13 March 2008
Subjective wellbeing and families: Issues of measurement and data interpretation
Professor Robert A. Cummins, Professor of Psychology, Editor-in-Chief, Journal of Happiness Studies, Deakin University
Abstract, slides and audio of presentation
Edited transcript
The following audio presentation is brought to you by the Australian Institute of Family Studies as part of our monthly seminar series in which we showcase national and international research related to the family.
The seminars are designed to promote a forum for discussion and debate. They are open to the public and free of charge.
Seminar facilitated & speaker introduced by Dr Daryl Higgins.
Professor Cummins:
Good morning to you, ladies and gentlemen. I put up this first slide which I must say I always find personally disturbing for some reason, but of course this kind of question is being debated by philosophers since people started thinking. Essentially it hasn't got very far, but over the last 30 years the area has been well and truly taken over by science and there's now about 30 years of systematic research into this area which is starting to become very interesting because it's becoming theoretically cohesive.
Three disciplines involve themselves in this broad area of life quality: economics, medicine and the social sciences. There are, in my view, very deep problems in the way that economics and medicine conceptualise life quality, and if any of you would be interested to see my critiques of that, then do contact me and I'll be delighted to send my publications on the topic.
Today I'm taking, of course, the position from the social sciences, and the social sciences is really the area which is making strong theoretical advances in terms of the understanding of happiness; what it is and what we do with it. Most fundamentally in this whole area of life quality, is the idea that there are two universes of wellbeing. There is the objective, familiar, physical world of wealth and health. And then there's how people actually feel about that, and these two do not happily relate to one another at all.
One cannot determine someone's level of happiness by looking at their wealth or seeing that they are in a healthy body. This healthy mind and healthy body stuff really irritates me, it's so wrong. We can't determine almost anything about psychological wellbeing from the physical parameters other than the measurements that are made. If we want to know how people feel about themselves, then we need to ask them, and then we find that many people who are quite severely disabled will report a reasonable level of subjective well-being.
Let me now move to the term happiness. There are all sorts of unfortunate and dreadful words in the English language. Happiness, unfortunately, is one of those. We normally associate happiness with a short-term state, an emotion, an emotion towards something, having a cup of tea on a hot day, something like that. That's not what this area of subjective wellbeing (SWB) is concerned with.
SWB concerns happiness in terms of it being a trait, a mood state that we carry around with us all the time, a basic attitude, if you like, to a feeling about ourselves. This mood state, which is very strongly based on contentment, has, we believe, a very strong genetic component. So it's this mood state, this trait level of happiness that I'm going to be talking about.
How do we define this subjective wellbeing? The definition that we go with is this: it's a positive state of mind that involves the whole life experience. It's not a part of life, it's not feeling happy with your job; it's feeling happy about the whole of your life.
How do we measure this? It can be measured through a single question, and this question has been asked now in national surveys through the western world for at least 30 years. It's a very curious question. "How satisfied are you with your life as a whole?" No-one has the foggiest idea what it means and people answer it in less than a second, and I'll explain to you why they do that shortly.
Its actually a good question, it gets pretty close to what we think is subjective wellbeing. But as I am sure you will all know, single items are not much good as scales, and therefore we have spent many years trying to devise a scale that breaks this question down into its essential components.
How do we recognise the minimum number of those components? The way that we do it is by setting up a rule. Now down the left-hand side you'll see our current eight domains as we call them, which are the broad areas of life. We say to people "How satisfied are you with------", and then, ------your standard of living, your health and so on down the list.
In order to be included in this list, each item must make a unique contribution, it must contribute unique variance to satisfaction with life as a whole. I'm not aware of any other scale that has been constructed along these lines. All the domains hang together like a little factor really strongly, but they only get a ticket in the team if they make this unique contribution. So we're constantly looking for new domains to add to the team, but they're becoming very hard to find.
