A school journey of 11 years began and ended in 2015 – just one of the concerns about think tank’s ‘Annual Report’

Janet Downs's picture

 Graphs, tables, percentages, rankings – CentreForum’s Annual Report is full of them.  But how sound is the data especially as much of it compares school performance against a hypothetical ‘world standard’ which may or may not be valid?

CentreForum admitted few countries in 2015 would have reached this ambitious goal for 2030  when it launched its measure in January.    Despite it being just a proposal, CentreForum decided to apply the suggested standard retrospectively to results for the 2015 GCSE cohort.

Unsurprisingly, CentreForum's calculations discovered its recommended 'world standard' wasn't met by the vast majority of secondary pupils in England's state schools.  Only children from a Chinese background came close.  

The report’s Key Findings said 'The relative performance of White British pupils falls as they progress through school’.   David Laws told the Guardian that the poor performance wasn’t ‘embedded’ at the start of their school ‘journey’ because data showed them doing ‘relatively well’.

This implies CentreForum had tracked the progress of the GCSE cohort in 2015 throughout their school life.  But it did not. Instead, it compared the 2015 results for 16 year-olds with the 2015 Early Years Profile of 5 year-olds.  This was not comparing results within the same cohort.  The only valid comparison for a particular group’s progress over time is to compare later results with the same group’s earlier ones*.

When challenged, CentreForum found a 2006 DfES report which showed White British pupils doing 'relatively well' in 2004.   White British pupils were ranked fifth on the Early Years Communication scale (based on a 10% representative sample).  

Comparing the 2015 GCSE results as given in School Performance Tables (not the hypothetical ‘world standard’) to this Early Years ranking, White British pupils fell five places.  Nevertheless, 57.1% of White British pupils achieved the benchmark 5+ GCSEs (or equivalent) A*-C including Maths and English in 2015 - the same proportion as 'Any Other Ethnic Group'.   That proportion is exactly the average for ALL pupils in England’s secondary schools.  

White British pupils make up 72.2% of children in English secondary schools.  Yet David Laws described the performance of the majority of 16 year-olds in 2015 as ‘very bad…versus other ethnic groups'.  But there were eight ‘other ethnic groups’ where the proportion reaching the GCSE benchmark was lower than that of White British.   If Laws’ logic is applied, then the performance of White British pupils versus these  eight groups would be ‘very good’.

The ‘other ethnic groups’ each comprise just a small proportion of pupils in English secondary schools.  Pupils from a Chinese background, for example, account for just 0.4%, Indian 2.8%, White Irish 0.3%, Any Other Ethnic Group 1.5%.   This raises the question of whether contrasting small groups with a very large one results in a statistically valid conclusion.  And that’s without considering other variables such as parental occupation, socio-economic background and maternal education achievement.

I’m not a statistician but there are sufficient concerns about the CentreForum report, its theoretical ‘world standard’, its retrospective application of this measure and its faulty cohort analysis to undermine David Laws’ assertion that White British pupils perform very badly – a statement that has led to hysterical claims that White British parents are to blame and that the education system has betrayed ‘white’ pupils.

*ADDENDUM 13 April 2016 08.57.   Tracking a cohort's progress over time is not without problems.  The cohort isn't stable: pupils come and go.  But if a group's 'journey' is to be tracked over time to GCSE, then it makes more sense to compare the cohort at age 5 and age 16.



Share on Twitter Share on Facebook

Be notified by email of each new post.

Add new comment

Already a member? Click here to log in before you comment. Or register with us.