Tackling the Behaviour Challenge in England’s Schools

For several decades, successive governments prioritised school standards following years of relative neglect. Their core and shared focus on accountability and rigour led to some successes. English pupils are, on average, among the best in Europe for reading, mathematics and science, bucking the global trend of declining international test scores.^[_] Yet serious flaws remain^[_] and deep inequalities cast a long shadow. The learning gap between disadvantaged pupils and their peers, for instance, now stands at 19 months by age 16.^[_]

Schools in England now also face a burgeoning crisis of behaviour and safety, which threatens both to erode the progress that has been made and to further compound the problems that remain.

The scale of the issue is daunting and suggests that an alarming number of schools are not the safe bastions of learning they ought to be. More than a quarter of secondary-school pupils report being bullied in the past year. More than a third of teachers experienced physical abuse or violence from pupils in the same time. Around a fifth of teaching time is lost to disruption in the classroom. And 30 per cent of teachers have had a parent undermine their sanction.^[_]

The ensuing damage is equally striking. A lack of grip over behaviour appears to be having a corrosive effect on retention, at a time when a third of teachers leave the profession after five years.^[_] More than twice as many teachers who occasionally or never feel supported to deal with disruptive behaviour consider leaving compared to those who feel always or mostly supported (41 per cent versus 20 per cent).^[_] When asked which factors are most important when considering leaving, behaviour is a top-three concern for teachers in state-funded schools.^[_] Without improving retention, the government’s pledge to hire 6,500 teachers will be like trying to fill a leaky bucket.

Poor behaviour is also crushing learning outcomes. Time lost to disruption is time lost to learning, with a knock-on effect to the UK’s future prosperity that may amount to as much as 1.5 per cent of GDP in the long run. And schools that fail to get a handle on behaviour perform worse on attainment measures such as GCSE results and Progress 8.

Pupils themselves say safety is a core requirement of a good school experience; our research shows it is more strongly linked to both attendance and aspirations for future study than other known factors such as family background or gender. In this light, it is all the more concerning that so many pupils are missing school: the rate of persistent pupil absence is almost twice what it was in 2018 and 1.4 million pupils are now missing 10 per cent of sessions or more, while 142,000 pupils are missing 50 per cent or more sessions.^[_]

Clearly, the scale and consequences of poor behaviour are unacceptable and demand a suitable response. Fixing this crisis of behaviour is foundational to improving educational outcomes and to realising the government’s “opportunity” mission. Yet the emphasis placed on it falls far short of what is required. It is time to tackle the problem head on, as a matter of urgent priority. The government’s response should comprise two sets of interventions.

First, the government should take steps to shift the current balance of power in teachers’ favour when engaging with parents and agencies beyond the school gate. This would help address the pressing issues with classroom disruption many teachers are facing today.

Many of the root causes of challenging behaviour lie outside the school gates, where the powers of schools to act are limited. This challenge has become more acute since Covid; the “unwritten agreement” between schools and parents has fractured, and more parents see adherence to school policies as optional.^[_] To make matters worse, the other agencies that schools could involve in dealing with severe issues are not always responsive and schools lack the power to drive multi-agency working themselves.

The relationship between schools, parents and agencies therefore requires a reset:

• Schools should be given the power to compel families to come to the table and agree an action plan when a pupil’s behaviour becomes severely disruptive.

• Where families refuse to engage, teachers need to be dealt a stronger hand. If parents fail to attend disciplinary meetings or engage with an action plan, schools should be able to escalate concerns to other relevant agencies on safeguarding grounds.

• The government should consider if any legal powers similar to those used to address repeated absence without a good reason should apply when parents refuse to cooperate repeatedly and without substantive grounds.

• Schools should also be given a stronger statutory role in driving multi-agency working, with powers to convene, co-ordinate and lead meetings where appropriate.

Second, the government should launch a National Behaviour Challenge to provide much better insight and targeted, contextually relevant support to schools that might be struggling with their approaches to behaviour. This would drive longer-term improvement and raise the standards for how the education system tackles the issue of safety and disruptive behaviour.

In the public imagination and on social media, debate around this issue often boils down to a simple choice between behaviour policies such as “silent corridors”. But no one kind of policy is a silver bullet.^[_] Experience and evidence show that clear expectations, consistent communication and fair enforcement of policies are what matter. Many of the tools that schools need to do this are in place already, but schools often struggle to make use of them or transpose them into approaches that work. And many teachers feel they are not sufficiently backed.

This is primarily a question of school management. In the past, systemic problems in education have been tackled successfully to achieve a rapid turnaround in schools’ fortunes. Focused, data-driven, peer-led programmes – such as the London Challenge, which helped transform state education in the capital – have achieved impressive results. A system-level intervention is now needed to tackle the behaviour crisis. A National Behaviour Challenge should be set up to cut classroom disruption by half and improve measures of wellbeing. Sponsored by the Department for Education and reporting into the “opportunity” mission board, it should:

• Collect aggregate data on behaviour and classroom disruption on a regular basis, via surveys and integration with school-management tools, to build an accurate present-time picture of behaviour and wellbeing in schools.

• Become the first user of a national data infrastructure centred on a digital learner ID, built and maintained by an independent designated data body. The aggregate data can also be used to inform the government’s new school “report cards”.

• Build strong relationships at the regional and local level, using data to inform a peer-led improvement and support system in which schools that struggle with behaviour are benchmarked and paired with others successfully tackling similar challenges – for example, via the newly announced Regional Improvement for Standards and Excellence (RISE) teams.

• Invest in generating evidence on what works in behaviour management in schools, making it widely available (in collaboration with the Education Endowment Foundation) and ensuring it feeds directly into both the initial teacher-training framework and the provision of continuous professional development.

• Implement data-driven ways of coordinating the provision of support to children outside the school gates, with much better multi-agency collaboration. Quality data and its effective, timely sharing should be at the core of safeguarding.

In its first King’s Speech, the new Labour government rightly sought to place a greater focus on wellbeing in schools. Where children do not feel safe, there can be no wellbeing. A good school system that unlocks opportunity for all pupils would combine high and rising standards of attainment with a safe learning environment free from disruption and fear of violence. Neither can be delivered without addressing the rising tide of challenging behaviour. This should be tackled as an urgent priority and made a central pillar of education policy.

It is not easy to define challenging behaviour objectively, as opinions about what this constitutes invariably differ. However, there is a point at which most objective observers would agree that certain behaviours are impeding pupils’ quality of learning. Viewed from multiple perspectives, as we show later in this report, it is hard not to conclude that this point has been far exceeded.

Pupils and Teachers Are Concerned About Personal Safety

Perhaps most pressingly, many schools cannot secure their pupils’ personal safety – the most fundamental requisite of any learning environment. According to the Department for Education’s (DfE) most recent National Behaviour Survey (NBS), which covers data in England up to May 2023,^[_] just 39 per cent of pupils reported feeling safe at school “every day” in the past week and more than one in four secondary-school pupils said they had been bullied in the past year.

Other sources reveal similar tendencies. For instance, according to a recent analysis by survey tool Teacher Tapp for BBC News, 30 per cent of teachers had witnessed pupils fighting during the week in which they responded. Two in five teachers had witnessed aggressively violent behaviour that needed an intervention in just one week.^[_]

It is not only pupils who are unsafe; many schools cannot ensure the wellbeing of their teachers either. A recent study by NASUWT – The Teachers’ Union found that in the 12 months prior to being asked, 37 per cent of the 6,586 teachers and headteachers it consulted had experienced physical abuse or violence from pupils, while 90 per cent had experienced verbal abuse or violence from pupils.^[_]

To provide a sense of some of the incidents recorded by NASUWT, in one case a chair was thrown at a teacher who was seven months pregnant. One teacher was sprayed in the face with an aerosol, triggering an asthma attack. Another was kicked unconscious. One teacher with a severe perfume allergy (anaphylaxis) was deliberately sprayed on three occasions.