The spirituality/religion one is the most recent addition. When we use that in Australia it never works, it doesn't make a unique contribution because Australia is such a Godless society. Even for the people who have spirituality and religion in Australia the domain still doesn't make a unique contribution. However, in countries where religion is much stronger, like Islam in Algeria or Roman Catholicism in Columbia, we find that this domain does make a unique contribution. So because the PWI is now part of an international collaboration, we decided about 18 months ago that we needed to add it to the other domains. When we add it into the Australian index, as I say it doesn't work very well, but neither does how satisfied are you with how safe you feel. In Again, however, while in Australia this doesn't contribute unique variance, it does in other countries. So what we're looking at here is a scale that is a bit compromised. It has one or two of these domains that actually don't work as intended in Australia, but we keep them in. I don't think it interferes with the functioning of the scale. It's like they're little 'sleeper' domains. People respond to them ; they don't find it hard to do that, but they just don't do anything very interesting as part of the scale. However we keep them in there because of this international collaboration.
So here is the personal well-being index. As you can see it's a very highly theoretically and empirically derived scale. It's not a constant scale. We're looking for ways to improve it all of the time and that, I think, is one of its interesting features.
When we ask people to respond to these questions, we do not use Likert scales. Ladies and gentlemen, Likert scales really are dreadful. They make just about every error in the book. Here is a beautiful little interval scale, one, two, three, four, five, , and then you put adjectives above the numbers that actually don't correspond to an interval scale at all. Nothing like it in fact. Not only is that a source of confusion, but it doesn't provide enough discrimination.
Why five points or seven points? It's really odd. The answer is because we run out of adjectives. You don't need these adjectives. Try zero to 10. Even little children know zero to 10, they all watch Sesame Street, and people don't need to have the numbers described to them. Zero to ten, defines the ends of the scale and then you let people do their own thing in terms of the number they choose. They can do this extremely reliably.
This scale was invented about 20 years after Likert by the amazing Thurston.
So this is the scale that we use. Everything we measure, we get people to use zero to 10. Then we re-jig our data to be like on a temperature scale zero to 100. This is such a good thing to do. It means that instantly your data have meaning. You can just look at it and you can see what's happening. A mean of 64 means 64 points on a zero to 100 scale. It has instant meaning, whereas a mean of 64 on a scale that runs from 12 to 140 or something, what on earth does that mean?
So we convert everything. All the data that you will see has achieved this conversion, and if you want to find the formula for the conversion, it's in the manual for the scale which is on our website and can be downloaded at no cost, of course. So zero to 10, zero to 100.
Now this area of subjective well-being has been researched systemically for about 30 years. What has emerged from all of this research is the fact that whenever we look at groups of people, either whole populations or demographic subgroups, the groups that have the high levels of well-being are the ones that do well, on almost any measure of the human condition. Groups with high levels of subjective well-being, happiness, are healthier and live longer, they're friendlier, they create social capital, they solve problems and they're resilient. When bad stuff happens, they're able to resist it and to maintain their well-being.
With these ideas and the basic scale of the personal well-being index in 2000, we knew that we had the capacity to do something quite different, and so we got together with our partner, Australian Unity. We set out to create a quarterly index of SWB for the Australian population, and we have done this ever since. We're now in our eighth year of partnership with Australian Unity, which is a bit of a milestone because it's longer, as you would know, than the average marriage in Australia .
The idea is to measure the well-being of Australians as a nation and then to report those quarterly in the media, in the same way as GDP in order to get a comparison between the two very different forms of measurement.
We did our first survey in April 2001 and our last most recent one, the 18th survey, was conducted late last year. Every time we do one of these, we sample 2000 people across Australia on a geographically proportional basis. So the numbers that I'm going to be showing you will be based on very large sample sizes because for most of what I'll be using the cumulative database.
What do we find? Okay, up on the left-hand side here we have the zero to 100 scale of subjective well-being. Down here we have all of the 18 surveys, and is the date of each survey. Then because the line goes up and down, as these lines tend to do, I look at each change and make up little stories about why it's going up and why it's coming down. I'm not quite sure I ever believe those very much, but they're the best we can do at the time.
What I want to draw your attention to is something perfectly extraordinary. Each point on the graph represents the mean of 2,000 randomly chosen new respondents. Amazingly, the total variation is just three percentage points, between 73 and 76 points. The standard deviation between these 18 measures of subjective well-being is less than one percentage point.
