Multiple choices

If SATs can widen access, we owe it to disadvantaged students to try them out, says Peter Lampl

Last year, the Schwartz committee, of which I was a member, documented the mushrooming of additional admissions tests for university entry. While this proliferation is understandable given the inflation in A-level grades, it acts against the transparency that formed the backbone of the committee's recommendations. Worse, it risks restricting access for applicants who are put off by the complication and stress of taking a different test for every university they apply to.

The committee considered the results of a trial by the National Foundation for Educational Research (NFER) of the US SATs in British schools in 2000.

More than 1,300 pupils from 70 secondary schools completed a short version of the test and their scores were compared with their GCSE and A-level results. The pupils came from three groups: low-attaining state schools, high-attaining state schools and selective independent schools. Despite not having done this type of test before, the students scored comparably to their US counterparts, suggesting that the test is easily transferable to Britain. Other key findings were that the SAT measures a different construct from A levels and GCSEs, and can identify those whose academic potential is not reflected in exam results.

Particularly interesting was the finding that in the below-average performing schools, 30 students out of 600, or 5 per cent of the sample, scored well enough in the SAT to be considered by a top US university - yet only one of them achieved the three As at A level required by our most selective courses. Given the proliferation of tests and as a result of the promising results of the NFER trial, the committee recommended an operational trial of a national test of potential, citing American SATs as a possible universal admissions test to be used in conjunction with A levels.

The launch of a £1.6 million longitudinal study, conducted by the NFER in partnership with the Department for Education and Skills, the Sutton Trust and the College Board, was announced last month. Approximately 50,000 students - about a fifth of all A-level candidates - will sit the SAT in November. Their scores will be compared with their GCSE and A-level results, and later with their performance at university, to assess, among other things, its usefulness as a predictor of degree success. No one would argue that the SAT is a panacea to the current ills of the admissions system.

However, some recent criticisms of the trial are unfounded. Chris McManus's suggestion ( Times Higher , October 7) that SATs used in a 1967 trial was "identical in form, structure and rationale" to the current exam couldn't be further from the truth. That trial was of a UK-developed aptitude test, not the SAT, and one of the objects of the research is to assess the SAT's recently added essay component.

Likewise, even the staunchest defender of the standards of today's exams would not make a straight comparison between the A levels of 1967 and their current equivalents. In 1967, only about 10 per cent of the cohort took A levels and they had access to a similar level of schooling via grammar, direct grant or independent schools, so achievement to date as measured by A levels could be taken as a fair measure of potential for university study. A levels were an academic exam for specialised university courses; many failed them and only 10 per cent achieved A grades. Today almost half the cohort take A levels. The pass rate is 96 per cent and 23 per cent get A grades. Significantly, today's students experience vastly different levels of schooling, from good independents to state schools in special measures. We need to reconsider the crucial question of the predictive validity of A levels and SATs in the current context.

There are other key issues that this trial will seek to address. In the US, students from certain ethnic groups and from low socio-economic backgrounds tend to score poorly, and we need to ascertain how far this is the case in the UK. Coaching has also become a significant issue and the subject of much research, which the NFER has reviewed as part of the earlier work.

This showed that the average improvement gained was small and most took place in the first ten hours. If the trial results are promising enough for SATs to be introduced, then some familiarisation should be built into the examination procedure to ensure a level playing field.

We must be sure that the SAT - or any test proposed for university admissions - is fit for purpose, and this trial will seek to answer that question.

Sir Peter Lampl is chairman of the Sutton Trust, which provides educational opportunities for children from non-privileged backgrounds.