was in 2003, not long after he won his Nobel Prize. We sat next to one another on a short bus ride to a conference dinner or something. Within the five or ten minutes that we were chatting, he invited me to Princeton. So I went. That got me interested in both the behavioral science of psychology and also the happiness research, which he was getting much more into at that point.
When we think about stories and narratives, one of the things that we don't like, as part of the human condition, is luck as an explanation. Luck has no agency. If we tell a story of a film, we don't tell a story about the randomness of what happened. There's a narrative from the beginning and the middle through to the end. The human condition doesn't like luck as an explanation because it has no agency. But in my own case, and in so many other people's cases, luck is hugely important. I was just lucky to meet Danny.
One of the interesting things for the use of happiness measures in policy is that we should probably not use the term happiness. Kahneman has made this point, too. Misery and suffering are much more salient terms in policymaking. If you say to a policymaker, "The purposes of policy should be to maximize happiness," many of them will look at you and think you're silly and think it's trivial. If you say the purpose of policy is to minimize misery and suffering, they'd be pretty pathological not to agree with you. Even though sometimes the measures that we might use for those things would be precisely the same, the language that we use as presentation for policymaking matters. I try to remind myself, although it's against my natural temperament, to be thinking and using terms like misery and suffering, rather than happiness and well-being.
A few years ago we wrote a report for the cabinet office called MINDSPACE, which was a nine-letter mnemonic, a checklist, for behavioral science interventions that influence us largely through unconscious and automatic processes. Of the nine elements, three come from economics, three come from cognitive psychology, and three come from social psychology. The behavioral science that I do captures each of those areas of activity. One would add in, although I don't do this research myself, neuroscience as another. There have been huge advances in neuroscience over the last decade or two, so I guess the behavioral science arena covers those areas.
It's a good term to use because economics is often put off by psychology. Psychologists don't understand economics as well as some of them think they do, so it's a nice encompassing term for the interdisciplinary interface of those disciplines. I'm interested in a way that economists are not typically interested. They take data and they do regression analysis, but they take the data as given and as available. Most of the data and measures we have of happiness are life satisfaction reports. I'm interested in thinking about whether those kinds of questions reflect and capture happiness.
Think about that question for a second: Overall, how satisfied are you with your life these days? It would take you quite a long while to answer that question because it's quite cognitively demanding. Even if you did, what does it reflect? It's reflecting much more the degree to which you feel like you've achieved the things that you wanted to achieve, and that your preferences have been satisfied. It's nothing to do with happiness or mental states in the way that we would think about it.
Also, it takes people about three seconds at most to answer that question, so they're clearly not answering it. They're giving some heuristic or some response that says, "I'm probably about a seven or eight out of ten," which is a much easier question to answer than thinking about how satisfied you are with your life. I'm much more interested in understanding how people are feeling as they experience their lives day-to-day, moment-to-moment. The important contribution that I've made is to think of those experiences not just in terms of hedonic states of pleasure, pain, misery, anxiety, and excitement, but also in terms of purpose.
Of course, purpose has been talked about since Aristotle and for the last two and a half thousand years, but always as an evaluation, as a story. Does my life have purpose? Does my life have meaning? I'm not interested in the meaning of life. I'm more much more interested in the meaning of moments. I'm interested in whether this conversation has purpose in the experience, whether it feels fulfilling, worthwhile, meaningful, as I engage in the activity and not just as I reflect upon whether it's been worth it either afterwards or in some global sense.
My meaning of life, my purpose as a father, doesn't show up when I reflect upon being a dad. It's in the experiences that I have with my children, in the stuff that sometimes is not particularly pleasurable. Listening to them read the same story again, helping them tie their shoelaces, teaching them their times tables, there are so many more things that would be much more fun than that, but they feel like they're fulfilling activities.
My contribution is to say that life contains these elements of both pleasure and purpose in our experiences, and that happy lives are therefore ones that contain the right balance for that individual, which won't be the same for everybody. That's what happiness is. That's my conceptualization of it. That's a contribution—to see purposes in experience is a different way of thinking about happiness.
There are some very interesting intertemporal and interpersonal comparisons of happiness that raise a huge number of questions. Dealing with the intertemporal ones, that is, changes of happiness over time, and even whether happiness was something that people thought of as being something that they were motivated by or cared about in any sense. Feeling good, in whichever sense we try to capture and measure that is part of a desire in the human condition. No one would knowingly seek out misery. Of course, we make all sorts of mistakes, especially when you add purpose into the mix. There're lots of things that we do that might not be particularly pleasurable, but they give our experiences value and worth.
