Can understanding society's concerns help ensure a future of wellbeing?
Presented by Dr Terry Flynn
UTS, The Great Hall, Tuesday 25 May 2010

Attila Brungs:

Good evening everyone and thank you very much and a warm welcome here to UTS. My name's Attila; I'm the Deputy Vice-Chancellor Research and it's my pleasure to introduce Terry Flynn a bit later on.

Firstly however, the University of Technology Sydney, acknowledges the Gadigal and Guring-gai people of the Eora Nation, upon whose ancestral lands UTS now stands.

Tonight is the fifth presentation in the 2010 UTSpeaks lecture series and I'm particularly looking forward to some very interesting discussion and debate on the topics that Terry will raise, and your questions at the end of the lecture.

If I could ask you to please be aware that tonight is being digitally recorded and also videotaped to be later aired on the ABC2, Big Ideas television program. So if you could turn off your phones completely to ensure that no electronic interference occurs and also if you have to come and go, could you please enter and leave quietly, and close the doors.

So, first of all, now let's get to Terry. Terry comes from one of the most impressive and impactful research centres that we have here at UTS, CenSoC - the Study of Choice. I'm relatively new at UTS and the CenSoC study - the Research Study, was one of the most, as I said, impressive centres to really start to understand what drives people - what it is that drives consumer behaviour, in a fundamental and a rigorous manner.

Recently they just published two studies: one - and independent public enquiry into Sydney's transportation system, which showed, despite the fears of politicians, that the majority of residents are actually pay the amounts it will take to overhaul the public transport system to make it work for them so they can actually use it.

Later, they also demonstrated that the Australian public were willing to pay for one of the options on offer in the emission trading scheme, indeed, using their techniques, they understood that the government should go ahead with a plan, whether or not other countries did so.

These are very valuable insights that allow policy makers to make decisions that affect our lives on a day to day basis.

If I can ask - talk about Terry. Terry heads the Social Policy and Economic Evaluation Stream in this centre. Terry's work has been incredibly influential internationally. His diverse research addresses the major public policy issues such as quality of life, nationally and internationally, what Sydney residents are willing to pay for public transport, what value Australians derive from surface and ground water dependent eco-systems.

Before moving to Australia, his work influenced the UK policy, with respect to the guidelines on which social care related to quality of life should be valued by the public. His work on choice modelling has been recognised by the top UK funding bodies and he collaborates on major British studies, to value social and health related quality of life.

So, without further ado, I'd ask Terry. Thank you.

[Applause]

Terry Flynn:

Okay, thanks Attila and thanks for Robert and UTS for giving me the chance to speak tonight.

Just a little bit more background, first of all. I moved to Australia just over a year ago, and it wasn't just for the weather. I've been coming here on and off for a few years. I started working with Jordan Louviere, the Director of CenSoC, about six years ago, and he was very interested in using choice models to value quality of life. He was always very encouraging and I would always have three or four manic weeks every year over here, trying to get as much work done as possible.

Eventually he created this job for me and I jumped at it. When I saw the diverse areas that CenSoC work in, it was a just way of expanding my knowledge and experience across public policy.

Indeed, one of the first projects I was involved with which Attila has mentioned, was work for the Sydney Independent Transport Enquiry and this was really very interesting to work on and we were really very flattered that the Herald chose to lead with the work that we had done - that CenSoC had done, eliciting preferences amongst Sydney residents for potential improvements to the Sydney public transportation systems.

This was the article that was on the front page in February this year. What was really interesting and had the high impact was that we used choice models to find out how much people were willing to pay for potential improvements.

Choice models essentially present people with hypothetical but realistic options. They tend to be specifications of a good or service. So we presented alternative improvements to the public transport network, and the road network, along with associated costs - things like paying additional taxes, property taxes, paying additional fares, a potential congestion charge driving to the CBD.

This allowed us to elicit how much people were willing to pay for these big improvements. As we always find with choice models, people are different. There's no one single view out there. But what was very encouraging was that almost two thirds of residents were willing to pay the amounts necessary to bring the Sydney transport system up to the standards in many other countries.

There was a smaller minority of people who were willing to pay, but they wanted the roads to be improved. The smallest group of people didn't want to pay for anything. They were quite happy to accept higher congestion and all the problems that would be associated with a larger population that is inevitable in Sydney.

Now, this is an example of a choice model in public policy. Choice models are now increasingly being used in other areas, and quality of life is the area that I've been working in.

I thought I'd give you another media example here to illustrate how things have been done a lot of the time to date. This is from my local rag, the Mossman Daily and claimed that North Shore residents are the happiest in Sydney. The results from the Cumberland Courier Community Pulse Survey no less, revealed that the happiness score out of 10 from Mossman residents was 7.5. Yes, and?

