A scheme to raise achievement in urban schools has had little effect on GCSE results, a report suggests.
The scheme is cost-effective in results for 14-year-olds, the report says
Excellence in Cities has been more effective at raising test results for 14-year-olds particularly in maths, the government-commissioned report says.
It said learning mentors, piloted under the scheme, had proven effective, and raised the performance of black boys.
The government said the research was based on out of date results and the scheme was now working.
The Excellence in Cities programme provides extra funding for challenging urban schools in England.
The report, commissioned by the government and carried out by the National Foundation for Educational Research, said that in 2003 there was "little evidence to suggest pupils in Excellence in Cities areas were making more progress at key stage 4 (up to 16 years old) than similar pupils in other areas".
It found that 14-year-olds raised their performance in maths as a result of the scheme, raising their attainment by between one and two percentage points.
The report concludes that the impact of the Excellence in Cities scheme is greater on the most disadvantaged schools and that mentors can help raise the self-esteem and confidence of pupils.
It says early mentoring of all pupils at Key Stage 2 can raise results later on.
The scheme also had a positive impact on attendance, the research found, with attendance in participating schools improving by one half day per pupil per year.
Teachers were generally positive about Excellence in Cities, saying it increased diversity, widened opportunity and promoted inclusion, the report said.
However, some teachers expressed concern that the scheme did not directly affect the majority of pupils.
Excellence in Cities - set up in 1999 - is designed to raise standards in urban schools through partnerships formed between several schools. Funding is allocated to each partnership, which can decide how the money is spent.
One third of secondary schools now participate in the initiative.
The scheme piloted learning mentors and learning support units for pupils facing barriers to learning, support for gifted and talented pupils, city learning centres providing computer equipment, and set up action zones to enable schools to find solutions to local problems.
The report looked at the performance of the pilot schools between July 2001 and July 2003.
Results at key stage 3 (14-years-old) improved where there was a high level of commitment to the partnership, it said.
Participating schools were given £120 per pupil.
But the report questions the effectiveness of the initiatives in all schools if rolled out more widely because it is not aimed at all pupils.
A spokesperson for the Department for Education and Skills said a look at GCSE results from the last two years showed that Excellence in Cities was working.
"In 2005, the rate of increase in GCSE performance for EiC areas is nearly twice that of other areas for the fourth year running - up three percentage points (compared with 1.7 points in non-EiC schools) on last year," he said.
He said the gap in those gaining five or more good GSCEs between EiC and other areas narrowed from 12.4 percentage points in 2001 to 7.6 percentage points in 2005.