Friday, 22 May 2015

The facts really do matter and aren't difficult to find out

I've just been listening to some very interesting podcasts from the LSE of the day long engagement with the work of Thomas Piketty. There are great talks by Bob Rowthorn, Tony Atkinson and David Soskice all of which are genuinely interesting and insightful. 

There are also some talks which can best be described as utterly confused nonsense. One Professor of Sociology, who shall be nameless, makes the bold and shameless assertion that, and I quote: "...your occupation is a less powerful predictor of your income than it used to be." The authority for this statement seems to be a paper by two of his colleagues Friedman and Laurison, who, as far as I can see, make no such claim. In any case the paper which he appears to be referring to is about earnings not income. 

Let's make the charitable assumption that what he meant was earnings rather than income. In fact the distinction probably doesn't matter too much in this case because earnings from employment is such a large proportion of most people's income. 

The fact of the matter is that the truth is quite the opposite of what the good professor so confidently asserted. In Great Britain no matter what  level of aggregation  you look at, between  1976 and 2008 the proportion of variation in log earnings accounted for by occupation  increased. This is comprehensively demonstrated in a paper written by my former doctoral student Mark Williams who is now at the University of Surrey. 

There is no room for doubt about it; in as far as we can determine them this is what the facts of the matter are. The paper appears in a mainstream journal not in some esoteric place so it beggars belief that someone who claims to be an expert on inequality could a) fail to have read it and b) so confidently assert the complete opposite of the truth.

We expect newspapers to publish corrections when they get the facts wrong or simply make things up. Shouldn't we require the same of academics who either knowingly or through ignorance make false pronouncements in a public arena?

Or do LSE Professors simply not see it as part of their role to get the facts right?

Tuesday, 19 May 2015

Anybody out there?

I guess I'm as intellectually lazy as the next fellow. From time to time you get into conversation with somebody from another discipline and  you can see they are trying to figure out what kind of beast you are. Often on these occasions, if the conversation turns to my department, I resort to a rather crude  heuristic and explain that in the UK there are departments like mine and then there are the other lot, the "body sociology" departments and usually go on to say that we don't have much to say to each other. This strategy works, at least in the sense that my conversation partner normally either nods knowingly or just walks off.

Sometimes I've wondered what I would do if I was called out and asked to give some actual examples of "body sociology". I can't say it has given me sleepless nights, but there was a bit of a chink in my armour there. Well now I can relax, because the esteemed British Sociological Association flagship journal Sociology has come to the rescue by publishing just the thing I need.  And there is a bonus because, as I learn, the sociology of bodies is now firmly connected to the sociology of names and identities. Science just grows and grows every day! You can read the abstract here. The rest is pay walled but I think the abstract tells you  all you need to know.

If you feel gloomy after that, here's Satchmo to cheer you up with, yes, you guessed it: I ain't got nobody.

Monday, 11 May 2015

Pollster that got it right

Obviously there is a lot of soul searching going on in the polling industry right now given what  looks like a rather serious failure to get close enough to the target that mattered - the percentage voting for the  two largest parties. I've no more insight into what went on than anyone else. The industry has set up an inquiry to be chaired by Professor Pat Sturgis. He's a good man for the job and that gives me some confidence that  if the answers can be found they will be. We'll all await the verdict with interest.

Meanwhile, here's some food for thought. Damian Lyons Lowe CEO of Survation revealed that they conducted a  poll on the 6th of May which was very close to the actual result. However, they felt it was so out of line with all the other polls that they didn't offer it to the Daily Mirror.

A couple of things strike me about this, one trivial, the other possibly more significant. The trivial thing is that clearly I don't entirely understand the polling industry's business model. I assumed that pollsters carry out polls for clients ie nobody asks any questions unless someone is paying.  One interpretation of  Lyon Lowe's account is that  I'm wrong about this: "We had flagged that we were conducting this poll to the Daily Mirror as something we might share..." suggests that the Mirror was not formally the client for the poll.

Much more important is the implication that the polls, or at least what we the punters eventually see of them, are not independent probes of the underlying reality. In this case we have a poll judged too far out of line to enter into the stock of publicly available data and thus the precision of what we actually get to see is over estimated.

I wonder - and I hasten to add I don't know the answer to this question - how far this is just the tip of the iceberg? In polling what is observed is not what you get, by which I mean the raw data has to be weighted to produce estimates that, hopefully, approximate what you would have got had you been able to sample from the target population. Weighting is unavoidable, but inevitably it involves judgments about things that, by definition, are unknown. Now here is my thought. I wonder about the extent to which pollsters are looking over their shoulders at each other when they make weighting decisions? Polling is a commercial activity. If you are certain you are right being an outlier might be good for business. If not, in a crowded market place where reputation really matters, staying near to the average might look like an attractive strategy.

Friday, 8 May 2015

Election post mortem

So, it looks as though the Tories will get a majority or something close enough. Not an outcome that gives me much pleasure, but there you are. 

Personally I'm sad for Vince Cable in Twickenham. He has been a spectacularly good constituency MP and is a decent man. I've voted for him in three general elections even though I don't support his party. His party leader led him in the wrong direction & what could he do but follow? The gallant 57 galloping through the Valley of Death have been reduced to single figures.