They talk about, oh yes, well I only need to go down to Balmoral Beach, look around to see all the smiling faces, said Mr Palmer. Okay, maybe I'm just a silly Pom, but isn't the beach where you go to have fun? I don't really see many people crying down there. Apart from a few locals who are pissed off with the traffic charges and parking et cetera - and parking charges, but you can never please Mosmanites.

But, it's actually not quite as simple as that. This headline result conceals some other disturbing facts. It showed that residents were workaholics - 32 per cent more than anywhere else in Sydney, saying they worked more hours compared to last year. People concerned with the economy - economic related issues troubled 68 per cent of respondents.

So it occurred to me, well these happiness scores, I'm not really sure we're getting the full picture here. Indeed economists who've begun these really, should know better. If they'd looked in other disciplines, they might have had a little more reticence about endorsing some of these measures.

Indeed it's well known that they provide problems in certain contexts. I actually prefer the quote from Henry Louie Mencken, from the early 20th Century in America, who was quite a popular commentator. He said,

"Explanations exist. They've existed for all time. There is always a well known solution to every human problem: neat, plausible and wrong."

I think that the happiness scores are an example of this. Mencken made very controversial statements. I think he was deliberately controversial, but a lot of politically incorrect statements were made by him, but he did have some nice quotes.

Now, there are reasons why we should be suspicious of happiness scales. Happiness scales are an example of what marketers call a rating scale and it's well known in the marketing literature that people answer rating scales differently. Your 7 out of 10 is not necessarily the same as my 7 out of 10.

There are various reasons and various phenomena that have been spotted in the literature, particularly in International comparisons. The phenomenon of lucky and unlucky numbers. The number 4 is considered unlucky by some Chinese people. Other people just don't use the extremes full-stop. Everything's going on in the middle of the scale.

You also get some respondents who indulge in yea-saying, perhaps they're keen to please the interviewer, if it's an interviewer administered survey, or just to give the right answer. But most worryingly, these figures don't tell us where people want to be - what they value. They only purport to tell us where they are. So, what about the missing 2.5 of the Mossman residents? We don't know. Finally, a problem with these scores is that they're typically calculated at the level of the population, or just a level just slightly lower. This leads to problems, what we call an ecological fallacy, where we use inference at the level of the population to make conclusions at a lower level, say the individual level. Sometimes these are erroneous, and that's an ecological fallacy, because a lot of these happiness scores simply cannot give you robust differences at a high level of disaggregation. I want to know really what I and my peers value, not what the average person in Australia does.

Now in fact, I've got data which I think shows that people are using different parts of the scale. Bristol City Council, where I was - I was based in Bristol before I moved here. They administer a quality of life survey every year, in October. They're interested in a lot of issues, people's perceptions of the locality, fear of crime, aspects of their house, an all sorts of issues that may influence city council policy.

As well as a number of individual questions about individual circumstances, people are asked about their quality of life and they administered a quality of life questionnaire that my team was involved with developing, along with a happiness question. Now, we looked at the data and the first thing I did was rescale the happiness scores which were out of ten, to be on a percentage scale to be comparable with the quality of life scores, we got from our measure. Now I'll describe how we got our numbers later, but what's interesting is looking at the happiness scores, when you look at the averages in 5 year age bands.

The red dots here are the average happiness scores, with the age along the bottom. The unhappiest people appear to be teenagers. I'm not sure I needed a survey to learn that. People in their 20s seem to do reasonably well. There's this big trough in middle age, and that's when life seems to be pretty crap.

It then rises quite steeply really. There's a suspicious little peak here at age 65. I don't know why that happened. It falls down then a bit, and then starts rising again into advanced old age.

Now, all well and good. I looked at respondents other answers, and a lot of these people up here are very, very sick. Some of them are house-bound, very lonely and they're saying 9 out of 10. Something just doesn't seem right here.

Looking at our quality of life instrument results, we ask about 5 key dimensions of quality of life, and we use methods which I'm going to come onto to give us a percentage score, summarising their life based on their tick box answers - how much independence they feel they have, how good the relationships they have are - and we saw some comparability for the early part of the age spectrum.

The trough is here, but it happens a little bit later in quality of life and the rise again is nothing like as steep as it appears to be in the happiness scores. These are data from the same people.

We've got the peak at age 65, but then quality of life falls gradually and then steeply into old age. That's far more consistent with the specific questions we asked people about their health and living circumstances.

So there's a very different picture here, between happiness and quality of life.

