When it comes to opportunity, we won’t entrench the advantages of the fortunate few, we will do everything we can to help anybody, whatever your background, to go as far as your talents will take you.
Theresa May, 13th July 2016

So what should we expect to see in the education sphere if this rhetoric is to be turned into reality?   This is a difficult question to which researchers do not yet know all the answers.  The mechanisms through which educational disadvantage is transmitted are many and complex.  It is unlikely that there is a single solution waiting to be uncovered.

However, we do know quite a lot about what won’t work or isn’t sufficient to close the gap, and should be wary of a programme that is restricted to well-worn assumptions and minimises the importance of developing better skills and human capital.

  1. More grammar schools could make the gap worse

Since 1998 it has been illegal to open any new selective schools, but recently a satellite site located nine miles from an existing grammar school was approved by DfE, under the aegis of the expansion of successful schools. Nominally, this is not a new school, but many are keen to see further expansions, or to change the legislation to allow new grammar schools to open.

Very few disadvantaged children are admitted to grammar schools.  This is not because of the limited number of grammar schools; it is because disadvantaged children are already almost 10 months behind other children in development by the age of 11.[1]  A child in a wholly selective area who is not eligible for free school meals at the start of primary school is three times as likely as a free school meals child (30 per cent vs. 10 per cent) to achieve in the top quarter of pupils at age 11.[2]  Clearly, intervention needs to start much earlier than this.

Expansion of selective education is more likely to increase attainment inequality than reduce it.  While selection generates small gains to the most-able children, it creates larger losses to less-able children.[3] And the effects reach beyond school.  People who grew up in selective areas experience greater earnings inequality, after accounting for individual and area differences.  The gap in earnings between those at the 90th percentile and those at the 10thpercentile is £10.93 an hour in comprehensive areas, but £13.14 an hour in selective areas.[4]

  1. Getting parents into work hasn’t made the gap close

Universal Credit is designed around the principle that you should always be better off in work than out of work.  Other elements of the system include personal work coaches and tax-free childcare allowances to encourage starting work and increasing working hours.

Employment is a positive outcome, but it is not sufficient to close the gap.  We know this because the percentage of children in workless families has decreased dramatically over the last twenty years, but the attainment gap is stubbornly persistent in spite of this.

The IFS released figures showing that 23 per cent of children lived in workless families in 1994-95 and this has fallen to 15 per cent in 2014-15.  The decreases were concentrated in the bottom two fifths of the income distribution for families with children; in the lowest-income fifth, the decrease was from 60 per cent to 37 per cent.[5]

Perhaps the problem is families in which nobody has ever worked?  But only 1.8% of children in the UK live in non-student households where no adult has ever worked.  A top-end estimate for the number of these children in the 2015 GCSE cohort in England is 10,000 (it is likely to be fewer because worklessness is lower in England than the rest of the UK, and younger children are likely to be overrepresented).[6]

To close the gap in pupils achieving 5 good GCSEs including English and maths would have required an extra 42,000 disadvantaged pupils to reach this standard in 2015.[7]  There are not enough children who have never seen a parent go to work to close the gap, even if every family entered work, and every child in those extra working families went from not passing to passing.

  1. New investment in ‘universal’ childcare isn’t reaching the poorest

The government currently plans to implement the 30-hour childcare offer for children aged 3-4 in 2017.  The additional hours, available to parents earning above an income threshold equivalent to 16 hours per week each at the national living wage, are expected to cost over £1bn per year by 2020, but our analysis shows it will not benefit the most disadvantaged children.[8]  The policy is effectively a cash transfer to better-off families many of whom already access additional paid childcare, under the guise of universal provision.

Offering additional childcare, which presumably holds some educational value, to all but the neediest, at significant cost to the tax-payer, isn’t the worst of this policy.  The hourly costs paid by government may well be too low to support an expansion of places, resulting in pushing disadvantaged children to the back of the queue, and/or damaging the quality of the 15 hours they are entitled to.[8]

Disadvantaged 2-year-olds are eligible for 15 hours of free childcare, representing targeted offer for poorer families. But by January 2016, the proportion of eligible children who had taken up this offer was only 68 per cent.  DfE statistics suggest that almost 80,000 eligible children are missing out on this entitlement, even after a decrease in the number of children who are eligible of 24,000 since 2015.[9]

  1. The social mobility case for more academies is limited, especially for current plans

Having stepped back from forced academisation, the previous secretary of State intended to bring about near-universal academisation by means of performance and viability thresholds for local authorities, below which schools would not be allowed to remain local authority maintained. This is in addition to the Education and Adoption Act, following which all inadequate schools will be taken over and action taken will be taken in ‘coasting’ schools.

