Philip Hensher: How odd - every year, exam results keep getting better

'We don't know if standards are improving. The measurements have been tampered with'

Imagine the following situation. A meteorological office reports a rise in the average summer temperature of half a degree on the previous summer. No one thinks anything of it, nor when the same thing is recorded the next year. But the meteorologists carry on reporting the same rise year after year, for 19 years. It doesn't feel any hotter, but the figures are clear. Faced with a steady improvement with no blips, there is only one course of action to take: go and see if something is wrong with the recording apparatus.

That, pretty well, is the situation in education today. It is completely plausible that educational standards can rise over time. If they did, we would expect to see an overall climb, interspersed with drops and troughs from time to time. A bad year would be outnumbered by the good years. But in no circumstances would we expect to see a straight line; that is, to claim that every year of school leavers is better educated than the year before. If this extraordinary situation were to arise, at the end of 19 years the extreme cleverness of this new generation would have become obvious to everybody.

Instead, after 19 years of improvements in A-level results year on year, we have a mass of anecdotal evidence that appears to show that school leavers are far more ignorant than 20 years ago. Numerous university departments are finding that knowledge they once could take for granted in their first-year students now has to be taught by rote. That remains anecdotal and impressionistic. The instrument of measurement continues to insist that standards are steadily improving. The truth is that we simply don't know whether educational standards are improving or not. The instruments of measurement have been tampered with.

I'm not much of a one for conspiracy theories, but it seems very clear from this point at the end of the history of A-levels, that educational policy has quietly been diverted by the Education Department. Most people, 20 years ago, thought that A-levels represented too high a level of specialisation too soon. It was a mistake for schoolchildren to do only three subjects after the age of 16, and something closer to the French baccalaureate would result in a better, wider education.

There is something to be said for this, although I do think that the old-style A-levels stretched pupils in more ultimately fruitful ways than the hurly-burly of AS-levels. What is deplorable, however, is the way the transition seems to have been carried out. Steadily lowering the demands of A-levels, if that is what is being done, has lead to a situation where the most able pupils will take five or six A-levels.

It has also, however, lead to a widespread doubt about the value of the final certificate in the outside world. Employers and universities can no longer take even an A grade at A-level as a guarantee of outstanding ability. It is quite unacceptable that the best universities are now having to chose between applicants with the same perfect scores, when on further investigation their abilities differ widely. Once, a candidate with four As was indisputably in possession of a first-class brain. Now above a certain level, everyone gets an A, and the certificate loses all value.

The point is that A-levels represent both a means of education and a means of measurement. Employers and universities need to be able to trust them, as weathermen trust temperature gauges. We need to know about the general standard of learning in schools, just as we need to know about literacy levels. To downgrade the demands of A-levels over time may produce a more rounded education in the end, but it also has the effect of removing any reliable means of answering this vitally important question.

The new AS-levels seem to have brought confusion and despair to schools, as any new system is apt to do, and the exams clearly need a great deal of rethinking. But there is one crucial issue which must be borne in mind, and that is the issue of examinations as a guarantee of quality. If we are to have examinations at all, they must be reliable as a statement of ability. To that end, concerted efforts must be made to ensure that an AS-level in 20 years' time means much the same as an AS-level now. Furthermore, there must be much more discrimination made at the top end of the marking scale. At present, 17 per cent of entrants get an A, and that is no use to man or beast. If, say, only 4 or 5 per cent of pupils got the top grade, that would clarify the situation enormously.

I don't necessarily want to return to A-levels, though I do think that they produced a higher level of education. But what anyone who cares about the duty to educate children will expect from all examinations is a measure of rigour. Without that, school leavers are left helplessly clutching pieces of paper of no value. It is irresponsibility of the highest order to pretend that things are getting better all the time. To do so only inculcates the belief that everything is going to the dogs.

hensherp@dircon.co.uk