What's going on? For Monty Python fans, here's the black knight from Monty Python and the Holy Grail, getting his limbs chopped off and saying, it's just a flesh wound.

I think the older people in Bristol are putting a brave face on. I think there's an element of Blitz spirit. The very old will have lived through the Second World War. Bristol was bombed flat by the Germans, being one of the two major ports where all the Lend-Lease stuff came in from the US, and I think their attitude is that well if the Jerries can't get me, nothing else will.

So, I think they've essentially changed their frame of reference that - so their answers are just not comparable with those of younger people. Indeed, from the marketing literature, there are plenty of examples of people using different parts of the rating scale.

Other people have criticised happiness scores and two particular streams of work, which have got quite a different take on things, both have criticised them.

One is from the work of a particular scientist called Matthieu Ricard, who gave up science, became a Buddhist Monk and wrote a very good book, Happiness: A Guide to Developing Life's Most Important Skill, and he tried to really get the essence of what these meditation techniques were - written for the scientist - the Western scientist who's quite - can be quite suspicious of all this religious mumbo jumbo.

He made some very nice statements, one of which is

"Someone who enjoys inner peace - true happiness - is no more broken by failure than he is inflated by success. He understands that experiences are ephemeral and it's useless to cling to them."

Now, his work really is consistent with us own having - each of us having our own internal happiness rating. We have a default setting. It can be changed, but it requires many, many - and we're talking hours of meditation here, to try to really immunise oneself to what life throws at you, both good and bad. He certainly doesn't see happiness as something that we can get from a rating scale.

Another criticism of this work comes from a guy called Oliver James, a Psychologist, who wrote a very popular book called, Affluenza. Certainly very popular in the UK. London was one of the case studies. Sydney was another one of the case studies actually, as was New York and he was another person who really didn't think that these happiness scores are worth the paper they're written on.

In fact, he's very dismissive of these think positive techniques that are increasingly advocated that they don't actually equip people to deal with what life throws at them and indeed he argues that happiness is just something akin to pleasure. It's something that you'll get every now and then. It's not something that you should be trying to change by public policy.

The things that give lasting long term benefit to people largely come in four areas. He talks about these in terms of: one is connectedness. It's about relationships, feeling part of a community. A second one is about autonomy, being independent, master of your own life, not being buffeted by events, experience - feeling that you can make a difference, you have this value in what you are doing, and finally security, both in terms of materially and emotionally. A lot of psychologists will say that those are the kind of human needs that are important. Not the ones that tend to be emphasized in a lot of these think positive courses, because these are concentrating on materialism and things that really just do not give long term benefit.

So, the question arises, well how do we get to here with all this stuff about happiness and did we take a wrong turn somewhere in trying to elicit well being.

Perhaps it's useful to go back to one of the great names of the 18th Century; Jeremy Bentham whose work was very influential in the development of economics.

He says, "Nature has placed mankind under the governance of two sovereign masters, pain and pleasure. It's for them alone to point out what we ought to do as well as to determine what we shall do. On the one hand the standard of right and wrong. On the other the chain of causes and effects are fastened to their throne. They govern us in all we do, in all we say, in all we think."

Now there is this idea of a spectrum, pain at one end and pleasure at the other. But he is saying that there are other things that people may be interested in, that they can be valued somehow on this spectrum. Indeed other great writers of the time were acknowledging too that it wasn't just about some very simple measure of happiness. Indeed a quote from the second section of the US Declaration of Independence is, "We hold these truths to be self evident. Men are all created equal, that they are endowed by their creator with certain unalienable rights, that among these are life, liberty and the pursuit of happiness."

So the issue of liberty here and other things in addition to happiness was coming out. Indeed it's interesting some of the debates that went on at the time. Ben Franklin and others scrubbed out the original wording of giving people the rights to acquire property and that's what government should be about and substituted happiness in its place and this has led to a lot of controversy in economics about the role of land. But land's another issue in itself.

These issues were very instrumental in the development of economics and the idea that there is this latent scale - this unobserved scale that we can actually place things on according to how much pleasure or pain they give the individual. This is largely - it was developed in the 19th Century and has triumphed when it comes to a lot of evaluation of public policy. Essentially what we do and how we value things, is we can't know exactly where things are on this latent scale. It's inside the person's head, but we can make inferences about that and the positions of things on that by the choices people make. In particular, in private markets, the amounts that people are willing to pay give an indication of the value that they place on the good or service they're buying.

Now, welfare economics in which developed all this has been very influential in public policy. It's generally the framework used to value goods in Transport and Environmental Economics. Indeed we quoted willingness to pay figures for the Sydney Transport Enquiry in terms of how much people are willing to pay in additional taxes, fares et cetera for a given improvement.

