Its revised “Criteria and Guidance for National Vocational Qualifications”, launched on January 24, claims to put more emphasis on knowledge and understanding, allow wider use of options, and improve assessment. All sensible moves in the general direction that the Government, the CBI and the Further Education Funding Council, among others, are wanting the NCVQ to go.

But will the revised criteria - many of which are a mere re-arrangement of the words - make much real difference? In the past the NCVQ has shown itself to be adept at deflecting attempts to get it to make real changes. A notable example was Kenneth Clarke’s requirement, when he was Education Secretary, for external tests and grading in General National Vocational Qualifications.

The NCVQ’s ducking and weaving then led to the odd form of those tests, and grading becoming a separate exercise in which the tests played no part. Sir Ron Dearing and the School Curriculum and Assessment Authority are still grappling with the legacy as they try to sort out assessment for GNVQ Part I - the new scheme to offer vocational courses for 14-year-olds.

The NCVQ’s resistance to change comes from an almost religious conviction in its own rightness. The basic method, as “Criteria and Guidance” tells us, depends on occupational lead bodies setting standards by progressive analysis. This is reflected in the structure of NVQs which are sequentially sub-divided into units, elements, and performance criteria - of which there can be several hundred.

Candidates for NVQs have to collect evidence that they can meet the performance criteria across specified ranges. Assessment is passfail without grading. Superficially attractive, this fragmented approach makes it difficult to set out clearly what learning is required and leads to unreliable and bureaucratic assessment.

The greater stress on “knowledge and understanding” that we have been promised will count for little if it comes out as itemised “knowledge specifications” rather than a coherent body of learning which will give adaptability and flexibility in a changing working world. Wave upon wave of assessors and verifiers, as envisaged, will not rescue an assessment system which is so loose that it can go off in any direction John Hillier, chief executive of the NCVQ, in his foreword to “Criteria and Guidance”, boasts that “the fundamentals” are unchanged, yet it is in those fundamentals that the problem lies.

It would not, however, take much to put them right. The NCVQ needs to recognise that job analysis is only the first stage of setting out a vocational qualification and not the point at which to freeze the process.

Having decided what a job consists of, it is then necessary to go beyond that to identify the essential knowledge, skills and understanding, and settle on how to assess if they have been achieved. The NCVQ is fond of citing the driving test as its inspiration, yet curiously it has adopted the driving-test model without the driving test.

In terms of the structure of NVQs, a more integrated approach would mean retaining the units, but moving beyond “evidence indicators” to ask awarding bodies to devise an appropriate mix of practical, oral and written tests, and workplace assessment, to enable levels of performance to be judged.

As Alison Wolf suggested on this page last week, there could be different modes for those already experienced in employment and school-leavers. The schematic range statements (the example in Criteria is “Materials: plastic, card, metal”) would be replaced by programmes of essential content.

Framing vocational qualifications in this way would be truer to the notion of “occupational competence” since there is no guarantee that aggregating numerous individual “competencies” will amount to skilled overall performance. Being able to dribble and do headers, for example, do not make a footballer. It is the way they are put together that matters.

Operating at the level of the unit would also allow for sensible assessment. Just as all the individually-assessed “statements of attainment” nearly sank the 1988 national curriculum, so the gathering of evidence in relation to hundreds of performance criteria is a severe strain on NVQs. The current emphasis on breaking vocational qualifications down into numerous little bits gets in the way of practicable, fair and valid assessment.

When I first heard that the new NVQ criteria represented a considerable revision, I was delighted because NVQs are not delivering all that was hoped of them. A comparison of the 1995 criteria with those of 1991, however, suggests that, in spite of a six-fold increase in the number of pages (through the addition of guidance), the changes are less significant than might have been supposed.

My fear is that the new criteria are really a holding operation in the face of increased pressure for change - as well as the scrutiny of the top hundred NVQs already announced, all NVQs will now have to come under review by 1996. When the results are available it would not be out of character for the NCVQ to say to the Government that, as several millions more students had now signed up for NVQs, it would be inappropriate to make any changes at all.

Meanwhile, the new criteria have left the door ajar. The lead bodies must take the chance to set standards which reflect skilled overall performance and the awarding bodies must devise assessment methods that match them.

Professor Alan Smithers is director of the Centre for Employment and Education Studies at Manchester University