More than 600 "failing schools" in England have been identified by the government and told to improve, or face closure. There's just one problem says Michael Blastland - some of the "worst" are also some of the "best".
The best state school in England is a borderline underachiever. Who says it's the best? The government. Who says it's a borderline underachiever? The government.
Confused? You might ask if so is the government.
In announcing this week that 638 English secondary schools should either improve standards or risk being taken over or closed, Education Secretary Ed Balls produced what seemed conclusive and damning statistical evidence. In these schools, 70% of pupils, or more, failed to obtain five good GCSEs (including maths and English).
Many in the media described them as "failing schools". And so, it seemed, they were justly said to need radical attention.
But for one fact: in the judgement of Ed Balls' Department for Children, Schools and Families (DCSF), some of these same schools add more value than almost any other in the country.
There are broadly two main measures of schools' performance. The government encourages both. Unfortunately, they often produce inconsistent results.
Thus one school named as underachieving badly on GCSEs (the first measure) is by the second measure well into the top 1% of all state schools. Some are poor on both counts, but more than 30 "failing schools" are in the top 5%.
One of the key criteria judges schools on exam passes
This second measure is contextual value added (CVA). CVA looks beyond raw GSCE scores to how well the child was performing on arrival at the school, compares this with performance at the end then measures the difference. That's the "value added". What makes it "contextual" is an amendment to these scores to take account of the child's background, so that schools are judged, it is hoped, on what they achieve given the raw material that comes through the door.
One of these contextual adjustments is based on a child's economic circumstances. So although Ed Balls expressed a determination to "break the link between poverty and poor performance", his department publishes an evaluation of schools that asserts just such a link - known in the CVA methodology as the "income deprivation affecting children index".
While declaring that there could be "no excuses for poor performance", his department's 11 contextual adjustments are arguably just that. Whether these excuses/contexts are valid is another matter. That the government uses them is undeniable.
Top of the tables
The results on the two measures are often poles apart. For example, the Castle Community College in Kent, which achieves five GCSE passes in maths and English at grade C or better for what we might be inclined at first sight to judge a measly 23% of its pupils, and could conceivably now be threatened with closure, has a phenomenal CVA score of 1066 (the baseline is 1,000).
This ranks it the seventh best school in the country, comfortably ahead of every grammar school. In a truly glowing report, Ofsted, the schools inspector, describes it as "outstanding". Joseph Rushton Technology College in Lincolnshire, with 13% (desultory?) of pupils achieving five A-C passes at GCSE in English and maths, scores 1057 (miraculous?) on the CVA measure.
Steve Davis, a spokesperson for the school, says if you look at the "many awards Joseph Rushton has received in the past two years, and look at our Ofsted report which is 'good' with a lot 'outstanding', I don't know how anyone could say this was anything other than an excellent school".
Contextual value added looks beyond raw GSCE scores
He pointed out that on GCSEs, without specifying English and maths, the school was the second best in Lincoln. On yet another version of the CVA tables, using slightly different methodology, it is the second best in the country. He said of the inconsistencies: "It's all a bit silly."
Even the very best school in England on the CVA measure, Moreton Community School in Wolverhampton, with a stratospheric CVA score of 1090, only achieves the 30% GCSE target by the skin of its teeth. The year before it would have failed.
If you are said to be that bad, and also to be that good, what are parents supposed to make of you? As for the schools themselves, told to improve or risk closure, it might be asked "how?" when their teachers are already supposedly getting more out of their classes than almost anyone else in the country.
Mr Balls, who has set the 638 schools a National Challenge to achieve five good GCSEs for at least 30% of their pupils by 2011, conceded that "GCSE success is not the only measure of how a school performs, but it is critical - teenagers need these qualifications to go on to further study, work and prosperity.
"A young person with five good GCSEs will almost always earn considerably more than a teenager who leaves school with no qualifications. Employers expect these qualifications as a minimum."
Schools Secretary Ed Balls: 'Five good GCSEs are expected as a minimum'
He added: "Of course National Challenge schools face real challenges but no child and no school is on a pre-determined path to low results. There are many schools in communities of high unemployment and low aspirations where children achieve excellent GCSE results."
He has allocated £400m from existing resources to helping the schools improve, an average of about £630,000 each between now and 2011.
A spokesperson for the DCSF added: "We have not doubted that some of these schools have done very well on CVA. The point of the National Challenge is to provide the support and help they need to get over 30%."
The GSCE and CVA measurements tell us something interesting about a school, but different, and in this case are incommensurate. If you want to assert failure or underachievement based on one of them, you have to ignore the other, which would be a significant policy U-turn given that CVA has been 15 years in the making, the result of progressive refinement of raw GCSE results and performance tables.
