Foreword and executive summary

The Education Policy Institute’s flagship annual report has now been tracking changes in the attainment gap between disadvantaged children and young people and their peers for a decade.

To begin with, the story it told was one of relative success. The gap was slowly narrowing, year-on-year. That progress was uneven, but heading steadily in the right direction.

However, in the years that followed, progress began to falter. In 2019, we estimated that it would take 500 years to close the disadvantage gap. In 2020 – based on data from before the pandemic – we reported that the gap had stopped closing, and in fact had started to widen.

Then, of course, Covid-19 hit, with its devastating impact on children and young people – particularly the most disadvantaged. Research we undertook in 2022 reached the dismaying conclusion that much of the hard-won reduction in the disadvantage gap over the last decade had been reversed during the pandemic.

The years since have seen some progress but, worryingly, this year’s report tells a different story again. Across all of early years, primary, secondary and post-16, the gap has at best plateaued, and in some cases started to widen once more.

In this context, the government’s ambitious target to halve the disadvantage gap, as set out in its schools white paper Every child achieving and thriving, could not be more important – or EPI’s tracking of the gap more essential.

Narrowing the disadvantage gap, and improving the lives and life chances of those children and young people facing the greatest challenges, is the defining educational imperative of our age. I hope this tenth edition of EPI’s annual report will serve as a rallying cry to government, to schools, colleges and everyone who dedicates their lives to children and young people to continue to do everything we can to rise to that challenge.

Julie McCulloch, Chief Executive, Education Policy Institute

 

Executive summary 

For this year’s report, we have undertaken our usual forensic analysis of the disadvantage gap as a whole, across different regions, and for children and young people with particular protected characteristics, vulnerabilities and challenges. We use these findings to make recommendations for where policy action is most urgently needed, and to suggest areas where more research could improve our collective understanding of the reasons for the gap, and how it can be addressed.

Methodological note

This year’s report introduces an updated measure of disadvantage based on whether a learner has ever been eligible for free school meals (‘Ever FSM’), replacing our previous six-year measure (‘FSM6’). This provides a more stable basis for tracking disadvantage over time, while having little impact on the size of the gap in earlier years. 

This change reflects growing limitations in the FSM6 measure, which has become less reliable for tracking trends following changes in free school meal eligibility with the roll-out of Universal Credit. Using Ever FSM provides a more robust basis for monitoring whether changes in the gap reflect changes in outcomes rather than changes in the measure itself. Importantly, this change in the measure does not explain the widening gaps seen between 2024 and 2025 which is evident across multiple attainment measures. 

Top-level findings

  • Any post-pandemic recovery appears to have stalled or reversed. Even before the pandemic, the long-term trend of gap-narrowing had faltered and has since gone into reverse. By 2025 (the data on which this report is based), disadvantage gaps remained wider than before the pandemic at every phase, but particularly in the early years. Compared with 2019, the disadvantage gap increased by 17 per cent in the early years, 9 per cent at key stage 2, 5 per cent at GCSE and 2 per cent in 16-19 education. Concerningly, any apparent post-pandemic recovery appears at best to have stalled: between 2024 and 2025, gaps widened again in the early years and at key stage 4, and plateaued at key stage 2. The slight fall in the 16-19 attainment gap is less positive than it might appear, masking as it does the fact that increasing numbers of young people are not captured in this measure due to rising non-participation after age 16.  
  • Geography matters. Where disadvantaged children and young people grow up continues to have a profound effect on their educational outcomes. While London’s success for its disadvantaged learners is well established, the West Midlands also stands out: it had the second-smallest gap at every phase in 2025 and has demonstrated remarkable resilience since 2019, reducing its disadvantage gaps at both key stage 2 and 16–19 against a backdrop of national gap-widening. 
  • Fewer disadvantaged young people are participating in post-16 education. Although the measured 16-19 attainment gap remains close to pre-pandemic levels, this masks a widening participation gap, as disadvantaged young people become increasingly less likely to remain in education or training after age 16. Whilst participation rates for non-disadvantaged young people have remained relatively stable in recent years, almost one-quarter of disadvantaged young people are now not studying towards a substantial qualification or apprenticeship after key stage 4.

