The government is being urged to use data it holds about pupils' backgrounds to produce better secondary school performance tables for England.
Assessing schools is important in raising standards, report says
A report from the National Audit Office says there are risks in assessing schools solely on their exam results.
It acknowledges the move to reflect pupils' previous achievements, but says this does not go far enough.
It urges all concerned to come up with a better measure of social deprivation than entitlement to free school meals.
The audit office says better assessment of performance would allow support to be targeted better on schools most in need of it.
Head teachers are delighted with the report, which they say reinforces what they have been saying for years.
This year's primary school tables are due out next Thursday, followed by secondary schools' national curriculum test results, with the GCSE and A-level results due in January.
The office got the National Foundation for Educational Research in effect to re-run last year's tables, taking into account extra data on more than a million pupils in more than 3,100 state-run secondary schools.
This can make a big difference in the ranking of individual schools, it says.
"For example, of the 621 schools ranked in the bottom 20% of performers in terms of academic achievement at GCSE level in 2002, just 272 remain in the bottom 20%".
Sixty of the schools move up to the top 20%.
"Conversely, the ranking of some highly ranked schools falls".
The new use of "value added" data, which compares pupils' final results with those of pupils with similar previous achievements, is "an important step forward", says the report.
"However, the assessment took no account of other external influences on performance, particularly pupils' economic, social and cultural backgrounds."
Available data meant it could consider such things as pupils' ages, gender, ethnicity and eligibility for free meals.
But it had no access to other likely influences such as school funding, the condition of school buildings, and parental occupations and education.
The audit office did find that some types of school generally beat the average, once external factors were taken into account.
These included selective schools, specialist schools, faith schools, beacon schools and single sex schools - both boys' and girls'.
The only ones that went up by the equivalent of a grade per subject at GCSE when external factors were considered were specialist schools and the "very small number of Jewish and 'other faith' schools".
"Our analysis does not explain the reasons for the associations," says the report - with some specialist schools getting a relatively low ranking, for example.
As critics of league tables have long pointed out, prior academic achievement had the strongest association with current achievement.
Eligibility for free school meals had a strong negative association - but was "a fairly imprecise indicator of the economic position of a pupil's family."
The chairman of the Commons public accounts committee, Edward Leigh, said: "Many schools face substantial barriers. These are typically the social and economic deprivation of their pupils and their low previous academic attainments.
"Yet some of these schools make a bigger difference than many schools in kinder circumstances."
More info for parents
Using information on pupils' backgrounds "would highlight many schools which, contrary to expectations, are doing an excellent job in adverse circumstances".
"But it would also put into the spotlight those schools that are not doing so well by our children.
"Parents across the country will look forward to this better information on schools' performance being published", he added.
The general secretary of the Secondary Heads Association, John Dunford, said the existing tables did not present a true picture of the performance of schools.
"They discourage schools from admitting pupils with special needs and the sooner they disappear the better so that we can devise fairer performance measures for schools that give parents the information they want."
He wanted to see more data provided locally but not in national performance tables - more like the situation in Northern Ireland, Wales and now Scotland.
The Department for Education and Skills said 'value added' information, plus the usual performance indicators in the tables, did provide parents with "a rounded picture" of the performance of pupils and the progress schools helped them make.
"We welcome the report's findings showing what makes a good school," a spokesperson said.
"These include having a clear vision and ethos, effective leadership and management, high quality teaching and effective encouragement of pupil attendance and good behaviour.
"This is why the government is placing such importance on these areas."