Every child achieving and thriving: data drivers and our response
Blog / February 26, 2026
The Schools White Paper released this week has been widely discussed in the media for its plans related to the provision or otherwise of children with additional needs, special educational needs and disabilities.
Less discussed are its mentions of data and technology driven policy and practice, and we consider them here under 7 themes:
1. The definition of disadvantage is under review, with a view to adopting a household income-based measure in future.
2. New data collections, including a new ‘data spine’.
3. Individual support plans including its data linkage, with greater integration between education and health.
4. The EHCP status review and changing labels on children.
5. Claims made on data characteristics of “white working class”.
6. Artificial Intelligence and Technology.
7. Oversight and scrutiny, appeals and redress (and subsequent information flows).
In summary, there are lots of things it says. We have thought about and interpret from the available information, what those may mean. We’ve also considered potential unintended consequences, and with horizon scanning, we suggest what might be to watch out for as risks or that need safeguards and further information.
As the ‘Every Child Achieving and Thriving’ paper itself makes reference to policy “still in the making”, and a further consultation forthcoming in the summer, these are only comments on the skeleton outlines in the 118-page White Paper so far. We will be following up with these subjects as more detail emerges.
1. “We will develop and test a new model for targeting disadvantage funding to help schools narrow the attainment gap.” (Part 1. Chapter 3, p.44-45)
What it says: “How funding rates could vary according to the level of disadvantage, so that children with more entrenched disadvantage attract more funding. Using income data rather than the binary metric of eligibility for free school meals would enable schools to receive funding, for example using a stepped model, reflective of the level of disadvantage in their cohort. The model could take into account how low family income is, and for how long this has been the case. It would also remove the need for families to choose to take up the offer of a free meal in order to be eligible for deprivation funding, and reduce the administrative burden on schools. We are also considering whether to target funding based on the place a child lives, as well as their individual family economic circumstances. We will consult on our proposals in Summer 2026, inviting views from the sector and local communities on how we better allocate funding and address the impact of entrenched disadvantage on the education outcomes of children.” (p.45)
“The definition of disadvantage is under review, with a view to adopting a household income-based measure in future.” (p.105)
What it means: The personalisation of pupil funding, specifically pupil premium for those from less wealthy families, (“PP”) tying how much money a school receives to the individual and identifiable pupils in their classrooms, will become further entrenched. This moves away from past models, prior to academisation and the government of 2010, when funding was allocated to Local Authorities to distribute with regional knowledge. Today, calculating the money a school receives depends on how many children are on roll, or its estimates at a point in time. A school may have a minimum requirement to teach X subjects with Y number of teachers across a curriculum, and have a fixed infrastructure of Z number of classrooms to heat, light and maintain. But if fewer children live in that catchment area year-on-year (for example as housing becomes more expensive and excludes families with younger children) the school might receive less and less money from central government. If the money follows the pupil, not the infrastructure requirements of a setting, this exacerbates the financial problems for schools who have a minimum core provision cost, but ever less money to provide it with. There are various references to other block and high-needs funding and how all this fits together needs to be set out in detail. The fragmented educational financial distribution landscape where some schools are academies, and some still operate under the Local Authority, and some money follows the pupil is complex.
Unintended consequences? This personalisation of pupil funding as a household measure, could have all sorts of far-reaching implications for intrusive reach into families’ incomes, and their HMRC, DWP and potentially banking details, and its history. Will it ask only the same information of how pupil premium is calculated and FSM ever-6 duration today and that parents offer, or ask for the local authority or school to validate in terms of Income support and welfare benefits that trigger the PP? It sounds like more. Who is a child’s income connected to, and how far does their “income” get measured in “a household”, or how is “the place a child lives” tied to multiple households over a term, in shared parenting etc. Real-time income volatility could have substantive impact. Will overall funding be reduced to support only real-time provision? For pupils who move a lot, sometimes some of the most deprived, this could have complex bureaucratic effects for assessment, its length of verification processes and knock-on effects for pupils who move in and out of new schools? How will this tie into data drops from the school census, and money awarded retrospectively? How will it affect the stability of schools’ budget planning over time? It must not become mis-used in secret for more intrusive national surveillance with punitive outcomes.