Socio-economic disadvantage 

  • The socio-economic disadvantage gap remains one of the defining challenges facing England’s education system, with disadvantaged children and young people continuing to have substantially lower educational outcomes than their peers, and the gap widening steadily as children progress through every stage of education. 
  • The recovery from the pandemic has gone into reverse. By 2025, disadvantage gaps remained wider than in 2019 at every phase, with particularly large increases in the early years – the only cohort not in education at the time of the pandemic – consistent with concerns highlighted in previous EPI research about the impact of the pandemic on young children. Progress also appears to have stalled, at best, in the most recent year: between 2024 and 2025, the gap widened further in the early years and at key stage 4, and remained unchanged at key stage 2. 
  • The relative stability in the measured 16-19 attainment gap masks a widening participation gap. Among young people who finished GCSEs in 2023, 23.0 per cent of disadvantaged young people were not studying towards a substantial qualification or apprenticeship at the beginning of year 12, compared with 9.7 per cent of their non-disadvantaged peers. This gap has widened from 10.5 percentage points in 2020 to 13.3 percentage points in 2023. 
  • The move to ‘Ever FSM’ in this year’s report marks a transition towards more robust measures of socio-economic disadvantage. For the first time, we no longer report on our persistent disadvantage measure because changes to FSM eligibility have made it increasingly difficult to interpret changes over time, though recent EPI research reinforces why this vulnerable group remains of major concern.

Change in disadvantage gaps since before the pandemic

Disadvantage gap by phase Months of learning Grades
Early years Key stage 2 Key stage 4 16-19
2019 4.2 9.2 18 3.3
2024 4.7 10.0 18.8 3.4
2025 4.9 10.0 19.0 3.4
Change, 19-25 (% change) +17% +9% +5% +2%
Change, 24-25 (% change) +5% 0% +1% -1%

Special educational needs and disabilities  

  • Children and young people with education, health and care plans (EHCPs) have some of the widest attainment gaps in England’s education system, while those receiving SEND support also remain substantially behind their peers. This is despite long-term progress in narrowing these gaps at every phase except in the early years. 
  • The SEND gap at age 5 remains a particular cause for concern, with many children with SEND beginning school already significantly behind their peers. Despite narrowing at older ages, gaps for pupils on SEND support in reception year remained above pre-pandemic levels in 2025 and, in the case of pupils with EHCPs, remained at their highest levels on record.  
  • Alongside rapid increases in the proportion of children and young people with SEND over the last decade, there have been substantial changes in need types, with an 8 percentage point increase in the share of pupils with autism and a 12 percentage point decrease in the share recorded with moderate learning difficulties since 2016. These shifts are likely to be influencing trends in overall SEND attainment gaps, with new EPI research showing that pupils with different types of need experience widely varying attainment outcomes. 

Change in SEND gaps since before the pandemic

SEND gap by phase Months of learning Grades
Early years Key stage 2 Key stage 4 16-19
EHCP SEND support EHCP SEND support EHCP SEND support EHCP SEND support
2019 19.7 11.8 28.1 18.4 41.1 24.4 7.2 4.2
2024 20.1 12.6 27.2 16.8 39.6 21.8 6.8 3.5
2025 20.1 12.3 27.2 16.2 38.9 21.1 6.8 3.4
Change, 19-25 (% change) +2% +4% -3% -12% -5% -14% -6% -20%
Change, 24-25 (% change) 0% -2% 0% -3% -2% -3% 0% -3%

Ethnicity  

  • Attainment continues to vary markedly by ethnicity and these patterns shift substantially as pupils progress through the education system. Although most minority ethnic groups start school behind White British pupils, many catch up, and by the end of secondary school outperform them. Black African, Pakistani and Bangladeshi pupils stand out, having already overtaken White British pupils by age 11. 
  • Once again, Chinese and Indian pupils are among the highest-attaining groups throughout education, while Gypsy Roma and Traveller of Irish Heritage pupils remain the lowest, followed by Black Caribbean and Mixed White and Black Caribbean heritage pupils.  
  • Between 2019 and 2025, most ethnic groups improved their attainment relative to White British pupils, though there has been no progress in gap-narrowing among White and Black Caribbean pupils at any phase. More recent changes, however, show that some of these gains began to reverse at GCSE between 2024 and 2025. 
  • Disadvantage affects ethnic groups differently. White Irish and White British pupils consistently experience some of the largest disadvantage gaps, while the attainment of Chinese, Black African, Pakistani and Bangladeshi pupils is much more resilient to the effects of disadvantage. The performance of disadvantaged Chinese pupils is especially striking: at both key stage 2 and key stage 4, they outperform non-disadvantaged pupils from every other ethnic group.