There are also indications that violence and abuse may be rising even further: 89 per cent of teachers surveyed by NASUWT said the number of pupils exhibiting violent and abusive behaviours had increased, while 93 per cent reported that the number of pupils verbally abusing staff members had increased.^[_]

Disruption Levels Are High

It is not just the most egregious forms of abuse and violence that are apparent; persistent low-level disruption routinely features in schools, too. Viewed more generally, pupils’ behaviour is frequently reported as inadequate.

For example, according to government data, in May 2023 just 43 per cent of pupils (and 55 per cent of teachers) reported that pupils’ behaviour was either “very good” or “good” in the past week. In June 2022, the proportion of pupils (47 per cent) and teachers (64 per cent) reporting “very good” or “good” behaviour was higher, indicating behaviour may be deteriorating.^[_]

As a result, schools are too often fraught, turbulent environments. According to the same government data, just 54 per cent of pupils reported that their school had been calm and orderly “every day” or “most days” in the past week. This represents a decline on 2021/22, when 70 per cent of pupils reported that their school had been calm and orderly “every day” or “most days” in the past week.^[_]

Exclusion Rates Are High and Rising

Expulsions and suspensions are not always a reliable guide to pupils’ behaviour and should not be relied on exclusively to benchmark behaviour. This is because some schools appear to exclude pupils too readily to maintain more favourable performance statistics^[_] and, conversely, some schools hold back on exclusions due to fears they might undermine Ofsted judgements.^[_] However, exclusions remain a useful, albeit incomplete, bellwether when it comes to behaviour.

The figures make for sobering reading and further reinforce the urgency of the problem. According to official statistics,^[_] the annual number of permanent exclusions in state-funded schools is now the highest on record.^[_] Suspensions, too, are now also strikingly high. There were 787,000 suspensions in the 2022/23 academic year, while the rate of suspensions in the same year was almost 10 per cent overall and nearly 20 per cent in state-funded secondary schools.^[_]

While the pandemic caused permanent exclusions and suspensions both to fall, the most recent data show that rates in each case have rebounded strongly since.

Serious misconduct and routine disruption have no place in our schools. Keeping children safe in education is a fundamental requirement of the system and the prevalence of challenging behaviour is a serious threat to this. But challenging behaviour’s pernicious effects extend beyond safety alone; it also serves as a handbrake on opportunity.

Constant interruptions rob pupils of learning time and schools that struggle to get a grip on behaviour tend to perform worse on key current measures such as attainment, progress and inspection results. This loss of learning, as the analysis later in this report shows, also threatens the economic prospects of young people and risks costing the UK economy dearly for decades to come.

No less importantly, safety and behaviour are key components of student wellbeing and happiness; pupils who suffer the most from challenging behaviour in their schools are more likely to miss school, contributing to the attendance crisis. They are also less likely to want to go on to further study. In fact, TBI research explored in more detail below suggests wellbeing may be more strongly associated with both attendance and aspiration than family background.

Challenging Behaviour Correlates With Poor School Performance

It does not take much, or many pupils, to stall a classroom – even persistent low-level misbehaviour obstructs teaching. The frequency of disruption is very high. NBS data show that in May 2023, 76 per cent of teachers said misbehaviour had interrupted learning in at least some of their lessons in the week prior to being asked, while 37 per cent said all or most lessons were interrupted.^[_]

Disruption wipes out learning time. In England, on average, for every 30 minutes of lesson time, 7 minutes were lost due to misbehaviour in May 2023 and 5.5 minutes lost in March 2023. This means that during this period around a fifth of lesson time was lost due to poor behaviour on average. In some cases, the loss was particularly extreme; 25 per cent of teachers reported that more than 10 minutes of teaching time was lost due to pupil misbehaviour per 30 minutes of lesson time.^[_]

Inevitably, this has an impact on learning outcomes. For this report, TBI commissioned Edurio, an education-technology company that works with schools and multi-academy trusts to run surveys of pupils, teachers and parents that inform school-improvement strategies, to analyse the impact of different aspects of the school environment on academic performance.

According to these data, 69 per cent of teachers report that their work is disrupted by poor pupil behaviour sometimes, often or constantly.

Edurio’s analysis of its data suggests that various elements of poor behaviour (pupil behaviour, safety and bullying) are strongly linked to lower levels of attainment in major performance measures in the education system (GCSEs, Progress 8,^[_] Ofsted grades and expected primary levels of literacy and numeracy). The less safe pupils feel and the less clarity there is about expectations around behaviour, the worse the schools perform.

This analysis, published here for the first time, has three main implications. First, although the DfE runs an annual survey of behaviour in schools, data on the school environment and how pupils, teachers or parents perceive it are not collected routinely, which means that understanding of the scale of the problem is limited and often out of date. Data presented here demonstrate that it is possible to collect and analyse this information at scale, in deeper detail and with greater timeliness.

Second, and perhaps most importantly, this analysis demonstrates clearly that a link exists between perceptions of safety and behaviour in the school environment and that school’s performance across the key current measures of accountability. Safety and behaviour should be seen not as an isolated issue but rather as an important strand of school standards.

Third, the analysis begins to show a path forward for a different approach to school improvement. Of course, the relationship presented here is one of correlation and not necessarily causation. However, even correlation can act as an early warning signal. Much of the accountability and improvement system relies on lagging indicators, so that by the time a problem becomes apparent and can be addressed, it is too late to help pupils whose results were affected by that problem. Evidence of challenging behaviour, given this relationship with attainment, can be seen as a leading indicator of school quality and used to direct improvement support more effectively.

The Economic Impact of Challenging Behaviour Is Significant

As previously noted, challenging behaviour is choking off learning time at frighteningly high levels. To understand what this amounts to in practical terms, we modelled a speculative scenario where average levels of classroom disruption were reduced from current levels of around one fifth of teaching time^[_] to roughly half of that. The latter figure is based on the rates typically enjoyed by teachers who lost comparatively less time in the DfE’s National Behaviour Survey, thereby serving as a proxy for a reasonable low-disruption environment.^[_]

When disruption driven by misbehaviour falls, teachers claw back productive teaching time. Academic evidence across different countries and age groups suggests that more interactive teaching time improves pupils’ educational outcomes.^[_] Therefore, pupils in a low-disruption environment are likely to learn more in their school years and subsequently enjoy higher productivity when they join the labour market. In the UK, the Education Endowment Foundation has also concluded that effective behavioural interventions can lead to improvements in average student attainment of up to four months of additional progress.^[_]

In our model – described in more detail in the technical annex – halving the amount of time lost to disruptive behaviour could boost average educational attainment provided this low-disruption environment is sustained across a pupil’s primary and secondary schooling. The economic consequences of these gains could be significant as a better educated workforce is also likely to be more productive and hence support economic growth.

It would of course take several years for pupils educated in low-disruption environments to join the workforce and for the economy to realise the associated productivity gains, but over time these are likely to become increasingly significant. According to our model, cutting the time lost to disruptive behaviour in the classroom could add up to 1.5 per cent of GDP in the long run through higher workforce productivity.

