Furthermore, there is data on teachers' perceptions of the tests and how different children perform. A 1994 survey covering 32 schools in four English education authorities showed that just under half the teachers interviewed were concerned that bilingual children, along with poor readers, had probably lost marks because of difficulties in understanding the questions.
A follow-up survey conducted after the 1995 tests told a different story, however. That survey, which was also carried out as part of the National Assessment in Primary Schools Project, revealed that 17 of the 30 teachers questioned had received permission from their headteachers or LEAs to make special arrangements for bilingual pupils and poor readers.
Eleven of the teachers said this had been done to allow questions to be read to the children, sometimes individually and in separate rooms. A few had received permission for extra time. But at least 20 teachers made time before the tests to practise working to time limits and test formats with these children.
As a result, 22 of the 30 teachers thought the 1995 tests had not disadvantaged children in any way. One of the remaining eight was undecided, while the other seven mentioned language problems (either for pupils with special needs or bilingual children or both). The single point of unanimity was the unfairness of strict time limits.
But there are educationists who contend that there is a more fundamental problem with the tests that requires a more radical approach.
Hugh South, chair of the National Association for Language Development in the Curriculum, and manager of a Section 11 project in Hertford, feels English testing is inappropriate for children with little or none of the language. "The national curriculum was conceived as a monolingual curriculum and children at an early stage of English acquisition are not taken account of in the tests. "
This means that there is likely to be a disparity, as has been anecdotally reported, between bilingual children's achievement in English and in maths and science. One reason is that bilingual children receiving language support are entitled to interpretative help with maths and science tests at key stages 2 and 3. That help may come from either Section 11 teachers, bilingual classroom assistants or a combination of the two. Those supported by a mother-tongue-speaking helper appear to fare better than pupils who are reliant on monolingual Section 11 support.
The question of the allocation of language support in the tests can be a tricky one, however. What happens when a child who needs help cannot receive it in their mother tongue - or at all? Mr South gives the example of a classroom in which Urdu-speaking children are supported by a mother-tongue bilingual assistant while the one or two Sylheti-speaking children miss out because there is no staff member who speaks their language. He says: "There is clear evidence that those children with first-language help can show their understanding in the tests; without it, their assessment may well show them at Level 1."
Janet White, of the School Curriculum and Assessment Authority, reckons there are often not enough people to help those children who need it. Lack of local authority funding is an on-going issue, but other measures are being explored that could work to the advantage of bilingual children.
SCAA, under Ms White's direction, has set up a working party to produce policy guidelines on teaching English as an additional language which will include the issue of tests. In addition, SCAA has been collecting precise performance data on the achievements of different groups of children.
The Office for Standards in Education is introducing a new framework for inspectors that removes a separate section on equal opportunities and special needs and replaces it with a recommendation that inspectors "permeate" their reports with references to these considerations. Included in the inspectors' observations should be information on performance differentials between racial or ethnic groups.
However, this can only be done if schools collect and offer this data, which is not being demanded by the inspectorate. The decision is in the hands of individual schools.
In the meantime, there are practical strategies for supporting children, depending on the financial and staffing constraints. In English, the choice of book by teacher and pupil for the reading test is crucial. Ms White says: "Although publishers have become more sensitive to multicultural books, they often don't think of language structures most useful to English-as-an-additional-language children. Predictable structures and repetitive refrains are more important than setting books in obvious multicultural contexts such as the Notting Hill Carnival or Diwali."
Another important factor is pacing the tests. Key stage 1 is flexible, with possibilities for class or group activities and spreading the one-to-one interview over several weeks but key stages 2 and 3 are less so. Even then, the lower levels of both these tests give teachers leeway to allocate more time.
Lastly, having well-prepared mother-tongue support available to children in the tests is the most fundamental and valuable help schools can give children. With continual assaults on Section 11 and other funding for schools, the equality of opportunity enshrined in the national curriculum is being more and more determined, for children with English as an additional language, by budgetary factors.