But health economists have been very uneasy with this. Health economists in Australia, New Zealand, Canada, most of Western Europe rejected the idea of willingness to pay - the idea that we value something in monetary terms, because there is the very strong feeling that willingness to pay may be influenced by ability to pay and health care should not be made available on the basis of wealth.

But, whilst health economists did abandon money as a metric, they did recognise that in the real world we have to make decisions. We have limited resources and there are competing claims on our time and resources. So we should be asking people to make choices and we should observe the choices they make and infer from those what people value, and that brings in the whole field of choice modelling and stated preferences.

To make this a little bit more concrete for you, I'm going to talk through the development of the first of the quality of life instruments, the team I've worked in, have developed. I was working in the ICEPOP Team, between 2001 and 2009 - acronyms of course abound these days. The original remit was investigating choice experiments for the preferences of older people. We were interested in several strands of work. But the main one was to develop a quality of life instrument for older people.

Now, although we were funded by the UK Medical Research Council, the remit was not to develop yet another health related quality of life instrument. General quality of life is of interest, because older people tend to require a mix of health and social care interventions.

So we started off by doing several rounds of iterative, qualitative research with older people asking what was important to them, what gave value to their lives and we pursued the answers they gave, and tried to distil the essence of what was driving these. After several rounds we elicited 5 dimensions.

Now, we use particular academic terms that we wouldn't dream of presenting to people in a questionnaire, but I'll just mention these so you can see them in the slides that come.

So we talk about Attachment; we actually ask people about love and friendship. We ask about Security and this is all about worries about the future and the extent that people can be free of these. The Role attribute or dimension is all about doing things that make you feel valued. Enjoyment is enjoyment and pleasure and Control is all about independence.

Now, to value quality of life using an instrument such as this, we make a distinction that's common in health economics between - we call it measurement and valuation, because the problem with the happiness scales is that people are using a different metric in their head. They're using numbers in a different way. But before they get to that stage, although they don't explicitly think this way, they're making a judgement in their mind about how impaired they are in the areas that matter to them.

So in essence we call that measurement, and then when they try to distil that into a number, that's valuation.

So what we do for the measurement stage is a respondent in a clinical trial or in a survey, will fill in the instrument - this is the ICEPOP capability instrument, Version O for older people - and they would tick a box to indicate which most closely approximated their life at that point in time for each of the 5 dimensions.

The dimensions each have 5 levels and the levels are generally phrased so that they are bounded by all of the amount that you want and none, although it does vary a little by dimension. Now, once a person has ticked boxes, we could take their 5 answers and put them together to give a description of their life - that's a measurement of their life. We're interested in getting a value that summarised how impaired that life was compared to the top state if they had ticked say the top level for each of the 5 dimensions, and the bottom state if they ticked the bottom level for each of the 5 dimensions.

One way of doing this is just to add up the scores - 1, 2, 3, 4 - for each dimension. We didn't want to do that. That's not what economists generally do because we're interested in eliciting people's preferences, how good or bad these impairments are. It's far from clear that everyone would consider extreme loneliness to be the same as having no independence. I wouldn't and my views may be different from yours as well.

So we don't just want to sum score these numbers. We actually want a percentage score that reflects how bad these various states are.

To do this, we used a valuation exercise and health economists use valuation exercises to value health generally, but we used the same framework to value quality of life. What we are doing is essentially we're trying to build up a scale for each person and for each person find out how would they value all these impaired states. So what we have to do is to get them to think about living in these states, as we would in a choice model. We ask them to make choices, discrete choices about what's good and bad in these lives to try to build up a picture of just how bad all these impairments are.

So what we're actually doing is getting a person to stand in someone else's shoes, or see a life through their eyes. That's way we tend to explain to people exactly what it's all about. But in terms of the actual task, here's an example.

This is taken from the survey that has been in field with our industry partner, Pure Profile, to get Australian values for the quality of life instruments - the original valuation exercise in the UK was interviewer administered, because it was with older people who are not very PC literate - and we presented them with imaginary - hypothetical quality of life states and these varied according to the levels - the amount that each dimension had.

This particular example is given from the second ICECAP instrument - ICECAP A, for adults of any age. It has the same - essentially the same 5 conceptual dimensions, but the wording is quite different on a couple of them.

For this state the respondent has to imagine living in it and decide what would be best and what would be worst, for that living in it. So essentially the choice from best here comes down to the first three dimensions, because they're all pretty good. You're able to feel settled and secure in many areas of your life. You've quite a lot of love friendship and support. You're able to be independent in many things.