Wellbeing, Attendance and Aspiration

The evidence outlined so far makes it clear that the feeling of safety and freedom from disruption are important aspects of school quality, measurably associated with performance and attainment.

To better understand the relative importance of these factors to student wellbeing and their broader experience of school, TBI worked with the polling organisation Opinium to conduct a nationally representative survey of pupils in the process of studying for their GCSEs at schools in England. We found that pupils saw safety and freedom from disruption as a priority, with a particularly pronounced impact of challenging behaviour on those pupils who felt unhappy at school. We also found a robust link between pupils feeling safe at school and their attendance, as well as their aspirations for future study. All of this suggests that tackling challenging behaviour and classroom disruption is key to improving the wellbeing of pupils, addressing the attendance crisis and helping pupils develop a lifelong love of learning that would serve them well in the age of AI.

When asked their priorities for what makes a good school, feeling safe was a key factor identified by GCSE pupils. Of 13 different suggested characteristics of a “good” school, safety was ranked number two. Even more strikingly, those pupils in the survey who told us they feel unhappy at school were particularly likely to say minimising disruption would improve their school experience. For these pupils with poor wellbeing, both classroom disruption and safety are top-three issues.

We also found that feeling safe at school was key to pupils’ likelihood of attendance. While seven in ten of those who feel unsafe missed a day at school in the month prior to when the survey was fielded, the same was true for less than half of pupils that feel safe. In principle this is not surprising, but it is worth noting that feeling safe at school is more strongly associated with school attendance than any other factor we tested.

Similarly, when pupils were asked about their plans for further study (either through university or an apprenticeship), feeling safe at school was a key determinant. Over two-thirds of those who feel safe at school were planning to go on to further education compared to just under half of those who feel unsafe at school.

While the role of gender and parental income in driving attendance and access to higher education is well known, feeling safe within the school community clearly also plays a key role in determining the aspirations of young people. To test whether school safety was acting as a proxy for these other socio-demographic factors, we used a linear regression model and controlled for gender, region, eligibility for free school meals and their school’s Ofsted rating.

Pupils who reported feeling unsafe at school outside the classroom were 1.2 times more likely to have missed a day of school in the past three months, compared with their peers. We found similar results for aspiration, with pupils who reported feeling safe at school outside the classroom 1.15 times more likely to say they wanted to pursue an apprenticeship or university after school, compared with their peers.

This is robust evidence that pupils’ feelings of safety and the way behaviour is managed are not just subjectively important factors but are linked to objective outcomes for the GCSE pupils we spoke to across England.

Finally, we looked at the association between feelings of safety and happiness at school (as a proxy for overall wellbeing). The link between safety in school and happiness is clear: of those who feel safe at school, 72 per cent are also likely to say they feel happy. Fewer than one in five, just 19 per cent, of pupils who feel unsafe at school say they are happy. Less than half of those who have been bullied at school in the month of our survey said they were happy at school. In Figure 15, we have plotted the relationship between happiness and safety, feeling of belonging to the school community, bullying and feeling able to learn new things at school.

Managing disruption and creating a safe school environment has always been one of the core tasks of teachers and school leaders. The complex and varied roots of challenging behaviour invariably mean that the answer must involve action both by the school itself and by other agencies.

Effectively addressing these requires improved multi-agency collaboration, integrating social services, mental health support and community programmes. Equally crucial is strong school leadership to consistently, fairly and confidently apply behaviour policies, support teachers in doing so and provide them with relevant training. Together, these strategies tackle the root causes of behavioural issues and create lasting, positive change. However, evidence now suggests that English schools are struggling to meet this challenge.

The School System Has a Pivotal Role to Play

The actions schools take to understand and manage behaviour matter enormously. A substantial amount can be achieved, even in the most challenging environments, if teachers get the support they need to effectively implement appropriate behaviour policies and interventions.

There is no single definition of what “appropriate” means. School cohorts, as well as the nature and scope of the behavioural challenges they face, vary substantially. Despite the occasionally unhelpful focus of the debate in the press and on social media on the relative merits of different approaches, whether zero-tolerance “silent corridors” or “person-centred” Positive Behaviour Support techniques, there is no one right way. Schools with highly contrasting styles and philosophies have found success in equal measure, and a school’s ability to choose the approach that best fits its own culture and values is an important element of school autonomy and parental choice.

Instead, what tends to underpin success is a school’s ability to combine clarity of vision and policy, effective communication, consistent implementation and sufficient training. Whether due to strained resources, a lack of know-how or particularly challenging cohorts, many schools seem to be struggling to put these elements in place.

Behaviour Policies Lack Clarity, Communication and Consistency

Ultimately, it is heads and their senior teams who shape their school’s approach to understanding and managing pupils’ behaviour. One of the main ways schools have attempted to do this is to craft and implement behaviour-management policies (according to one recent study, 95 per cent of teachers report that their schools have one in place).^[_]

But for these approaches to work, three things need to happen. First, heads and their senior teams need a clear vision of the culture they want to foster and a system of rules, expectations and processes that are workable and appropriate to the context and values of their school. Second, they must be able to articulate that vision to staff, pupils and parents so that there is a shared understanding of expectations. Third, they should enforce their approaches consistently.^[_]

On all three counts, many schools appear to be falling short. In the first case, some schools’ approaches to behaviour management lack the clarity and workability needed. According to one study, less than a quarter of teachers rate their school or college’s policies and/or approaches to dealing with violence and abuse from pupils towards staff as good or excellent.^[_] According to Edurio’s analysis conducted for TBI, almost a quarter of teachers find it difficult (18 per cent) or very difficult (5 per cent) to get support with pupils’ behaviour.

Second, policies are often not communicated widely or clearly enough. According to one recent analysis, only 69 per cent of teachers report that behaviour polices are communicated to pupils, while the figures for teachers are 85 per cent and for parents 62 per cent.^[_] According to the government’s figures, meanwhile, 81 per cent of school leaders and teachers agree that there is a shared understanding among their staff of what “good behaviour” means, while 85 per cent of pupils know how their school expects them to behave.^[_]

Third, policies are not consistently enforced. Although 95 per cent of teachers report that their schools have behaviour-management policies in place, only 8 per cent report that their policies are always enforced. Meanwhile, less than half report that their schools’ policies are usually or always enforced and 40 per cent report that their policies are only enforced sometimes.^[_]

Teachers Lack Agency When Escalating Concerns and Engaging Parents

The lack of effective behaviour policies also feeds another crucial problem at the heart of the current approach – namely, that teachers ultimately lack the agency they require to manage challenging behaviour.

All too often, teachers feel their concerns are not heeded at senior level. For instance, 44 per cent of teachers do not feel supported by their school or college’s approach to dealing with poor pupil behaviour. Worse still, 45 per cent of teachers say poor pupil behaviour is seen as part of the job where they work and that they should expect to receive abuse from pupils, while 45 per cent report they are made to feel to blame if they have an issue with poor pupil behaviour.^[_]

Invariably, many teachers feel disenfranchised as a result: only 18 per cent of teachers, for instance, feel their setting’s approach to behaviour management empowers them as a teacher or leader.^[_] This lack of agency means that teachers often feel there is little point in reporting incidents – for instance, only 55 per cent of teachers report all incidents and among those who do not, around 73 per cent say this is rooted in the prospect of inaction.^[_] Even where teachers do report incidents, many feel underwhelmed by the resulting actions.^[_]

Teachers also lack agency when it comes to engaging parents directly. In many cases behavioural issues cannot realistically be resolved without exploring matters outside school, which in turn ultimately requires engaging parents. Given the contact they have with pupils, teachers are often ideally placed to do this well because they can spot signs that challenging behaviour may have deeper, complex roots and can help source appropriate support. However, their scope to carry out targeted outreach is limited.