What to watch out for: Summer 2026 consultation may be buried in the school, parliamentary, and media holidays. The Home Office getting involved to exclude migrant children from access to FSM or intervening in family life through data sharing like Operation BORTZ. Let’s hope it is well thought through with the necessary firewalls between departments and safeguards, and not the government’s intentions to trial its plans to “rewire the state” and “modernise digital government” on disadvantaged children and families, and already precarious educational settings.
2. New data collections and new data “spine”
What it says: “Access to high-quality data Schools have more data at their fingertips than ever before. Harnessing this can provide fast, accurate insights to improve children’s outcomes and help schools deploy their resources wisely. We want data to flow seamlessly, not be locked within individual systems, so insights can be put directly into the hands of teachers, leaders and parents. As the foundation to delivering this: We will develop a new ‘data spine’ that will create a secure, privacy-respecting and streamlined way to connect and share information across different systems in education.
“This will include data such as pupil records, attendance, progress and assessments. It will ensure a steady stream of connected information that empowers school staff, provides more immediate insight about the effectiveness of interventions (including AI and tech) and simplifies experiences for parents and carers. Along with updated safety standards and our planned sovereign education benchmarks, this will help support schools as they make decisions about what EdTech tools to invest in.”
“We will also ensure that new tools such as the recently launched Education Record app, and our plans for recording the needs of children and young people with SEND speak to one another and are informed by user research”(p98).
“In this world of change, children and young people need to feel a sense of belonging. Yet we have not prioritised this for our young people. Out of 27 European countries, the UK is last in how happy 15-year-olds are with their life.2 We want to give children their childhood back.“(p.7)
“By 2029, we expect every school to monitor pupils’ sense of belonging and engagement.” “We will publish a new Pupil Engagement Framework later this year to enable all schools to measure the key factors that determine their children’s engagement in education and make improvements” (p.62).
“We will also consult on how to better recognise schools that are successful in achieving progress for all children through the development of a new progress measure, in addition to Progress 8, focused on the progress made by children who start secondary school significantly behind their peers (see Chapter 3: Sidelined to included)” (p.100).
What it means: Who will build and host the new single unique ID that is the “data spine” and what structure that will take, where, is not defined. Details on what the (yet another) new digital platform to provide schools with more data analytics and benchmarking their performance against similar schools. More formalised data collection, administration, and bureaucratisation of processes that (should) already happen in schools. What any “new progress measure” will look like, is not specified. New obligations to monitor “pupils’ sense of belonging and engagement”.
Unintended consequences? Designing an administrative and datafied process about belonging is unlikely to engage well with the disengaged, and could lead to false sense of “doing something” rather than actually achieving anything. Increasing workload for the sake of national statistics is unlikely to add value and carries all sorts of risks. New and more profiling risks the same problems of the existing ones, pigeon-holing children onto expected “flight paths” of predicted linear outcomes. There is little detail to be able to comment on the aims “make decisions about what EdTech tools to invest in” or how these tie in to “a steady stream of connected information”. The new DfE pupil data “spine” could mean all sorts of things as yet undefined. The broader curriculum (in or outside Progress 8) will be welcomed by most, including improved digital citizenship, but will have to mean cuts in other things to make it all fit into the school day. It’s not clear yet what that will mean.
What to watch out for:
Creating a new ‘data spine’ to connect and share pupil information across different systems in education sounds very like the NHS data “spine”. The NHS Spine supports the IT infrastructure for health and social care in England, Wales and Isle of Man, joining together over 44,000 healthcare IT systems in 26,000 organisations. Since the DfE is “piloting the NHS number as a consistent identifier” and is mandating the use of a single unique ID in the Children’s Wellbeing and Schools Bill what might this mean? The unique single ID will need to exist for every child to use on demand in the event of it being needed in the moment for the care purposes permitted in the Bill. It is sensible to ask if and how this DfE “spine” and the NHS “spine” are intended to join up. If it is to be one and the same, we need to be asking several questions about how that, and indeed if that, should be done. The NHS number demands confidentiality and high security both lacking in educational settings, where data processing standards are very poor. The database state should not be used just to see into our lives, but to empower people with information about what others know about us and how it is used. Let’s see what this spine is intended to be for, used by whom, and with what oversight.
