The School Curriculum and Assessment Authority has suggested three simple methods, which have been accepted by Education Secretary Gillian Shephard for further evaluation.

An advertisement inviting tenders to run an 18-month pilot scheme is published in today’s TES. SCAA hopes that at the end of that time researchers will be able to recommend which of the three schemes - or an amalgam - should appear alongside league table information.

In the meantime, Mrs Shephard is considering how indications of how a school’s performance has changed over several years could be included in next year’s tables. SCAA has suggested two non-value-added approaches - either a three-year rolling average or an “improvement index” which would show how a school’s results had changed as a percentage. The subject will be discussed early next summer. This would allay criticism of the current form of league tables until a value-added system could be made to work - which probably could not be until 1998 or 1999.

The three methods selected for further consideration by the SCAA group, chaired by Dr John Marks, were chosen largely because they agreed with Mrs Shephard’s principles that they should be straightforward, intelligible to parents and based on pupils’ actual attainments.

Accordingly, the group rejected value-added methods based on social factors such as free school meals. SCAA chief executive Nicholas Tate explained that they believed such factors were implicit in the “baseline” of pupil attainments compared with later results. Moreover, SCAA chairman Sir Ron Dearing’s suggested terms of reference had asked the group to concentrate on academic measures of value-added.

The SCAA group’s report concedes thatboys and girls perform differently, but says it did not recommend that the sex of pupilsbe included as a value-added factor for the sake of simplicity and because “it did not wishto allow such factors to affect expectations”.

The simplest option put forward by SCAA is a scheme which would plot a school’s GCSE results against tests taken by pupils at 11.

More complicated is a suggestion which estimates the proportion of a school’s pupils achieving better results than expected. The figure for comparison would be compiled from the average GCSE results achieved by children rated as being in different ability bands at 11.

The third method would split schools into groups: those performing roughly as expected from the national figures and those doing significantly better or worse.

The report from SCAA stresses that “finely differentiated” measures of academic attainment will be required for such value-added methods to work properly. Dr Marks’s group is confident that new marking systems for key stage tests will provide enough detail.

New-style league tables would also have to include raw results, the estimated error for each school and the percentage of pupils qualified for accounting in the scheme - those present for the initial and final tests.