Thank you, and I definitely needed to hear it in this case. I am always nervous taking big public swings at senior peoples' opinions (never stops me though!).
I think many of the artifacts of conventional measurement methods can be avoided using the ACSA method, which combines qualitative with quantitative evaluation. It asks you first to reflect about the very best and very worst moments in your life (qualitative), then to situate your present level of wellbeing relative to those anchoring points (quantitative):
Bernheim, J. L., Theuns, P., Mazaheri, M., Hofmans, J., Fliege, H., & Rose, M. (2006). The potential of anamnestic comparative self-assessment (ACSA) to reduce bias in the measurement of subjective well-being. Journal of Happiness Studies, 7(2), 227–250.
Thank you for this reference Francis. The example in the paper of transplant surgeries is instructive. I think this approach can be superior for impact evaluation. However, I think it has unattractive qualities for subjective wellbeing studies more broadly. It would be ideal to have respondents imagine possibilities, not just anchor on things they have already experienced. This seems especially relevant for developing countries and young people. What I was originally interested in was studying self-actualisation longitudinally. I can see some applications of this anamnestic approach therein as people discover values or behaviours they are a better fit for them, but they need to imagine their life better under a new self first. Lots for me to think about here. My sadness with the field is that I find myself mostly critiquing the way we currently do things rather than getting to think about better ways of doing things to answer the questions I am actually interested in.
Scale norming on life satisfaction was one of the first things I heard at my first Pos Psych conference in 2015, so good to see someone's actually tackling the problem. Great article Mark, it left me thinking that it's overdue for Wellbeing Science to have a 'stock take'. Consider what were all of the assumptions/unknowns we were carrying into this space, which have been addressed, and (most importantly) which have we continued to hold as some assumptions have transformed to standard academic practice. It reminds me of someone finishing their PhD... "Knowing what I know now, I would have ...".
Phenomenal breakdown of the scale norming problem! Your long COVID example really drives home how measurement tools can obscure what they're supposed to reveal. I remeber trying to compare my satisfaction over a few years and realizing the whole scale had shifted, like you can't step in the same river twice. Makes me wonder if all this quantification is actualy making us worse at understanding wellbeing in the first place.
Loved reading this and love seeing you on Substack, Mark. Your walkthrough of well-being scales and the problems with accepting those numbers at face value still lingers in my mind as I read different studies. It's made me a better consumer and healthy critic of science!
(Also, I think you are being incredibly kind to Professor Kashdan here, and that maybe Professor Kashdan should consult his spiritual advisor of choice regarding the content of his social media posts. Scientists derived whatever authority they once had by not being just another bunch of hucksters.)
Unfortunately, psychological science has very much gotten into bed with commercial publishing and it is very damaging. I published a commercial book this past year and the pressure to make things snappy and take shortcuts with the science is very real, so I sympathise with my colleagues. But it can be resisted, and you can at least put a disclaimer up front that this is scientifically informed opinion, not science communication.
Maybe this is a naive statement, but can't you just... not write a commercial book? If the book you want to write is not commercial, you could just write a non-commercial book and put it on Substack! Though I suppose this does not solve the problem of other people writing bad commercial books.
TBH I think part of the problem is that the culture of academic psychology is much too agreeable. If researchers are writing bad pop science books, then their colleagues must be willing to publicly dispute them or the reputation of the field will be run into the ground.
Ironically, this may also solve the marketability problem; what is more commercial than interpersonal conflict?
Completely agree on the culture being too agreeable, in the same way that economics is too toxic and macho. Another problem is that psych has a culture of 10+ publications per year, and there's no way you can do that while being careful and thoughtful about theory. And a third problem is that the field's top journals all have very strict word limits, especially on conceptual frameworks. Tenured faculty have to take the lead on shifting these norms, but they're ironically often the least likely to because they've benefit from the system.
I have mixed thoughts on commercial book publishing. We can't leave the space to influencer types like the Let Them Theory. At the same time, commercial publishing (and consulting) is clearly poisoning the science - witness Ariely, Gino etc. There's far too much of a rush to publish commercially. Oishi has already got a pop book on psychological richness when the critical papers have just started appearing. It is now going to take an eternity to pull this stuff back out of the pop ecosystem, much as with Power Posing and the like. Commercialisation encourages us to find catchy slogans and terms, but these often lead to confusion. Every year I explain in detail to my students that a nudge must target a behavioural/cognitive bias, otherwise it's just psychological manipulation. And every year the students just understand nudges as 'small' policies regardless.
But then it's also true that if you don't do this pop stuff you have less reach and that means more ground seeded to the grifters.
You are certainly right that the main thing is to be firm about your scientific integrity if you're claiming to be doing science communication. My book has a disclaimer after the introduction that explains why it is isn't popular science even though I'm often talking about scientific findings.
What do you think about study designs where they ask participants to measure their at-the-moment happiness, on some scale, once an hour for several hours? It seems like this data wouldn’t suffer from many of the biases in summarizing your life satisfaction, and it would suffer less from several other ones.
