A charity’s claim that poor pupils are being let down in schools with low GCSE results has been called into question by both a leading educationalist, who described it as “unthinking”, and fresh analysis of last week’s results.

Stephen Gorard, professor of education research at Birmingham University, said the pressure group End Child Poverty, a coalition of more than 130 organisations including Save the Childen, should have much more evidence before making such a sweeping statement.

TES analysis has also revealed holes in the lobby group’s conclusions, which fail to take into account ethnicity as a factor in success.

Last week, the group attracted widespread publicity after compiling a league table of the rates of GCSE success of children eligible for free school meals in different local authorities. This purported to show the “best and worst places for the poorest students to go to school”. Kensington and Chelsea, in London, ranked top, with 59 per cent of free school meals pupils achieving five A*-C grade GCSEs. Eight of the top 10 local authorities were in London. Nottinghamshire finished bottom, with only 22 per cent attaining this benchmark.

The charity put London’s high performance down to extra Government cash, such as the pound;40 million London Challenge initiative.

Jason Strelitz, a spokesman for End Child Poverty, said: “This shows there is no excuse for failing the poorest students.”

However, Professor Gorard said that there could be a number of explanations for the disparities. For example, the number of free school meals children could make a difference, as areas with many poor pupils might find it harder to cope.

And some authorities might have many children only just poor enough to be eligible for free meals. There is no sign that this had been taken into account by the charity.

Also, it offered no evidence on how London authorities’ performance had changed over time, to justify its statement about the benefits of extra investment.

Professor Gorard said: “The absence of doubt and boxing off plausible alternatives considerably weakens this report.

“It’s not particularly bad, just normal unthinking behaviour for UK research and policy advocates.”

TES analysis of the results suggests that differences between the local authorities may be linked to the number of ethnic minorities they educate.

In End Child Poverty’s league table, 21 of the 30 top performing local authorities were also in the top 30 when ranked according to the greatest proportion of ethnic minority pupils they educate. Nationally, on average, white free school meals children perform worse at GCSE than other ethnic groups.

A spokesman for Save the Children, which helped compile the report, said the figures clearly demonstrated an apparent link between local authorities doing well at GCSE and those which had had extra Government investment. He added: “We do not pretend for a moment that this is a simple issue.”