Inspectors are to be instructed to take into account the social background of pupils and compare “like schools with like” in the interests of reaching fairer judgments.

From April, factors such as the number of children eligible for free meals, those with special needs and the educational level of the local population as well as the percentage of non-whites will be included in inspectors’ initial briefings.

According to an internal Office for Standards in Education report, by ranking schools using such indicators, inspectors will be able to make “fairer” comparisons.

The decision to go ahead with the new system follows work by OFSTED’s research branch which has used such methods on data for all schools and GCSE results from years 1992 to 1994. The unpublished work, based on a smaller-scale study by London’s Institute of Education, found that some schools do dramatically better than others with similar characteristics.

According to the OFSTED study, 20 per cent of single-sex comprehensives are rated “much better” than other similar schools, compared with 9 per cent of mixed schools. The report said a 1 per cent increase in the number of girls in a school will increase average GCSE points by 0.054, therefore an all girls’ school should have an average score of 5.4 more than an all boys’ school of similar intake.

It also found schools with higher percentage of pupils eligible for free meals scored below average, while those with more non-whites achieved significantly better results than others with similar characteristics. Schools without a sixth form, those in inner-London, and those that had opted out all did better than “like” schools at GCSE.

If schools were inspected in this way, then those at the middle or near the top of league tables might receive worse reports than expected if the inspectors discover their results to be below “like” schools.

Alternatively schools doing badly in the national tables could be found to be doing well if compared to others with a similar intake.

The Government has so far tried to play down connections between social deprivation and poor performance. But now such information will go into the “pre-inspection context and schools indicator” (PICSI) reports for comprehensives from April when the new inspections framework comes into force. Detailed maps showing other schools in the area will also be included.

OFSTED does not pretend comparing like schools with like is the same as value added which attempts to measure how much a school has helped its pupils progress. It does say it is a fairer way than at present to inspect schools and “will make it clear that some schools in less favourable circumstances do much better and some much worse . . . At the other end it may encourage inspectors to look harder at schools with average GCSE results but which should be doing better”.

There were hints of OFSTED’s change of perspective in the recent annual report of Chris Woodhead, the chief inspector, who said: “I am naming those schools that are not only achieving well, but are doing so in comparison to schools in similar circumstances.”

The report also warns that while a school may be doing as well as like schools, standards may be inadequate generally.

John Sutton, general secretary of the Secondary Heads Association, said: “It seems a sound idea in principle, but I am concerned that some of the criteria may be flawed. For example, the number of pupils with special needs statements often has no relation to the number who should have them.

“Our other concern is that OFSTED will not compare schools on the basis of differential resources.”