A Statistical Cat Fight

July 23, 2010 by Jonathan Rowson
Filed under: Citizenship, Education 

A quick post to alert readers to yesterday’s wonderful event about ‘The Spirit Level’ held in The Great Room.

Kate Pickett and Richard Wilkinson, authors of ‘The Spirit Level: Why More Equal Societies Always Do Better, were pitted against Peter Saunders, author of the Policy Exchange publication, ‘Beware False Prophets‘ and Christopher Snowdon, author of ‘The Spirit Level Delusion‘.

I have never seen anything quite like it.

The substance is hugely important. The question at stake is how much inequality matters, and the relative importance of increasing wealth or redistributing it more fairly. This is an old ethical question about distributive justice, which is the core turf of political theory, but ‘The Spirit Level’ attempts to cut through the ethical quandary with a statistical analysis of the world’s 50 wealthiest countries. The debate was therefore not so much about the importance of inequality, but about the uses and abuses of statistics to make political arguments.

And the style of the debate…. Matthew Taylor began by asking the speakers to seek the ‘transcendent moment’ in which the two sides avoid mud-slinging and assuming that the other side is incompetent or immoral, and instead try to find a way to agree what exactly it is that they disagree about.

That didn’t really happen. Instead we had what felt like a ’statistical cat fight’ with both sides accusing the other of missing points, statistical or otherwise.

(Image from National History Museum website)

You can listen to the audio, but I encourage you to wait for the video which will soon be available because there you can see the facial expressions on both sides, which reveal contempt, embarrassment, ridicule, anger and disdain. What I heard in the debate (not exact quotes) was something like:

“My regression line is better than you regression line”. “My outlier is more significant than your outlier”. “My research methods class is more worthy than your research method class”. “Just because I want to control for ethnicity doesn’t make me a racist”. “If this was a second year statistics assignment it would fail.” And so they went on, tearing into each other with technical details of statistical analysis coated in layers of moral invective.

I need to read all the relevant material to decide where I stand on the issue, but as I mentioned at the end of the debate, I want to believe in the thesis of The Spirit Level, for ideological reasons, but after this event I have serious doubts about whether the evidence stacks up.

  • Share/Bookmark

Related posts:

  1. Keeping up with the Joneses
  2. Inequality of Power
  3. Meso Soup
  4. First blog
  5. Global Consciousness?

Tags: , , ,

Comments

  • Bravo, Bros! keep going like this, more good info again.
  • Louise Thomas
    Good summary of the almost unbearably tense atmosphere in the Great Room as audience members supporting both sides held their breath hoping that their side's argument wouldn't be dealt a fatal blow! That's what I was doing anyway.

    Same as you, I don't have the statistical know how to understand fully the rights and wrongs with the way the statistics were presented.

    I did, however, draw two important points from the debate:

    1. That there is no 'absolute right answer' in statistics (although there may be absolutely held conventions about how to go about things) and that any regression line approximates reality in order that we make sense of it, and so is always a form of interpretation which requires judgement. Maths, like education, cannot compensate for society because our limited understanding needs shortcuts to deal with such huge complexity. It was telling that the apparent political allegiances on both sides matched the conclusions of the statistical methods each employed. And only you, Jonathan, of an audience of more than 200, changed your mind as a result of the debate. Makes you wonder whether statistics ever change anything, or merely reinforce existing opinions.

    2. The second thing I took from it was that the above point reinforces the arguments made by Peter Saunders that whether inequality matters is a moral and ethical argument that statistics have limited impact upon. If you strongly believe in equality as an ethical principle then a statistical analysis showing that unequal societies have lower murder rates is hardly going to convince you that inequality is a good thing, so we shouldn't expect it to work the other way. What the Spirit Level seemed to offer was a 'way out' for policy makers not, as Peter Saunders suggested, of making the ethical case, but of picking one side against the other, the poor against the rich, the working against the out of work. An analysis that showed that everyone would be better off in a less rich but more equal society is a fundamental challenge to the way the government goes about its business, and what the priorities, dilemmas and compromises that it faces are.

    What the publication of the Spirit Level did is to make the assumption that we need the rich to get richer so that we can all be better off became a bit harder to maintain so let's have more debate about that!
  • Jonathan Rowson
    Thanks Louise. I agree with both points, which overlap. The same issue repeats itself in neuroscience and education- the desire to by-pass ethical quandaries and seek warrant in the name of science.
    I hope to post again on the spirit level when the video goes up, but I agree the main point is that statistics tend to persuade rather than prove, and we are always more likely to pay selective attention to whichever statistics persuade us that what we already feel is right!
  • Steven
    Where is the video - please provide!

    Thanks
  • Stevenlongden
    Thanks

    Other than keep checking this page is there a way I can be nofified when the video is posted?

    Thanks
    Steven
  • Jonathan Rowson
    Hi Steven, Alas, I don't think so, but I know the video will be up later this week.
  • Hi Jonathan, when making your mind up don't forget to check out Kate Pickett and Richard Wilkinson's full written response to the critics - available on The Equality Trust website www.equalitytrust.org.uk from Monday!
  • ianduffy
    Sounds like as much fun as your excellent event "The Trouble With Physics" a year or so back. That went something like "it's only real physics if you can measure it" versus "we're operating on a much higher plane than you can imagine with your small brain".

    When is the next bunfight scheduled?
blog comments powered by Disqus

  • RSA Projects films



    View more RSA Projects films

  • RSA Projects

    RSA Projects put enlightened thinking to work in practical ways. We aim to discover and release untapped human potential for the common good. By researching, designing and testing new social models, we encourage a more inventive, resourceful and fulfilled society


    Find out more at www.thersa.org/projects

  • What we're thinking about

  • Recent posts

  • Comments

  • Archives

  • RSA Bookshop


    The RSA Bookshop, in partnership with Blackwell's the UK's leading academic bookseller, offers Fellows and friends an opportunity to purchase books and to explore our speakers’ ideas in greater detail. Visit the RSA Bookshop.

  • RSS Current RSA Events