Let Them Eat Shrubbery

by | Nov 19, 2008

Science is often at its most newsworthy when it proves the blatantly obvious: getting drunk makes you more likely to fall over; sunbathing is a risk factor for sunburn, etc. But science that ‘proves’ something that journalists, commentators and policy bods like to think was blatantly obvious can also command more than its fair of column inches. Last year, for example, we reported on how a paper published by the Royal Society had proved once and for all that The Great Global Warming Swindle really did get it wrong about the influence of the sun on global warming, just like the Royal Society had said it had. More recently, there was the news that human activities had been attributed directly to the rise in temperatures at the Earth’s poles:

there was not sufficient evidence to say this for sure about the Arctic and Antarctic.

Now that gap in research has been plugged, according to scientists who carried out a detailed analysis of temperature variations at both poles.

Their study indicates that humans have indeed contributed to warming in both regions.

Which presumably came as a surprise to Naomi Oreskes, Professor of History and Science Studies at the University of California, San Diego, according to whom the matter had been settled decades ago. In fact, the ‘fact’ that polar temperatures have been increasing faster than in the rest of the world was the prime ‘fact’ presented by Oreskes in her influential lecture, The American Denial of Global Warming, to justify her thesis that the whole issue of global warming and what we should do about it was done and dusted in the 1970s.

The latest example is to be found in the Independent, where Steve Connor reports on how his own personal prejudices about the restorative effects of green space are now supported by ‘the science’:

Proof at last: living near parks and woodland boosts health, regardless of social class.

Another whose assumptions are now vindicated scientifically is David Tibbatts from the charity GreenSpace:

“The study confirms what we have been saying for many years – parks are important for health and everyone should have access to high quality, beautiful and vibrant green spaces.”

The ‘proof’ is provided by a study published in last week’s issue of the Lancet (free registration required). It presents evidence that the health gap between rich and poor is narrower in areas where there is easy access to green space. Having made a good stab at controlling for the effects of socio-economic class – no easy task – the authors of the paper find that socio-economic health inequalities are nearly halved in those areas where greenery is most easily accessible compared to where it’s most sparse.

Of course, whatever the Independent might tell us, a single study rarely ‘proves’ anything, especially in such a noisy, contingent science as epidemiology. As co-author of the study Richard Mitchell of the University of Glasgow told us on the telephone:

It’s the first time anybody has done this, and inevitably it raises more questions than it answers.

Like how big the effect is compared to socio-economic class itself, for example? The answer to that one is far from clear. It’s certainly smaller, as Mitchell confirmed, but it’s hard to tell how much smaller. That’s because Mitchell and his collaborator Frank Popham (University of St Andrews) have used a narrow measure of health – mortality rate before retirement age – that does not compare directly with data on life-expectancy, which is the usual measure of health inequalities. Comparing like with like would require a separate study using different datasets. Life expectancy is approaching a decade shorter for those in the lowest social class compared to to the highest in the UK. Mitchell and Popham estimate that close proximity to green space means that an additional 1328 people live beyond retirement age per year. But it takes some imagination to expect that the latter makes much impact on the former.

And does it really follow that town planners just need to make more space for trees? Which is the take-home message of Connor’s piece:

Dr Mitchell, who is based at the university’s department of public health and health policy, said: “We would encourage the Government to consider carefully what their policy on green spaces is and to bear this research in mind when planning urban areas for the future.”

In these deterministic times – when we see ourselves as slaves to economies, climates and our evolutionary psychologies – these are questions that nobody feels the need to actually ask. Complex political questions are just too easy to ignore when simple mechanistic answers are handed out on plates.

It’s also striking that, although the paper concerned was published as part of a special issue on social inequalities in health, none of the other papers have received nearly as much attention. This paper is media-friendly precisely because it confirms the prejudices of a media machine that consistently fails to challenge the prevailing orthodoxy of environmentalism.

Neither do the press reports mention any of the caveats that Mitchell and Popham raise in their paper:

We undertook a highly powered population study with a simple approach, using robust health outcomes from reliable data sources. The study was hypothesis driven, and that hypothesis was based on findings from a large amount of research. However, the study did have several weaknesses. First, the measure of exposure to green environments was restricted. Although we knew the proportion of green space in the area of residence of people who had died, we had to assume that individuals living in areas with equal proportions of green space actually had equal access to that green space. Had appropriate data been available, we could have used a measure of distance to defined green spaces as a proxy for access, although we would still have had no data for whether populations living closer to a specified green space did actually access it to a greater extent. Furthermore, quality of green space could be a substantial determinant of use and activity within it, and we had no data for quality. No national dataset describing the quality of green space to which the population has access in England is available.

Second, our data were cross-sectional. We had no means of knowing the extent to which individuals had access to green environments throughout their life. Migration before death (eg, to access residential care) could have placed some people into a distinctly different environment from that in which their disease was acquired or developed. If such migration varied by income group, our results could be affected. Since we have no data for migration patterns, we were unable to quantify the effect of this factor.