Figure ES1: Nearly all ethnicities have improved their GCSE attainment relative to White British students between 2019 and 2025, though this has started to reverse since 2024 

 


 

Gender 

  • At age 5, girls outperformed boys in all seventeen early learning goals in 2025 and continued to outperform boys in most subjects at later phases. While this results in large gaps in girls’ favour, the gender attainment gap is unusual in narrowing as pupils progress through primary school, partly because boys overtake girls in mathematics by age 11. 
  • While gender attainment gaps are long-standing across phases, the recent progress boys have made in their attainment has resulted in the overall gender gap falling to a historic low at GCSE in 2025 and being wiped out at age 11. Meanwhile, girls’ attainment has largely plateaued at key stage 2 and slightly declined on all official key stage 4 measures in 2025, with girls’ Attainment 8 score reaching its lowest level since the measure was first introduced. 

Figure ES2: Girls’ Attainment 8 has slightly declined while boys’ has improved, causing the Attainment 8 gender gap to reach its narrowest point since the series began in 2019

 


 

 

 English as an Additional Language 

  • Attainment varies considerably by first language and time of entry into the English state school system. Pupils arriving late to the English state school system with English as an additional language (EAL) have large attainment gaps compared to their peers. These gaps have been on a long-term downward trend, though there are some signs this has plateaued or reversed in 2025. 
  • Among pupils in reception year with EAL, the gap further narrowed in 2025 to 1.4 months, its lowest on record since the start of the series in 2013. 
  • However, this progress was not replicated among older pupils arriving late to the English state school system with EAL, who fell further behind their peers in 2025 at both key stage 2 and key stage 4, for the first time since 2018. 
  • While this appears to be explained by recent changes in the ethnic background of the late-EAL group at key stage 2, even when we account for this at key stage 4, there is a marked deterioration in their GCSE outcomes of over one month. This aligns with broader patterns of declining GCSE attainment across minority ethnic groups in the 2025 data. 

Geography 

  • This report features interactive geographic tools to help users explore how disadvantage gaps vary across England. For each phase and geographic level, we compare the attainment of disadvantaged pupils locally to the attainment of non-disadvantaged pupils nationally, allowing comparisons to be made across England. 
  • Once again, London continues to outperform all other regions for its disadvantaged learners in 2025, while the West Midlands consistently has the second smallest gap across phases. 
  • Looking at changes in the gap between 2019 and 2025, the West Midlands has remained one of the most resilient regions, while pupils in the South East and South West have experienced some of the largest increases in the gap since the pandemic.  
  • London’s exceptional performance should not be taken for granted. Although it continued to have by far the smallest GCSE disadvantage gap, London’s gap has increased by more than in any other region between 2024 and 2025, contributing to the national widening in the gap.

Change in regional gaps since before the pandemic

Regional gap by phase Months of learning Grades
Early years Key stage 2 Key stage 4 16-19
2025 gap Change since 2019 2025 gap Change since 2019 2025 gap Change since 2019 2025 gap Change since 2019
London 3.4 +0.8 5.6 +0.4 10.8 +0.5 1.4 -0.4
West Midlands 4.4 +0.2 9.2 -0.4 19.3 +0.5 3.2 -0.2
East Midlands 5.3 +0.6 11 +0.3 20.9 +1.1 3.9 -0.2
North East 5.5 +1.4 9.8 +0.9 20.5 -0.1 4.4 +0.3
North West 5.6 +0.8 10.5 +1.0 20.6 +0.4 3.4 +0.3
Yorkshire & The Humber 5.4 +0.6 10.9 +0.6 20.7 +0.8 4.0 +0.3
East of England 5.7 +1.2 11.7 +0.9 20.7 +1.2 4.0 +0.1
South East 5 +0.8 12 +1.4 22.2 +2.0 4.1 +0.2
South West 5.4 +0.5 12.5 +1.5 21.3 +0.6 4.4 +0.2
  • Local authority disadvantage gaps are even more variable. By age 16, the difference between the best-performing local authority (Newham) and the worst (West Berkshire) is equivalent to almost two years of learning. Such wide geographic variation indicates there is scope to learn from the highest performers – including areas outside of London such as Birmingham – at a time when the government is trialling place-based approaches to improving educational outcomes in Scarborough, Hastings and the North East. 
  • In every local authority in England, the disadvantage gap widens between ages 5 and 16. Although some local authorities are remarkably successful in limiting this growth (such as Westminster and Newham), many more demonstrate how disadvantage compounds at each stage, even in areas with relatively few disadvantaged pupils (such as West Berkshire, and Bath and North East Somerset).   