The other 2 dimensions are pretty bad. You cannot achieve and progress in any aspects of your life. You cannot have any enjoyment and pleasure. So we can predict to some extent what people are going to be choosing as best and worst, but it's by no means certain, given that we use a statistical design matrix to ensure that the levels - the amounts that we present of each dimension have particular properties that will allow us to make inference about how good and how bad, relatively speaking, all these various dimensions are.

This is where the science bit comes in with choice modelling. What we essentially do is we look at people's choices to infer how good or bad the various levels of these 5 dimensions are. The idea behind all choice models is really quite simple. How much a person chooses something gives us an indication of how strongly they value it.

If the person keeps choosing the attachment attribute to its top 2 levels as best, no matter what it appears with, then they clearly think that that has a lot of utility - a lot of benefit to them. If they keep on picking a particular low level on one of the dimensions as worst, then they clearly think that's the most terrible thing to live with.

You notice we didn't ask people to rank things. It's well known that people are not very good at ranking lists. They don't tend to pay attention, particularly in the middle, when things get difficult. So we exploited this and just asked people about the top and bottom - the best and worst. This type of choice experiment was invented by Jordan Louviere and has been used internationally in a variety of areas now. It's called best-worst scaling. There are a few references here for those who are interested.

What we're trying to do is to elicit each respondent's norms - their values and essentially place on a measuring scale, all these various impairments, so we can compare them and see how bad or good they are compared to one another. So how bad is social isolation, for instance, compared to impaired independence? How bad are severe worries about the future, relative to being able to do things that make you feel valued? Those are the kinds of difficult decisions that we're trying to inform from the choice model.

Okay, we did the valuation exercise and we got some results and these reflect the average values - the average norms of British older people. In terms of interpreting this, we can simply look at the bars for the 4 levels, for each of the 5 dimensions. So for instance, this would be our look up table. It's a set of off the shelf scores that we can apply to summarise a person's tick box answers.

So if somebody ticked that they were experiencing the top level for all 5 dimensions, we'd be adding the blue bars. They have the particular property, if you add them up they sum to 100%. If somebody were ticking that they were in the bottom level, in a very unfortunate life, for each of the 5 dimensions, we'd be adding the red bars and they sum to 0.

All the other potential states that are defined by the tick box answers, lie somewhere in between. But it's interesting that not all impairments are valued the same. Having no independence is the worst thing that could happen to you. Having all the love and friendship you want is the best thing, and indeed even the third level of attachment, having most of it, is still better than the top level of the other 4 dimensions. So these people are really bothered about relationships and they're really bothered about losing their independence.

Those 2 dimensions punch above their weight. They each account for around 25% of quality of life and the other three dimensions have been squashed as a result, because these have the properties that they must sum to 100%, and these are the relative preferences people have for these.

Moving from the 4th to the 3rd level doesn't really generally impose a big impairment on quality of life. The distance tends to be quite small. Security is the only one for which that is not true, and that's because it's qualitatively different. We're going from no worries, to some worries. So we were expecting a larger decrement there.

So, we can use these as off the shelf scores to apply a person's tickbox answers and summarise value - how good or bad their life is, as seen through the eyes of the British older person - a slightly strange concept but it does mean that we have a common denominator here. There's no potential for my rating scale answer to be different from yours due to a different way of answering it. We're applying the same set of values to people's measured responses to the 5 dimensions.

So how would we use this? Well we've used it in trials and surveys and the first clinical trial it was used in, was in a trial of joint replacement in Scotland. People had a hip or knee replacement, they filled in ICECAP-O before they had the operation, they filled it in again a year post-operatively and we applied these values to their tick-box answers. So we could essentially look at the percentage quality of life score pre-operatively and one year post-operatively. This was the first test of longitudinal data of the ICECAP-O measure and we were chuffed to see that despite a relatively modest sample size, we had significantly improvement, not only overall, in quality of life, but in the 4 dimensions that we expected there to be some improvement in quality of life.

There was one dimension we were not convinced we'd see any change in and we didn't. Anyone have any guesses which one it is? Shout out someone.

Role? Any other answers? Attachment, yes. There was no change in the attachment scores. We didn't really see any good reason why there should be. Unless anyone was widowed, for instance, we weren't expecting any major changes there, and indeed there wasn't. But there were improvements on all the other 4 dimensions.

We've also - we also administered ICECAP-O which although was developed for older people, initially, it's interesting that 4 of the 5 dimensions correspond quite closely to the fundamental human needs that psychologists talk about. We hadn't yet developed ICECAP-A for all adults and so we only had ICECAP-O to go on. But we administered it in the Bristol survey and it was quite interesting that we were able to then look at average quality of life scores in the 35 electoral wards of Bristol.

