This raises the question of why it’s always “grade inflation” we hear about and never “grade growth” or “educational productivity improvement”. A while ago I responded to this Chris Cook piece about The One Where They Decided To Downgrade Everyone To Match Their Class Except For Subjects Only Rich Kids Do – remember that? it’s as if it had been buried under another pile of monstrosities – and I pointed out that in the circumstances of 2020, it doesn’t make any sense to think of norm-referencing exams as a way of preventing grade inflation.
Let’s work through the logic. This cohort of students have missed a lot of school, we believe schooling is good and exams measure it, and therefore by construction we expect this cohort to do worse on the exams. Imposing norm-referencing – grading on a curve, in American – is not necessarily going to help here. If you draw a curve and decide that the top x per cent of students will get an A, you’re essentially promising x per cent of A grades, no matter how good or bad the students are. If you set this target to last year’s percentage of As, assuming the students are in fact worse, you have done the exact opposite of preventing grade inflation – you’ve baked it in. An A is achievable with worse exam performance than last year. That’s the definition of grade inflation.
This leads us to the counterintuitive conclusion that the government was trying to find a way of accepting the grade inflation without looking like that was what they were doing, and perhaps reserving its benefits to the right kind of people, but they catastrophically screwed up…but honestly, looking at this lot, how counterintuitive is that?
That led me to wonder why, given that they’ve been whining about grade inflation literally all my life, nobody has seriously tried to go back to norm-referenced exams.
I conclude that there are two conflicting motives at work – first, there’s the classic conservative motivation to impose a zero-sum selection process. In this view, there are only so many credentials to go around, they derive their value from their scarcity, and the role of government is to maintain this scarcity so as to preserve hierarchy. Obviously, if you believe this, you’re going to be very worried about standards, inflation, and the like. If you’re honest with yourself, though, you must also be opposed to any actual improvement in education, not just to illusory progress through generous marking, as to grow is to change.
On the other hand, not many conservatives are seriously willing to deny the possibility of improvement, and especially not on the campaign trail, and a lot of them are strong believers in the Protestant work ethic.
This reminds me of something I read in, I think, Theodore Zeldin – a speech by Adolphe Thiers arguing for fiscal austerity in terms that admitted he was more worried about the potentially progressive social effects of growth than inflation as such. This completely transformed my understanding of classical and Austrian economics, and I should probably write about this, as it shapes how I think of the relationship between modern libertarians and neo-reactionaries.
If you can’t get the Prussian army to murder anyone who disagrees, though, this level of brutal candour is not career-positive. It’s therefore necessary to argue that the austerity is eventually good for growth. In the educational sense, this implies promising the possibility of self-advancement, boasting of improved exam results, but also insisting on the restricted supply of credentials. Also, if it’s possible for the exam results to get better, the secretary of state for education can use this metric to impose managerial authority on the teachers. The three factors, taken together, imply that criterion referencing is here to stay, but that conservatives will never be able to resist the temptation to shade the numbers – and that is precisely the nonsystem Michael Gove created, and that blew up so badly this summer.
nobody has seriously tried to go back to norm-referenced exams.
I don’t know what goes on in secondary education, but this would have an odd ring to anyone who had worked in a university department where quotas for both passes and ‘good’ marks (2.i+) were openly discussed and applied at the level of individual subjects, with remedial action being called for from any unfortunate academic whose students are seen to underperform. To anyone in that position, criterion referencing would sound like a very good idea.
I was going to write a long reply about why grading to a curve/norm-referencing was a good solution to the faulty black box of assessment – that it’s a very good way of making sure your outputs match your inputs, thus correcting for a lot of very bad (by which I mean, far too hard) assessments that teachers (at all levels) typically make.
But then i read the linked piece and realised that your critique was correct – grading to a curve as a solution to the asterix year that is COVID lockdown? Is he insane? Or just pushing an agenda?
If ever there was a year when the assumption of a normal distribution of student outcomes was invalid, it is this year, with the very-far-from-normal chaos of the lockdown. Grading to a curve requires an assumption of normality – it’s one of the biggest knocks against norm-referencing; one of the biggest strengths of rubric assessment is it requires no such normal distribution. 2020 is an outlier year: grading is going to be messed up this year, and grading to a curve cannot fix that.
+1 for the inherent tension between Conservatism and election-friendly endorsement of narratives of improvement.
I think there’s also an underdiscussed angle that there’s not much reason to believe that year by year every cohort will average out as being as smart/dumb as any other. This curve business is a statement of belief that this is the case, but AFAICT the evidence doesn’t match up, when you dig around for what few markers there are that might point to this.
Finally, I’ll just toss in that when you start looking, really looking, is an A* in History the same thing as an A* in Physics? Really not clear that it is (and you can make cases for superiority in both directions, so it’s not about that) but this gets you into real trouble once you look at the Chris Cook stuff and say “what is the statistical basis for norming in subjects with small uptake” and of course, the wider point about scarcity and norming and Conservative values.