By Gary Eason BBC News Online education editor

National Audit Office says new-style tables are better but still flawed

The two rankings, available side by side in the BBC News Online tables, will pose some interesting dilemmas for parents trying to choose the "best" local school for their children.

The compiling of the value added measure takes several pages of explanation in the official notes accompanying the tables.

But put simply, it involves working out how well each child did this year compared to the middle performance of all the others nationally who had a similar level of achievement four years ago.

Results for all the children in a school are combined into a number above or below 100.

Small schools

The national range turns out to be 95.2 to 105.5.

NEW CHALLENGE FOR TEACHERS Tests measure those achieving Level 4 - standard expected by end of primary education New targets set for those reaching Level 5 - standard expected of most 14 year olds 100% of pupils did so at Combe Church of England Primary School in Witney, Oxfordshire At least half of them were very bright. That encouraged the others Head teacher Barbara Jones

At the bottom was Manton Primary School in Nottinghamshire, with 58 pupils.

At the top: Delaval Community Primary School in Scotswood, Newcastle upon Tyne - though it did have only 12 pupils eligible for assessment.

That in itself poses a problem for the government in terms of the authority of its performance tables.

Normally it does not publish the results from "small schools" - those with 10 pupils or fewer eligible for the tests.

Twelve at Delaval qualifies it for inclusion, therefore - but in fact several of them did not sit some of the tests, being "absent or disapplied", in the jargon.

Other small schools might have done as well or better - but because their value added measure has not been published there is no way to tell from the official data.

The department cautions people about the reliability of the results from small numbers of pupils.

'Health warning'

So - as with any tables - these need to be read with a large pinch of salt.

Official statisticians reckon that for schools with more than 30 pupils covered by the measure, scores of 99.1 to 100.9 equate to "broadly average" performance.

For schools with 50 pupils or more, scores of 99.3 to 100.7 are broadly average.

Anything of 102 and above is in the top 5% nationally.

Anything of 97.9 and below is in the bottom 5%.

Best and worst

On the traditional benchmark - the "aggregate" score pupils achieved in all three tests in English, maths and science - the number of schools which achieved the maximum 300 this year was 142.

That was down by a fifth on last year's 178 schools and the 179 who managed it in 2001.

Four schools have managed 300 for each of the four latest years reported in the tables - something three did last year.

The "top" school - in having the largest number of pupils - was Werrington Primary in Cambridgeshire.

All 62 of its Year 6 pupils performed at or above the level expected for their age in all three subjects.

Showing how fickle the tables are, last year's best - Henleaze Junior School in Bristol - came well down this year with a score of 280.

Results quashed

The national average score was 234, the same as last year.

At the other end of the scale, the school with the poorest results this year was Moor Nook Community Primary in Preston, Lancashire - which managed a total score of just 66.

Four schools had some of their results annulled by the Qualifications and Curriculum Authority (QCA) for suspected cheating.

In English: Oxclose in Sunderland, Kneesall in Nottinghamshire and Southwold in Hackney - whose maths result was also quashed, as was the maths result at Ironville and Codnor in Derbyshire.

The school which got the worst results last year, Hothfield Village in Ashford, Kent, does not feature in the tables this year because it had only four pupils taking the tests - results of 10 or fewer pupils are not published.

Another alternative

On the average point score (APS) - intended better to reflect the achievements of all the pupils - the national average this year was 27.4 points, compared with 27.5 last year and 27.3 the year before.

The APS assigns points to those who otherwise would count for nothing - but is weighted to give most to those working above the expected level.

The worst on that measure was Northdown Primary School in Margate, Kent, with 48 pupils managing 20.5 points.

The best was Combe Church of England Primary in Witney, Oxfordshire, whose 12 pupils averaged 33 points.

It's part of the culture of the school to overcome barriers to high achievement Margaret Libreri, head of most improved school Schools on the up and up

Each year the Department for Education teases out of the data the top 100 "most improved" schools.

This lists only those schools whose combined results have improved consecutively since 2000.

Top of the list this year is Wellington Primary School in Bow, east London.

The aggregate of its English, maths and science results has gone from just 82 in 2000 to 279 this year.

Children with special educational needs (SEN) who attend mainstream schools are included in the tables.

The term can include unusually gifted children, but mostly refers to those with learning difficulties of one sort or another - in the official statistics, there is no way of telling.

The incidence of "statements" - for those with the greatest needs - ranges from nothing in the case of 5,955 schools, up to 35.1% in one school.

The average rate of statements nationally in schools for which results are reported is 2.6%.

On the broader measure of SEN - those without statements - almost all schools have some pupils with some reporting more than 80%. The national average was 19.6%.

Absences

No unauthorised absences were recorded during the year in 2,550 schools and the rate of truancy was negligible in almost as many others.

But the school with the worst record was Temple Normanton Primary in Chesterfield, Derbyshire, with 7.9% of sessions missed.

The national average truancy rate was 0.4%.

There is a higher average rate of authorised absence - 5.4% nationally.

But one school - St Weonard's in Herefordshire - recorded 29.1% of sessions missed with permission.