A pound;250 million Government scheme to transform education in England's most deprived areas has failed to improve results for most pupils and has even had a negative effect on some, an official independent evaluation has found.
Big spending on classroom facilities and support workers has not helped the poorest children to achieve better exam results and has even hindered the brightest pupils and those who are white and British, according to the Oxford University study, which was commissioned by the Department for Communities and Local Government (DCLG).
The pound;2 billion New Deal for Communities (NDC) programme aims to regenerate whole areas. Since 2002, pound;249 million has been spent directly on education - but according to the in-depth study, there have been no "statistically significant improvements".
There have also been "negative changes" for Indian pupils, according to Kate Wilkinson and David McLennan at Oxford's Social Disadvantage Research Centre.
The money was spent on extra teachers and resources, extended schools, laptops and transition programmes to help children settle into secondary school. The level of investment in each area, which varied heavily, made no difference to the success of the programme, the research shows.
Only limited "sub-groups" of children have made any discernable progress at all. These include the very poorest, the worst-performing in Key Stage 2 and 3 and pupils from black, black Caribbean and Bangladeshi ethnic groups. Science results in some areas have also improved.
But the report says: "There was little evidence of a programme-wide improvement in attainment outcomes."
Not all children benefited from the extra support or services in each area and this might explain the programme's lack of success, Mr McLennan and Ms Wilkinson said. Pupils living in the areas also attend a large number of different schools.
"Both of these factors could potentially hinder effective implementation of interventions to improve educational attainment," the report said.
Researchers looked at KS3 and GCSE results between 2005 and 2007 to analyse whether they had risen as a result of the programme. Some areas had seen improvements - there was a rise in KS3 science and GCSE passes in Luton, Tower Hamlets in east London and Sunderland.
The programme was least successful in Wolverhampton and Doncaster, where the number of children getting the right level in English Sats declined by 7 per cent and 12 per cent respectively.
In Hartlepool numbers at the right level in English at KS3 declined by 4 per cent and KS4 performance fell by 14 per cent. But those who run the project in the town say they have had better results in primary schools.
"Improvements have been seen more quickly there, probably because there are greater proportions of children from NDC areas there so it makes it easier to focus resources," said Ian Worthy, evaluation manager.
Another evaluation of NDC by academics at the Centre for Regional Economic and Social Research at Sheffield Hallam University - released at the same time - said there is "no evidence that the presence of the NDC partnerships has made a decisive difference" and schools in other disadvantaged areas did just as well.
A spokeswoman for the DCLG said: "The report clearly states that the research can identify changes in educational attainment, but it does not tell us that these changes are owing to NDC activities."