It has been noted many times that targets tend to provoke game-playing. Establish a departmental target for reducing the number of children excluded from school and, before long, schools are finding ways to make it happen.

Every year 15,000 children are suspended from school on five or more occasions. Temporary suspension means the school has it both ways - it meets the target for fewer exclusions but the disruptive children spend most of their time out of the classroom. A recent survey by The Times suggested that the number of primary pupils temporarily excluded rose 14 per cent in the past year to 27,000.

Yesterday Ofsted reported that some primary schools were excluding very young children because of their aggressive and sexual behaviour. It would be a hard heart that was too dismissive of the children concerned. The problems cited include an exposure to domestic violence, sexual abuse and a parent with a serious illness. One boy cited by Ofsted had seen his mother killed in a refugee camp. These are representative stories. Less than a tenth of the school population has a special need, but more than half of permanent exclusions. Being excluded from school vastly increases the likelihood of a child ending up in prison.

Still, a disruptive child ruins a class. The incentive for the school to exclude recalcitrant children is high. However, that simply raises the vexed question: where do they go? Ofsted found recently that only half of local authorities were meeting the target of alternative provision for excluded pupils. Pupil referral units (PRUs), which were set up to take children who had been expelled from school, are too often liable to fail themselves. Some PRUs are overwhelmed - the worst of them can offer only a couple of hours, once a week. Besides, there is an obvious flaw with taking all the least well-behaved children and putting them together. It is no great surprise that only 1 per cent of pupils in PRUs get five good grades at GCSE.

Fortunately, the Ofsted report does offer some useful counsel for what a concerned school might do. Poor behaviour is not a kind of sociological destiny, inevitable in poor areas. Ofsted found the social context of the school to be a poor predictor of the rate of exclusion. What schools do is more important than where they are. The schools that tended to have a high level of exclusions were almost always poorly led, with high staff turnover. They had always failed to intervene at the first sign of trouble. They also invariably had very poor relationships with parents. Where all of these things were improved, schools found it possible to keep children in school with the minimum of disruption.

But, even if schools were perfect, there would be some intractable issues needing attention at national level. Mental health support for young children who have suffered extreme trauma is all but non-existent but it is hard to imagine a successful exclusions policy without it. And there will always be children who resist even the best attempts at integration.

The issue of school exclusions still awaits innovative policies from new providers, paid according to the results they produce. The alternative is that rebellious children will carry on playing games in class and ingenious schools will play games with the targets.