The trouble with SATs was not that they were not a good idea, but that they were two good ideas rolled into one: diagnosis and grading. And whereas teachers were concerned to discover what each child needed, the Government was more interested in accountability. Kenneth Baker, Secretary of State at the time, was content to fudge the issue to get some tests off the ground, but the incompatibility of the two aims could not be ducked, and following the 1993 boycott it took Sir Ron Dearing's supreme skills to sort out the mess.
You would have thought we had learnt our lesson. But we are in danger of repeating the mistake with five-year-olds. It is not as if the School Curriculum and Assessment Authority had not been warned. At an early stage in its soundings, as its consultation document notes, doubts were expressed as to whether it would be possible to diagnose and grade through the same instrument. The SCAA took the view that the two were "not mutually exclusive", and has put forward three assessment schemes - essentially checklists - to serve both purposes.
It is ironic that while the teacher unions fought against the grading function of national curriculum tests, they appear to be supporting baseline assessment on the grounds that it will pave the way for value-added measures of school performance. This seems doubly perverse. The present tests, which they opposed, are raising some very interesting questions.
The results published last week showed, for example, that only about half of 11-year-olds are reaching standards in English and maths thought reasonable for that age. What of the rest: are they ready for secondary education and, if not, what should be done? It clearly does no good to push up pupils who are not properly prepared, and we should not be surprised if there is disaffection, disruption and truancy.
The tests also reveal a very widespread performance - equivalent to nine chronological years at age 11 if the scale is to be believed. Is this acceptable, or should we be looking for ways of enabling all children to master the essentials?
But whatever the merits of these tests, grading the rising fives, especially for value-added purposes, is pernicious. If numbers are attached to children on entry, any school worth its salt will strive to enable the high scorers to score even higher to demonstrate that value has been added. This will only serve to exaggerate the differences between children still further, even though we already have the most heterogeneous classes in the developed world.
This does not mean that we should not be testing five-year-olds. On the contrary, as the SCAA booklet states, it is important to "identify the child's strengths and learning needs". But this is very different from slapping numbers on them. The SCAA checklists also seem to be about the wrong things, focusing as they do on prior attainment. What we need to know about children at the start of their schooling is where they have got to in concept development - being able to tell "before" from "after", "above" from "below", "in front of" from "behind".
Our early years education tends to take these fundamentals for granted, or fatalistically assume, following Piaget, that you can only wait. Other countries, such as Germany, Switzerland and Hungary, are more interventionist. Children in kindergarten have a daily lesson in how to learn. Songs and exercises, involving raising the hands above the head and below the knees are used to drive home basic concepts. What the children need to learn in common is regarded as more important than the differences between them, and the whole class is expected to reach certain standards. Several local authorities, notably Barking and Dagenham, have been trying to adapt the best of European experience to the British setting.
But we do not have to look abroad for our inspiration. We already have some excellent examples of good practice in this country, but in the health service rather than education itself. The language units in Brent and Westminster, for example, have well thought-out strategies for developing concepts, such as time, in children to whom they do not come easily. The approach is based on diagnostic testing, an intensive interventionist programme and follow-up testing to see what progress has been made.
The speech therapists often achieve remarkable success, and many of their charges are re-integrated into ordinary schooling. Adapting their strategies could be the key to solving the problem of the long tail of under-achievement that afflicts this country. If we can work miracles with the thought-disordered, why should it be so difficult to teach children who are not so disadvantaged to read, add and behave.
SCAA is due to announce the results of its consultation soon. Can we avoid yet another unworkable compromise? Diagnosis, yes, but subordinating children to the demands of "value-addedness" is measurement gone mad. The term, baseline assessment, does not bode well, but is it too much to hope that for once we have learnt from our mistakes?
Alan Smithers is professor of policy research and director of the centre for education and employment research at Brunel University