People used to say oh, why would you bother measuring this stuff? People are happy one moment, they're sad the next. There's too much variation. Now people say to me, there's so little variation, why do you bother measuring it? It is extraordinary. It means that we can now get a random sample of people and, as long as there's nothing special going on, I can predict their level of well-being to within less than one percentage point. Why? How can it be so incredibly stable?
The idea that we've been working on now for many years is that we're looking at a managed system. A psychologically managed system, and we've stolen the term homeostasis from the physiologists. Of course, you're familiar with the fact that if we walk outside into the heat, our core body temperature remains stationary, as long as we don't stay out there for too longThis is due to the homeostatic control of our body temperature which involves a very determined set of physiological devices. As an analogy, we call this psychological management system, subjective well-being homeostasis. The way that we think it works is that each of us has a set point for our subjective well-being. We don't quite know what the normal range of the set points, but a best guess is between 60 to 90 points.
So even when it's low on 60, it's in the positive part of the range. People on 90 are a bit too high, very irritating people, constantly joking and laughing and bad stuff happens, they dip down for a while, and they're up again doing their thing. People on 60 are a bit low, they're not glum, they're not depressed or anything, they're just more introspective than people at the other end of the range. And the average is 75, which is the mean score we derive for the Australian population.
One reason for the stability of our population data is that people tend to lead remarkably boring lives. We don't like surprises. We do all sorts of restrictive things to avoid surprises. So we track through the day in a predictable way as far as we possibly can. Of course we get fluctuations around the set point to some extent. There's what we call a set-point-range, or the normal range of variation that we experience around our set point. We tend to oscillate around that set point.
Because people's SWB is normally located within their set-point-range, when we go to the general population, we are essentially measuring the average of individual differences in set points. That's why it's so stable. Of course, some people will have just experienced something very good or very bad that takes them out of this range, but that's unusual and it doesn't last very long, particularly the effects of good things don't last very long. We come back very quickly into the normal set-point-range for ourselves, whatever that is.
So we generally maintain ourselves in this positive state. But of course, if you go outside and stand in the sun for too long, you will no longer be able to maintain your body temperature. You will become hyperthermic, and temperature homeostasis will fail. That is so with any homeostatic system. It can only absorb a certain amount of challenge.
When we overwhelm this homeostatic system with negative experience, we lose our ability to maintain our normal positive sense of ourselves, we lose our positive affect. That is depression. Thus, the result of homeostatic failure is a tendency to depression.
However, this doesn't usually happen. We're pretty resilient, and we have a whole host of resources that we bring to bear when challenges are around to not let those challenges actually affect our sense of is subjective well-being. We divide these resources that we have to defend ourselves into two baskets. These are external and internal. The two major external resources are certainly relationships and money.
When we think of relationships, in this context we're talking about social support. We're talking about having someone to whinge to at the end of the day whenever things have gone wrong, someone to off-load to, someone who will protect us also to whatever extent they can against things that might damage or challenge our well-being. That is really what we refer to by social support. In 2005 we pursued this to some extent. We asked how much support people thought they received from their partner, the rest of their family, et cetera, here on the list.
Now when researchers ask this kind of question, they typically have the assumption that they're working with a linear system. And you know, this is one of the fundamental flaws of research, that we assume linearity. Indeed our statistics demand it, unless you're one of these strange people who understands curvilinear stats. We're all stuck with it and it causes us to think in the wrong way. In fact, nothing that we measure is linear, not in the world of psychology.
What I'm going to show you now is an example of this non-linearity. Here we have again our zero to 100 scale, of course, and here's the level of support. I've just taken support from partner as an example. The other sources of support show the same pattern. The yellow bar here is the normal range. That's where groups should be if they're operating normally, and what you can see here is we're measuring subjective well-being, so we've said to people how much support do you receive and then we're plotting that against subjective well-being. People who say they've got a supportive 10 out of 10 relationship of course do extremely well. They have a level of SWB that's getting fairly close to maximum.
We can't get groups of people to give us a well-being score much above about 82 points. This limit is caused by the genetic individual differences of set points causing a distribution within any group. Thus, no matter how good the group is going along, the distribution in the group will cause the mean to come down to about 82 at the top. So the 10/10s are doing extremely well.