We're driven and motivated, across generations and across time, to seek out things that we at least think will make us feel better rather than worse. I don't think that's ever changed. What has changed is the language that we might use to reflect and represent that. The avoidance of suffering and misery may have been language that we used more frequently in the past, and the celebration of happiness and flourishing might be language that would use now, but the basic desire and motivation for action hasn't changed.
The international and interpersonal comparisons issue is an interesting one. I'm not particularly interested in international comparisons of happiness data. I don't think it tells us very much. You can't translate the word into some languages. It has very different meanings if you can translate it. There's a whole range of cultural effects in the self-reporting of the data, which leads me to conclude that people are making quite a big deal of something that is problematic.
It draws me much more to measuring more directly people's experiences of pleasure and purpose. They might be more universal in some sense. When you get kicked in the leg and you feel pain and you report how much it hurts, that's probably more interpersonally comparable and culturally comparable than asking people how satisfied they are with their lives these days, which has all sorts of problems even if you can translate it. Two and a half thousand years of ethical discourse, we've not managed to resolve the question about what it is that we're substantively driving and striving for as individuals or as populations. We're not going to do it through the course of this interview.
What we do need to be clear about is the normative view that we adopt when we're approaching individual decision-making and public policymaking. We need to be clear from data, ideally randomized trials or big surveys, where the differences between our accounts of well-being will lead to different policy conclusions. That's what we need to flush out, whether these normative and philosophical discussions that have been going on for two and a half thousand years actually matter, and the conditions and circumstances under which they do matter. For example, it does matter which measures we use when we talk about mental and physical health. We talked about that earlier. It's important that we have some clarity about what we're striving for when we're intervening in people's lives to help improve them.
If you push me into a corner, the corner that I will be in is the one that says we assess the impact of all policy interventions on the experiences of pleasure and purpose and pain and pointlessness that people have as they go about their daily lives. That's the normative position that I adopt when approaching individual choices, choices that I make for my children, and when I'm informing policymakers about what they should do with scarce public money. That's the frame of reference that I have.
The self-report question in happiness research is a huge issue. At the moment, it's the way that's been much more widely used to understand how someone feels. That makes a lot of sense. If I want to find out how you're feeling, I'm not best placed to judge it. I can ask you how anxious you're feeling, how worried you are, how stressed you are, how excited you are, and how joyful you are. You're probably a good judge of it most of the time.
What is more challenging is asking people these reflective narrative questions, evaluative questions: Overall, how satisfied are you with your life these days? I don't go around routinely experiencing that, but I do experience the joy, the pleasure, the purpose, and the pointlessness, and the futility of the experiences that I have in my life. Self-reports are a good way to tap into them, but they're not the only way.
The academics are often quite slow on the uptake of new technologies, new methods, and big data, and we should be using that data much more than we are. There's more that can be done on conversational analysis, looking at tone and intent, the tone in which people are speaking to one another, the moods that people exhibit when they're engaged in conversation. You can do that without any self-report. That's one of the ways in which we're going to improve our measures of happiness and well-being.
The replication issue in psychology is probably the single biggest issue right now. How many of these studies are replicable? One of the things that we know from behavioral science, and neuroscience has helped us with this over the last decade or two, is the role of the unconscious and the automatic system in shaping and influencing behavior.
Our understanding now is that context matters so much. You only have to change something almost trivial in the environment, in the situational context within which someone acts, for it to have a significant effect on their behavior. It's little wonder that some of these studies are hard to replicate, because you can't precisely replicate the environment and the context of the situations in which those initial behaviors were first observed.
That opens up the whole issue around priming. There's no doubt that we're primed all of the time. Our situations cue us to behave in particular ways. Does that mean that we have a good understanding of precisely what those primes are, or when they operate, or how long they last for? No, it doesn't. It means that we need more research, and to sometimes accept that we're not going to replicate because of those situational factors that influence us so heavily.
There is no doubt that we're being primed and influenced and cued by our environments all the time in unconscious and automatic ways. The brain is looking to make life easy for us, and it's making associations all of the time to help us do that. Most of the time, it does a pretty good job of it. Of course, the academic and policy interests sometimes are where we make the mistakes and the errors. The subtleties of priming are so interesting and make the academic research so bloody difficult, because if you over-prime someone, then it's in their conscious awareness and it ceases to be a prime.
