There is a useful short history on the Wikipedia site. A-levels were introduced in the early 1950s to replace the Higher School Certificate, which was a single qualification, like the modern international baccalaureate. The idea was that pupils were failing the HSC because of a single weak subject, so the A-levels provided a separate qualification for each subject. Those supporting the IB have come full circle – they want to force people to study a rounded set of subjects in order to do well.
Its main function was to help universities with their admissions process, so the grading was determined by setting the boundaries to give A 10%, B 15%, C 10%, D 15%, E 20%, O (a fail, but equivalent to an O-level) 20%, F (outright fail) 10%. However, there is an argument that this is also the most reliable method of establishing an absolute standard. This is a complex and disputed area.
Take first an area like maths. A maths question is typically “Prove X”. An answer can be (1) right, (2) part right, (3) wrong. There is little room for dispute about (1) or (3). Even the appropriate marking for (2) can be established without much difficulty. Typically the proof will involve several steps, some harder than others. If the question has a total mark of 10, and 4 steps, you might give marks of 1, 3, 5, 1 for the individual steps, so someone who got about half way through the third step might get 1 + 3 + 2 = 6/10. Sometimes there will be more than one possible proof, but it is still not hard to get a reliable and fair method of marking.
The difficulty is assessing how hard the question is. In practice, it is common to find that given two questions judged to be of similar difficulty by the setters, one will be correctly answered by 80% of candidates, and the other by only 20%. Evidently the initial judgment was wrong. The first question was much easier. This is not easy to overcome.
Note that the situation is quite different for factual questions like “What is the capital city of Bulgaria?”. Short questions about facts which are in the syllabus can simply be marked right or wrong and given equal weight. So the paper might have 100 and your score is simply the number you get correct. Equally, it is not hard to pitch the difficulty of tests that measure speed of basic thinking (“IQ” type tests) by a large number of questions.
But A-levels are intended to test more than rote learning and thinking on your feet, and this makes assessing the difficulty of the questions much harder.
For the liberal arts the difficulty is slightly different. Many questions are essay questions. Here there is no “right” or “wrong” answer and the “difficulty” of the question is not something which has much meaning. What matters is how it is marked. Marking essays well requires a good deal of skill. Are these comments about Henry VIII insightful? Well supported by the facts? Or are they just drivel? Is this a brilliant exposition of an unfashionable view, or complete nonsense? Is that sound? Or mere parrotting of a fashionable view?
Even harder is getting consistency across markers. Nonetheless, by an elaborate programme of prior briefing and test checking of the markers, reasonable consistency can be achieved. But that still tells you little about absolute standards. Setting the same question as last year and putting a fair number of last years’ scripts through this year’s markers would help, but clearly has other problems.
So I tend to end up siding with the view that the most reliable way of grading is to use the original A-level method of setting the boundaries to give 10% of pupils As etc. The snag, of course, is that then there is no way of telling whether as a whole schools are improving or worsening, all you can do is to track a school’s performance relative to its peers. That, of course, is useful information. Indeed it is the only information which is of practical use to parents. Information about whether schools as a whole are improving or worsening is only of interest to policy makers – deciding whether to increase investment or whether to change teaching methods or training.
Unfortunately, that is not the way it went. In 1987 the floodgates were opened with the introduction of a new system to try to fix criteria for grades B and E absolutely (so that in principle all entrants might reach grade B) with the rest by interpolation. From 2000 all grades were fixed “absolutely”. The snag about this was that everyone involved in the process from politicians down had a strong vested interest in declining standards (see an earlier blog entry). In any case declining standards is certainly what we got. The comparison between my math A-level papers (1965) and more recent papers is dramatic and far too large to be explained away by tougher marking.
Note, incidentally, that the policy of hugely expanding the number of universities would have brought about a decline in standards even under the original A-10% system. Although the correlation was clearly imperfect and some stupid people went to university in the 1950s, on the whole it was the more able people who went and who took A-levels to try to get there. As the numbers mushroomed far more less able people were taking A-levels, so the average ability of the A-level entrant fell. So the average ability of the top 10% would have fallen too.
For those who want to burrow more deeply into some of the complexities, and perhaps get some insight into the kind of people in the modern educational establishment in the UK, there is an interesting 500 page book available free online on the ofqual site. It was published in 2007 by Ofqual.
Post a Comment