This new “steady stream” of data sounds remarkably similar to Google DeepMind’s streams project for which they got a rap over the knuckles, nearly a decade ago when London hospitals took the blame as data controllers, for granting a flow of data to DeepMind that enabled unlawful access to identifying patient data without consent. Is it in connection with them? The White Paper does say there is work in partnership “already” going on but what is unclear, and there are few answers what oversight exists or any way for the public to find out.
What the new progress measure look like and what data will it be based on, we will have to wait and see. We know that many schools already find “benchmarking their performance against similar schools” doesn’t work due to its contextual collapse. The norming is too generalistic, and outliers that can be explained by key individual children or cohorts and affect the whole set of data, are invisible. While the data is designed top-down, used to see-like-the-State, and not for the purposes that SLT need informed local decision-making, we continue to prioritise the wrong outcomes measured by numbers that do a bad job of measuring the human impact and holistic effects of school staff in a child’s and their family life. Many systems are offered by for-profit commercial providers today and they might not take kindly to the government intruding on their market, even though the DfE gives them the identifying pupil data for free to build their products with. Many of the current analytics systems are not fit-for-purpose by using the wrong inputs, so let’s see what the new one looks like.
The vague new monitoring obligations around wellbeing and belonging? What will this be used for? The idea of “belonging” is juxtaposed on p.7 with the UK ranking least happy for 15 year olds the EU, which is at best an unevidenced link, rather than concerns for climate, cuts to mental health services provision, unemployment and poverty, or any reasons from the delivery of education itself. The idea that “belonging” is a predictive measure of your lifetime pathway is contentious, and the White Paper makes reference to an American pysch study. Measuring “British Values” as did previous assessment of ‘belonging’ in children’s school engagement could be badly weaponised by all sorts of local, regional or national interests. Being told by the state or institutions what is an appropriate sense of belonging, and that there is a right and wrong way to do so, is political, highly complex and tied to some of society’s thorniest problems. The state doing so top down, is deeply authoritarian and removes agency just at a time when a backlash is brewing on what that means in schools, and what identity and belonging mean in the UK and beyond. Parents and researchers can have very different views on what is acceptable for researchers to record about children, and schools can fall foul of the law with substantive reputational consequences for them, risks to the children, and damage to trust in data collection and research.
Friction is not always a bad thing in data ‘flows’ and this sounds like plans to increase high risk processing moving data between places, when data access should be granted in safe settings to trusted users where the data stays in one place and is securely linked, rather than “flow” across or out of systems with safe and secure silo walls. The confidentiality of young people with SEND attributes on their records must be given priority as these labels can last a lifetime, and affect how they are treated, and we are drifting towards ever-weaker UK data protection laws and U.S. business models. How long until we have UK offers based on your results, such as in the U.S example where this car insurer discriminates its pricing against pupils with lower attainment?
3. “Individual Support Plans” (see Figure 6: Layers of support, Part 1. Chapter 3, p.50) and greater health interventions in school
What it says: “we propose introducing new Specialist Provision Packages, which will form the basis for an EHCP.””with input from education and health professionals”.
What it means: This is quite vague but could potentially signpost better integration of what is necessary and useful for working together in the best interests of a child. Formalising SEND support without the full ECHP could be improved. However, these professionals already make these decisions daily, and the barrier is rarely there but at Local Authority level when places of choice are rejected, or the plan not given.
Unintended consequences? The existing arrangements for data transfers between health and education systems are often weak and poorly defined. More data transfers can mean greater risks. More “data” of itself is useless without contextual meaning and accurate records. Both are notoriously weak in education, and health data risks being ‘downgraded’ in its handling outside the oversight of NHS-trained Caldicott Guardians.
What to watch out for: Individual care plans for direct care, for individual children must be done with families’ consent and shared understanding to support trusted collaborations. Bulk data transfers or those done behind the scenes without parental involvement, or where outcomes are disputed undermine these relationships and processes need clarified and secure and accountable infrastructures need built and/or improved. Better might mean that health data *already* in education records, starts to be treated with the due care and attention that health data demands in law and is missing today.
If there are more data and more “Experts at Hand” involved in decision-making and the appeals process, it is likely to increase data flows around and across multiple systems. Some of which are not interoperable today. The professionals mentioned for data sharing include police for example, but the education system has no control over those systems, or their funding, for example when Operation Encompass can take a couple of weeks to inform a school of a domestic abuse incident, something isn’t working. These infrastructure implications need de-conflated, specified, costed, and must not be underestimated.
