Making A levels harder, which appears to be the government's policy, will depress demand and achievement.

This isn't necessarily a bad thing, unless new generations of youngsters, obliged to stay on in education until they are 18, decide to flock to the vocational qualifications that provide a much narrower base for progression to higher education.

Or unless you are a parent, in which case you may be secretly hoping that the reforms will come only after your youngsters have secured their places at university.

The consultation on A-level reform published this summer by Ofqual, the qualifications and examinations regulator, raises interesting questions about the purpose, content, assessment and grading of A levels.

³ÉÈËVRÊÓƵ

For example, does the fact that 34 per cent of candidates who study A-level chemistry achieve grades of A* or A mean that it is easier than psychology, with its mere 18 per cent at A* or A?

Standards and comparability are important. With such wide variations in subject performance, and with the Russell Group publishing its own list of "facilitating" A-level subjects, perhaps only root-and-branch reform will do.

³ÉÈËVRÊÓƵ

But as well as the usual learned squabbles about content in various subjects, reform requires the application of deep assessment expertise. A consultation that does not discuss the popularity of subjects such as psychology or media studies (both of which are ahead of physics, geography and all languages) isn't getting to the heart of assessment issues.

Stepping aside from the polarising arguments about "real" or "Mickey Mouse" subjects, is it acceptable that the awarding bodies apparently cannot balance demand, assessment and outcomes between the two keystone subjects, English and mathematics? Just what is it about the assessment of English at A level that means that 7 per cent of candidates achieve A*, where 18 per cent achieve purportedly the same standard in mathematics? Is it because the attrition rate between AS and A2 is 50 per cent higher in mathematics than in English, thus upping the concentration of higher performers? What is the correlation with prior GCSE performance? Or is the assessment methodology flawed?

Incidentally, the attrition rate between AS and A2 across all subjects is 28 per cent. The value of more than 300,000 AS qualifications held in each cohort without the matching A2 is not discussed in the consultation.

Making significant changes to high-volume, high-stakes qualifications without adequate development time and piloting carries risk. Memories of the dramatic fall in demand for mathematics at A level following the Curriculum 2000 changes, deemed to have made it more difficult, have faded in the wake of a recent recovery.

³ÉÈËVRÊÓƵ

There should be caution about recalibrating the A level, which, by dint of its ubiquity, has currency that is well understood by universities, employers, students and parents.

It is worth thinking about the A* as a glass ceiling constraining the most able. Is it time to move from A*-E grading to a number-based scale, say 1-10, with 10 the highest? As well as leaving the currency of the current grades intact, this has the advantage of creating a finer scale for selection for competitive higher education courses and a smaller "discount" for near-miss offers. Grades 9 and 10, for example, could be pitched above the current A* to allow headroom for more able candidates. Ten years from now, when everyone has got brighter and teachers have learned how to teach the new specifications, the grading could be shifted up to 3-12, for example, with no devaluation of grades held from previous cycles.

A levels were designed as subject qualifications, originally supporting progression to mainly academic degree courses. Taking three or four subjects in any number of uncontrolled combinations will never guarantee a broad and balanced curriculum in the way envisaged by the International Baccalaureate or the nearly-was Tomlinson diploma. If universities want a reliable assessment of higher-order cognitive skills as well as deep subject knowledge, they may have to look elsewhere; it is difficult for a single qualification to perform many purposes. The Extended Project Qualification seems quietly to be doing rather well in this domain. If it were compulsory, or at least an entitlement, it might start to feature in admissions requirements - although this in itself might spoil the strengths that the extended project currently derives from being a low-stakes, mind-broadening freedom from the prescribed national "curriculum" of A-level subjects.