First, they lack the power to engage meaningfully with parents. Thirty per cent of teachers have had a parent undermine their sanction.^[_] Although parents are expected to engage with schools to support pro-learning behaviour (for instance, statutory duties require them to ensure their children attend school regularly^[_] and DfE guidance suggests they must support schools’ behaviour polices^[_]), teachers cannot compel them to attend meetings if they have concerns about their child’s behaviour.

Second, schools remain peripheral to the multi-agency ecosystem – which means that even where robust mechanisms exist for teachers to escalate matters internally, schools cannot necessarily exert appropriate action. While schools have a statutory duty to co-operate with safeguarding partnerships (including, for instance, by supporting Multi-Agency Safeguarding Hubs),^[_] and can put in requests for agencies to intervene, their role is largely collaborative and ultimately more deferential. For instance, education is not treated as a fully fledged safeguarding partner when it comes to multi-agency working^[_] and schools lack the power to initiate, convene or lead multi-agency meetings.

Together, these realities are damaging teachers’ morale and contributing to the retention crisis. For example, according to recent DfE analysis,^[_] 37 per cent of teachers who rated pupil behaviour as “poor” were more likely than average to consider leaving within a year, whereas the figure for those who rated behaviour as “good” was 21 per cent. Of the teachers who were considering leaving, those who occasionally or never felt supported to manage disruptive behaviour were more than twice as likely to leave than those who felt they were always or mostly supported (41 per cent versus 20 per cent). And when asked which factors were most important when considering leaving the profession in the last year, behaviour was a top-three concern for teachers.^[_]

Teachers Need Support and Training to Help Improve Behaviour

But it is not just the direction and backing teachers are given that matters; teachers must also be well versed in behaviour-management techniques if they are to manage classroom disruption. After all, there is little point in devising a top-notch strategy on behaviour if teachers have not been given the know-how and confidence to enact it in the classroom.

In 2015, the Carter review recommended the inclusion of behaviour-management modules in a reformed Initial Teacher Training framework,^[_] with the new Early Career Framework published in 2019 and statutory induction practices changing to reflect it in 2021.

Yet even now, teachers report struggling to access quality training in student behaviour and classroom management, with only 27 per cent recalling that their training covered this area since 2021.^[_] The most recent NBS also suggests training is not being tailored to the context of individual schools, with 31 per cent of teachers reporting they could not access training and development support for behaviour management relevant to their experience and needs.^[_]

In this context, it is perhaps unsurprising that so many teachers find it difficult to cope. For example, just 35 per cent of teachers feel very confident in managing misbehaviour.^[_] And according to analysis conducted by NASUWT, almost three-quarters of teachers do not feel they have the resources, support and knowledge to meet the behavioural needs of all the pupils they teach.^[_]

A good example of where laudable efforts to improve provision fall short is the DfE’s Behaviour Hubs programme, which partners struggling schools with schools they can learn from. Conceptually, the programme is on strong ground. But it is very small, with limited spaces (only 664 schools and multi-academy trusts (MATs) have been supported,^[_] while there are 24,453 open schools in England^[_]). It is also voluntary and not based on a systematic approach to early determination. And participants are not matched scientifically to a sufficiently wide pool of best-fit peers (only 50 lead schools and ten lead MATs are available),^[_] increasing the risk that the support provided by lead partners is at odds with a school’s vision, culture or policy.

An interim evaluation of the programme’s effectiveness has been recently published,^[_] though a full one is yet to come. The results suggest some improvements in staff perception of behaviour within schools taking part in the programme. This appears to be driven, among other things, by greater uptake of proven effective methods of behaviour management such as adopting a holistic approach to it, encouraging teachers to de-escalate in class and setting clearer expectations for students via definitions of good and poor behaviour. This suggests there is promise to the peer-led approach, but at the current scale it is not making enough of a difference.

There Are Limits to What Schools Can Achieve by Themselves

Behaviour Is Significantly Shaped by Factors Beyond the School Gates

Challenging behaviour in schools proliferates for many reasons. On a general level, pupils with lower emotional, behavioural and social wellbeing tend to disengage more with school and vice versa.^[_] Some elements of wellbeing relate to schools; whether the learning environment is stimulating or whether pupils have positive friendship groups there, for example. But much is influenced by factors outside school. For example, children growing up in tumultuous and/or traumatic home environments are more likely than their peers to struggle emotionally and socially. Their experiences can affect their brains’ stress responses and ability to self-regulate, triggering or exacerbating developmental challenges, leading to poor behaviour.^[_] External shocks can also make matters worse. For example, Covid-era lockdowns damaged some pupils’ socio-emotional skills, including their ability to regulate their emotions, maintain attention and engage positively in social situations.^[_]

Left unaddressed, these factors can result in disruptive behaviour that derails pupils’ classroom learning. But because its roots are often outside the school gate, not all challenging behaviour can realistically be managed by teachers. A fully comprehensive strategy to tackle it would, therefore, require complementary policies to improve various specialist support services and better connect them to what goes on in school.

This strategy would entail spotting problems early and sourcing specialist support for interventions targeted at and tailored to individuals where necessary. These might be external services, such as Child and Adolescent Mental Health Services (CAMHS) or wraparound initiatives such as the Supporting Families programme,^[_] or in-school specialist support, such as on-site counselling or special educational needs and disabilities (SEND) specialists.

Dwindling resources have made it harder to build and sustain the systems and links required. While the turn of the millennium saw substantial investment in preventative initiatives and early help, the opposite was true from 2010 onwards as sharp cuts set in. It has been estimated, for example, that funding for local authority services for children and young people fell by £2.2 billion between 2010–11 and 2018–19.^[_] The balance of resources has also shifted markedly away from early help to more reactive measures. For instance, while local authority spending on late intervention accounted for 58 per cent of all children’s services spending in 2010–11, that figure shot up to 80 per cent by 2020–21.^[_]

In this context, it comes as no real surprise that teachers are finding it hard to access external specialist support in a timely manner, with almost two-thirds of teachers and leaders now highlighting delays.

Multi-Agency Interventions Are a Necessary Component of Addressing Behaviour

Limited resources are not the only issue, however. Practice matters enormously, too. Perhaps most saliently, we know that interventions are generally far more successful when different safeguarding agencies work closely together.^[_]

One of the main reasons for this is that a child’s disengagement from education is often just one of many things that have gone wrong in their life. When information is scattered across service lines, practitioners lack the holistic view needed to make good decisions. Working together across different touchpoints (whether in education, health, the police, youth offending and children’s social care) gives practitioners more chance to succeed than if they explore their own spheres alone.^[_]

Sound multi-agency intervention entails, first and foremost, practitioners being able to draw on high-quality information from other agencies – the lifeblood of all good multi-agency work.^[_] That information first needs to be linked to individuals, with appropriate privacy-preserving guardrails in place, and then shared using case-management systems that are interoperable and underpinned by common standards.

However, the sharing of good-quality information alone does not guarantee improved outcomes. For that to happen, practitioners also need the skills to make the most of this data. These might include, for instance, the ability to interpret the data, decide what is relevant and draw quality insights from it, particularly when this relates to other fields. These also include the ability to understand and make the most of sharing powers that already exist, and the coordination to avoid duplicate efforts.^[_]

In addition, as several analyses have shown,^[_] it is not realistic to expect that all practitioners will foster good links with one another and share data effectively, as this tends to happen only at the margins. Instead, links must be purposefully cultivated by leaders to drive a culture of sharing.