Thanks for engaging. Mood measures like the day reconstruction method or experience sampling are more robust to scale issues - they're not on a scale! I think we've done good research with mood data. I especially like the mappiness project. But I don't think mood is wellbeing. It's got a big role to play and I think we should be using a variety of observables in our science, but its not the object of inquiry. Hedonists disagree of course but they're wrong and I think they implicitly admit that through their own behaviour.
Can you elaborate more on the important differences between extended mood and wellbeing? I guess it’s not super clear to me that wellbeing is so much more important
Happy to send you a PDF if you email me at my Warwick address.
Basically, mood, affect, and emotion are signals not destinations. If you have arranged your life in a way that is coherent for you it will generate an abundance of pleasant, fulfilling, and valuable experiences. But it will also be full of 'negative' mood, affect, and emotion like struggle, stress, fatigue etc. What's important is that these things make sense as part of the wider whole. A simple analogy is that many people dislike the effort and burn they feel at the gym, but persist with it because of how health, aesthetics, or fitness fit into their life and self image. You need to introspect on emotional signals in order to figure out what this good life looks like for you, but these emotional signals are not themselves the good life, which is why all the most prominent advocates of hedonism fill their life with self-expressive things that make them feel bad instead of just having sex all day.
I suspect you have a different problem, reification (not the Marxist sort). Happiness is abstract. You cannot treat it as a measurable concrete thing, because it isn't. Indeed, I do not think that it is a _thing_ at all. Happy is what you do. If your goals are to be able to walk longer and take the dog up hills, then as you get better at accomplishing them you get happier. And you might decide it is time for new goals -- finishing your PhD -- and then as you get better at accomplishing that, you will get happier. But it is not just your own goals that matter -- the goals of those others who matter to you also play a role. If your favourite sports team wins a game, or a championship, you do indeed feel happier because as a fan you share in the goals of the team to win games. If you don't care for the sport, or the team, you cannot share in these goals, and thus you are indifferent to the results of the latest game.
Now as a practical matter, there is a lot you can do to further the goal 'I would like it if people were happier' and getting them to either be less neurotic, or more capable of managing their own feelings, and setting good goals, and not being overwhelmed by difficulties, and knowing when to abandon goals you had at one time and when to persevere and all this is very much worth knowing and studying. But you don't get to discover which human thoughts and behaviours promote happiness if you start with the idea that happiness is a tank you need to fill up and thus you need to begin with knowing how full the tank is.
But I also think there is a lot to learn from noisy but numerical data on wellbeing. To take a simple example, this paper shows a disturbing upward trend in the number of people reporting that every day of their life is miserable:
We don't need to measure the thing perfectly to get insights that only quantification of some sort can provide. But we shouldn't mistake a reification/simplification for the actual thing we are trying to build an understanding of. Too much life satisfaction research concerns the study of scale responses rather than the study of wellbeing.
Yes but I don't think adaptation is as strong as currently believed and I think it's treated as more automatic than it is in reality. People adapt in part because they "do the work" - things like therapy and cognitive restructuring, not just because "time heals all wounds" or whatever. And certainly scale norming has more of a role to play than is currently appreciated. We know that from the response shift literature in medical quality of life studies.
I'm struck by the role stoicism has also played in the most popular therapeutic intervention (CBT). Many practitioners (excepting Ellis who explicitly acknowledged his debt to it) appear to be unaware of CBT's philosophical foundations.
Equally odd: CBT will explicitly - and fallaciously - cite Socratic questioning as a philosophical pillar to building well-being, and the practitioners seem equally unaware that Socrates' goal was to undermine people's thinking. Socratic questioning made so many people unhappy that Socrates ended up being put to death for it.
Thanks for engaging Steven. The paper you linked was very interesting, and the point about Socratic questioning making people depressed made me chortle.
I saw on your book blurb that you draw from all sorts of areas of psychology, but I didn't see psychophysics. Several of the issues you discuss sound like phenomena that arise in sensory perception of all sorts, and in psychophysics we have, like, actual proper theories of them! With math and everything! So I think you should consider adding psychophysics to the list.
To be a bit precise, your long COVID patient example reminded me of luminance perception. Our sense of what is "bright" depends on the ambient brightness our environment--a candle looks bright in a dark room, but not on a sunny day. Yet at the same time, we can easily acknowledge that it is brighter outside on a sunny day than it is indoors. This all falls out fairly easily from a model where we are trying to maximize the amount of sensory information we can get out of our local environment.
The long COVID patient wildly recalibrating her wellbeing scale sounds very much like a pair of eyeballs going from a bright outdoor scene into a dimly lit room. The question is, in this analogy what exactly is the "information" she's trying to extract?
Where would I go for a quick overview of psychophysics? Is there a handbook or something?
What you describe was studied quite a lot in the early happiness literature, notably by Kahneman in his colonoscopy studies. I think it's part of the explanation of why people report their life satisfaction in the way that they do, but not a defence of the accuracy of life satisfaction scales as measure of subjective wellbeing. That woman can understand her life today as worse than her life in the past. It's not that her *perception* is affected the way it is with luminance or hot water. Only her reporting is affected, and by the scale, not by her perception. We as researchers assume people's scale are static over time, and if their numbers don't change it must be because of adaptation, when in fact the numbers can mean completely different things.