Now, here I've got a couple of geographical information systems maps and these give on the left hand side, the city of Bristol with its 35 wards and quality of life is coloured. The more yellow it is, the higher the average quality of life. The darker and redder it is, the lower the quality of life and it shows there are some pretty bad quality of life in the wards in the centre and the south. Quite affluent areas here in the west.

We were interested to see the extent of agreement with another measure, that's produced by the Office for National Statistics in the UK: the Index of Multiple Deprivation, which uses some magic formula and plonks in the unemployment rate, rates of hospital admissions, crime rate, and house prices and tries to get an overall picture of how deprived electoral wards are. We'd expect some agreement, but not necessarily strong agreement, given that we don't know if the values in the magic formula reflect the real weights that you or I might assign to how bad crime is compared to unemployment.

But, on the right, we've got the IMD - the Index of Multiple Deprivation scores, again colour coded dark means deprived ward, light is a less deprived ward. There was a fair degree of agreement, but there were some differences and indeed the beauty of the ICECAP instrument is that we were able to look at the scores on the 5 individual dimensions to see where these are coming from.

So there's a ward just here. That's Cotham I believe, near the university, didn't do quite so well in terms of quality of life, and certainly compared to its index of multiple deprivation score. What was pulling it down was people's answer to the security question.

There were a lot of worries there and it's probably because all those wards had a massive house price boom. Everyone piled in there, took on enormous mortgages to try to get in there and the mortgages were already at the time, beginning to, given the house price falls, were looking distinctly worrying and they've only got worse since.

We were able to see that, because we have individual level data which was the real beauty of this. Now this could be done in Australia, there's the SEIFA which I can't remember off-hand what it stands for, but it's similar to the index of multiple deprivation. So we could make comparisons in Australia if we wished to.

Now that's the UK for you. You're all thinking what about Australia?

Well, Pure Profile administered the ICECAP-O instrument last year. It was just before we had ICECAP-A, so we didn't have that to go on at the time.

But it was administered to between 5000 and 6000 Australians, randomly sampled from the Australian population. We over-sampled old people a little, because most web panels slightly under-represent them. We applied the UK scoring, which was all we had at the time, to people's responses.

So although it's the UK scoring and I'll come on to show that we didn't go too far wrong with that actually. We had this as a common denominator. We were able to then look at differences in quality of life across Australians in various groups.

Now the average score we generally see, in general population samples, is in the order of 78% to 83%. To make this a little bit more concrete for you, for what kinds of values we expect to see with common impairments, poor health was associated with an 11% impairment in quality of life - sleep quality, a further 11 %. Now you may was ask why we asked about sleep quality. Sleep quality is often seen as a good proxy for mental health, when you feel you can't ask about depression, sleep quality is often the good next best thing.

So physical and psychological health, both associated with fairly large impairments in quality of life among Australians. What about income? Well the income effect was actually surprisingly small. Improving the income from the bottom category, under $20,000 per annum, to the $40,000 to $65,000 band improved quality of life only by about 1%. To get another 1% you had to go way up to $120,000. Now this doesn't mean that rich people didn't have higher quality of life, they generally did. But these figures are after adjusting for all other factors. So if wealth has enabled you to live in a nice area, to keep your health good, then the actual effects of income per se are not necessarily going to be that large.

But things became to get interesting when we looked at a whole load of other factors, such as socio-demographics and people's attitudes. Family and social life, well the first thing we found was that having children seems to be good for your quality of life unless you live in Sydney. I don't know why this is but anecdotally, I keep getting people who've got kids saying the schools terrible here and I'd much rather be in Canberra and send my kids to school there. I don't know.

But this was quite marked and the improvement of having children was quite modest generally, but it was significant. But yes there was this quite large decrement associated with having kids in Sydney. So if you've got kids, go somewhere else. If you don't, Sydney's the place to be.

What about divorce? Well it hits Aussie men hard but Aussie women seem to like it. It's generally well known that relationships are protective to your quality of life. It tends to boost your quality of life by 3% to 4%. Divorce, you lose that and for some people you end up worse than being single, for men in particular.

I'm not going to say anything about Aussie stereotypes and macho men and everything, but it was very interesting the difference here. Where the women were concerned the lower boxes they were ticking for relationships were compensated by much higher answers to the independence question. They were ticking the top box and this was really quite marked, particularly compared to the UK.

