There are in fact 17 local authorities that appear to have worse primary school results than Islington. Of the 101 local authorities that feature higher scores, only Hammersmith and Fulham in London has a similar proportion of primary school children on free schools meals. Just over 40 per cent of Islington’s primary pupils are eligible for free school meals.

However, while even the Government does not dispute that areas of deprivation are unlikely to produce the highest results, there are schools in the borough with a significant number of children on free school meals which produce results equal to those achieved by schools with a more favoured intake.

Within the borough’s education service, officials have been attempting to identify the range of attainment across schools that have roughly the same intake in terms of free school meals and the number of children with English as an additional language.

On the 11-year-old test scores, the highest performing is the Roman Catholic St Peter and St Paul, where all 26 pupils achieved the expected level in maths; 76.9 per cent achieved level four in English and 96.2 per cent in science. In that year between 35 and 39 per cent of the pupils were eligible for free school meals. Another school with slightly fewer 11-year-olds on free school meals - St John’s in Upper Holloway - has only 21.4 per cent of its pupils achieving the expected level in maths.

There appears to be just as wide a variation in the results of schools where deprivation is particularly high. At Newington Green junior school almost three-quarters of children are on free school meals, yet 64.3 per cent of 11-year-olds achieved level 4 - the same score as St John’s in Highbury, where fewer than 20 per cent of 11-year-olds are on free school meals.

The problem for policy-makers is working out why schools that appear similar can produce strikingly different results. As well as examining the impact of deprivation, officials in Islington have analysed results on the basis of the ethnic background of pupils.

The breakdown suggests the highest maths scores are achieved on average by children from Irish and Chinese backgrounds. They also seem to suggest that groups underperforming are from African, Caribbean and Bangladeshi backgrounds. The exercise is limited by the fact that numbers in some ethnic groups are relatively small. However, the findings also show that children fluent in English, but not from an English-speaking background, tend to outperform children who speak only English.

Parents of 11-year-olds in Islington are being given a booklet that contains the test results and teacher assessments of every school in the borough. It also shows the percentage of pupils on free school meals, the percentage non-fluent in English and the percentage of 11-year-olds that have been in the school for less than three years.

Within the service, the next stage is to attempt to account for the differences between schools. The problem in Islington is that around 42 per cent of parents with secondary-age children educate them elsewhere. Any improvement at primary level may boost results in Camden’s schools and leave the borough at the bottom of the secondary-school tables.