It comes down pretty fast, though. At nine out of 10 and eight out of 10 they're still okay, but by the time it comes down to seven out of 10 support, their level of subjective well-being is no different from people who've never married or lived alone.
Both of those groups have levels of well-beingbelow the normal range, and then it goes into a catastrophic nosedive. The reason that it falls down so much is that relationships are supposed to be reciprocal: you give support; you receive support. They're supposed to be a resource for you to use to defend against low levels of well-being, but if somebody isn't reciprocating, then they're a black hole for you. You support them; there's nothing coming back. It's a negative relationship for you, and so it's draining your resources and so naturally you might think this is bad for well-being. Indeed, it is.
Then as levels of support get even less, down towards 0, well these people are just totally disengaged. They may or may not be physically separated from their partner, but they're certainly not putting any energy into them.Because of this, the negative influence of the partner on SWB is reduced, and so their well-being comes up to some extent.
So it's very non-linear and obviously if we just apply linear stats to it, we'll get a result, but we're going to lose the story if we do that. We lose the understanding of what's really going on here.
It's interesting to observe that about 80% of people in this sample have a level of well-being from 10 to eight in the normal range or above. So presumably people are making a choice and if things are too bad, they're getting out of it and, of course, that is a key to this whole story.
If you want to maintain your well-being and you're in a damaging relationship like this, then the only thing to do is to find a better partner. That then is what's happening in terms of relationships.
The other major external resource is income. It's not that income gives you happiness when you buy a new pair of shoes. It might do, but it will be very fleeting. Thatkind of purchased happiness is an emotional response. Emotional responses by their very nature are short-term.
The role of money in terms of subjective well-being is different from that. Money is a flexible resource. So if you're a rich person and you don't want to do something like cleaning your house or picking up the kids from school or whatever it is, you can pay someone to do it. That potential source of challenge to your well-being, can be disposed of by throwing money at it and getting someone else to do what you don't want to do.
So the power of money is evident, most evident, in the low income ranges, and you can see subjective well-being rising significantly. The little stars on top of the bars indicate that each mean is higher than the previous mean.
People with very low incomes are, of course, pretty naked in the face of the things that happen to them. They haven't got the financial resources to buy their way out of stuff, and so their well-being suffers as a consequence, but as income rises, so does our capacity to defend ourselves until it gets up to about, say, $100,000. This is gross household income, not individual. Once it gets up to around $100,000 , on average the income effect peters out, exactly as you would expect. You can only buy your way out of a certain amount of trouble.
If you've got a difficult partner, well you've got to take different action. So there's a great limit to the extent to which money can do its job of defending well-being.
Let me try and put this together for you. Here we're looking at again income and subjective well-being, and we're looking here at a situation where someone is living only with their partner. I say to my students, who are young enough to not yet have committed most of the errors in life that most of us have committed, , look carefully at this graph. This is a group of people doing really, really well, even when they haven't got much money. They're still strongly within the normal range. Find a nice person, I say, settle down with them and don't make babies, or if you do make babies make sure you have enough money.
This is what happens when we add children to the mix. Children of course draw the resources away from their parents, financial and emotional.
Now once we get up to a certain level, in fact the median level of income in Australia, this effect dissipates and the well-being of the parents is no different from people who don't have children. But in the lower income groups it is clear that the resources of the family are not sufficient to properly sustain the family.
This pattern becomes exaggerated when we look at single parents. Of course, single parents very often lack both external resources. They are low income and they lack a partner, and so we just see an exaggeration of the trend.
It's not true, however, that single parents necessarily have low well-being. As you can see once their income gets up towards the $90,000 mark, their well-being comes back into the normal range. It's a question of resources. They have regained, at that level the resources, the ability to manage their own well-being.
When the external resources fail, all is not lost. We also have very strong internal mechanisms that again prevent us usually from feeling bad for very long. We call this device by various names, but the generic term is cognitive restructuring.
Here's poor old Todd, got up in the morning, made his cup of coffee, oh no!, immediately he feels bad, well-being plummets. Then he thinks, well, I can't be good at everything and I'm good at so many other things. Carrying coffee is stupid, everybody would make a mistake like this once in a while. So he takes the event, recasts it in his head, thinks, well it's not important for my self-esteem that I can't carry a cup of coffee, and starts to feel good about himself again.And that's what we all do.