4. Review of EHCP Plans (Part 1. Chapter 3, p.50)
What it says: “Children with an EHCP [label/ on record] in the early years will retain it until they complete the primary phase, when it will be reviewed.”
What it means: The EHCP status (and where a child will attend, associated provision of services, support, and funding) will be uncertain before the decision outcome and any appeal process is made. Today, the most common types of appeal are thousands of disputes about the contents of a plan and where the place will be, (64%), rather than the process of getting a needs assessment or securing a plan (33%).(Special Needs Jungle).
Unintended consequences? In today’s system, injecting this re-assessment workload and timing delays into pre-secondary, will be untenable. There is already too much demand to be able to assess EHCPs and provide approvals or rejections in anything like a sensible time period. A child can wait months even years. In England today, parents and carers who appeal to the SEND (Special Educational Needs and Disability) almost always win, but it takes time in which their child may not be in any or the best school for them, may have inadequate support, and the setting has inadequate funding to provide the correct provision. In 2024-25 academic year, 99% of appeals that went to a full hearing were decided in favour of the parents/appellant).
What to watch out for: Timing. This comes across and feels like the plans for access to information and its re-assessment are too late in a child’s life stage, rather than “at the end of primary” the timing in the process needs to work backwards from taking into account the EHCP plan label assessment, its ‘validation and approvals process’ (already extraordinarily long, taking months and years in some cases), back to prior to the term in which a family must make decisions on Admissions applications to the setting for the next stage of education, often in late autumn term or January of the year ahead of a September start. Realistically that means a child moving in regular circumstances needs to know their (re-)assessment outcome by the half-term of autumn term in year 6 at the absolute latest and that gives very little time for families to consider and often visit, potential settings, especially necessary for those with SEND. The decision really needs to be clear by summer in the year prior to transitioning into 6, giving families a term to make any available choices which, in current capacity, means the re-assessment starting in year 4 or early 5 and that’s unlikely to account for appeals — when a plan may only have just been put in place that was asked for in Reception or Year 1 or 2. There are fewer than 3 years by 2029 (p.58) to fix this substantial structural capacity issue and consider its squeezed timing between the first designation and its re-assessment.
5. Data underpinning the statistics in the claims made on “white working class”
What it says: “Survey data also shows that white working-class children are more likely to find the transition to secondary school more difficult than their peers.” (p.31)
“Other groups of FSM eligible children also have particularly low outcomes, especially White Irish, White and Black Caribbean, and Black Caribbean children. And disadvantages overlap and compound: outcomes for children both with SEND and eligible for free school meals are far below their peers who face neither of these challenges. Research suggests that half of children who do not secure a pass in their English and maths GCSEs were judged to be behind at age 5.” (p.40)
Unintended consequences?
What to watch out for: (page 63) “The government’s Race Equality Unit will work with us to undertake further research on the issue of racist abuse in schools.” Who is on that unit affects the outcomes. Don’t forget, the findings of the Sewell Commission report were controversial and contested by experts, and that the current party and individuals in government, will change.
We should also be horizon scanning for what an increase in personalisation, linking health and education and SEND, linking ethnicity and educational outcomes, and seeing educational costs like-the-State means or might be used for — what the State thinks your education returns to the Treasury (HMRC) or costs the State (DWP) is already being measured by LEO data. They join up all the available data they have from each individual’s lifetime education and income and the welfare state. Let’s speculatively add in all the data that is already available or will be soon, from health systems about every baby and child too.
As the Education Select Committee and the DfE have more recently considered, and some of its commercial partners are actively working on, some people consider genomics and/or ethnicity matters in debates of ‘nurture vs nature’ especially for children with SEND. Some believe in scientifically-debunked claims how education and health are linked, suggesting connections between DNA, race, and IQ.