This month there has been a wave of new research into the effectiveness of academy trusts and the programme as a whole.  Three separate reports are unanimous in their conclusion that performance by academies is highly inconsistent. EPI’s School performance in multi-academy trusts and local authorities found that the performance of schools in MATs is no better on average than that of schools maintained by local authorities.[10]

Looking at the performance of academy chains on disadvantaged pupils in particular, the Sutton Trust’s Chain Effects 2016 reported that 13 out of 39 chains performed significantly below average for their disadvantaged pupils, and the proportion of chains performing above-average for these pupils falling from year to year.[11]

Further LSE research commissioned by EPI confirms that converter academies that were good or satisfactory on their Ofsted rating prior to conversion (as opposed to sponsored academies and those already rated outstanding) show no evidence of improvement resulting from conversion to academy status.[12]

Positive impacts were found for sponsored academies opened under the Labour government, when the policy was more focused on disadvantage, and funded more intensively over smaller numbers of schools.  Research is underway to test the effectiveness of primary academies and of sponsored academies that opened after 2010.  Concerns about the latter have been voiced by key providers; in evidence to the Education Select Committee, Lucy Heller, Chief Executive of Ark, commented:

We have to remember that many of the MATs have taken on, as we have, some of the most challenging schools in the country. It can take a long time to turn around. It is much tougher now than it was at the beginning, simply because of the change in funding. I would certainly want to say that you have to look carefully at each situation. It is very nuanced, but if there is consistent underperformance then there is a question about why that MAT should continue to be operating those schools.[13]

  1. 90 per cent Ebacc entry is trying to run before we can crawl

In November 2015, the government consulted on its proposal for a 90 per cent entry target for the English Baccalaureate, which requires pupils to enter at least seven GCSEs spanning five mandatory subject areas.  No government response has yet been published.

The Sutton Trust has analysed 300 schools that increased their EBacc entry rates the most immediately following the introduction of the measure to school performance tables.[14]  These schools were by definition unusual, and with the fairest wind possible, the effect on the English and maths GCSE gap was a reduction of 6% of one grade in each subject.  The gap in English and maths was equivalent to 1.2 grades at the time these effects were measured, so it would not be credible to claim a decisive role for the EBacc in closing the gap even if this effect could be replicated in all schools, which is extremely unlikely in the current teacher recruitment climate.[15]

The report’s author, Dr Rebecca Allen, recommends that the 90 per cent entry target is reconsidered.  The 300 most responsive schools increased their entry to an average of 48% without any evidence of negative impacts on overall achievement.  This tells us little about what would happen at 90%, given that this would take many schools deep into the percentage of pupils who do not currently achieve grades A*-C in English and maths, or do so only by the smallest margin.

To be clear, expectations for disadvantaged pupils have been set too low, too often and for too long.  Increases in meaningful outcomes including achievement of the EBacc are welcome, but the 90 per cent entry target is trying to run before we can crawl.  Fewer than half of the 35,000 disadvantaged pupils who entered in 2015 achieved the EBacc.[16]  An accountability measure is not a magic wand.  If we want all children to have a language GCSE, that’s a legitimate aim, but let’s begin by teaching foreign languages from key stage 1, not by entering pupils into qualifications in the belief that we have only to wish it to make it so.

So if these policies do not provide the answers, where should we look for better chances of success? It’s important not to pretend that we have all the answers, and to acknowledge that the devil is almost always in the implementation of even the best-evidenced interventions.  However, there are some clear threats and opportunities that suggest better options for where to put the next pound of educational investment in the quest to improve social mobility.

 

We recommend more research to support policy expansion in the following areas:

  1. Targeted early years provision focused on educational quality and parenting support

Around 40 per cent of the disadvantage gap at age sixteen is already present by age five; while the gap at this age is falling, it still stands at over 4 months of development as measured by the early years foundation stage profile, and is much larger when measured in terms of vocabulary.[17]

Economic research from around the world indicates that targeted investments in disadvantaged children during the early years of childhood are the most cost-effective form of public investment in human capital.[18]  Lasting improvements in cognitive ability as measured by IQ can be achieved if disadvantaged children are targeted for educational support before the age of three.  We also know that programmes which include support for better parenting skills are the most effective at improving outcomes for children.

