The Dearing studies found mathematics and science to be relatively “difficult”. In English, by contrast, there has been a rise in the numbers and achievement of candidates at A-level in recent years, following a similar success at GCSE - now accepted as a genuine improvement in standards. More students have come through to A-level English through the wider reading and writing in the GCSE coursework schemes and are better prepared for the more advanced work. Coursework schemes at A-level have built on that success.

It has proved all too easy to misrepresent coursework as an unreliable soft option. What actually matters is the quality of work produced. Quality of work has hitherto been the sole criterion for the award of grades. The only official pressure on the boards has been to insist on just that - with the added stipulation that the same quality of work must qualify for the same grades from year to year.

This is enshrined in the mandatory codes of practice for both the GCSE and AAS-level examinations (presumably now to be re-written). It is also monitored by the School Curriculum and Assessment Authority and OFSTED and recognised by Dearing. If candidates produce better work under one scheme of assessment than another then so much the better - the principle being that candidates should best be able to show “what they know, understand and can do”.

This principle has been undermined on a political whim in the savage cut-back in coursework in GCSE English from 1994. The 5 per cent reduced entry against the general rise in GCSE entries overall that year barely managed to maintain the same grade percentages, and these slipped for the first time in the partially restored entry for 1995. It is the beleaguered 1994 GCSE cohort which first faces the similar coursework cutback in A-level English this summer.

So, how to make things worse at A-level - for that is what Dearing is proposing - and how to reduce the numbers, in the grades and overall? The same quality of work might be required, but with further hurdles to inhibit performance or recognition of performance. This could mean more demanding papers, less generous marking, and more unsuitable forms of assessment. Exams can, after all, be made quite pointlessly difficult for simply perverse reasons.

The best exam schemes are those which interfere least with the kind of performance to be assessed. The standards are what emerge at the end of the process, the judgements in the final awards about the kinds of work - however produced - qualifying for the grades.

Dearing concedes that generalised grade descriptions play little effective part in awarding, so if previous standards in observed quality work are to be abandoned, the only recourse must be to statistical norms. No doubt these will be based on statistical comparisons with performance in other subjects but, crucially, it will mean departing from the principle of criteria- based awarding on which the present system rests.

Dearing recognises that this is setting a dangerous precedent, and it would leave us with a system operating different awarding procedures in different subjects. What, then, about “standards” and comparability, especially within “English”, which comprises a range of 41 different linear and modular A-level and AS syllabuses in English Literature, English Language, English Language and Literature and even English? There is a strong case for reducing syllabus numbers, but that is to miss the point. Norm referencing is going to make a nonsense of any standards in English, let alone higher standards.

Dearing also touches on other issues affecting standards in A-level English. Modular schemes of teaching and assessment are no doubt attractive, but there are obvious problems of maturation and synoptic assessment in a subject like English where the syllabus is not so much linear as cumulative. Students bring their increasing experience of reading and writing to bear on all their reading and writing. The particular sequence of units of work perhaps hardly matters, but standards of achievement ought to improve on the performance in early units. Otherwise we are all wasting our time.

This all makes the Dearing plan for a reformulated AS exam, called the Advanced Subsidiary, particularly difficult for English. It would represent “the first half of the A-level syllabus” and be “graded in the same way”. That takes no account of the cumulative nature of the English syllabus. It looks like a core and options model: first-year AS core English, second-year A-level literaturelanguagemedia options. And what of one-year intensive courses in further education?

Keith Davidson, a former examination board senior officer, is a consultant in English and assessment. He chaired the national group that drafted the original GCSE national criteria for English and English Literature.