Third, and perhaps most importantly, the measure of green space might be associated with other risk factors that we have not controlled for in our models. One of the difficulties of exploring the effect of physical environments on health is that access to good physical environments is strongly associated with the socioeconomic position of individuals. Residual confounding is therefore a threat to studies of this type.

Mitchell and Popham suggest that a combination of increased opportunities and motivation for physical activity plus a reduction in stress-related health problems might account for their results. This is supported, they say, by the observation that inequalities in early death due to any cause and due to circulatory diseases were reduced by close proximity to greenery while class differences in deaths due to lung cancer and suicide were not. The kind of stress that leads to circulatory disease is, according to Mitchell ‘a different sort of stress’ to that which makes smoking attractive or leads to self-harm. That maybe. But that’s certainly a claim that requires scrutiny from the scientific community.

When we spoke to him, Mitchell made it clear that planting trees is no substitute for wealth redistribution:

No government in the UK is ever going to get elected on the kind of really radical redistribution you would need to close the gap that way. The extent to which the gap shrinks in the greenest areas is more than has been achieved by current policy in the UK for quite a long time though, so this is a sizable effect. It’s really important not to overlook the redistribution of wealth. But, while we’re waiting for the revolution, it’s good news that environment can exert what seems to be a substantial influence.

We share his frustration, if not his defeatism. Wealth inequalities might have increased under the New Labour government, but that’s surely something to revile rather than run with. Mitchell makes a similar point in the press release issued by the University of Glasgow:

“Obviously, resources must still be ploughed into trying to narrow the inequality gap between rich and poor, and with that will come advances in the population’s general health.”

But Mitchell himself isn’t always so circumspect. In an article he wrote for the Sunday Herald about this research, socio-economics doesn’t get a mention. But he does sign off with a call to action:

So, the next time there is a planning permission argument in which a local green space is at risk, bear in mind that the green space is worth protecting because it is making a contribution to your health, and those of the people around you.

It reminds us of our conversation back in July with marine biologist Jason Hall-Spencer, who told us:

When you’re interviewed, you can give your own personal point of view, but when you publish a paper that goes up for rigorous peer review, then it’s got to have the caveats and everything else.

Well, yes, you can. But it doesn’t follow that you should.

In support of his case that policymakers should heed the message of his study, Mitchell also told us that:

the other thing is, so far as we know so far, green spaces don’t do any harm.

Which could read like a mission statement for environmentalism. As we tend to argue quite frequently on this site, environmental policies introduced in the absence of any better ideas can and do do harm. Green spaces and policies compete by definition for space and resources with amenities and policies that might address class inequalities per se. And when basic services such as refuse collection and sanitation are being reinvented under the banner of environmentalism, tree planting is a convenient and attractive option for cash-strapped town planners.

We feel a bit mean picking on Mitchell. The bulk of the blame certainly does not fall on his shoulders. We don’t doubt that he’s doing good science, or that his intentions are noble. But he has a career to look after like the rest of us. And as a scientist, looking after one’s career means constantly justifying one’s research in policy terms. Then there is the tyranny of the news peg – the reporting of newsworthy single studies in the absence of context or caution – to consider. And the tyranny of ‘relevant’ science. And of ‘evidence’-based policymaking. And of ‘science communication’ policies that require scientists to seek to popularise their work as part of the job description. Combined, the result is that any tin-pot political philosophy can be ‘proven’ scientifically, which does an injustice to both politics and science.

7 Comments

  1. geoff chambers

    The oddest quote in your article is surely this, from Dr Mitchell:
    “No government in the UK is ever going to get elected on the kind of really radical redistribution [of wealth] you would need to close the gap…”

    Why on earth should this be announced as a self-evident truth, given that there are many more poor voters than rich ones? A huge redistribution of wealth could be achieved very simply by going back to the income tax levels which existed under the government of that arch-egalitarian Edward Heath.
    But that would be politics, of course, which is far less deserving of discussion in a quality paper than a study which scientifically “proves” something, however trivial.
    There are many reasons why I’d prefer to be rich than poor, but living ten years longer isn’t one of them. I somehow feel that what I do between the ages of 80 and 90 is unlikely to be as much fun as what I’m doing now, however rich or poor I may be.

    Reply
  2. Janice A

    “The study confirms what we have been saying for many years – parks are important for health and everyone should have access to high quality, beautiful and vibrant green spaces.”

    This is obviously written by someone who does not live in certain parts of the Southwest USA. Out here we say that someone has been here too long when they can tell when the desert turns green in the Spring.

    The up side is, that we have a lot of emptiness between towns, and the desert does have a special beauty all its own.

    Reply
  3. Jim Clarke

    “Obviously, resources must still be ploughed into trying to narrow the inequality gap between rich and poor, and with that will come advances in the population’s general health.”

    I am afraid the West is doomed! When the educated among us can make such irrational statements and readers nod as if the lunacy was self evident…we are doomed!

    The inequality gap between the rich and the poor has absolutely nothing to do with the populations general health! The populations general health is determined, obviously, by the wealth of all of its citizens, relative to poverty! For example, if you take the wealth of all citizens and reduce it to the poorest levels, than all citizens will be equally poor (no gap at all) and the populations general health will blummet. Even the health of the poorest will diminish as they no longer benefit from having wealthier citizens supporting the infrastructure of society.