There Are Pockets of Good Practice in Targeted Multi-Agency Intervention

So, how does the current system fare against the standard we have set out? Over the last two decades or so, a slew of national policy initiatives and legislative changes have paved the way for better information sharing between safeguarding agencies, and by extension better multi-agency working. These initiatives and others have set the tone for the change and, to some degree, have provided the powers to realise this. As a result, more proactive and adequately resourced local partners have made improvements on the ground.

For example, multi-agency safeguarding hubs (MASHs) have proliferated. These local entities co-locate practitioners from agencies across health, social services, the police and education, with the aim of sharing data and making joint decisions about which children and families ought to be prioritised. Research by Ofsted has found that MASHs have generally improved communication between different local agencies and have boosted the expediency and quality of their decisions.^[_]

Notable, too, are the gains realised under the auspices of the Supporting Families programme (formerly known as the Troubled Families Programme), which employs key workers to engage with families with multiple challenges, in tandem with local authorities and agency partners. As part of the ongoing evaluation of the programme, a data-maturity model has been built to assess how good local partners are at managing and sharing data, including to inform better risk assessments and decisions.^[_] Generally, the results have been positive.^[_] In addition, targeted programme initiatives have been put in place to improve local systems of sharing that are particularly weak – for instance, by working with local trade associations to design more dynamic case-management systems.^[_]

More generally, several local authorities have taken it upon themselves to improve data sharing by integrating newer technologies into their case-management systems. Although their methods vary, they all aim to help practitioners access relevant information from other agencies, with a view to building a more rounded view of a family’s circumstances. They tend to draw on a mixture of existing identifiers to link disparate data (for example, the NHS number for health data, the LiquidLogic ID for children’s services, the Unique Pupil Number for education data) and in some cases have bolted on predictive analytics to help practitioners assess risk and prioritise their resources.^[_]

Effective Multi-Agency Interventions Remain More an Exception Than the Rule

While some local authorities and agencies are innovating, the pace of progress more generally remains slow and its scale limited.

Part of the problem is that government still lacks a consistent ID with which to match different data sets to pupils – a crucial first step to collecting more rounded information about a pupil’s context. The closest current surrogate is the NHS number, which since the introduction of the Health and Social Care (Safety and Quality) Act 2015 has been used as a single identifier to link data across health and adult social care. But it omits children’s services, including education. In the absence of a coordinated approach, it also creates unnecessary risks to children’s rights and privacy.^[_]

Even if there were better ways of collecting and matching data to children through a single identifier, many local agencies lack the ability to share this information seamlessly with one another. This problem has several dimensions; chief among them is that local systems often lack interoperability. Instead, each agency has its own case-management system, which is seldom compatible with other systems used by other local agencies.^[_] Even where local partners have managed to circumvent some of these issues by sharing new, more integrated platforms, their systems are often not compatible with agencies in other local-authority areas – a significant problem when attempting to support children who move in and out of different geographies.^[_]

Adding to the overall inertia is the fact that agencies must negotiate data-sharing agreements before they share information, as existing approaches are often inconsistent. This process can be long and laborious, even extending to years in some cases.^[_]

Other problems abound. Even within agencies, sharing is not as fluid as it might be. As research from the Children’s Commissioner has found, information is not always diffused internally to those who might make the most of it. Instead, oversight tends to vest in certain positions – for instance, in Local Authority Designated Officers and health safeguarding leads – which limits how the information could be used by frontline practitioners to help more children.^[_]

Schools also face several practical constraints. Although some use systems such as CPOMS and MyConcern to share safeguarding concerns within their own schools, this information cannot be directly accessed outside school systems. In addition, if a child moves to another local authority (or to another school), their case history is not automatically transferred. And while some schools have ways of uploading certain data onto local authority systems (for example on the Liquidlogic system used by social workers), there is no visibility of what happens next and no ability to access the data recorded by other agencies.^[_] All of this leads to a lack of targeted, coordinated interventions for individual pupils, and can create duplicate efforts that waste limited resources and cause unnecessary frustration for families.

But it is not just the lack of quality data and systems for sharing that are hindering multi-agency partnership. As noted, sound multi-agency working also rests on human capabilities^[_] and there is ample scope here for improvement. One of the most pressing concerns across safeguarding agencies is that there appears to be widespread confusion about what the law permits practitioners to share and under what circumstances, which in turn appears to drive risk aversion and an underutilisation of existing powers.^[_]

More generally, practitioners are not always clear about the information that might be most usefully shared with other agencies and vice versa.^[_]

Lastly, good multi-agency working is almost always built on a culture of sharing, which must be purposefully cultivated by leaders who have the inclination and power to do this. Yet although this level of leadership and direction is evident in some local cases, this is far from ubiquitous. Where leadership of this sort is lacking, the dynamic between local agencies can instead be rather brusque – agencies become more mutually sceptical about their views being heard and more likely to disagree on which families need to be prioritised, all of which can lead them to pivot further inwards.^[_]

As this report has demonstrated, clawing back lesson time lost to poor behaviour would have a significant impact on attainment, attendance and aspiration. Achieving this goal should therefore be a major political priority for government and command greater emphasis. The government’s response should comprise two sets of interventions.

Intervention 1: Teachers should have far more agency when it comes to engaging parents.

As we have noted, challenging behaviour often stems from factors outside the education system. While it is imperative that the right culture of consistent enforcement and clear communication is sustained in schools, there are several reasons why teachers should also be able to act beyond the school gates, to better understand and address the root causes of challenging behaviour.

Given their level of contact with pupils, teachers are uniquely placed to spot behaviour issues and any ancillary needs early, and therefore to help source the right kind of support. In addition, parents are arguably more likely to engage with schools than with some other agencies – not least, where parents’ impressions have been tainted by prior interactions. A more central role for schools would also give them more agility by reducing the inefficiency of lengthy referrals and subsequent chasing, while providing teachers with confidence that their concerns will not simply fall through the cracks.

However, schools’ scope to carry out targeted outreach is limited.

First, they lack the ability to engage meaningfully with parents. For instance, 30 per cent of teachers have had a parent undermine their sanction and since Covid, the “unwritten agreement” between schools and parents has fractured as more parents see adherence to school policies as optional.^[_] Although parents are expected to engage with schools to support pro-learning behaviour,^[_] teachers cannot compel them to attend meetings if they have concerns about their child’s behaviour.

Where a child’s behaviour reaches a certain threshold of severity, teachers should have the power to compel parents to attend meetings with them. Parents should be expected to engage meaningfully and agree subsequent steps that are proportionate in context. If parents refuse to engage and have no substantive grounds to support this, schools should have the right to bring in other safeguarding partners where appropriate to do so.

Recommendation 1: The scale and depth of challenging behaviour in schools demands a re-set of the parent-teacher relationship. Where it is reasonable and proportionate to do so, teachers should have the right to compel parents to engage with them and to escalate to other agencies where parents refuse to cooperate. The government should consider if any legal powers similar to those used to address repeated absence without a good reason should apply when parents refuse to cooperate repeatedly and without substantive grounds.