There's another aspect to this which is the difference between adaptation by 'working through it' and adaptation by acclimatisation. When I lost my hair I was quite sad for a while, but I gradually got used to my appearance in the mirror. This is closer to the luminance example. In contrast, we see a lot of adaptation through things like therapy, e.g. one interviewee says something like "it means getting over my grief at not being able to have a child and finding some other outlet for those yearnings, like emergency foster parenting". There is way more effortful cognitive restructuring going on here than in luminance. Again, it's not the perception that changes, but the judgement, and in deliberative ways.
I think this raises further questions about whether making yourself 'happier' through the various cognitive tricks of positive psychology is improving people's 'wellbeing'. Perhaps prudentially we should be more interested in people's reflective judgements rather than their immediate perception and the tricks they overlay on it.
I think the (a?) major insight from the perception and decision-making side of the field is that you really need a very rich theory of the respondent to make sense of scale use. What information are they trying to extract, and what information are they trying to convey? What reference point, what baseline environments are they assuming and why, etc? Frankly, it all seems very difficult!
Re: overview of psychophysics, that's a great question that I don't have a great answer for. A lot of this stuff comes from vision, which prides itself on being intensely technical, and the idea of applying it to subjective judgments is relatively more recent. I'll reply in a note so I can provide some links.
This an oldie, but a great introduction to the "statistical decision-theory" way of thinking about the brain and behavior. (the specific claims about parietal cortex neurons are contested, naturally)
Thanks for taking the time Seth, this is super helpful. I'm generally not a fan of decision theory as an empirical claim (as opposed to a modelling tool) but I haven't explored it from the vision side so this might be enlightening.
Decision theory is an extremely good empirical model... for simple perceptual judgments. Also, for simple value-based choices. Basically, any situation where all the relevant pieces can be assembled into a neat little self-contained box.
Decision theory struggles, of course, with fuzzy open-ended problems; the kind of problems that people are mostly concerned with. It is sometimes concluded, therefore, that decision theory is no good as a model of people.
But I think that's wrong. *Everything* struggles with fuzzy, open-ended problems! Decision theory is the best model we have of human beings, imho, because it fails in similar ways that humans fail.
Thanks Seth, this is a good perspective on decision theory that I hadn't considered. It will make me less mad when reading and that's always a good thing. I also thank Felice below for their comment because indeed most of my grievances with decision theory come from economists and analytical philosophers who can't fit things into decision theory but then seem to treat this as a problem for reality rather than a problem with decision theory.
@Daniel Greco I wonder what you’d think about this entire thread (starting from OP, which discusses philosophy as well as…”humanists”).
Seth - fwiw, in a vacuum, when I hear “decision theory”, I think more econ and not, like, wet-lab problems w/cleaner empirics than stuff involving H. economicus.
Really loved this until I saw you make reference to McGilchrist and the notorious left brain/right brain neurobollocks. Which is a shame because your whole point here is about myth-busting and against oversimplification! Would love to restack but would you consider leaving the left brain/right brain myth out? I love the elephant and the blind men metaphor.
Do you have a cite for it being mythical? I have asked neuroscience colleagues whether they thought it was bunk and all of them said to me that McGilchrist is reasonable, just not settled science. He's quite well received at Oxford. He goes to a great deal of effort in his works not to oversimplify it as left/right, I'm just writing a sentence here in a small blog about very different things.
The idea of hemispheric specialization of complex traits and emotions has been abandoned ages ago (other than speech and orientation and even that is not clear cut). I am curious to know who are those neuroscientists who say he is legit?! Shall I DM the refs?
Yes please do, though I’d also be happy for them to just be posted here if you don’t mind so others can also follow.
McGilchrist also says that it was abandoned long ago and that he’s not seeking to revive the old model. But he thinks that the baby was thrown out with the bathwater, as it were. He uses left/right as a shorthand and there’s dozens of pages of caveats and footnotes.
In any case, my concern is more about what might be called linear habits of mind or something like that. Maybe I’ll just use that phrasing in the future, but I like the wider analysis that McGilchrist is putting forward, especially about seeing the whole vs the parts and things like that.
I understand you use it as a shorthand but to me, that is how the myths become cemented and enshrined in culture. Of course you can leave it in- it’s your piece! Just saying it’s a shame as I would restack it otherwise as you provide very knowledgeable and nuanced insight on topic that my audience would love to hear about. For the refs, please see below, we go through the whole shebang
Thanks, that was very interesting. Is your concern with flippant use of the left/right distinction or with McGilchrist's work specifically? I don't think he would disagree with anything you've written, especially not this notion being dumb:
"By drawing a physical line down the middle of the brain, it seems to give tangible, material proof of this cognitive-emotional divide. And beyond that, it echoes an even older narrative: the age-old split between soul and intellect."
