Media pans parents for white British pupils not hitting ‘targets’. But the ‘targets’ were a 2030 proposed benchmark not actual exam results
‘Pupils let down by their parents: White British children are overtaken by 10 other ethnic groups by the age of 16 as alarming report shows their families do not do enough to support them in school’
‘Pupils with English as second language 'outperform' white British at GCSE’
‘Disaffected parents to blame as white pupils lag behind other ethnicities’
‘White children 'falling behind other groups at GCSE'. Lack of parental support blamed for poor performances at age 16 despite high early attainment at age five.’
You’d be forgiven for thinking GCSE results in 2015 showed white British pupils had plummeted down national exam league tables. But that’s not so. The ‘targets’ regurgitated by these papers were not actual tests. They were a proposed world standard, a goal for 2030, devised by think tank CentreForum whose executive chairman is David Laws, ex-schools minister and a part-time, paid adviser to ARK, the academy chain.
CentreForum’s proposed world standard for 2030 is 50+ points on the Attainment 8 scale, the new secondary school accountability system being used from 2016. The results of applying this proposed standard to 2015 test scores were given in a CentreForum publication, Education in England: Annual Report 2016.
I asked CentreForum if it was correct to say their researchers had taken their proposed world standard and applied it retrospectively to Attainment 8 in 2015. The answer was:
‘Not quite no.’
CentreForum explained the proposed world standard was an ‘ambition’ comprising:
1 The standard the think tank wanted individual pupils to achieve (50+ points on Attainment 8);
2 The proportion of pupils needed to achieve the proposed world standard if ‘we’ (ie England) ‘are to be world leading in 2030.’
It’s been recognised before that children in English schools whose parents came from China and parts of Asia outperform other pupils. The success of London schools has been partly attributed to large numbers of such pupils. But it’s rather a leap from this to claim, as David Laws did in the Telegraph: ‘We know that we've got this very bad performance of white pupils versus others.’ The 2015 GCSE results using the present benchmark of 5+ GCSEs A*-C (or equivalent) including Maths and English do not validate the ‘very bad’ claim.
Nevertheless, the media swallowed the ‘very bad’ judgement. And papers were in no doubt whose fault it was – feckless parents. The Mail’s front page went further: it was a ‘betrayal of white pupils’ by the education system.
The media claimed the performance of White British pupils had fallen in 2015 from their performance at age 5. But CentreForum has just confirmed to me in an email it was NOT comparing the same group of children. The data for performance at age 5 and age 16 in CentreForum's report (and which were reproduced in huge, colourful chunky graphs in the Mail) were for 2015. Claims that performance has fallen over time and that early competence has not been sustained can only be made, surely, by comparing results of the same cohort.
In criticising this ‘alarming’ report and the hysteria surrounding it, I am not saying the education system in England is not letting pupils down – it is. It’s letting them down because of the excessive emphasis on exam results. The OECD highlighted the dangers of this extreme focus in 2011. High stake exams encourage shallow learning rather than the deep understanding needed for future progress. The emphasis on academic subjects squeezes out creative ones.
Now this focus on exam results is becoming greater. The Government wants baseline tests for 4 year-olds. CentreForum proposes a world-class standard based on exams taken at 16. But most of the developed world do not have exams at this age. Where such tests exist, they are few in number and are used to decide progression to upper secondary school not for judging schools.
The Department for Education ‘welcomed’ the report saying it showed the ‘stark choice’ education in England faced. It’s only a matter of time before CentreForum’s analysis is used to justify wholesale academization.
UPDATE 6 April 2016 16.45 CentreForum tells me a 2006 DfES paper confirms White British pupils were doing ‘relatively well’ at age 5 in 2004. This report gives Early Years Foundation Stage Profile data for 2004 based on a 10% representative sample. Although there isn’t a graph which shows relative performance in all targets, the one for Communication (Fig 18) shows White British 5 year-olds ranked fifth (they were placed third at age 5 by CentreForum using 2015 data).
The 2004 performance profile for 5 year-olds, therefore, is not the same as the 2015 one. If we look at actual GCSE results (instead of CentreForum’s proposed ‘world standard’), we can see White British pupils were at 10th place ranked by proportion of pupils who reached the benchmark. This would support CentreForum’s conclusion that the relative performance of White British pupils as a group has fallen since they were aged 5. It is, however, an exaggeration to say this drop in five places is a ‘very bad’ performance.
CORRECTION AND CLARIFICATION 9 April 2016 08.15 The Update has been corrected. I originally wrote, '... we can see White British pupils were at 8th place ranked by proportion of pupils who reached the benchmark.' This should have read '10th place'. I have corrected this error.
The drop of five places from 5th (on the Communication scale) in 2004 to 10th on the GCSE benchmark ranking in 2015 was on a scale of 20. White British pupils are, therefore, at the midway point with 57.1% reaching the benchmark. This is exactly the average proportion of all pupils in state secondary schools in England (57.1%).
White British pupils comprise 72.2% of pupils in English state secondary schools. The ethnic groups which outperform White British pupils comprise a tiny proportion of secondary pupils: pupils from a Chinese background, for example, comprise just 0.4%, Indian: 2.8%, White Irish 0.3%. Is comparing a very large sample with very small ones valid? The answer is, I don't know - I'm out of my depth. An internet search for unequal sample size baffled me further. I'm not a statistician - a question for Radio4's 'More or Less', perhaps?