There's all sorts of tricks that we use to turn what is an obvious negative into something which is either not us or a positive. We get fired from our job, feel dreadful for a while, then we start to think, well that was a crappy job anyway, I hated it. Oh, the boss was such a miserable so-and-so. Now I'm going to find a much better job.
So you reconstruct the whole thing in your own favour. It's terribly important that we do this. Here's poor old Todd, drops his coffee, feels bad, but then has a little think about his life and
realises it was a minor event and up his well-being comes again.
Of course, if Todd was a rich person, he would hire a Butler and none of that would have happened in the first place.
So how does this homeostasis work as a psychological system? Well, the conventional wisdom in this area looks like this. There is a genetic component for the well-being, for the set point, and traditionally that's been cast in terms of personality. Of course, psychology for a very long time has regarded personality as being largely a genetic pre-determined characteristic, and two of the five personality dimensions, extraversion and neuroticism, are particularly thought to be powerful determinants of subjective well-being. If you like, extraversion positive, neuroticism negative and somehow these creating a balance.
We're consciously unaware of all of that personality stuff, but we become aware of our self in interaction with the world through our sense of control, how much control we think we have in a general sense, how we feel about ourselves and how we feel about the future. We call these conscious buffers because they're pretty hardy, they don't change all that much, and this is where the experience meets our awareness of ourselves.
We we measure the output from this system as subjective well-being. That is the conventional view.
In 2003 a remarkable American researcher, James Russell, who will be known, I'm sure, to some of you, who has spent his lifetime largely in the area of affect research, published an extraordinary article in psychological review. This took our imagination enormously. What Russell did was to define a new psychological construct, and he called this construct Core Affect. He had some trouble defining core affect, but he did so in the following terms: he said it's consciously accessible as a non-reflective feeling not tied to an object. It's not an emotion; it's a mood without any link to something else. Like body temperature, it's always there, it's primitive, universal and irreducible and it's ubiquitous, it's throughout the entire brain.
So we looked at this and thought this was an interesting idea and wondered if this, in fact, was really the essence of subjective well-being.
We published our findings at the end of last year, with enormous difficulty. The article is interesting to read because it's accompanied by a very fierce critique and our rebuttal. In this article we make the determination that core affect can be characterised by these three adjectives: contented, happy and alert, and that these, in fact, describe subjective well-being much better than personality.
So the model that we're currently operating on in terms of how all this stuff works looks like this. That at the centre of the universe we've got core affect. This is produced by a little generator in the brain under genetic determination that throws out positive affect, a mixture of contentment, happiness and alert, and that this drives everything else. Itdrives personality, it drives the buffers, and keeps them at 75 points.
If I say to you, how much control do you have in your life, like a general question, you'll tell me 75 points on average. So you're actually reporting core affect, and this is what we measure. We think through subjective well-being.
Core affect allows us to explain why people are so adept at responding to satisfaction with life as a whole. You remember this general question that's been used so often. It's actually a very neat question because it's very personal and it's very abstract. People answer it by referring to core affect, and, in fact, we find that it shares about 60% of the variance with core affect.
When we use the personal well-being index, which is rather more specific and rather more cognitive, the amount of shared variance with core affect drops down to about 56. So we've got a bit of a trade-off with the personal well-being index. We're a bit further away from core affect, but we've got a more powerful and interesting scale because of the component domains.
If core affect is, as we say, such a central determinant of everything, then why aren't we aware of it? Well, perhaps we can be. Perhaps we can be aware of it through meditation. Perhaps when we meditate this is where we go.
Here is a meditation master, and this is his description of the process, his version of meditation. Essentially it involves, of course, cutting out the cognitions and the emotions that are interfering with our sense of ourselves. If we cut those out, cut those away, where we go to then, I suspect, is core affect.
I have looked at various definitions of how people describe themselves in the meditational state, and here are the first three that I came across: calm, peace and moments of great joy. Remember we're looking here at a description of core affect as being contented, happy and alert. Peaceful, positive, loving and experiencing clarity. They're quite reasonably consistent with it and, indeed, I'm now becoming a bit intrigued by this possibility.
