Lies, dammed lies and politicians

The recent spat between the Conservative Shadow Home Secretary Christopher Grayling and Sir Michael Scholar, head of the Statistics Authority, seems to be a perfect example of what would happen if we were to spend too much time denigrating basic social science descriptive spadework. The 'facts' would be constructed by every Joe Six in bar room debate or, even worse , by politicians: enormous whoppers would routinely go unchallenged.
The latest sickening round of half-truths and misrepresentations is dissected on the BBC Home Editor Mark Easton's blog.

Users for sale

As everyone knows the Economic and Social Research Council (ESRC) makes a great thing out of research funding applicants demonstrating "user involvement" and "user benefit". I wonder if they are aware that in doing so they are creating a nice little earner for some folks. Yesterday a colleague told me the following story.
They are currently putting together a grant proposal in a fairly applied area. So, you would think, no problem to find users willing to say they are interested in the work. They approach a contact in a well known (partly publically funded) think-tank to talk about the proposed research, find out if it is the sort of thing the think-tank might find useful etc - basically something that could be done over lunch at my colleague's expense. Imagine their surprise when they got a response along the following lines: "Talking to me for 1 hour will be charged at £70 + VAT".
The justification was that the think-tank get a lot of requests of this sort and therefore need a mechanism to bring demand and supply into equilibrium. I guess from their standpoint the logic is impeccable, but surely this is not the kind of thing ESRC intended when they began to make user involvement one of the criteria to be taken into account in funding decisions. In some circumstances ESRC also demands input from user referees. I wonder if they should start to have a section on their grant application forms where the applicants have to declare the "consultancy" fees they have been obliged to pay to users and user referees have to declare any income they have received from the applicant. Can anyone think of any good reasons why such information should not be in the public domain? After all, if we are going to have a market let's have an efficient mechanism for disseminating the going rate. Anyone got anything to hide?

Bernstein comes to Harvard

Check this out http://www.youtube.com/watch?v=TTLbqQPJeYs it's quite amusing. Clearly these guys have too much time on their hands and you could write a whole dissertation on the sub-text underlying it (please don't). Despite my intellectual sympathies I think it might be more fun to hang out with the Quals.

In Praise of Description - more on the NEP report

Anyone reading the terms of reference for the National Equality Panel report http://www.equalities.gov.uk/national_equality_panel/publications.aspx
will see that they are limited to aims that are largely descriptive. Speaking personally this was a great relief which I imagine the Chairman shared. Getting 10 people to agree about the facts of the matter and the most appropriate way to present those facts is difficult, but with goodwill and a sense of proportion - which is what I experienced in all our meetings - doable in the time we had available. Given the existing evidence base, trying to move beyond description and say sensible things about causes and by implication how to produce change, would have required heavy reliance on assumption and speculation to fill in the large gaps in our empirical knowledge. Clearly it would have been absurd to attempt this in an official report. I'm reminded though of how often "mere description" is held up for contempt by those who think it is trivial and that what they prefer to do is intellectually deeper. Well, maybe it is, but the irony I find, is that those who take this view often rely entirely, and frequently naively, on the very "fact grubbers" they distain to provide the evidence to calibrate their speculations or furnish exemplars for their just-so stories. Recently rereading Stanley Lieberson's Making it Count: The Improvement of Social Research and Theory and now slightly less impressed by some of his arguments, I still think he basically gets it right about the value of description (and a number of other things). Forgive the long quotation but I think it is worth setting out in full (pp213):

"...one of the really valuable functions of empirical social research is a descriptive one. Social scientists who criticize activities simply because they are not theoretically driven are using warped and convoluted reasoning. In the stereotyped world of contemporary social science, there is a tendency to view such work not as a brick contributing to the construction of a great temple, but as a symbolic statement about one's disposition towards being a stark naked logical positivist. In point of fact, nobody is better prepared than contemporary social researchers to measure, describe and summarize such features of society. Is there something to be gained by deprecating such an activity? Is the society better off if such such questions are either left to speculation or addressed by less competent researchers? All too often, there is a propensity for those unsympathetic to empirical research to respond to solid empirical fact-finding not for what it is, but as if the researcher had implicitly said, 'Here is what all social scientists ought to do.'"

Tell me it isn't so...

Cold water from Caldwell - the NEP report

Last week the National Equality Panel published An Anatomy of Economic Inequality in the UK (AEIUK). It received a decent amount of press exposure and most of the straight reporting was fair as were most of the opinion pieces - where one might expect a bit more rough and tumble. The exception, I think, was Christopher Caldwell's piece in the weekend FT.

http://www.ft.com/cms/s/0/c5ef14de-0d0a-11df-a2dc-00144feabdc0.html?nclick_check=1

Anyone who has regularly read Caldwell's column will know that he expresses the sort of robust opinions that in a UK context appeal to the political right. I've actually always rathered enjoyed reading him; perhaps it is a version of the sort of frisson that the old left got from reading Hayek - you know you shouldn't encourage it but it is splendid astringent stuff. On this occasion I thought his piece was just otiose; a stock ideological response revealing a contempt for reason and evidence. At this point I have to reveal an interest: I was a member of the NEP and though my part in the production of the report was minor, when I read Caldwell's piece I did, for a moment, take it rather personally.
"One must read between the lines to discover the source of Ms Harman's alarm...It is oppression that Ms Harman's academics are looking for" opines Caldwell. One would have to look very hard indeed between the lines of the report to find the dreaded O word because it isn't there either literally or metaphorically. Neither is it true, as Caldwell asserts, that the report: "...takes 'inequality' as a synonym for 'prejudice and discrimination'". This is in fact precisely what we took some pains not to do except where there was particularly robust experimental evidence of differential treatment of, for instance, ethnic groups, that could not plausibly be accounted for in any other way (see for example pp 234-5 of AEIUK).
By the middle of the article Caldwell is clearly so splenetic that whatever tenuous grasp he had on reason completely deserts him. He apparently believes that part-time work is in some, unexplained, sense, less productive than full-time work and therefore quite rightly attracts a lower wage-rate. Here is his reasoning:
"If work is part-time, then either the demand for it is less pressing or the supply for it is less reliable. A milkman who delivers milk a few days a week on a flexible schedule is less valuable per delivery than one who delivers it regularly". Er, no Christopher, not unless there are frictional costs associated with flexibility or a payroll tax that is based on a headcount rather than the hours worked. As a consumer it makes no difference to me whether Monday's milk is delivered by John Doe and Tuesday's by Jane Doe and ceteris paribus in the absence of a flexibilty or tax penalty a part-time Jane should be as productive for her employer as a full-time John and therefore get the same pay-rate. It is of course an empirical question as to whether some sorts of work contain indivisible elements that defy flexibility and therefore attract a part-time penalty. However it is difficult to believe that delivering milk would be a good example of this.
Finally when it comes to social class Caldwell loses touch with reality altogether and smites us with what he clearly regards as a killer argument: "But the inequalities that exist are obviously not the programme of a self-conscious class". Indeed not, and nowhere do we say they are. Followed by: "The class problems that progressive governments make it their business to manage have mostly been solved. The problems that remain are problems of meritocracy, of which inequality is a natural result." Clearly Caldwell skipped the various bits of the report that show in detail how on average children from working-class backgrounds have to show more merit (in the IQ+effort) sense to achieve the same economic outcomes as those from more priviliged social origins. Since it is just a fact that this is the case I wonder in what sense Caldwell believes that we are in the fortunate position of having simply to deal with the problems created by meritocracy?