Chuffed though for my colleague Stephen Fisher who along with John Curtis, Jon Mellon & Rob Ford (all ex-Nuffield)  were responsible for the BBC exit poll. I can just see them having a whip-round to buy the hat that Paddy Ashdown is going to eat. Great job boys, Bayesian models rock for prediction!

Wednesday, 6 May 2015

Election Special

So a day to go and everything as clear as mud. There are two must reads today, Simon Wren-Lewis on the ludicrous message the Independent is sending to its (thankfully limited) readers on the legitimacy of SNP participation in UK government and Stephen Fisher on coalition politics.

I voted (tactically) a couple of weeks ago by post and since then I've not gone out of my way to follow the campaign. I make no secret of the fact that I'd prefer a different government to the one we currently have without being overwhelmingly enthusiastic about the alternative. 

Looking at the forecasts, the bookies odds and the electoral arithmetic my best guess (which I don't make with much confidence) is that I'm going to be disappointed and that the Conservatives, after winning the most seats, together with the rump of the Lib-Dems will patch up a  deal with the DUP and blunder on to the next crisis on the horizon - the EU referendum. How the Conservative Party will look after that is anyone's guess.

Given the abysmally low expectations the boy wonders Milliband and Balls fought a  largely gaffe free campaign (everyone seems to have forgotten Ed B's memory loss about the names of actual Labour Party business donors). But their report card says they could have done a lot better. 

They were not nearly aggressive enough in confronting the media-macro fantasy that the last Labour Government was responsible for global recession. I know the Great British Public have difficulty focusing on any argument that requires more than two sentences but was it so difficult to get the message across that Gordon Brown was not responsible for the crash of Lehman Brothers?

They have saddled themselves with a daft promise of no deals with the SNP a party which arguably advocates more of the things that traditional Labour Party supporters want than the Labour Party itself. Again they should have been more aggressive with a hostile media.

They have also saddled themselves with a policy of reducing university tuition fees. It is incomprehensible to me how a party, supposedly of the left, could be advocating that. 

Instead of offering a bribe to the electorate why not  explain the facts of the matter? The arithmetic is not that difficult. Unless you give enormous weight to the impact on mature students all the numbers say that this policy amounts to redistribution towards the (future) middle-class. Maybe that's what the Labour Party wants. But if it does, it should have the courage to say so clearly.

Monday, 13 April 2015

In which I break a promise...

A while ago I promised, more or less, not to bring up the subject of the Great British Class Survey in these precincts again. So let's just call this a contribution to information diffusion. 

The data are now deposited in the UK Data Archive and you can read the documentation here. And very interesting reading it makes too. If you have a stomach for it you can also download the data. This is where it gets really interesting. 

Remember the GfK molehill that supported the GBCS mountain? If you want to know how solidly attached to the ground that molehill  is you really need to look at the marginal distributions of household income and wealth. There is something primitively satisfying about having one's intuitions confirmed. In Sunday school we used to sing a chorus about the foolish man who built his house upon the sand...

The terms of the agreement you have sign to get the data are themselves interesting. I've never seen anything like it. Here they are in full:

BBC Great British Class Survey, 2011-2013
Additional Agreement on Conditions of Use
The BBC Data Collection shall not be used in a manner which:
  • distorts the original meaning of the BBC Data Collection, for example by changing the context;
  • discriminates against any specific social group or otherwise exploit vulnerable sections of society;
  • promotes, encourages or facilitates violence;
  • promotes, encourages or facilitates illegal activity;
  • promotes, encourages or facilitates terrorism or other activities which risk national security;
  • promotes the tobacco, armaments, alcohol or pornographic industries;
  • encourages hatred on grounds of race, religion, gender, disability, age or sexual orientation;
  • promotes, encourages or facilitates anti-social behaviour;
  • might be perceived as damaging the BBC's reputation for accuracy or impartiality; or
  • otherwise brings the BBC into disrepute.

Can anyone tell me what the first of these means? And the last looks pretty scary too. Who knows, in the eyes of some corporate lawyer or other, just by writing this I may already have broken the conditions of use. I could go on about the dangers of public bodies confounding their commercial interests with their public service duties, but I've probably said enough already. 

I'll let you know if a solicitor's envelope arrives on my doorstep.

Friday, 20 March 2015

Keeping public policy makers in the dark

Vikki Boliver alerted me to a very important report released by the Social Mobility and Child Poverty Commission on the issue of administrative data sharing. It covers, among other things, the damage done by  UCAS's decision to discontinue making  anonymized micro-data available (for a fee) to bona-fide researchers. The consequence of this particular piece of bone-headed lack of civic virtue is that the evaluation of  public policy goals with regard to equality of access to tertiary level education has been rendered all but impossible. 

Nice one UCAS, of course we trust you to always do the right thing even when it conflicts with your commercial interests...Then again a government that actually cared about such matters could always legislate to remove any potential conflict of interest.

But let's not beat up exclusively on UCAS. The issue is wider than just the use of administrative data as Stephen McKay & I point out in comments on the SMCPC web-site. One arm of government is charged with promoting social mobility, while another constructs an obstacle course that makes it difficult for the interested citizen to find out what the facts of the matter are. You couldn't make it up. On second thoughts, you may as well because there is hardly any way to find out if you did.