There was something I'm still not entirely sure about that the - to use an old phrase - living in sin. Women don't seem to like that as opposed to being married and I'm not entirely certain this is a true effect. The phrasing of the relationships question in the pure profile panel doesn't match the ABS one and I'm doing a bit more investigation to see if this is really the case. But most of the benefit of being in a relationship seemed to occur through marriage rather than simply living with your partner, amongst women.

Middle age is the worse time of life. Early 20s and retirement were best. The difference between the trough and the peak is about 3.5%. So it's there; it's not massive, but it's there. The image of the gregarious Aussie is bounded in reality. Socialising does provide a big boost and we're talking 3% to 5% if you're socialising generally during the week as well as at the weekends. So for the single people who generally have the lower quality of life than their partnered peers, get out there and socialise; you can partially offset that.

There's something we call social empowerment. Now these questions I asked because they were asked in the Bristol survey and they try to get an insight into people's perceived mistake in democracy and how much engagement they feel with the local community.

So there were some statements made and people were asked to agree or disagree with them. There was something that seemed to emerge, the disenfranchised Aussie male, looking at the people who disagreed with the statement; "I can have an influence on decisions that affect my local area."

Now people who disagreed with that generally were lower quality of life, we expected that. But again there were as a gender difference. For a given category of disagreement, women tended to be less worse off than men, apart from in Sydney.

Men and women were equally worse off, in terms of the disenfranchisement. I wasn't sure what's going on here, is this something like metro-sexual man that the men here are less - their identity is less bound up in being the macho man who provides and has the influence. Well we didn't see this in Melbourne, so I discounted that.

Another possible explanation is that Sydney men have just revised downwards their expectations. They don't expect to make a difference here and - possible. I mean how many Premiers has New South Wales had in last couple of years? Maybe...

University education seemed not to give any benefit. Now again this is like income, to the extent that it enables people to get better jobs, it probably does have some benefit. But when you adjust for those, the benefit is pretty much zero.

What about lack of trust. The statement, "I can trust many or most people locally." Those who disagreed with that had about 5% lower quality of life and this was very marked in Sydney. It was more like 11%. So I don't know yet why that might be. People, who have views about the go-getting attitudes and greater markets in Sydney, may have something to say about that.

Okay that was using British scoring. What about the Australian scoring? I finally got this about five days ago and it doesn't look that different from the UK.

These are the values of the average Australian. Now of course the average Australian might not exist. But leaving that aside for the moment, the pattern across the 5 dimensions is really quite similar to the average British older person.

The only difference is that they seem to emphasize more and be more bothered about attachment. This is largely an age effect. These are Australians of any adult age. The British sample were people aged 65 and over. We tend to find that older people worry more about loss of independence, due to the effects that poor health are likely to have. So the fact that we have a younger sample here is largely responsible for this slightly larger emphasis on attachment. So having none of the level of friendship you want is the worse thing. It's not having no independence is the worst thing.

So applying these scores, these norms, to the tick box answers doesn't really change the picture that much. But is there such a thing as the average Aussie? Going back to another Monty Python film the best one, Life of Brian, when he's trying to get the crowds to think like individuals and say, "yes you're all different. Yes we're all different", and then the guy cries out, "I'm not".

He's our average person and there isn't really an average person it turns out. We've averaged over apples and oranges. There're actually three, at least three, Australian types. We're still collecting data here. The final story I'm sure there'll be more. But I think these are broad categories which I think are fairly robust.

There're young people looking for love. They have an extreme version of the last but one slide. There it's all about relationships. People in middle age are not so bothered about that. They seem to be more worried about security and control - independence but particularly security. So I think they're worried about not being the no-worries Aussie. Worries to them are particularly dreadful.

Then the old people are much more worried about independence. We saw this in the British data as well and there is also some suggestions of a marital effect here as well. But once you lose a partner for whatever reason, you're much more bothered about independence, partly probably because you've actually regained some independence and it assumes greater importance for you.

The one instance where that was not the case in the UK data were widowers. Widows were really bothered about loss of independence. Probably because there's no longer someone to look after them if they get poorly. If they lose their health they've got to go into residential care which they universally loath the idea of.

The widowers were much more like their married counterparts. They were still worried about being lonely, probably because British older people rely on their wives for the social networks - Doris talking over the hedge to Vera next door, sort of thing. But I wonder if there's an effect here too.

So what happens then when we use the scores, the norms, applicable to a more relevant group? We're no longer applying the values of some average Australian or average Brit, but we apply the scores that are applicable to our age group.

Well Sydney does even better if you're childless. Which was interesting and even worse if you don't have them; so the effect here is being magnified. Interestingly, being widowed isn't so bad - and I know why this is going on - because people who are widowed are typically older. Therefore they are more worried about loss of independence.

