Higher achievement under Blair is down to smoke and mirrors rather than effective reforms.

• Apparent improvement in the last ten years has not been real: the results look better but pupils know less
• A 'Pre-Blair A-level' is worth, on average, a whole grade higher than the 'Post-Blair A-level'

Blair's claims of school improvement over his decade in power add up to little more than shameless moving of the goal posts - according to a briefing published today by independent think-tank Civitas.

The report exposes the dramatic and unprecedented rise in all major national test results as a sham. It reveals that improvements have been largely manufactured by lowering test standards, narrowing the school curriculum and forcing teachers to teach only for the next barrage of tests - leaving the real prospects of pupils ignored.

By setting targets, issuing never-ending guidance from central government and using Ofsted as a draconian enforcer of government policy, Blair has cajoled maintained schools into following a path of pursuing good-looking results regardless of the real costs to our children's education. This has led to paradoxical situations:

• The government's own figures admit that primary schools use four months of teaching in an academic year to coach pupils for a single set of exams.

• Schools are encouraging pupils to take an IT GNVQ in sending emails, worth four good GCSEs in the league tables.

• Pass-marks have become so low that a Business Studies GCSE paper in 2005 required only 47% to gain an A*.

• A-Levels have dropped so much in real value over Blair's decade that they should be regarded as different qualifications. Independent comparisons with the International Test of Developed Abilities suggest that the 'Pre-Blair A-level' is worth, on average, a whole grade higher than the 'Post-Blair A-level'. Blair's legacy here has been to inflate results so much that today's B grade is only as good as a C of ten years ago.

These shortcuts have allowed the government to reap a regular crop of official statistics demonstrating better results. But they have undermined trust in tests, harmed the morale of teachers and pupils alike and made it impossible for universities and employers to judge the actual abilities of their applicants.

'While Blair inherited a poor education system, he has resorted to slapping a thick layer of paint on the rotten edifice and hoping that good presentation will do the rest. This particular "legacy" would be best forgotten.'

'In ten years Blair has failed to sever the link between academic achievement and social background' say the authors of the report, 'instead he has destroyed the link between passing a reading test and actually being able to read.'

From the report:

Key stage 2 SATS

In 1997, only 63% of pupils at Key stage 2 (age 11, final year of primary school) managed to achieve level 4 in the literacy SATS tests and only 62% in the numeracy SATS.

By 2006, these figures were 79% and 76% respectively. But independent tests carried out at the University of Durham annually since 1997 found no evidence of a rise in literacy levels. This is hardly surprising since to achieve the expected level (level 4) in the 2006 Key Stage 2 literacy SATS papers, pupils were required to accrue just 43 points out of 100. In the 2006 Key Stage 2 numeracy SATs papers, a level 4 pass needed only 46 points out of 100.

GCSE

In 1997, 46.3% of GCSE students achieved 5 or more A*-C grades while 6.6% had no passes. By 2006 58.5% of pupils achieved 5 or more grades A*-C, while only 2.2% did not achieve any passes.

But when English and maths were included in the five, fewer than half of pupils (45.8%) qualified. Hence, 21.7% of pupils who managed to obtain what counts as five 'good' GCSEs did so without having reasonable knowledge of maths and English.

Dr Robert Coe of Durham University used independent data to show how this allowed pupils of similar standards to achieve different results over the years. A student who scored 45 (just below the average) on the YELLIS test could expect to achieve D grades in French, Maths and History at GCSE in 1996, but by 2005 would be receiving C grades; enough to push a number of students into achieving the 5 grades at A*-C government 'benchmark'. Taking an average of 26 subjects, pupils of the same YELLIS standard could generally expect to achieve around half a grade higher in 2005 than they could in 1996.

In other words when the government claims that over half of pupils now achieve the expected standard, this is because the standard has shifted to include those that did not reach the same level in previous years.

A-Levels

In 1997, 87.2% of A-level entries were passes. By 2006, this was the case for 97.2% of the entries. The percentage of A grades rocketed from 16.1% to 24%.

From 1997 to 2006, the number of A-level entries has increased by over 50,000. However, this increase has not been reflected in traditional subjects. In fact, many have declining numbers of entries: physics, French, German and mathematics have all registered reductions of between 3,000 and 10,000 over the last decade. By contrast, psychology has increased by 26,000; Media & Film Studies by 14,000; ICT by 12,000 and PE by nearly 10,000. This is not just a case of students opting for easier subjects. Schools have been discouraging pupils from taking subjects that are deemed more challenging and are therefore less 'safe' an option for league table purposes.

'Artificial Achievement: Blair's School Legacy is a Sham' by Anastasia de Waal and Nicholas Cowen is available online here