There is one plausible (though speculative) argument that might lie behind the inconsistency, but politically it would be difficult to admit without accusations of intrusive social engineering. The government might have concluded that it is the concentration of academically challenging pupils in a small number of schools that causes the results to be lower than they could be.
The thinking would then run as follows: too many needy pupils in one place and it does not matter how good the teaching is, the only remedy is to spread them around and perhaps close the school where they used to cluster. Whether or not this squares the circle where the teaching can be good, but the outcome is still poor enough for us to demand change and threaten closure, it is not an argument the government is making. What else it might be trying to say about school standards is hard to fathom.
Michael Blastland is the author, with Andrew Dilnot, of The Tiger that Isn't - Seeing Through a World of Numbers.
Below is a selection of your comments:
It is ludicrous to measure a school on its GCSE passes. Those schools where there are supportive educted parents (eg middle class) will always do better. The CVA table is the only meaningful measure.
Geoff Lawrence, Wokingham
When will this government realise that qualifications aren't everything. There will always be a need for people to do jobs which don't require good academic skills. We can't just rely on overseas workers coming in to pick up these jobs - it simply isn't sustainable in the long term!
I think the article raises some interesting points. Schools in deprived areas will acheive lower pass rates than schools in more affluent areas. This is simply that parents with above average intelligence will tend to have children with above average intelligence. Generally more intelligent people tend to stay in education for longer and do more degrees - people with degrees tend to get paid more money than people who leave school at 16. People with more money tend to live in more affluent areas. Obviously there are plenty of exceptions but it's hardly surprising that schools in nicer areas do better in terms of % of good GCSE's attained. It doesn't help when grammar schools filter off the brightest pupils leaving everyone else to attend comprehensives. This automatically reduces the average IQ in a comprehensive school to below average for the local population. None of this takes into account the quality of teaching. It is perfectly possible to acheive poor GCSE results and!
still be providing excellent teaching! The school i attended (bog standard comp in Birmingham) had many excellent teachers but but when i checked teh tables it has failed to achieve the 30% benchmark. This doen't really surprise me as when i attended out of 6 sets in Maths not even everybody in the top two sets got grades A-C.
Excellent article. There is nothing intrinsically wrong with using different measures of performance for a single assessment - after all both raw scores and "CVA" sound reasonable but different things to measure. Provided they are satisfied that CVA and raw scores are both valid measures then the obvious thing to do would be to focus on those schools that score badly on both. The authors concerns are spot on if the government has now decided to focus exclusively on raw scores and it will likely lead to a further fall in standards of education standards in state schools.
Also, it has to be said that arguably on the first measure it could take 5 years to improve. This would be based on a full academic cycle from arrival to taking GCSEs. In other words, 'three years to improve' gives you only 60% of the time available with the pupils to bring about improvement.
I despair, as a teacher, of ever being able to satisfy the governmet or the press.
We have been constantly denigrated since Sir Keith Joseph started the tinkering with education, and OFSTED took its legions of failed heads and teachers and taught them how to criticise but not help. I retire in six weeks and will no longer have the daily battle of turning the chav's sow's ears into silk purses; thank God.
dexey, Birmingham, England
"with 13% (desultory?) of pupils achieving five A-C passes at GCSE in English and maths, scores 1057 (miraculous?) on the CVA measure"
Desultory means jumping from one thing to another at random. Grade E for English for this author!
The question is how are these schools achieving such high CVA scores when a relatively low number percentage their pupils achieve grade C or above in GCSE English and Maths? The answer is by changing their curriculum and delivering non-GCSE, vocational courses which carry the same points value such as Btech diplomas which do not require the same ability to analyse, evaluate and reach reasoned judgements as GCSEs. Additionally, high CVA scores are generated by entering pupils for large numbers of qualifications - a Btech is 'worth' up to 4 GCSEs. English and Maths above grade C are important, so is the double Science above grade C measure of attainment. Why not combine CVA and these measures but cap what is measured to 8 GCSE equivalents in total, i.e. English, Maths, Double Science plus four others, then we might get a true picture of how our schools are performing.
I really do worry about the intelligence of people (and the government), who judge schools on 5 A*-Cs, rather than Value Added.
I also hate the term 'good GCSE'. It implies that anything below a grade C is 'bad'. I have taught countless new arrivals to the country who have made more progress than most students in the country, to get a number of GCSEs (below a grade C). The message from the govenment, media and public, is that their GCSEs are 'bad' and worthless. It is so unjust and their 'Herculean' efforts are not being recognised.