Indeed, iirc, he opens by saying that this way of thinking about left/right was dumb. That's what I mean by 'bring it back but not as it was before'. He always starts his talks by saying that both hemispheres are involved in everything and the situation is immensely complex. His books are sooooo long for this reason. If there was a pithy description of the difference between the hemispheres in there it would be that the left tends to think in parts and instrumentally, and the right in the wholes. I can see how this would drift into things that give you grief, like "right to right brain communication is where therapy happens", so my sympathies. I'd be happy to remove the left/right reference but I'm reluctant to purge McGilchrist because I think his critique of linear, instrumental, look at the parts thinking is extremely important, even independently of the neuroscience element.
We wondered about something similar in our "How people know their risk preference" study. And yeah, you can find examples of incommensurable scales when you look at introspection. But isn't it interesting that it somewhat works at all (i.e. aggregated ranks kind of make sense as in the "feelings integers" paper), including in some cases across countries? I also wonder whether the introspection given after the response is all that truthful vs sort of confabulated.
Thanks for engaging. It's definitely interesting that it works at all. On the introspection, we found a couple of people doing what seems like confabulation, but honestly most people had quite reasonable frameworks for thinking about their lives and why they were going well or badly. Two of my coauthors were expecting a lot of confabulation and were pleasantly surprised at how sensible people were. One of the most valuable aspects of the study was just hearing how non-academic people think about stuff and express themselves. I am more sceptical than any ever of the speed with which scientists accuse people of being 'irrational'.
I agree re walking back psychology's tendency to frame things as irrational. I'm a big fan of Mercier's framework and writing on this.
Still, I don't think we can tell whether people have access to the process that produced the response by whether it's reasonable? Maybe if you choice-blindness-style switched up their response for something else, they'd justify that reasonably too?
That's why I had outside raters guess the risk preference of the authors of the introspective statement. I found some intersubjective validity. Not a ton, but these statements were often very short. Are you planning on doing that?
Ah I see. I think I had the wrong interpretation of your question. I'm perhaps not familiar enough with the literature you're operating in to give a tight answer. We certainly haven't gotten to the point yet where we are pinning down parts of their cognitive process using experimental conditions or something like that and I don't have plans to do that. Some of my coauthors might be interested though.
Speaking more simplistically, people answer these questions very fast, which might indicate a more affective response. Some people explain their reasons in very affective terms - "yeah life is good, I'm in a nice relationship, I'm happy". But others have very elaborate frameworks for thinking through their life, and we suspect that the reason they can answer so quickly is because they check in with that framework occasionally. They can call this framework to mind quickly too. There are others who are more rambly in their justifications, but there was much less of that than we expected.
I better understand what you've done and found now. To confidently come to the conclusion that the justifications describe the process, I would need more evidence. My guess would be that it's somewhat right, but people don't spend a lot of cognitive effort answering survey questions. So, I wouldn't be surprised if quite a few people might change their next numeric response a bit after going through an in-depth interview of their circumstances just because they've reflected more.
non of this is experimental though. I haven't done choice blindness work either and I don't know how solid that literature is. I think more solid than the "weather colors your life satisfaction" work but that's not saying much. This paper https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1598254/full seems to say there's trouble with the CB paradigm.
I'll have a quick look, thanks. And thanks for the clarification.
As it happens, the last question of our interviews was "after all the conversation we've had, would you like to change your scale response?", and we had very few people say yes. 5 if I remember correctly, and all upward. The ones that I interviewed all said they'd been too grumpy at the beginning and reflection had made them see more bright spots or communicate a bit more positively (the latter perhaps reflects on the interview experience and the interviewers).
I also don't have full confidence that the justifications describe the process. But I also don't have confidence in using these scale responses while we have 0 understanding of the process that generates them. Our paper opens with an analysis of how the standard statistical techniques in this space (i.e. ordered probit) assumes things about the underlying data generating process that have never been tested. We also say that our paper is a not test, but only necessary preliminary work for getting to a test of some sort.
That's really interesting. I find it kind of reassuring for self report stans. Maybe I was overindexing on the choice blindness stuff. Either way, I agree we should better understand how subjective reports come about.
I'm still puzzled that men and women have almost exactly equal life satisfaction in most EU countries, despite differing on many inputs like income and personality. Like, how does that happen?
INCREDIBLY BANGING WORK MATE (as you know, but never hurts etc)
Thank you, and I definitely needed to hear it in this case. I am always nervous taking big public swings at senior peoples' opinions (never stops me though!).
Thanks for the kind words. Just reposted with my thoughts. Some good stuff in here.
Thanks Todd
I think many of the artifacts of conventional measurement methods can be avoided using the ACSA method, which combines qualitative with quantitative evaluation. It asks you first to reflect about the very best and very worst moments in your life (qualitative), then to situate your present level of wellbeing relative to those anchoring points (quantitative):
Bernheim, J. L., Theuns, P., Mazaheri, M., Hofmans, J., Fliege, H., & Rose, M. (2006). The potential of anamnestic comparative self-assessment (ACSA) to reduce bias in the measurement of subjective well-being. Journal of Happiness Studies, 7(2), 227–250.
