A Conservative borough is claiming to be the first to use so-called baseline assessment of four-year-olds to discover how effective its schools are.
Researchers in the London borough of Wandsworth have compiled a value-added league table for primary schools by comparing the results of tests children took at ages four and seven.
The researchers, led by Dr Steve Strand, discovered that seven-year-olds in the seven most effective schools were doing better than their peers at other primaries despite the two groups having achieved the same results at age four.
The six schools the researchers identified as the least effective were considered to have favourable intakes from affluent areas, but not to be stretching their pupils.
Dr Strand told the British Educational Research Association conference in Lancaster last Friday that: "We can conclude that schools do make a difference to pupils' progress: schools with intakes of similar attainment and with similar compositions can and do achieve very different key stage 1 results. "
Dr Strand's research compares two similar schools which are producing strikingly different results: "Both schools serve quite disadvantaged populations with 72 per cent and 64 per cent of pupils entitled to free school meals - the borough's average is 36 per cent.
"The majority of pupils from school A are attaining higher levels at KS1 than would be expected from their baseline [test results at four years old]. In contrast, the majority of pupils from school B are attaining lower levels at KS1 than would be expected."
However, Wandsworth does not know why the top seven primaries are so effective, and plans to monitor all its 57 primary schools to find the secret of their success.
The borough introduced baseline assessment for 1,699 pupils starting reception class in 1992 after two years of development, piloting and in-service training.
Teachers assess motor skills, social and emotional development, attainment in English, maths and science, and emergent literacy. Information is collected on pupils' background, ethnic group, home language and previous education experience.
But Dr Strand warns that because of the borough's transient population, 17 schools had only between 10 and 20 pupils who were eligible for analysis.
The research also shows that girls are doing better than boys even at four, and worryingly, the gap widens between the ages of four and seven. Dr Strand's team divided the four-year-olds into the highest and lowest achieving groups. He discovered that 14 per cent of boys and 21 per cent of girls were in the top group, and that 24 per cent of boys and 15 per cent of girls were in the bottom group.
The achievement gap also widens for those children on free meals who are more likely to start with lower baseline assessment scores than other children.