Second, schools lack the powers required to drive multi-agency working. Schools have a statutory duty to co-operate with safeguarding partnerships and support them,^[_] but their role is largely collaborative and ultimately more deferential. They are expected to refer cases to relevant agencies, share information with them and support them, and they can put in requests for agencies to intervene. But they do not have powers to initiate, convene or lead multi-agency meetings and they are not treated as fully fledged statutory safeguarding partners.

Schools should be given new powers to drive multi-agency working where it is reasonable and proportionate for them to do so. They should, for instance, be able to initiate and convene multi-agency meetings with safeguarding partners and agencies, co-ordinate responses and help make appropriate decisions. To maintain coherence and manageability, a school’s Designated Safeguarding Lead would ultimately oversee the powers at school level, but teachers involved in any given matter should have a right to be informed of next steps and attend any subsequent meetings to contribute.

In cases where either parents and teachers have agreed actions, or where multi-agency working leads to an agreed plan, data should be used to benchmark progress and determine whether a plan is meeting its aims.

Recommendation 2: Schools should be given a stronger role in driving multi-agency working, with a statutory partner role in safeguarding.

Intervention 2: Much better insight and support should be available to schools that might be struggling with their approaches to behaviour.

Within the school gates, leaders and teachers must be able to create an environment in which behaviour policies are based on best practice, articulated clearly, communicated consistently and implemented fairly – all while preserving their freedom to act in accordance with the specific challenges they face, rather than being prescribed any one purported tonic. Key to this is identifying schools that may be struggling with challenging behaviour early on, providing timely improvement support that complements the school’s context and philosophy, and tracking progress over time.

There is precedent for this – not specifically in relation to behaviour but regarding school performance more generally. At the turn of the millennium, London was home to some of the country’s worst-performing local-authority areas. The complex and varying nature of their problems demanded an approach that could adapt to the specific challenges each school faced. By generating meaningful insights and enabling peer-led support in context, the New Labour government was able to transform state education in the capital through its “London Challenge” programme.

Today’s government needs to grip the current behaviour problem in the same way. Rather than search for an elusive “silver bullet” solution implemented in every school, it should set up a “National Behaviour Challenge” programme aimed at tackling the idiosyncratic factors that drive behaviour in different schools. And because it now has access to technology that was not available two decades ago, it would be able to do this with far more weight, scale and precision.

Recommendation 3: The government should set up a new programme, sponsored by the DfE and reporting into the “opportunities” mission board, to focus on urgently addressing challenging behaviour in schools and its consequences. The National Behaviour Challenge should be set an overarching objective of improving reported measures of wellbeing, raising attendance and cutting classroom disruption by half by the end of this parliament, with safeguards against the gaming of those objectives.

The first step of this new venture should be to collect much richer and more up-to-date data than are currently available, and to start diffusing these data so that schools and their trusts, local authorities and other relevant agencies can begin to make use of them.

Since the 2021/22 academic year, the government has taken tentative first steps towards collecting better data on behaviour in schools by setting up the National Behaviour Survey. Although encouraging, its remit remains very limited; for example, it is based on a sample of teachers, takes place termly and offers a headline temperature check for the system overall rather than a granular analysis of the state of play in each school. The information that is collected therefore lacks the scope, depth and real-time qualities needed to support all schools at an individual level.

Instead a far more scientific, comprehensive and systematic approach that allows policymakers and schools to broach behaviour management strategically is needed. Regular surveys of pupils and staff, with appropriate privacy-preserving guardrails in place, should be conducted in each school. Data should be collected on the school environment, including on staff and pupil safety and low-level disruption, but also on a wider set of indicators of staff and pupil wellbeing and happiness at school.

At a school level, these data should be available to school leaders, and local authority and academy trust improvement teams (with number suppression for any entries with low response rates). Following the example of the NHS Model Health System,^[_] the data should be benchmarked to schools with similar pupil-population mixes, so that judgements about the absolute and relative level of disruption and wellbeing can be made within an appropriate context. In the case of behaviour, a further analysis can be carried out of a school’s behaviour policies (using generative AI tools such as those being piloted by The Key Group^[_]) so that it can be matched to others with similar values, culture and approach.

With better data, policymakers could radically improve the current nascent approaches to peer-led support such as the Behaviour Hubs programme. Building on the approach adopted for the Education Endowment Foundation’s Families of Schools database^[_] or the FFT Education Datalab’s Schools Like Yours tool,^[_] schools could be benchmarked to peers so that contextually appropriate decisions could be made about suitable pairing. The data should also be shared securely across relevant agencies so that pupils and families with complex needs can be identified for proactive interventions.

Recommendation 4: Put in place data-collection tools and practices to provide an accurate present-time picture of behaviour and wellbeing in schools. Share data with schools, local authorities and academy trusts to inform improvement-strategy development and implementation. Use these data to inform a revamped peer-led support system in which schools are benchmarked more precisely and contextually suitable decisions can be made about pairing them with others successfully tackling similar challenges – for example, via the newly announced Regional Improvement for Standards and Excellence (RISE) teams.

While improving data collection and beginning to use these data to broker peer-to-peer support, the government should contemporaneously start building a national data infrastructure for education.

TBI has previously argued that the development of a national data infrastructure for education, centred on a digital learner ID, could transform the education system’s approach to accountability and improvement.^[_] Infrastructure of this kind would provide pupils (and their parents) with much greater control over their data and help address the shortcomings of the current ad-hoc processes for this type of data collection and processing.

It would also allow more stringent privacy guardrails, thereby overcoming recent issues encountered by government even when attempting to collect more rudimentary school-level data in present time. One example of this is the daily sharing of pupil-attendance data, launched as a pilot in 2022 and made mandatory for all schools from September 2024.^[_] A big advantage of this approach is that schools with management information systems from the most popular suppliers can automatically share data via an application programming interface (API), minimising the workload burden. Data are then made available to school leaders and, where applicable, academy trusts and local authorities to inform improvement strategies and track their impact. However, the initial implementation was criticised by the Information Commissioner’s Office for a lack of clarity around privacy protections and whether the data collection was proportionate to the stated purposes.^[_]

This can be addressed if a dedicated system is put in place to replace the current ad-hoc approach. With appropriate privacy-preserving guardrails, regular surveys of pupils and staff could be conducted in each school. These surveys would draw on pupils’ and teachers’ experiences and would in turn drive an aggregate model, allowing the nature and scale of disruption across the country to be mapped. Information could be collected via prompts in a digital learner ID app or via other digital tools used by the school – here, too, an interoperable and secure national data infrastructure is key to minimising the data-collection burden.

Aggregate, anonymised present-time data could then be used to spot performance trends across a broad range of important metrics. This could include survey data from teachers, learners and parents to build a fully representative picture. Analysis for this report shows that data on behaviour and safety can be collected at scale and can serve as a leading indicator of performance, enabling proactive improvement support. Some schools, local authorities and academy trusts do this already; a national data infrastructure for education would make this the norm.

The availability of these aggregate data would also significantly simplify the proposed change from lengthy and often outdated Ofsted inspection reports and single-word judgements to more nuanced school “report cards” proposed by the new Labour government. The data would also be useful in supporting the Labour government’s focus on improving wellbeing in the education system by helping to build an accurate and nuanced picture of the current state of the system and the impact of future policy interventions.

TBI has previously estimated that introducing such a system in the UK would require upfront investment of a little over £50 million, with running costs of approximately £10 million a year.^[_] Given the economic cost of failing to address disruptive behaviour identified in this report, this is an investment that – if used to support interventions that reduce disruption in the classroom – would more than pay for itself in the long run.