There's a spiritual community in India that I'm collaborating with now to see if we can quiz some experienced meditators a bit more closely to see how they report their states. Watch this space.
Of course, homeostasis can fail, and when it fails we lose our ability to maintain our well-being and, as I've said, that results in depression. How does this all relate to the well-being of families?
Well, we do produce special reports as well as our national reports. We also look at the well-being of groups in Australia, and last year we combined with Carers Australia, who have recently been in the headlines again in their fight with the Rudd Government.
This, of course, is a special purpose sample, and we mailed questionnaires to about 12,000 carers, and we got 4,000 back, which wasn't a bad effort considering the very poor state of mind that most of these carers are in.
To give you an impression of what we found, here again we've got the personal well-being index on the left-hand side, zero to 100, and these are the groups that we have found through a search of our own demographic database. These are the lowest groups that we found in Australia. Essentially they involve unemployment most particularly in middle-aged males, low income and living alone. When we start to combine these three together, and here are the combinations, we find very low levels of well-being. People who live alone and are unemployed here you'll see have a level of well-being of 60, which is extremely low.
When we went to these 4,000 carers in Australia, to my absolute horror we found they were even lower, 58.5. That actually registers as the lowest group that we've yet found, and given the size of the sample, the data are just nothing short of horrific.
Abnormally low well-being, as I indicate, is a measure of a high risk of depression, and in this study with the carers we actually used the depression scale from the DASS, the depression, anxiety and stress scale that I'm sure will be familiar to many of you in the room.
Here are the horizontal bars of the cut-off points of the various levels of depression, and here on the bottom line we've got mild depression, and this is the cut-off point. When we looked at where our carers were on depression, consistent with our measures of subjective well-being, the mean, of this enormous group of carers, according to the DASS, were, in fact, exhibiting a moderate level of depression.
We also asked, 'are you the person who provides most of the care?' Are you the primary caregiver? We've asked this question before and the data that we found confirmed a previous study. Here you can see the people who said yes, they were the primary caregiver and those who said no, they were just another member of the household. Their wellbeing actually differed slightly but the difference is swamped completely by the fact that even the people who were not the primary caregiver had extraordinarily low levels of well-being. In other words, the situation of the caregiving, depressing a member of the household who's the primary caregiver, which in many cases is mum, actually has this pervasive affect on other members of the household, to give them a very poor sense of their well-being as well.
So one purpose of monitoring this subjective well-being is really to detect people and groups, individuals and groups, who have well-being that's below where it should be. When this happens, then we know that they need more resources.
As far as I can see, the vast majority of depression concerns resources, either personal resources, like social resources or money, or the resources of just being able to restructure their own cognition so they can think their way out of something that's causing them to lose their well-being. That, I believe, is probably able to be taught.
The other two are more difficult to achieve, higher income and better relationships, but if we can find these resources, then all we're doing is providing what people need in order for them to then return to their normal levels, their natural levels of homeostatic control.
So subjective wellbeing can be easily measured. If it's lower than normal, it indicates that the resources have failed and it's a high risk of depression. It is possible that this is actually a better measure of depression than most of the depression scales, particularly those ones that use psychosomatic sequelae as part of the scale. Then, after having made that kind of measurement, of course, the professional responsibility is to provide or attempt to provide those resources so that the person can regain their normal natural levels of positive wellbeing.
Daryl, ladies and gentlemen, thank you.
-- end --
IMPORTANT INFORMATION - PLEASE READ
The transcript is provided for information purposes only and is provided on the basis that all persons accessing the transcript undertake responsibility for assessing the relevance and accuracy of its content. Before using the material contained in the transcript, the permission of the relevant presenter should be obtained.
The Commonwealth of Australia, represented by the Australian Institute of Family Studies (AIFS), is not responsible for, and makes no representations in relation to, the accuracy of this transcript. AIFS does not accept any liability to any person for the content (or the use of such content) included in the transcript. The transcript may include or summarise views, standards or recommendations of third parties. The inclusion of such material is not an endorsement by AIFS of that material; nor does it indicate a commitment by AIFS to any particular course of action.