Thank you for this reference Francis. The example in the paper of transplant surgeries is instructive. I think this approach can be superior for impact evaluation. However, I think it has unattractive qualities for subjective wellbeing studies more broadly. It would be ideal to have respondents imagine possibilities, not just anchor on things they have already experienced. This seems especially relevant for developing countries and young people. What I was originally interested in was studying self-actualisation longitudinally. I can see some applications of this anamnestic approach therein as people discover values or behaviours they are a better fit for them, but they need to imagine their life better under a new self first. Lots for me to think about here. My sadness with the field is that I find myself mostly critiquing the way we currently do things rather than getting to think about better ways of doing things to answer the questions I am actually interested in.
Scale norming on life satisfaction was one of the first things I heard at my first Pos Psych conference in 2015, so good to see someone's actually tackling the problem. Great article Mark, it left me thinking that it's overdue for Wellbeing Science to have a 'stock take'. Consider what were all of the assumptions/unknowns we were carrying into this space, which have been addressed, and (most importantly) which have we continued to hold as some assumptions have transformed to standard academic practice. It reminds me of someone finishing their PhD... "Knowing what I know now, I would have ...".
Phenomenal breakdown of the scale norming problem! Your long COVID example really drives home how measurement tools can obscure what they're supposed to reveal. I remeber trying to compare my satisfaction over a few years and realizing the whole scale had shifted, like you can't step in the same river twice. Makes me wonder if all this quantification is actualy making us worse at understanding wellbeing in the first place.
Loved reading this and love seeing you on Substack, Mark. Your walkthrough of well-being scales and the problems with accepting those numbers at face value still lingers in my mind as I read different studies. It's made me a better consumer and healthy critic of science!
Nice to hear from you Danny! Hope you’re flourishing.
(Also, I think you are being incredibly kind to Professor Kashdan here, and that maybe Professor Kashdan should consult his spiritual advisor of choice regarding the content of his social media posts. Scientists derived whatever authority they once had by not being just another bunch of hucksters.)
Unfortunately, psychological science has very much gotten into bed with commercial publishing and it is very damaging. I published a commercial book this past year and the pressure to make things snappy and take shortcuts with the science is very real, so I sympathise with my colleagues. But it can be resisted, and you can at least put a disclaimer up front that this is scientifically informed opinion, not science communication.
Maybe this is a naive statement, but can't you just... not write a commercial book? If the book you want to write is not commercial, you could just write a non-commercial book and put it on Substack! Though I suppose this does not solve the problem of other people writing bad commercial books.
TBH I think part of the problem is that the culture of academic psychology is much too agreeable. If researchers are writing bad pop science books, then their colleagues must be willing to publicly dispute them or the reputation of the field will be run into the ground.
Ironically, this may also solve the marketability problem; what is more commercial than interpersonal conflict?
Completely agree on the culture being too agreeable, in the same way that economics is too toxic and macho. Another problem is that psych has a culture of 10+ publications per year, and there's no way you can do that while being careful and thoughtful about theory. And a third problem is that the field's top journals all have very strict word limits, especially on conceptual frameworks. Tenured faculty have to take the lead on shifting these norms, but they're ironically often the least likely to because they've benefit from the system.
I have mixed thoughts on commercial book publishing. We can't leave the space to influencer types like the Let Them Theory. At the same time, commercial publishing (and consulting) is clearly poisoning the science - witness Ariely, Gino etc. There's far too much of a rush to publish commercially. Oishi has already got a pop book on psychological richness when the critical papers have just started appearing. It is now going to take an eternity to pull this stuff back out of the pop ecosystem, much as with Power Posing and the like. Commercialisation encourages us to find catchy slogans and terms, but these often lead to confusion. Every year I explain in detail to my students that a nudge must target a behavioural/cognitive bias, otherwise it's just psychological manipulation. And every year the students just understand nudges as 'small' policies regardless.
But then it's also true that if you don't do this pop stuff you have less reach and that means more ground seeded to the grifters.
You are certainly right that the main thing is to be firm about your scientific integrity if you're claiming to be doing science communication. My book has a disclaimer after the introduction that explains why it is isn't popular science even though I'm often talking about scientific findings.
What do you think about study designs where they ask participants to measure their at-the-moment happiness, on some scale, once an hour for several hours? It seems like this data wouldn’t suffer from many of the biases in summarizing your life satisfaction, and it would suffer less from several other ones.
Thanks for engaging. Mood measures like the day reconstruction method or experience sampling are more robust to scale issues - they're not on a scale! I think we've done good research with mood data. I especially like the mappiness project. But I don't think mood is wellbeing. It's got a big role to play and I think we should be using a variety of observables in our science, but its not the object of inquiry. Hedonists disagree of course but they're wrong and I think they implicitly admit that through their own behaviour.
Can you elaborate more on the important differences between extended mood and wellbeing? I guess it’s not super clear to me that wellbeing is so much more important
I wrote a book about it!
https://global.oup.com/academic/product/a-theory-of-subjective-wellbeing-9780197635261
Happy to send you a PDF if you email me at my Warwick address.