Recommendation 5: Introduce a national data infrastructure centred on a digital learner ID, built and maintained by an independent designated data body, with an initial focus on harnessing the data outlined in Recommendation 2, so that the National Behaviour Challenge acts as the first user of this new infrastructure. (Over time, a fuller suite of education data should follow.^[_]) The design should include mechanisms for opting out of certain data-collection practices and provide full visibility of where a pupil’s data are held and for what purpose. A data-interoperability standard, designed with privacy protection as a core requirement, should also be introduced so that external software providers can share the data they hold in real time. The aggregate data can also be used to inform the government’s new school “report cards”.

Collecting and utilising granular, present-time data on behaviour and wellbeing also opens up new opportunities for system-level improvement. At present, many of the metrics used to inform improvement interventions are lagging indicators. This creates two problems: first, by the time an intervention is proposed, the situation may have changed. Second, even if the intervention is successful, it comes too late to help the cohort of pupils whose experience led to, for example, a poor Ofsted judgement. At the same time, existing metrics are often too broad to be useful guides to the specific improvement interventions required. A more proactive and targeted system is needed.

TBI has previously proposed reforming Ofsted to give it a greater role in coordinating the provision of contextually appropriate improvement support. Under our proposed system, Ofsted staff would build close ongoing relationships with school leaders based on a shared understanding of the challenges faced by each school, informed by the data collected and benchmarked via the national data infrastructure for education. These data could act as early warning indicators – a bellwether to determine which schools most urgently need help.

In the first instance, the job of identifying areas of improvement and strategies to tackle them would fall, as it does today, to school leaders, local authorities and academy trusts, as well as school governors. However, where the situation continues to deteriorate over time – for example, the amount of time lost to disruption in lessons steadily increases – and breaches an agreed threshold (based on peer benchmarks), schools would be asked to seek peer-led support to address the challenges they face. The National Behaviour Challenge can act as a pilot for testing this approach and implementing it around behaviour support as a first priority.

In the NHS, data from the Model Health System are used to inform interventions by the Getting It Right First Time programme, where groups of expert practitioners visit hospitals to investigate the causes of reduced performance and propose improvements based on best practice.^[_] A similar model could serve as an extension of the Behaviour Hubs model, with leading schools or trusts within each “peer family” matched to others in the same group and circumstances that need support. Partner schools would need to be incentivised and resourced appropriately to provide peer support. Clear milestones should then be agreed and progress tracked using the same data sets for consistency.

The government has recently announced the establishment of a new regional support structure: the RISE teams that provide signposting to best practice or targeted support depending on the level of challenge faced by the school (with structural interventions for a small number of those struggling the most). The National Behaviour Challenge should work closely with RISE teams to quickly direct appropriate school-level support based on the data collected through the steps described earlier. Given the foundational nature of safety and good behaviour to learning and wellbeing, these interventions should be prioritised.

Evidence suggests that struggling schools benefit from suitably matched peer-to-peer support and counsel, including from accomplished external leaders with a track record of success in similar contexts.^[_] Positive outliers are also known to buck difficult trends and so there are likely to be exemplars whose know-how can be shared. For example, a recent Nesta analysis^[_] – in this case, specifically in relation to non-attendance – showed that some schools managed to buck a very stubborn, pervasive spike in non-attendance. Schools of similar size, and with similar characteristics in terms of cohorts receiving free school meals, with special educational needs and with English as an additional language, had wildly diverging absence rates (for example, severe-absence rates differed by a factor of eight), which implies that agency makes a substantial difference and that schools in similar circumstances could learn from one another.

A national programme of behaviour- and wellbeing-improvement support, based on data insights into the nature and extent of the problems, should be available to all schools. Insofar as possible, this should be based on voluntary participation and a relationship of trust with successful peers. However, where schools fail to improve over time or refuse to engage with the support provided to them, existing sanction powers (such as academisation or rebrokering) should remain available.

Recommendation 6: The National Behaviour Challenge should build lasting local relationships with schools, local authorities and trusts. Using benchmarked data to identify the most suitable peers that are successfully tackling similar changes, they should work with RISE teams to coordinate the provision of peer-led improvement advice and targeted school-level support for those struggling to tackle behaviour, safety and wellbeing issues.

To help teachers better tackle challenging classroom situations, improved training provision is required. As noted earlier in this report, there have been positive developments in the initial teacher-training framework, which now includes dedicated modules on behaviour and classroom management. The government should continue to invest in the development of high-quality training in this area, recognising that there is no one-size-fits-all approach and instead covering a range of different methods to improve behaviour so that early-career teachers are prepared for work in schools with different policies and values.

In addition to initial teacher training, greater attention should be paid to the high-quality provision of continuous professional development (CPD) so that experienced teachers can receive appropriate support and information about best practices. There is a strong argument for funding a greater number of hours of CPD in recognition of the evolving requirements of the teaching profession, such as the growing importance of digital pedagogy and education technology, and any such extension should include dedicated time on classroom management and behaviour.^[_]

Opportunities offered by emerging technology for scaling up quality CPD provision should be considered. For example, the Incubator for Artificial Intelligence in the Department for Science, Innovation and Technology should explore ways that generative AI tools can support higher-quality training, such as creating classroom simulations that allow teachers to safely experiment with different approaches to behaviour management.

Recommendation 7: The National Behaviour Challenge should invest in generating evidence on what works in behaviour management in schools and making it widely available, in collaboration with the Education Endowment Foundation. It should work together with the DfE to ensure that both the initial teacher-training framework and the provision of continuous professional development include a focus on effective ways of managing challenging behaviour and classroom environments.

As well as giving schools more sway over multi-agency working, there also needs to be a step-change in the quality of the interactions that take place between safeguarding partners and agencies.

As our earlier analysis shows, one of the key components of success here is the ability to share, access, understand and act on data. The education system remains one of the main touchpoints for most young people and is well-placed to serve as a hub for this activity. The National Behaviour Challenge should therefore also dedicate resources to empowering better multi-agency working so that improvements in school-level policies can be supported with a more personalised approach to pupils exhibiting disruptive behaviours. Additionally, aggregate data on behaviour in local schools can be used at the regional and/or local-authority level as an effective, near-real-time proxy to track the overall effectiveness of multi-agency working.

It is imperative that safeguarding agencies work closely together seamlessly. Greater resources and better leadership are part of what is required to achieve this. But the quality and availability of data is ultimately the lifeblood of multi-agency working; without improving both, children will continue to slip through the net. Although there has been no shortage of government initiatives to gather and share better data, they have been too piecemeal and narrowly focused.

What is needed is a more comprehensive and interlinked data infrastructure and the tools to use it to its full potential. Relevant data should be linked to the unique digital identifier we proposed earlier in this report. This information should be shared between agencies through interoperable systems in ways that meet common data-sharing standards. Practitioners should be able to draw on better visualisation tools to give them a more comprehensive impression of a pupil’s circumstances. AI and predictive analytics should be employed to identify early on pupils who may be at high risk, choose the best likely course of action in given circumstances and prioritise resources according to need. All of this should be underpinned by suitable privacy-preserving guardrails and access rights that flex according to the data user’s role and seniority. For information-governance purposes, the National Behaviour Challenge should assume responsibility for coordinating and prioritising relevant data-sharing agreements between agencies to speed up the provision of support.

Recommendation 8: The National Behaviour Challenge should dedicate a team and resources to radically improving close-knit working between specialist agencies, putting its secure data infrastructure at the heart of individualised support for children and families where the root causes of challenging behaviour require specific expertise.

