Among those likely to be excluded are children in reception classes - four-year-olds just embarking on their school career.

Although the total number of excluded primary children remains small, the rise will heighten fears that behavioural problems are emerging earlier.

The figures, which also show that a disproportionate number of pupils from ethnic minorities are being excluded, will provide the first reliable picture of permanent exclusion rates for many years.

Previously, information on exclusion relied on a voluntary reporting system - the new data is based on a national census of all schools conducted in January. Overall, contrary to popular belief, the figures will not show a dramatic increase in total numbers of exclusions - the growth seen in previous years is stabilising at around 11,000.

The figures will also show that there is a huge variation between schools in the use of exclusion as a sanction.

Two surprising findings were the large number of exclusions from special schools and the fact that grant-maintained schools are excluding proportionately fewer pupils than LEA schools. The latter figure is likely to encourage the Government, though Labour will argue that it is not particularly surprising given that GM schools are allowed to be more choosy about their intake in the first place.

Presenting the trends at a conference of the Association of Educational Psychologists on Sunday, Peter Thorpe, head of discipline and attendance at the Department for Education and Employment, was keen to dispel the impression, fuelled by press coverage this week, that school discipline is collapsing.

“Individual cases are presented as a chaos symptomatic of all the nation’s schools . . . The thing that comes over most strongly from the data is the difference between the exclusion rates of different schools - schools which are nevertheless in very similar circumstances. Some schools are too trigger-happy. ”

Meanwhile, the Office for Standards in Education is also due to publish a report on exclusion in the next few weeks. The report, based on evidence from school returns, the OFSTED database and individual case studies, will examine different approaches to discipline in an attempt to establish what sanctions work best.

David Moore, the HMI who is writing the report, has confirmed many of the findings made by the DFEE, namely that there has been a rise in primary exclusions, although the numbers involved are small - 0.47 per cent of pupils permanently excluded.

OFSTED has also found that 20 per cent of all primary exclusions are of children in the reception class, suggesting either that children are becoming more disruptive at an earlier age, or that primaries are taking the opportunity to get rid of likely troublemakers as early as possible.

David Moore also found that 25.7 per cent of all exclusions are from minority groups and that two-thirds of those excluded are below average ability. Pupils most likely to be excluded have a reading age of between eight years five months and 10 years, suggesting a strong link between bad behaviour and poor literacy. The most vulnerable group consists of children who just escape being categorised as having special educational needs.

Children with SEN statements are relatively safe from exclusion, according to OFSTED, though this is directly contradicted by the DFEE figures, which, said Peter Thorpe, found a “large number of exclusions of children with SEN statements and from special schools - which certainly surprised me.”

He told the educational psychologists that the OFSTED report was likely to draw strong conclusions about the link between literacy and discipline, and it also drew attention to the large number of children in care being excluded (about two-thirds of all exclusions).