But where does that cusp come? Where is that tipping point between something being in your unconscious environment and something being in your conscious awareness? So little wonder, again, that the replication stage is so difficult, because sometimes they'll be in the unconscious environment and other times they'll be in the conscious awareness. It's very difficult, from the experimental methods, to discern and understand those differences.
My sense in policymaking, in the public sector and the private sector, is that we have almost an inability to get over ourselves. We need to accept the fact that we are driven so heavily by unconscious and automatic processes, and embrace that and celebrate that. People like Kahneman are a bit more pessimistic about our abilities to deal with those biases and effects in ways that are welfare‑enhancing. I'm optimistic that we can. If we understand the mechanisms that underpin our behavioral responses to particular situational stimuli, then we can design environments and organize society in ways that make it easier for people to do things that enhance welfare and harder to do things that don't. I'm quite optimistic that our inability to make rational, thoughtful, considered decisions is harmful. I think it's good.
This is a real arrogance of economics, or maybe just people generally, to think that a little bit of training and a bit of knowledge and understanding can somehow de-bias you from these effects that may have served as an evolutionary advantage for billions of years, that you can have a degree in economics and suddenly you're de-wired and de‑biased from these effects.
You know how in families siblings will fight one another, almost hate each other, and then if someone else from outside the family starts having a go, the siblings will gang up and fight the person from outside the family? Sometimes, I feel a bit like that about economics. I quite like having a go with my siblings in economics, but then you get people from outside having a go, and I stand up for economics a bit. Like most things in life, with individuals, as well as in societies, some of our biggest strengths are also our biggest weaknesses.
One of the greatest strengths of economics is it tries to predict. It's not a descriptive science in the sense of psychology. It is prescriptive about how the world ought to be and what the world would look like if certain levers, normally financial ones, were pulled. That prediction makes it incredibly useful and valuable for policymaking. The trouble is, of course, it often gets it wrong. Its great advantage is also its great weakness. When you make mistakes in prediction, of course you can cause problems. But the fact that we are willing—I'm now back in the family—to make those prescriptive predictive claims, is why economics has the ear of policymaking, much more than psychology does.
My earlier work as a health economist has had an impact on health policy. The trouble is, most of that work wasn't particularly good now through the lens of happiness. I'm trying to recast how we evaluate policy impact, using happiness as the lens. Academics bleat on about how policymakers don't listen to us and they don't pay attention to any evidence. Well, academics don't present evidence in a way that makes people listen. It's not particularly compelling. When it's loads of regression models and graphs, who cares? Tell a story, have a good narrative that sits alongside the good data, and people will pay more attention. It's incumbent on policymakers to listen to evidence more, but it's also incumbent on academics to present evidence in ways that makes policymakers listen.
One of the interesting things, and it's not a contradiction, it's a consistency, is that we need to tell fewer stories to ourselves and to each other about all sorts of things, including the sorts of lives that we should lead. In order to get scientific evidence listened to and taken seriously by policymakers, we need to tell stories.
We need to have good narratives that back up the evidence and the science. As we all know, all policy is really about narrative. We need austerity. Well, that doesn't mean anything. There's just a huge story and narrative that underpins austerity. Everything in policymaking is about the narrative. Academics need to understand that a little bit better than they do currently in order to have an impact on public policymaking—understand the narrative, understand the story, and be able to talk in a way that makes policymakers listen more than they do currently.
From health to happiness to stories, personally, I've had opportunity to do more public-facing activity, not just talks, but television and things like that. I'm quite liking the opportunity to take science to the public. The public appetite for any good science presented well is quite strong. There's an assumption that people don't care and don't want evidence or science or knowledge, but that underestimates people. My working‑class background is one where there was little formal education. In fact, none of my family had formal education beyond the point at which they had to stay at school, but they were interested in stuff.
Finding people that could present interesting stuff to them in a language and in a way that would appeal to them, is something that I'm now realizing is where my public-facing life might go. I've learned to do that better. I've filmed TV that is for the masses, as it were, and that's taught me that even though I thought I was a pretty straight academic, and I spoke in a lay language compared to how most other academics speak, I still spoke like an academic. I've managed to find ways of communicating more effectively with lay audiences.
Just the same way as academics need to learn to effectively communicate with policymakers, we need to understand much more about the people that we're talking to than we do currently. Maybe that's the main message. It's an obvious thing to say, but most interesting things probably are obvious. Understand the audience that you're communicating to, and do it in a language that is accessible to them.