Some who get into power believe genetics outweighs teaching. Let’s not forget that Genomics England’s Newborn Genomes Programme (NGP) is the first national newborn screening programme in the world to incorporate whole genome sequencing, and our Minister for Health plans to provide a genomic test for every newborn baby by 2035. The same year, that “As reforms become embedded, we expect the number of children and young people needing a Specialist Provision Package, and hence an EHCP to access support, will return to around today’s level by 2035.” (p.57) Some people believe such predictive personalisation offers appealing and definitive routes to pre-emptive interventions based on polygenic risk scores of children’s needs and outcomes, and package these as in children’s best interests, and others into the economic best interests of the state. But these generally collapse into eugenicist myths and lead to very dark outcomes under scrutiny. It doesn’t actually work like that.
We need to be very careful that the State is not building the infrastructure for such systems and outcomes, without any oversight and scrutiny.
6. AI and Technology
What it says: A lot, in fact AI has nearly sixty mentions. (p.78) “We will help teachers harness AI, including Oak’s lesson assistant ‘Aila’, in ways that amplify their impact, streamline their workload, and preserve their irreplaceable role in shaping learning, as covered in further detail in Chapter 7: Enabling innovation and ambition.”
“As the education technology (EdTech) sector continues to develop rapidly, we must ensure that evidence of impact on outcomes for children drives decisions on which tools schools use. We must learn from the schools who are already innovating to maximise the benefits that technology can bring while carefully managing the risks.” (pp.91-92)
“Alongside this, we want to cultivate a vibrant marketplace of innovative, commercial offers by encouraging schools to select offers that can help them the most.”
“We will build a new one-stop-shop digital platform for schools and signpost them to quality assured support to empower smart investment in data-driven, evidence-led self-improvement.”
“The new digital platform will provide schools with data that diagnoses their performance strengths and needs, allowing them to benchmark their performance against similar schools.”
“We will work with the EEF so that more school leaders and teachers access its high-quality research on what works to improve academic progress and attainment, while supporting networks and organising conferences to share best practice. We will continue to work with the EEF so that more schools benefit from their expertise and vital role in championing and implementing evidence-based practice.” (p.93)
“We are working with the school sector to set out clear principles for the safe and evidence-based use of AI in education. We will work with AI and sector experts to establish safety standards and efficacy assessments for EdTech and AI tools used in education, including by updating our AI product safety standards to account for emerging risks around mental health and social, emotional and cognitive development, and to help make sure that AI products used in education do not cause harm.”
“In 2026, we will build on the content store (which holds trusted content from Oak, EEF and others) so that AI tools are underpinned by rigorous, sequenced content that supports real progress for children.”
“We will work with industry to drive adoption of our standards and digital National Curriculum to improve the safety and quality of tools. This includes working with frontier labs – organisations such as Google DeepMind and OpenAI who are already working with government on initiatives designed to develop sovereign solutions to the UK’s hardest problems, including in education.”
“We will invest an additional £23 million to expand the EdTech Impact Testbed pilot into a 4-year programme and develop our wider research strategy to understand the effects of AI on learning and wellbeing.”
“Following DfE’s initial £1m funding in 2024, Innovate UK (part of UKRI) has now committed an additional £1m to developing promising AI feedback and formative assessment tools.”
“Arming young people with secure and well-founded knowledge is central to media literacy, and making citizenship compulsory in primary schools will introduce key content to children earlier, such as law and rights.” (p.26)
What it means: Tech is taking on a substantive emerging role in this package. Who will these companies be other than BigTech mentioned? TBD.
Unintended consequences? The public-private data partnership between Google DeepMind and London hospitals broke UK data protection law through DeepMind building its Streams App using patient data without a consent basis. The challenge with any company thinking they are “too big too fail” or above the law, is they are likely to do so again. What does that mean for children?
Consumer rights organisations such as the BEUC, have called for OpenAI to be investigated. Others have banned its products in Higher Education, and other countries are embracing customised products in their public service delivery at scale. What it means for the UK depends on our investment risk and how integrated and embedded we allow this single company to become. We are at grave risk of losing control over significant parts of our sovereign systems if we outsource their infrastructure and controls to third parties. Given the current geopolitical instability, doing this with U.S. based companies is at best naive.