Basically, mood, affect, and emotion are signals not destinations. If you have arranged your life in a way that is coherent for you it will generate an abundance of pleasant, fulfilling, and valuable experiences. But it will also be full of 'negative' mood, affect, and emotion like struggle, stress, fatigue etc. What's important is that these things make sense as part of the wider whole. A simple analogy is that many people dislike the effort and burn they feel at the gym, but persist with it because of how health, aesthetics, or fitness fit into their life and self image. You need to introspect on emotional signals in order to figure out what this good life looks like for you, but these emotional signals are not themselves the good life, which is why all the most prominent advocates of hedonism fill their life with self-expressive things that make them feel bad instead of just having sex all day.
Very interesting. I might check it out some time!
I suspect you have a different problem, reification (not the Marxist sort). Happiness is abstract. You cannot treat it as a measurable concrete thing, because it isn't. Indeed, I do not think that it is a _thing_ at all. Happy is what you do. If your goals are to be able to walk longer and take the dog up hills, then as you get better at accomplishing them you get happier. And you might decide it is time for new goals -- finishing your PhD -- and then as you get better at accomplishing that, you will get happier. But it is not just your own goals that matter -- the goals of those others who matter to you also play a role. If your favourite sports team wins a game, or a championship, you do indeed feel happier because as a fan you share in the goals of the team to win games. If you don't care for the sport, or the team, you cannot share in these goals, and thus you are indifferent to the results of the latest game.
Now as a practical matter, there is a lot you can do to further the goal 'I would like it if people were happier' and getting them to either be less neurotic, or more capable of managing their own feelings, and setting good goals, and not being overwhelmed by difficulties, and knowing when to abandon goals you had at one time and when to persevere and all this is very much worth knowing and studying. But you don't get to discover which human thoughts and behaviours promote happiness if you start with the idea that happiness is a tank you need to fill up and thus you need to begin with knowing how full the tank is.
I'm sympathetic to what you say - one of my first papers was on the doing bit:
https://link.springer.com/article/10.1007/s10902-019-00141-7
But I also think there is a lot to learn from noisy but numerical data on wellbeing. To take a simple example, this paper shows a disturbing upward trend in the number of people reporting that every day of their life is miserable:
https://pmc.ncbi.nlm.nih.gov/articles/PMC7483105/
We don't need to measure the thing perfectly to get insights that only quantification of some sort can provide. But we shouldn't mistake a reification/simplification for the actual thing we are trying to build an understanding of. Too much life satisfaction research concerns the study of scale responses rather than the study of wellbeing.
Do you think the hedonic adaption findings still hold?
Yes but I don't think adaptation is as strong as currently believed and I think it's treated as more automatic than it is in reality. People adapt in part because they "do the work" - things like therapy and cognitive restructuring, not just because "time heals all wounds" or whatever. And certainly scale norming has more of a role to play than is currently appreciated. We know that from the response shift literature in medical quality of life studies.
I'm struck by the role stoicism has also played in the most popular therapeutic intervention (CBT). Many practitioners (excepting Ellis who explicitly acknowledged his debt to it) appear to be unaware of CBT's philosophical foundations.
https://pmc.ncbi.nlm.nih.gov/articles/PMC10175387/
Equally odd: CBT will explicitly - and fallaciously - cite Socratic questioning as a philosophical pillar to building well-being, and the practitioners seem equally unaware that Socrates' goal was to undermine people's thinking. Socratic questioning made so many people unhappy that Socrates ended up being put to death for it.
Thanks for engaging Steven. The paper you linked was very interesting, and the point about Socratic questioning making people depressed made me chortle.
I saw on your book blurb that you draw from all sorts of areas of psychology, but I didn't see psychophysics. Several of the issues you discuss sound like phenomena that arise in sensory perception of all sorts, and in psychophysics we have, like, actual proper theories of them! With math and everything! So I think you should consider adding psychophysics to the list.
To be a bit precise, your long COVID patient example reminded me of luminance perception. Our sense of what is "bright" depends on the ambient brightness our environment--a candle looks bright in a dark room, but not on a sunny day. Yet at the same time, we can easily acknowledge that it is brighter outside on a sunny day than it is indoors. This all falls out fairly easily from a model where we are trying to maximize the amount of sensory information we can get out of our local environment.
The long COVID patient wildly recalibrating her wellbeing scale sounds very much like a pair of eyeballs going from a bright outdoor scene into a dimly lit room. The question is, in this analogy what exactly is the "information" she's trying to extract?
Hi Seth
Where would I go for a quick overview of psychophysics? Is there a handbook or something?
What you describe was studied quite a lot in the early happiness literature, notably by Kahneman in his colonoscopy studies. I think it's part of the explanation of why people report their life satisfaction in the way that they do, but not a defence of the accuracy of life satisfaction scales as measure of subjective wellbeing. That woman can understand her life today as worse than her life in the past. It's not that her *perception* is affected the way it is with luminance or hot water. Only her reporting is affected, and by the scale, not by her perception. We as researchers assume people's scale are static over time, and if their numbers don't change it must be because of adaptation, when in fact the numbers can mean completely different things.