Policy recommendations 

The scale and persistence of the gaps highlighted in this report underline the need for a coherent strategy to reduce educational inequality from the early years through to post-16 education. The government’s ambition to halve the disadvantage gap is welcome, but meeting it will require sustained investment, more sophisticated use of data, and targeted support at the points where disadvantage is most likely to become entrenched.  

Our recommendations focus on the areas where policy action is most urgent: early development, SEND, post-16 participation and funding, persistent disadvantage, and the measurement of disadvantage itself. They focus either on further changes we believe are required, above and beyond those to which the government has already committed in its recent schools and skills white papers, or on ensuring current proposals are implemented in the most effective way.  

Early years  

The early years remain the critical point at which educational inequalities become embedded. With recent EPI research showing that 40 per cent of the age 11 disadvantage gap is already established by age 5, supporting children’s early development – including in literacy, mathematics and communication – will be central to achieving the government’s ambition to halve the gap.  

The government’s recent focus on early years is therefore welcome. This should be strengthened as follows:  

Recommendation 1: Equalise access to funding for early education and care, ensuring that all children receive the same entitlements to funded hours regardless of their family circumstances. This is a vital first step toward fixing the inequalities in the current funding system, which has shifted away from universalism and prioritised working families to the detriment of children in disadvantaged families with parents who are out of work. 

Recommendation 2: Refocus attention on prioritising quality in the early years to ensure all providers in the complex, ‘mixed market’ system deliver developmentally appropriate and nourishing care and education. Specifically, the government should take steps to mitigate the risks of early years expansion relying on poorer quality private providers and address the shortcomings of the current Ofsted regulatory system. The Competition and Markets Authority (CMA)’s review of the early years market is a welcome first step toward much-needed systematic improvement underpinned by child-centred care, open-ended play and holistic learning in the early years.  

SEND   

Reforming SEND support is critical – both in its own right and as part of a credible disadvantage gap strategy. Previous EPI research has shown that SEND is under-identified in many vulnerable groups and, where it has been identified, is making a central contribution to the disadvantage gap.  

Again, the government’s current focus on SEND reform is welcome. We would specifically encourage the government to:  

Recommendation 3: Invest sufficiently in the workforce required to support and develop children and young people with SEND, both within and beyond school and college. Specifically, introduce foundational knowledge of child development and how to respond to atypical development, particularly among children with neurodevelopmental conditions, as a mandatory part of initial teacher training and early career development; increase the number of educational psychologists and speech and language therapists; and improve access to paediatric services and Child and Adolescent Mental Health Services. 

Recommendation 4: Introduce proactive screening for SEND, targeted at groups known to be at risk of unmet or late-identified SEND needs, including children with social workers, children looked-after, children with emerging attendance problems, those who have been suspended during primary school, those who have moved school frequently, and those who speak English as an additional language.  

Recommendation 5: Introduce a ‘strengths and needs profile’ approach to SEND assessment, informed by research into how needs cluster within and between developmental domains. Our ‘Rethinking SEND’ research suggests that this would help proactively identify how children’s needs group across different areas of development, including where children have needs in more than one area, rather than relying on specific diagnoses. 

Recommendation 6: Build on the recent Curriculum and Assessment Review by undertaking a thorough evaluation of where curriculum, assessment, qualifications and school accountability measures are acting as obstacles to meeting children’s needs and securing better long-term outcomes, particularly for the most vulnerable. This should ensure age-related expectations, assessments and qualifications better reflect child development, recognise progress for children with SEND, support personal, social and emotional development, and create meaningful pathways into post-16 education and adult life which are recognised within the key stage 4 accountability system. 

Post-16   

The gap in participation in post-16 education between disadvantaged young people and their peers requires urgent action. The recent interim report of the Milburn Review into young people and work frames rising NEET numbers not as a failure of aspiration, but as a failure of the systems around young people. Yet the transition into post-16 education, where risks such as low attainment, absence, SEND and poor mental health can have lasting consequences for young people’s future education and employment, is also the point at which targeted disadvantage funding falls away. 

The government should therefore:  

Recommendation 7: Introduce a 16-19 student premium: a targeted funding entitlement for disadvantaged young people in the 16-19 phase. As set out in EPI’s 2024 report, Closing the Forgotten Gap, this would address this funding cliff edge and support providers to act before educational disadvantage becomes long-term labour-market exclusion. This proposal, for which we and others have been calling for many years, is now urgent. 