  1. Providing teachers with time for CPD, professional collaboration and research engagement

Researchers from Cambridge University have found that English teachers work more hours than teachers in other countries and are less satisfied with their jobs.  On top of this, more advantaged schools were found to have teachers that are slightly more experienced, more qualified and spend more time teaching as opposed to undertaking classroom management and administration.[19]

Turning to how we might tackle the uneven distribution of teacher quality across schools, it is evident that policies to spread high quality teaching need to address the development of existing teachers as only around 10 per cent of teachers exit and enter the system each year.  Almost half of teachers surveyed by NfER reported that more time for continuous professional development (15%) or fewer contact teaching hours (33%) would encourage teachers to choose jobs in challenging schools.19

  1. SEN support, particularly targeting developmental delay and behaviour difficulties

Disadvantaged children have an elevated risk of experiencing a number of environmentally-linked types of special educational need. Without appropriate support, it is very likely that their educational achievement will be compromised.

For example, the ONS mental health of children and young people in Great Britain survey in 2004 found that 58 per cent of children with conduct disorders lived in low-income families compared with 33 per cent of other children.[20]  Research by Save the Children using Millennium Cohort Study data found that 18 percent of children in the lowest income quintile had language development at least one standard deviation below the mean (a reasonable proxy for developmental delay), compared with just 3 per cent of children in the top income quintile.[21]

According to the latest DfE statistics on special educational needs in England, social, emotional and mental health needs and speech, language and communication needs were among the most common types of SEN, but were among the least likely to result in an education health and care plan (or statement under the old system), with 85 per cent and 86 per cent of cases receiving lower level support.[22]  This may or may not be ‘the right level’, but this is a question well worth investigating.

 

 

[1] Perera, N. and Treadaway, M. (2016), Education in England: Annual report 2016.  Education Policy Institute.

[2] Author’s calculations from National Pupil Database.

[3] Cook, C. (2016), Why not bring back grammar schools? BBC news: http://www.bbc.co.uk/news/education-36662965

[4] Burgess, S., Dickson, M. and Macmillan, L. (2014), Selective Schooling Systems Increase Inequality. Institute of Education Working Paper 14-09.

[5] Belfield, C., Cribb, J., Hood, A. and Joyce, R. (2016), Living standards, poverty and inequality in the UK: 2016. Institute for Fiscal Studies.

[6] Office for National Statistics (2015), Working and Workless Households: 2015.

[7] Department for Education (2016), Revised GCSE and equivalent results in England: 2014 to 2015. SFR 01/2016.

[8] Johnes, R. (2016), Widening the gap? The impact of the 30-hour entitlement on early years education and childcare. Education Policy Institute.

[9] Department for Education (2016), Education provision: children under 5 years of age, January 2016. SFR 23/2016.

[10] Andrews, J. (2016), School performance in multi-academy trusts and local authorities – 2015. Education Policy Institute.

[11] Hutchings, M., Francis, B. and Kirby, P. (2016), Chain effects 2015: The impact of academy chains on low-income students. The Sutton Trust.

[12] Silva, O. et al. (2016), The Impact of Academies.  Presentation slides available at:  https://epi.org.uk/wp-content/uploads/2018/02/academies-summit-presentations.pdf

[13] House of Commons Education Committee (2016), Oral evidence: Multi-Academy Trusts, HC 204.

[14] Allen, R. and Thompson, D. (2016), Changing the subject: How are the EBacc and Attainment 8 reforms changing results? The Sutton Trust.

[15] Department for Education (2015), Measuring disadvantaged pupils’ attainment gaps over time (updated). SFR40/2014.

[16] Department for Education (2016), Revised GCSE and equivalent results in England: 2014 to 2015. SFR 01/2016.

[17] Perera, N. and Treadaway, M. (2016), Education in England: Annual report 2016.  Education Policy Institute.

[18] Heckman, J. (2016), The economics and econometrics of human development and social mobility. CEMMAP and CPP.

[19] Ilie, S., Jerrim, J. and Vignoles, A. (2016) Teacher attitudes and characteristics in English schools: Evidence from TALIS. And NfER (2016), Polling data.  Available at: http://www.summit-bestinclass.com/report/

[20] Office for National Statistics (2005), Mental health of children and young people in Great Britain, 2004: Summary report.

[21] Law, J. et al. (2013), Early language delays in the UK. Newcastle University and Save the Children.

[22] Department for Education (2016), Special educational needs in England: January 2016. SFR 29/2016.