There's another aspect to this which is the difference between adaptation by 'working through it' and adaptation by acclimatisation. When I lost my hair I was quite sad for a while, but I gradually got used to my appearance in the mirror. This is closer to the luminance example. In contrast, we see a lot of adaptation through things like therapy, e.g. one interviewee says something like "it means getting over my grief at not being able to have a child and finding some other outlet for those yearnings, like emergency foster parenting". There is way more effortful cognitive restructuring going on here than in luminance. Again, it's not the perception that changes, but the judgement, and in deliberative ways.
I think this raises further questions about whether making yourself 'happier' through the various cognitive tricks of positive psychology is improving people's 'wellbeing'. Perhaps prudentially we should be more interested in people's reflective judgements rather than their immediate perception and the tricks they overlay on it.
I think the (a?) major insight from the perception and decision-making side of the field is that you really need a very rich theory of the respondent to make sense of scale use. What information are they trying to extract, and what information are they trying to convey? What reference point, what baseline environments are they assuming and why, etc? Frankly, it all seems very difficult!
Re: overview of psychophysics, that's a great question that I don't have a great answer for. A lot of this stuff comes from vision, which prides itself on being intensely technical, and the idea of applying it to subjective judgments is relatively more recent. I'll reply in a note so I can provide some links.
LOL I forgot you can't embed links in notes. Seems like an oversight! Anyway here are a few links you might find interesting:
Gold & Shadlen (2002) Banburismus and the Brain. https://doi.org/10.1016/S0896-6273(02)00971-6
This an oldie, but a great introduction to the "statistical decision-theory" way of thinking about the brain and behavior. (the specific claims about parietal cortex neurons are contested, naturally)
Schwartz, Hsu, & Dayan (2007) Space and time in visual context. https://doi.org/10.1038/nrn2155
Overview of how the statistics of the local environment, across both time and space, influence the brain and behavior.
Glimcher, P. W (2022) Efficiently irrational. https://doi.org/10.1016/j.tics.2022.04.007
Describes some more recent work connecting the idea of "adaptation to local environment statistics" to choices based on subjective values.
Palminteri, S., & Lebreton, M. (2021). Context-dependent outcome encoding in human reinforcement learning. https://doi.org/10.1016/j.cobeha.2021.06.006
Similar except with particular application to learning.
Thanks for taking the time Seth, this is super helpful. I'm generally not a fan of decision theory as an empirical claim (as opposed to a modelling tool) but I haven't explored it from the vision side so this might be enlightening.
Decision theory is an extremely good empirical model... for simple perceptual judgments. Also, for simple value-based choices. Basically, any situation where all the relevant pieces can be assembled into a neat little self-contained box.
Decision theory struggles, of course, with fuzzy open-ended problems; the kind of problems that people are mostly concerned with. It is sometimes concluded, therefore, that decision theory is no good as a model of people.
But I think that's wrong. *Everything* struggles with fuzzy, open-ended problems! Decision theory is the best model we have of human beings, imho, because it fails in similar ways that humans fail.
Thanks Seth, this is a good perspective on decision theory that I hadn't considered. It will make me less mad when reading and that's always a good thing. I also thank Felice below for their comment because indeed most of my grievances with decision theory come from economists and analytical philosophers who can't fit things into decision theory but then seem to treat this as a problem for reality rather than a problem with decision theory.
@Daniel Greco I wonder what you’d think about this entire thread (starting from OP, which discusses philosophy as well as…”humanists”).
Seth - fwiw, in a vacuum, when I hear “decision theory”, I think more econ and not, like, wet-lab problems w/cleaner empirics than stuff involving H. economicus.
Really loved this until I saw you make reference to McGilchrist and the notorious left brain/right brain neurobollocks. Which is a shame because your whole point here is about myth-busting and against oversimplification! Would love to restack but would you consider leaving the left brain/right brain myth out? I love the elephant and the blind men metaphor.
Do you have a cite for it being mythical? I have asked neuroscience colleagues whether they thought it was bunk and all of them said to me that McGilchrist is reasonable, just not settled science. He's quite well received at Oxford. He goes to a great deal of effort in his works not to oversimplify it as left/right, I'm just writing a sentence here in a small blog about very different things.
The idea of hemispheric specialization of complex traits and emotions has been abandoned ages ago (other than speech and orientation and even that is not clear cut). I am curious to know who are those neuroscientists who say he is legit?! Shall I DM the refs?
Yes please do, though I’d also be happy for them to just be posted here if you don’t mind so others can also follow.
McGilchrist also says that it was abandoned long ago and that he’s not seeking to revive the old model. But he thinks that the baby was thrown out with the bathwater, as it were. He uses left/right as a shorthand and there’s dozens of pages of caveats and footnotes.
In any case, my concern is more about what might be called linear habits of mind or something like that. Maybe I’ll just use that phrasing in the future, but I like the wider analysis that McGilchrist is putting forward, especially about seeing the whole vs the parts and things like that.