Persistent disadvantage  

Although changes to the way in which persistent disadvantage (children who have been eligible for free school meals for at least 80% of their school career) is identified have meant we are unable to track this measure in this year’s report, the underperformance of pupils in this group remains a major concern.  

Our previous reports make it clear that persistently disadvantaged pupils consistently demonstrate some of the lowest educational outcomes in the system, falling, on average, nearly two years behind their peers by the time they sit their GCSEs. And while there was progress in gap-narrowing for the headline gap leading up to the pandemic, there was no long-term progress in narrowing the persistent gap, which has remained stubbornly high over successive years.  

We welcome the proposal, in the schools white paper to use household income to vary funding rates according to the level of disadvantage, so that children with more entrenched disadvantage attract more funding. How this proposal is implemented, however, is crucial to its success. The government should:  

Recommendation 8: Introduce an enhanced pupil premium, targeted specifically at persistently disadvantaged pupils. This must be in addition to the existing pupil premium, not a redistribution of it. At the same time, real-terms cuts to the pupil premium over the last decade should be reversed, to bring it back in line with inflation. This could be funded through savings made by falling pupil numbers.  

Data and measurement     

It is essential that the system as a whole is able to track disadvantage attainment gaps over time. The widely used FSM6 measure of disadvantage is no longer a reliable way to do this. The government’s plan to allocate disadvantage funding using household income rather than FSM eligibility is welcome and should provide a more accurate basis for identifying need, but it is crucial that this change doesn’t make it more difficult to assess progress in narrowing the gap. It is also important that schools and colleges have timely access to all relevant information on their learners to better support their educational journeys.   

In addition, the schools white paper signals a broader conception of educational success, with all schools expected to monitor pupils’ sense of belonging and engagement by 2029. This is a welcome development, given the links between pupils’ experiences of school, attendance, participation and attainment. It is important that we are able to track progress against this ambition at a national level. 

We encourage the government to ensure the proposals it has made in these areas are effective by taking the following actions:  

Recommendation 9: Develop robust statistical measures of educational inequality – including for persistently disadvantaged learners – that reflect the proposed changes to the way disadvantage funding is allocated, so that progress towards halving the disadvantage gap can be assessed transparently and consistently over time. 

Recommendation 10: Ensure schools and colleges are able to access detailed and timely data about their learners to enable early identification of at-risk learners and to implement more targeted support, particularly at the key transition points of 11 and 16. As a first step, ensure the Get Information about Pupils (GIAP) service is updated in a timely manner, and encourage its consistent use by schools and colleges.  

Recommendation 11: Develop wider measures of wellbeing, belonging and engagement alongside academic outcomes, particularly for vulnerable learners, to inform future assessments of educational inequality.

Future research

This report identifies important trends in educational inequalities, but also highlights areas where further research is needed to understand the drivers of these gaps and inform future policy. Priorities include: 

  • Understanding how changes in the profile of SEND over the past decade are influencing attainment gaps for SEND learners, including how outcomes vary by type of need and level of support.  
  • Understanding the relationship between unmet SEND, absence, suspensions, exclusions and school mobility, to better gauge where earlier intervention could prevent pupils from falling behind or disengaging from education. 
  • Understanding the drivers of the post-16 participation gap between disadvantaged and non-disadvantaged young people, including the role of peer and institutional context. Further research should examine the extent to which differences in participation reflect prior attainment, SEND, absence, exclusions, local provision, qualification pathways and wider barriers to participation, as well as the schools, colleges and peer groups that young people experience. This would help to identify where policy intervention is most likely to prevent disadvantaged young people from falling out of education or training after age 16. 
  • Understanding why disadvantage gaps vary so widely between places, and what this means for the design of place-based education policy. This should include assessing the nature of disadvantage gaps in areas currently prioritised by government, including Scarborough, Hastings and the North East, and evaluating the impact of these interventions over time. Further research should distinguish the role of local demographics, including ethnicity, first language and SEND, from factors that may be more amenable to policy, such as early developmental gaps, pupil absence and access to specialist support. 
  • Monitoring whether the recent decline in GCSE attainment among some ethnic minority groups represents the start of a longer-term trend. Further work should also consider whether this is linked to London’s marked GCSE gap-widening between 2024 and 2025. 

 

Explore the report