Resolving the burgeoning crisis of behaviour and safety in English schools should be a top priority for this government. As the evidence in this report highlights, poor behaviour has a corrosive effect on teacher retention, school performance and the wellbeing, attendance and aspirations of students whose learning is disrupted. To address this, the government must prioritise a dual approach: in the immediate term, empowering teachers to engage parents and external agencies more effectively; and implementing a National Behaviour Challenge to systematically reduce classroom disruption with better data, targeted support and more effective and contextually relevant interventions. Fostering a safer and more supportive school environment ensures that the teaching profession continues to attract and retain the best and brightest, and enables all pupils to thrive academically and personally – ultimately contributing to long-term prosperity and growth.

Methodology

For our work with Opinium:

A representative sample of 1,000 14- to 16-year-olds in England studying towards their GCSEs were surveyed by the research organisation Opinium. Fieldwork was conducted between 5-16 February 2024 and the survey was undertaken with parental supervision. Representative quotas were set for age, gender and region, and schools were coded by Ofsted and Progress 8 scores using Department for Education data. The full data set can be found here.

For our work with Edurio:

We partnered with Edurio to run analysis on their extensive parent, pupil and staff data sets. Using data collected from three surveys during the 2022/23 academic year, Edurio reviewed responses from 73,000 pupils, 16,000 parents and 13,000 staff members to look at the relationship between the pupil experience and elements of school success. The measures reviewed were Ofsted rating, Progress 8, the proportion of pupils achieving expected grades in reading, writing and maths (for primary schools) and the proportion of pupils achieving grade 5 or higher at GCSE (for secondary schools). In addition to this, we looked at elements of the pupil experience relating to another, more qualitative measure of school success: how happy pupils felt at school. The findings were collated and reviewed before being summarised, with some examples, in this report.

Challenging behaviour is associated with poor attainment in school.^[_] It takes time away from teaching, which leads to poorer educational outcomes.

Reducing the time teachers spend on managing disruptive behaviour would boost the time they spend on teaching. In turn, this would boost attainment, as academic evidence^[_] across different countries and age groups suggests that more interactive teaching time improves pupils’ educational outcomes. Pupils who learn more in their school years are likely to enjoy higher productivity when they join the labour market.

Our model reflects these assumptions. It computes:

1. the potential time savings;

2. the improvements to education; and

3. the boost to GDP

that would ensue from reducing disruptive behaviour, and therefore the blow to future economic prosperity that would result from inaction.

Potential time savings

To calculate potential time savings, we modelled a speculative scenario where average levels of classroom disruption were reduced from current levels of around a fifth of teaching time by half.

The current figure was derived as follows. According to the Department for Education’s (DfE) National Behaviour Survey (NBS),^[_] for every 30 minutes of lesson time, 7 minutes were lost in disruption on average in May 2023 and 5.5 minutes in March 2023. We used the average of these figures to create a proxy value for 2023. This worked out as about one-fifth of total classroom time (which was also similar to findings in the previous year’s NBS).

The reduction in classroom disruption by half was based on the rates typically enjoyed by teachers who lost comparatively less time in the DfE’s NBS, thereby serving as a reasonable proxy for a low-disruption environment.

According to the survey’s May 2023 data, 3 per cent of teachers experienced no class disruption at all,^[_] which suggests that a “no disruption” scenario is possible. However, achieving these levels uniformly is likely to be unrealistic. We therefore used data from another band of “low disruption” respondents to inform our model instead; almost a third of respondents in primary and secondary schools lost between 1 and 4 minutes every 30 minutes (equivalent to between 3.3 per cent and 13.3 per cent of classroom time). We consider the mid-point figure within this bracket – 2.5 mins every 30 minutes (8.3 per cent of classroom time) – and conservatively round this to 10 per cent (the basis to our low-disruption figure).

Our assumption, therefore, was that on average, 10 percentage points of teaching time would be reclaimed by reducing disruptive behaviour. Given that teachers spend on average 24 hours teaching per week out of an average total of 50 hours in a working week,^[_] a 10 percentage-point boost means that teachers would gain about 2.4 hours per week in interactive teaching.

Improvements to Education

Our approach to calculating improvements to education was driven by a similar methodology to the one we constructed in our recent paper The Economic Case for AI-Enabled Education.

In short, we mapped “effect sizes” (which show how pupils’ educational attainment could change in response to a given intervention, expressed as a standard deviation around average grades) onto overall percentage improvements in educational attainment. This is based on a UK DfE study, according to which one standard-deviation improvement in GCSE grades resulted in an almost 20 per cent improvement in average lifetime earnings.^[_]

According to Wedel (2021),^[_] on average, one extra hour of teacher instruction leads to an increase of 0.03 standard deviations in test scores. Based on these findings, the 2.4 hours of extra teaching time we arrived at above suggests a 0.07 standard-deviation increase in test scores. Using our mapping method, this is equivalent to a 1.7 per cent improvement in overall educational outcomes.

For a detailed breakdown of the methodology we used, please see pages 6-9 of The Economic Case for AI-Enabled Education.

Boost to GDP

In our model, we assumed that better educational attainment would be the sole channel through which GDP would be boosted by reducing classroom disruption. There are of course other potential channels – for instance, boosts to the number of subsequent years of education pupils undertake due to higher attainment – but we did not model these.

We used a growth accounting model^[_] to gauge how improvements to attainment could translate into higher human capital and therefore boost GDP growth. We assumed a similar time lag as we did in the model we used in The Economic Case for AI-Enabled Education, which allows for various set-up steps (two years for initial planning and infrastructure, one year for piloting and one year for full-scale rollout). We also assumed a similar lag-time to the one we employed in that model when it came to enjoying the full weight of human-capital effects (that is, once all primary and secondary school pupils enjoy the benefits of a low-disruption scenario); we assumed that the average returns per year of education would rise by 1.7 per cent and that this effect would compound as more pupils progressed through the education system.

For a detailed breakdown of the methodology we used, please see pages 12-18 of The Economic Case for AI-Enabled Education.

Limitations

The GDP boost we have modelled is based on a speculative scenario that may understate or overstate the economic impact that reducing classroom disruption might have. On the one hand, the model assumed that better educational attainment would be the sole channel through which GDP would be boosted by reducing classroom disruption. In practice, boosting classroom time would likely lead to other positive multipliers – for example, reduced social costs that would likely flow from higher overall educational attainment or increases to the number of subsequent years of education pupils undertake. On the other hand, this model relies on the model being fully effective and rolled out equally across all schools.

The NBS, whose underlying data form the premise of the assumptions we make about the time teachers lose to classroom disruption, provides helpful insights into the scale of this loss – but it is not a fully comprehensive mapping exercise depicting the precise time teachers lose. For instance, the figures we draw from its underlying data are an average between two analyses conducted at two different points in 2023, in each case charting respondents’ experiences over a week-long reference period. It is therefore feasible that a margin of error exists in the data that flow from this exercise and these data should therefore be treated as illustrative.

In addition, our model assumes that extra hours of non-disrupted learning map cleanly onto higher educational attainment based on Wedel (2021). However, her analysis was an aggregate analysis of additional teacher time applied. By contrast, if there is a link between disruptive behaviour and innate educational potential, then reducing disruption may have less of a large boost to attainment levels than suggested in her study.

Although the UK education system is devolved and this report relates to education policy in England, the underlying data used to drive our model are only available at UK level. Its findings should be interpreted accordingly – and so the figures we derive assume that similarly decisive action, and therefore gains, were realised in the UK’s other devolved education systems.