I understand you use it as a shorthand but to me, that is how the myths become cemented and enshrined in culture. Of course you can leave it in- it’s your piece! Just saying it’s a shame as I would restack it otherwise as you provide very knowledgeable and nuanced insight on topic that my audience would love to hear about. For the refs, please see below, we go through the whole shebang
https://open.substack.com/pub/neuroscienceandpsy/p/so-happy-together-deconstructing-ede?r=4bkvkm&utm_medium=ios&shareImageVariant=overlay
Thanks, that was very interesting. Is your concern with flippant use of the left/right distinction or with McGilchrist's work specifically? I don't think he would disagree with anything you've written, especially not this notion being dumb:
"By drawing a physical line down the middle of the brain, it seems to give tangible, material proof of this cognitive-emotional divide. And beyond that, it echoes an even older narrative: the age-old split between soul and intellect."
Indeed, iirc, he opens by saying that this way of thinking about left/right was dumb. That's what I mean by 'bring it back but not as it was before'. He always starts his talks by saying that both hemispheres are involved in everything and the situation is immensely complex. His books are sooooo long for this reason. If there was a pithy description of the difference between the hemispheres in there it would be that the left tends to think in parts and instrumentally, and the right in the wholes. I can see how this would drift into things that give you grief, like "right to right brain communication is where therapy happens", so my sympathies. I'd be happy to remove the left/right reference but I'm reluctant to purge McGilchrist because I think his critique of linear, instrumental, look at the parts thinking is extremely important, even independently of the neuroscience element.
We wondered about something similar in our "How people know their risk preference" study. And yeah, you can find examples of incommensurable scales when you look at introspection. But isn't it interesting that it somewhat works at all (i.e. aggregated ranks kind of make sense as in the "feelings integers" paper), including in some cases across countries? I also wonder whether the introspection given after the response is all that truthful vs sort of confabulated.
Thanks for engaging. It's definitely interesting that it works at all. On the introspection, we found a couple of people doing what seems like confabulation, but honestly most people had quite reasonable frameworks for thinking about their lives and why they were going well or badly. Two of my coauthors were expecting a lot of confabulation and were pleasantly surprised at how sensible people were. One of the most valuable aspects of the study was just hearing how non-academic people think about stuff and express themselves. I am more sceptical than any ever of the speed with which scientists accuse people of being 'irrational'.
I agree re walking back psychology's tendency to frame things as irrational. I'm a big fan of Mercier's framework and writing on this.
Still, I don't think we can tell whether people have access to the process that produced the response by whether it's reasonable? Maybe if you choice-blindness-style switched up their response for something else, they'd justify that reasonably too?
That's why I had outside raters guess the risk preference of the authors of the introspective statement. I found some intersubjective validity. Not a ton, but these statements were often very short. Are you planning on doing that?
Ah I see. I think I had the wrong interpretation of your question. I'm perhaps not familiar enough with the literature you're operating in to give a tight answer. We certainly haven't gotten to the point yet where we are pinning down parts of their cognitive process using experimental conditions or something like that and I don't have plans to do that. Some of my coauthors might be interested though.
Speaking more simplistically, people answer these questions very fast, which might indicate a more affective response. Some people explain their reasons in very affective terms - "yeah life is good, I'm in a nice relationship, I'm happy". But others have very elaborate frameworks for thinking through their life, and we suspect that the reason they can answer so quickly is because they check in with that framework occasionally. They can call this framework to mind quickly too. There are others who are more rambly in their justifications, but there was much less of that than we expected.
Does that answer your question?
I better understand what you've done and found now. To confidently come to the conclusion that the justifications describe the process, I would need more evidence. My guess would be that it's somewhat right, but people don't spend a lot of cognitive effort answering survey questions. So, I wouldn't be surprised if quite a few people might change their next numeric response a bit after going through an in-depth interview of their circumstances just because they've reflected more.
If you don't know it, Steiner & Frey have done some nice work getting into the weeds of "process tracing" https://psycnet.apa.org/doiLanding?doi=10.1037%2Fdec0000127
non of this is experimental though. I haven't done choice blindness work either and I don't know how solid that literature is. I think more solid than the "weather colors your life satisfaction" work but that's not saying much. This paper https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1598254/full seems to say there's trouble with the CB paradigm.
I'll have a quick look, thanks. And thanks for the clarification.
As it happens, the last question of our interviews was "after all the conversation we've had, would you like to change your scale response?", and we had very few people say yes. 5 if I remember correctly, and all upward. The ones that I interviewed all said they'd been too grumpy at the beginning and reflection had made them see more bright spots or communicate a bit more positively (the latter perhaps reflects on the interview experience and the interviewers).
I also don't have full confidence that the justifications describe the process. But I also don't have confidence in using these scale responses while we have 0 understanding of the process that generates them. Our paper opens with an analysis of how the standard statistical techniques in this space (i.e. ordered probit) assumes things about the underlying data generating process that have never been tested. We also say that our paper is a not test, but only necessary preliminary work for getting to a test of some sort.
That's really interesting. I find it kind of reassuring for self report stans. Maybe I was overindexing on the choice blindness stuff. Either way, I agree we should better understand how subjective reports come about.
I'm still puzzled that men and women have almost exactly equal life satisfaction in most EU countries, despite differing on many inputs like income and personality. Like, how does that happen?
I'm looking forward to the paper.