    Furthermore, if one doubles the wealth of the lower and middle class, the populations general health will improve dramatically, even if you quadruple the wealth of the upper class, making the ‘gap between rich and poor’ much larger than ever!

    The goal should be to move citizens away from poverty. The ‘gap’ is irrelevant…except to the media and college professors!

    We live in a post-rational world where the things that have become obvious to so many, are so obviously wrong!

    Reply
  4. Derek D

    The fundamental flaw in all of this is mixing the terms “study” and “science”, both directly and contextually.

    Taking a dataset and statistically manipulating it to draw conclusions is nearly on par with burning books. Folks. THIS IS NOT SCIENCE IN ANY SHAPE OR FORM. It proves 0.00000000%. I include the decimal places because that is the degree of certainty that I have that all of the above “conclusions” are totally baseless. And while they may or may not be true, statistical manipulation does not increase the measure of proof they provide by even the furthest decimal percent.

    Statistics work when a dataset is carefully crafted. When diligent effort goes into WHAT data will be captured, what it will represent, and the relative quality of any given piece of data. Then that data must be carefully gathered through valid scientific method. If these two requirements are satisfied, THEN the statistical manipulations can be valid, if earnestly and honestly done. But somewhere along the lines, people lost sight of this. Whether it is temperature data, CDC data or whatever, someone just decided one day to just gather random groupings of data, make their manipulations and draw conclusions. In the world of medicine, climate study, and epidemiology, this has become an epidemic, and in some cases has been used to force huge financial burdens on us. And all of these statistical trappings are completely WORTHLESS. Furthermore beyond being meaningless, they require more dependence on the statisticians ability to make assumptions to fill in the many gaps that come with the many unknowns.

    The studying how simply living near a park affects different demographics, involves so many degrees of freedom, no statistical treatment could ever draw remotely accurate conclusions, if a supercomputer even existed that could to the calculation in the first place.

    It all lies and politics. The laws of physics require no statistical treatment. The laws of chemistry require no statistical treatment. The laws of thermodynamics require no statistical treatment. String theory requires quite a bit, and not surprisingly is neither well received, nor progressing at any notable rate. In fact the quality of any conclusion is usually proportional to the amount of statistical treatment it took to arrive at it.

    The more people are fooled into believing that statistics represent any true scientific truth, the more our hopes for civil and intellectual enlightenment sink further and further into darkness.

    Reply
  5. Kunoichi

    using robust health outcomes from reliable data sources.

    In other words, this “study” was nothing more than a data dredge.

    In other words, they could make it reach whatever conclusion they wanted it to, since they never actually went and looked for themselves.

    Handy, that.

    Reply
  6. geoff chambers

    Derek D and Kunoichi are being unfair on Dr Mitchell. Data dredging conducted by an artist like Durkheim can tell you a lot. Dr Mitchell’s study maybe isn’t in that league, but he’s obviously trying to help. He seems to be arguing: “I can’t persuade the authorities to abolish poverty, maybe I can persuade them to plant some flowerbeds if I can prove it makes people feel better, live longer, even stay alive long enough to vote the authorities back in.” What’s wrong with that?

    Our Editors’ objection is that it’s a distraction from the Big Issue (poverty, injustice, whatever), and so it is. So is almost everything. Some distractions, like planting trees, are harmless; others, like destroying our energy infrastructure in pursuit of a scientific illusion, are suicidal. Given that the world’s scientific and political establishment are not going to yield to a frontal assault from Climate Resistance and company, maybe a bit of harmless distraction is in order. My own data dredging in the environment pages of the British press suggests a move away from saving the planet towards saving butterflies, encouraging bicycle paths, etc. There must be environment editors who can read a temperature graph, but, as you say of scientists, they have to look after their careers, too. Maybe with a little discrete prodding they can give up their computer models predicting disaster and go back to promoting safety ladders for hedgehogs. That’s where the readers’ hearts, and votes, are.

    Reply
  7. Editors

    Kunoichi and Derek D,
    No, this wasn’t a data dredge. They hypothesised a priori that deaths due to circulatory diseases would be more sensitive to proximity to green space and that lung cancer and self-harm would not, and their analysis supported their hypothesis. We are less interested in the quality of the science on offer, and more in the process by which a single, tentative study becomes headline news and transmutes into established fact along the way. Our post is not about bad science, but about the mis-use of perfectly fine science.

    Geoff,
    Our main complaint is not that this research distracts from the big issue. It does, of course, but that’s only because it’s so attractive to those charged with doing something about big issues when there is so little faith in the political system’s ability to do anything about them. The real problem is that this is science being used *instead* of politics. That criticism can be levelled at all sorts of science – when science is expected to be relevant, it’s a difficult trap for scientists to escape – but this study is worth highlighting because its very take-home message is that science can do what politics cannot.

    Reply

Submit a Comment

Your email address will not be published.