So losing a partner, although clearly was bad for them in terms of quality of life, it didn't have such an enormous weight that it had applied to it, when we used these average scores.

The income effect was very similar and the higher education effect again was non-existent. If anything the opposite and I wondered if there is a student debt effect being picked up here. I don't know; we need to look at that in more detail.

Poor health and poor sleep was even worse if you use our own values, probably because the groups who were most vulnerable to this are the older people. They're worried about loss of independence.

So people who are in poor health have got a lower score, using these new scores that have a very, very low value for the bottom categories of independence, with regard to the social impairment variables: no changes.

So where are we going with this then? Well I said we're still recruiting for the survey where we're finding out the types. So the choice experiment that we're doing; Pure Profile have got it in the field at the moment. We're still building up the sample sizes in order for us to get scores for both instruments.

I've talked entirely about ICECAP-O so far because it was the only instrument we had available last year. We've now got ICECAP- A and it's in field. But just eyeballing the results we have so far, I don't expect any major problems.

The concern we had was that ICECAP-O was developed for older people. Are we asking the right questions if we're administering it to young people? Well I think we're not going to go too far wrong. The conceptual dimensions are essentially the same. We've changed the wording of the security one and the role one but we're capturing the same sorts of concepts.

So we'd like to look more to see what are the Australian types? As I say, the suggestion from what I've seen so far is that, we can probably get a more nuanced view of these three Australian types. There are probably sub-groups within those that will give us a richer idea of what Australians value.

So far these seem to be related to age and marital status but they almost certainly are related to other things as well, most notably your own experience. What about lonely people? Are they really bothered about - is loneliness the worst thing that they can envisage? Maybe they try to compensate on one of the other dimensions and the idea of losing their independence is the worst thing to them.

We can test that because we've got people's tick-box answers and see if lonely people have got different values from people who are not lonely.

We want to administer both instruments in surveys and trials. ICECAP-O is already been administered in at least one survey in another unnamed Australian state. I'm not at liberty unfortunately, to divulge which at the moment. But we've got additional data from that, which has enabled us to kind of cross reference with the Pure Profile data. Longitudinal data is our real holy grail that we want here.

The values we've got from the middle age and older people may not be the same ones that I have when I get to that age. Seeing how people's quality of life changes over time will be important. Seeing how their values, their norms, change over time will be important, to see and to make predictions for the future about what people value, because if governments want to provide the services that provide the most benefit to people; and if companies what to provide their goods and services that people want most; we need to have a good idea of how these things change over time.

Ultimately what I'd like to have is for everyone to have a personalised scale. I've put up charts with average values. Ultimately, and this is the real draw for people in the survey - they tend not to be that keen answering choice experiments unless they're explained to them what the purpose of it is, they're inclined to say, these all look the same these questions. You have to stress to them that, no, this is why it's important - we're seeing what trade-offs you're making. How your answers are switching when we start switching the amounts.

If people do that, we can give everyone their own ultimately, personalised scale. All these impairments are on there. They could post it, I don't know, onto Facebook - make comparisons with their peers and friends. We could say who you're like, who you're unlike and relate it to people's own characteristics or not as the case may be.

I think that will be a real draw for people and these values; they come from the choice experiment. They represent frequencies, how often we've picked something. We know what the mathematical properties of these scores are.

Just to round off, with what else I and colleagues are doing in this area, one of my PHD students developed a measure for carers, for carers of older people in the UK.

Carers are increasingly being relied upon in industrialised countries with aging populations and they often have a pretty poor experience when they're having to be on call 24/7 for an aged parent or a spouse. The experience that they have is something which is neglected in many health economic evaluations.

So my PHD student developed an instrument for them. It's a 6 item instrument. We've got UK scoring for that, again coming from a bestworst choice model and it would be great to repeat that in Australia.

I'm involved with a study of getting scoring for a kid's quality of life measure in Adelaide and we're hoping to get funding from the NHMRC to do that. I'm involved with a big project in the UK to value social care related quality of life. That's quite similar to the work we've done with the ICECAP.

Some collaborators at the University of Western Australia are using best-worst methods to score a personality questionnaire, Schwartz's list of values, which I think will be quite useful in time, in helping us understand differences in the norms, in terms of quality of life amongst Australians.

I'd just like to thank you all for listening and I'd like to thank all my collaborators, particularly Jordan Louviere and Jo Coast who've led up the work in the quantitative and qualitative side; Pure Profile for making a lot of these surveys possible and for really wanting to take this to next level and various researchers who have helped me with the analysis and of putting it together and my project manager Gail and people at Bristol City Council; and thank you for listening.

End of transcript

Back to top