TESTING seven-year-olds is controversial. Nine out of 10 teachers say that they do not like it. The Assembly in Wales is to abolish it. And many middle-class parents worry that it is spoiling childhood. But not only should the Government in England continue with testing, it should make more use of the data it provides to help pupils to improve. Many of the concerns ignore crucial differences between the way in which younger and older children are tested.

The national curriculum expects children to be able to listen carefully and speak clearly by the age of seven; to know how to tell simple stories, write in sentences and spell familiar words. Children should know how to do easy sums, recite times tables, and understand basic shapes and how to tell the time. Most parents expect their seven-year-olds to be able to do all this. Yet those from middle-class backgrounds are more likely to have mastered such skills than those from poorer families with less home support. And unless such disadvantage is tackled early it will reinforce inequalities of opportunity later on.

This is why the terms of the debate are so worrying, since those who take such learning for granted in their own families often try to deny it to others. The literacy hour and the daily numeracy lesson are making sure more children get those opportunities. And since 1999, there have been improvements at key stage 1 of 3 per cent in reading, 6 per cent in writing and 11 per cent in maths, based on those achieving the expected level (2B or above).

But they need to be assessed objectively, which is where the KS1 tests are important. They are the only national measure of what children learnt at infant school. And with teacher assessments and inspections, they can identify which schools need to improve most.

Most objectors claim the tests place undue pressure on children. Yet most children enjoy being challenged. Where stress exists, it is usually transferred from parents or teachers. So long as the tests are sensibly integrated into lessons, there is no reason why children should not be challenged in this way. It could damage their development were they not so stimulated.

Yet Iam sure these are the pressures which persuaded Jane Davidson to scrap tests for seven-year-olds in Wales. Nevertheless, I fear she has made a big mistake which will be to the detriment of Welsh children, particularly those from poorer families.

The benefits of testing are significant. Teachers have much more information to help them address weaknesses early. The tests are vital to the autumn package of information which gives teachers much more detailed data on performance, including the chance to benchmark their results with other comparable schools. With individual pupil data and value-added information, they can address problems much earlier than before. The KS1 tests can be taken over several days and should form part of normal lessons. They last less than three hours. And they are not used to compile performance tables, though with value-added data, parents should certainly be able to see how their school compares with similar schools.

If there is a problem with the KS1 tests, it is the broad spectrum included in level 2. When national data are published each autumn, it is confusing. It would make sense to re-classify level 2 to cover only those achieving the level for their age.

But such technical change aside, tests, tables and targets are improving the education of millions of children. KS1 tests are a crucial part of that process. Abolishing them would profoundly damage future progress - and could reverse recent gains.

Conor Ryan was special adviser to David Blunkett at the DFES, 1997-2001