The method employed by the Department for Education and Skills involves comparing the performance at 14 of all pupils who achieved each combination of levels in the English, maths and science tests at 11. Less than 20 per cent of pupils achieve at least level 5 in all three subjects at age 11.

Grammar schools normally take a wider range of ability than this and therefore admit many pupils with at least one level 4.

However, they don’t need to admit all these pupils, so in order to achieve better discrimination than can be provided by these very broad levels, decisions are based on marks achieved in whatever tests are used for the selection process. This means that only the more able out of all the pupils in the area with a mixture of levels 4 and 5 will get into grammar school.

Not surprisingly, their results in the key stage 3 tests are better than the average for all the pupils who achieved the same combination of levels at age 11. This, of course, gives them a better value-added score, but only because they are being compared with a broader group of pupils, many of whom were not clever enough to have been selected for grammar school. It is not a like-for-like comparison but one which is skewed in favour of the grammar schools.

Unfortunately, this means that parents are being given misleading information and that the country as a whole will draw false conclusions about the benefits of selection. If the value-added calculations were based on marks achieved in the key stage 2 tests instead of levels, this problem would disappear. We would then be able to see whether pupils in grammar schools really do make more or less progress during key stage 3 than similar pupils in non-selective schools.

John Critchlow

19 Oak Tree Drive

Bedale, North Yorkshire