Education Viewpoint: Yesterday's tests that fail the needs of today's pupils

LAST year, John Major said that there was too much teacher assessment and its role in GCSE was to be reduced. At about the same time, Kenneth Clarke, then Secretary of State for Education, declared that new types of external tests being tried out for national assessment were unacceptable and we would have to go back to traditional short tests. I believe these changes will have a damaging effect on pupils' work in schools.

Why should school results rely so heavily on external tests, with the stress and luck elements they involve? In the world of employment, promotion depends on achievement in one's work - is there a company that ignores its own work records and uses a short written test instead?

Research indicates that the results of the short tests now being used for national assessment at age 14 could be very much a matter of chance. It is easy to see why. If a test were to rely on only one question, some able pupils might do badly because they misinterpret the point or overlook the particular topic involved. Several questions, varied in style and content, are needed. So how long a test is needed for a reliable picture? More importantly, are the tests being used for national assessment sufficiently reliable? Unfortunately, it seems there will be no research to check on the new tests to answer these questions. That leads me to the conclusion that parents and pupils should ignore the results and trust their teachers' judgements.

So why not use teachers' assessments? Perhaps the Government believes that teachers cannot be trusted to assess their pupils. If this were true, then they could not be trusted to teach them either, for you can hardly teach effectively if you cannot check frequently on what is being learnt. Perhaps it is believed that teachers cannot align their standards between schools. This is an important difficulty, and I believe that external tests are needed to help calibrate teachers' assessments, but not to replace them entirely.

However, external tests are also dangerous because they can distort teaching. This is why Mr Clarke's decision is harmful. Teachers are bound to train pupils to score well. For short tests designed to cover several topics, this means training them to answer collections of short questions that test knowledge and skills in isolation from one another. But research on learning shows that children taught like this do not retain information for very long, and cannot apply what they memorise because it did not make much sense to them in the first place. So tests should involve using knowledge and skills in relationship to some interesting task. Such 'tests' take longer, but they can reflect and encourage good teaching, as well as giving a fairer picture of what pupils can do.

This is what the tests that Mr Clarke rejected were designed to achieve. His decision, with Mr Major's, in effect reversed the policy decisions on national assessment which, following the advice of an expert group, Kenneth Baker set out in Parliament in 1988. There has been no systematic research into the new assessment methods or teachers' opinions of them. It appears that the Government has been guided by private advice from those who are not experts and whose opinions are based on nostalgia for old methods.

Earlier this year, the editor of the American Journal of Educational Measurement made this comment about US experience: 'If they are unaware of new research findings about how children learn, policymakers are apt to rely on their own implicit theories, which were most probably shaped by the theories that were current when they attended school. . . . If policymakers proceed to implement outmoded theories or tests based on old theories, they might actually subvert their intended goal - of providing a rigorous and high quality education for all students.'

This seems to me an uncannily accurate account of the decline of assessment in this country.

The author is Professor of Science Education at King's College London.