Black children will perform significantly worse under a new system to assess the educational performance of children up to the age of 5, widespread trials have suggested.
The early years foundation stage profile, which rates all children on a range of factors at the end of reception year, is being radically overhauled to cut down the number of assessments and toughen up what is needed to achieve certain grades.
But research from the Department for Education's Standards and Testing Agency has discovered that a stricter definition of what constitutes "good" performance has had a disproportionate effect on certain ethnic groups.
In a pilot study, just 41 per cent of children were judged to be at a good level under the new standards, compared with 64 per cent under the old system. But for black children the proportion fell from 61 per cent under the old system to 29 per cent - a drop of 32 percentage points, compared with an 18 percentage point drop for Chinese children, 22 percentage points for white children, 24 for Asian children and 26 for mixed-race children.
All ethnicities apart from the Chinese found achievement in writing and "numbers" - which includes counting and simple sums - the most difficult. But black children in particular struggled to reach a good level in numbers: just 46 per cent achieved the grade, compared with 59 per cent of white children and 72 per cent of Chinese children.
The research also included a survey of 204 teachers, in which just 58 per cent agreed they were able to complete the new assessments accurately. The most common complaint was that the content of the goals was too broad or ambiguous, making it hard to assess children exactly.
The researchers were unable to explain why certain ethnic groups had performed worse under the new assessments. They said the results should be treated "cautiously" because they were carried out before the introduction of the new curriculum to accompany the tests.
But Professor Gus John, honorary fellow of the Institute of Education, University of London, was concerned by the results. "The report is saying these are the results, but it is apparent to me that this particular test itself has to be questioned," he said. "If this test is leading to certain desired outcomes for some students and not for others and those others are defined by race, they have to be very careful.
"It has previously been the case that people tried to associate levels of intelligence to ethnic background. People publishing these results have a responsibility to be aware of those implications. We need to be drilling down into those statistics in order to try to establish a cause."
Under the old system, children were expected to be able to count to 10 by age 5. Now they are expected to count to 20, add and subtract two single-digit numbers and solve doubling, halving and sharing problems.
Megan Pacey, chief executive of charity Early Education, said it was too early to draw firm conclusions about the pilot results. "It will take a year or two to really see what happens on a national scale," she said. "One thing that has come out is that practitioners found it a challenge and so some rough edges may be smoothed in the next year or so."
A DfE spokesperson said: "The goals under the new early years foundation stage profile are based on what pupils should be achieving by age 5. This will ensure all children, whatever their ethnicity, are given the best chance to reach their full potential."
The early years foundation stage profile is completed for all children at the end of reception year.
In the past it assessed children in 13 different areas on a nine-point sliding scale in each - 156 goals in total.
In response to complaints that it was too cumbersome, the profile will be significantly slimmed down this summer to a total of 17 areas, with just three points on each.