What about the “wider research strategy to understand the effects of AI on learning and wellbeing.” We need to be very careful what is going to be measured and by whom. There is a huge swell of industry who claim to be able to assess wellbeing but the Ts&Cs state they are not diagnostic tools. In doing so, these avoid MHRA oversight and regulation. These are not the apps we are looking for to interact with our children’s most personal thoughts and feelings. Only this week, the Scottish government and researchers have deleted data from a controversial census which asked school pupils as young as 14 about their feelings and wellbeing (including sexual experiences), thanks to hard won concessions from parents in Scotland. Parents in England will not stand for similar and the backlash to years of uncontrolled data grabs is growing. They are already coming for EdTech, so teachers who will be stuck in the middle of these contrasting values and asks on what “good” edTech looks like and what is wanted, need the training asap.
The new process of establishing “minimum expectations for home-to-school partnerships, with clear communication expectations, tools for parents to support learning at home, and an improved complaints process” will therefore need to be as usable for edTech as much as for SEND. The new citizenship including law and rights sounds interesting and we hope children’s own rights will be in scope.
What to watch out for: (p.81) The detail is still to come. The training offer on technology seems to be narrow about using AI, not about AI itself. This is an opportunity to offer a far more holistic approach, catching up what’s been sorely missing for the last decade for staff and be about the digital environment as a whole. This needs to include data literacy, and law and ethics for the digital environment, such as data protection, privacy, plagiarism, and research impacts, especially when common edTech involve vast amounts of data processing, analytics, benchmarking, automated decisions or profiling pupils.
There is very little in the public domain on the Content Store and no processes around it to ensure scrutiny or oversight of its inevitable scope creep. What’s going on with Oak National is being fought in the courts.
The risk is that the government is more incentivised to promote its aim of cultivating, “a vibrant marketplace of innovative, commercial offers” than scrutiny or stopping the use of products that are found to be detrimental to outcomes or are unsafe or unlawful (case in point, see support for Google DeepMind).
Answers need to be more forthcoming on what the DfE “working with […] organisations such as Google DeepMind […] who are already working with government […] including in education,” really means (see 4 above). We know the content store exists (with Faculty AI) to enable AI training and product development. We know the government has no new plans to plan to inform the public if their data held in public administrative datasets is used for AI development; and the mechanisms by which departments publicly say they are already using public data for AI development do not include them all, and therefore are not (yet) working well. The duty is to inform the public, not to “publish a privacy notice” that does not inform the public of something they do not know happens so cannot know to proactively look for a notice about joined-up use of data they do not even know is collected or shared by government departments in the first place. There are potentially opportunities for new ways of doing that that could work well, and we could be discussing those. We should be looking at how the tools that we have, or those emerging such as the new App to inform pupils of their exam results, the recently launched “ Education Record app (p.98)” can be routes to integrating better access to information how those data about you are used, and by whom (in this case your grades, SEND, and it will soon be more).
We welcome the remarks on working, “with the EEF so that more school leaders and teachers access its high-quality research on what works to improve academic progress and attainment,” and, “implementing evidence-based practice.” (p.93) (In fact we said just this about EEF evidence, only last week at a panel at the LSE. Catch up here.) But we have been waiting a long time already for this government that committed to an ICO Code of Practice in Education from the Dispatch Box over a year ago, in the Data Use and Access Bill. That turning evidence into policy and practice could reduce workload for staff by sharing informed data impact assessments for example. And prevent unnecessary harms and costs by stopping schools from procuring products that cause harm, are ineffective, or simply don’t work for pupils. We’d love to see that happen.
7. Oversight: A new remit for the Children’s Commissioner to provide oversight and scrutiny
What it says: “There will also be a new remit for the Children’s Commissioner to provide oversight and scrutiny of SEND reform implementation.” (p.51)
What it means: as it says on the tin. Detail TBD. Some are worried that there may also be a reduction of rights from Tribunal compared with today. It’s unclear if the formal individualised plans (ISPs) that are will be on a similar legal footing as ECHPs, and so could be appealed or objected to, if the same Tribunal process does not apply.
Unintended consequences? The role and terms of appointment for the Children’s Commissioner for England are defined in Part 6 and Schedule 5 of the Children and Families Act 2014. The Commissioner is not an Ombudsman and, in general, cannot conduct investigations into the case of an individual child. The SEND Tribunal does that today.
What to watch out for: The Commissioner is toothless to act if these reforms do harm to children, and it risks creating an artificial impression of oversight and scrutiny that simply does not exist, unless put on a statutory footing. The role is separate in Wales. There is a weak appeals process and that is what parents need. Will the government put these powers on a statutory footing?