Beyond eligibility for free school meals (eFSM): data review (summary)

As part of its commitment to addressing inequality and supporting those facing barriers to their education, it is important that the Welsh Government (WG) understands and measures the relationship between socio-economic disadvantage (SED) and learners’ outcomes. The current and most notable proxy measure for SED is eligibility for free school meals (eFSM).

In recent years, however, there has been some concern that eFSM data is subject to limitations, including that it underestimates poverty levels and does not fully account for a range of other relevant factors beyond income alone. In response to recommendations from previous reports including the Review of School Spending in Wales and Developing a new data and information ecosystem, WG is therefore considering how SED is measured among learners. This report presents the findings of an initial phase undertaken by Wavehill Ltd.

The core aims of the study were to determine the range of needs the Welsh Government and other stakeholders have with regards to learner SED data, including the purposes for which they use it, and to provide a comprehensive understanding of how they currently use eFSM data. It also seeks to gather perspectives on eFSM’s limitations and recent challenges, including ways to overcome these, and to explore what other forms of data stakeholders are using or feel they require to better meet their needs.

The research methods involved two components of primary data collection, undertaken mainly between March and August 2024, consisting of a proforma exercise and qualitative research in the form of interviews. Overall, 33 proforma responses were received, including 13 from the Welsh Government policy and analysis teams, 19 of the 22 local authorities in Wales and the Welsh Local Government Association (WLGA).

In total, 39 online interviews were conducted involving 55 participants, including the Welsh Government, local authorities, school leaders, Estyn, Careers Wales, Universities and Colleges Admissions Service (UCAS), Qualifications Wales and the Children’s Commissioner for Wales. The interviews sought to explore emerging findings from the proforma analysis in greater depth or to gather insights from a wider group of stakeholders who are also important users of data on learner SED.

The need for data on socio-economic disadvantage

Stakeholders have diverse evidence needs regarding the use of data on SED among learners, as part of their shared goal of addressing educational inequality. These can be categorised as:

• For informing policy development across the Welsh Government and local authorities, for forecasting budgets, and for wider poverty monitoring activity.
• For calculating and distributing funding provided to local authorities, schools, and families in a way that is equitable and reflective of need.
• For providing targeted support to those learners who are most in need.
• For statistics and monitoring of outcomes, such as attendance and attainment at the national, local, school and individual levels.
• For school improvement, including as a means to identify learners with additional needs and to design and implement targeted support interventions.
• For managing school catering in terms of modelling levels of demand for meals and the associated resources that should be allocated to catering.

The data is widely used to inform the development and implementation of a range of policies at the national and local levels. This includes contributing to national milestones and the wellbeing of future generations (WBFGA) assessments, but mainly to strategic and financial planning within government and local authorities regarding educational disadvantage and improving outcomes for learners.

The research has confirmed the scale of hypothecated and unhypothecated funding allocated by WG on the basis of eFSM to local authorities via the Local Authority Education Grant (LAEG) and the Pupil Development Grant (PDG), and the Local Government Settlement. It also highlighted the variability of its use in local authority formulae for the Individual Schools Budget (ISB) when distributing funding to schools.

The data is used by local authorities to provide targeted support to vulnerable learners, including through Vulnerability Assessment Profiles (VAPs) to identify them as part of the Youth Engagement and Progression Framework.

For WG, eFSM plays an important role in the production of statistics that examine the relationship between socio-economic disadvantage and key outcomes, including attendance and attainment. The data is also used to produce the All-Wales Core Data Set (AWCDS) and publicly accessible tools such as My Local School.

The data available at the individual learner level is also important to the work of the Administrative Data Research Unit (ADRU), which coordinates data linking via the SAIL databank, and to those working in post-16 statistics through Medr (the new body for tertiary education) for tracking long-term outcomes for individuals.

School improvement activities undertaken by local authorities and schools use eFSM data to monitor school- or individual-level outcomes for disadvantaged learners and to benchmark these using AWCDS. This enables priority planning and evaluation. It is also used by schools alongside other data to identify learners vulnerable to the impact of SED and to inform the design and delivery of interventions or support.

In addition to these, data is used for practical purposes relating to catering, informing budgeting and resource allocation to understand and meet demand for food in schools, and informing a range of work delivered through arms-length bodies, including the Children’s Commissioner for Wales, Careers Wales, Estyn, and Qualifications Wales.

Stakeholders recognise the value of eFSM data as an individual-level measure that provides insight into learners’ household circumstances, and consider it crucial for meeting many of the purposes for which they require information on disadvantage.

It is clearly defined and well-understood, making it valuable for local authorities delivering services to a range of audiences, whether allocating grants to schools and settings or simplifying the FSM application process for parents.

It is frequently collected and made available annually through PLASC, making it relatively up to date, while a related benefit is that local authorities are the ‘data owners’ (as they gather the data), facilitating data sharing between departments and with external stakeholders. The criteria for eligibility and the data collection approach have also remained relatively stable throughout the years, enabling trend-based analysis of the prevalence of eFSM. Equally, the approach in Wales aligns broadly with approaches in other UK nations, particularly England, allowing comparisons across borders.

The research highlighted several technical challenges in the process of collecting and recording data that may compromise its accuracy. This relates to the variability in approach across local authorities, the potential for error in the manual handling of data, and the effects of recent issues posed by Transitional Protection (an order implemented to ensure that learners previously entitled to FSM did not lose their eligibility during the move to UC). Stakeholders also pointed to an underestimation in learners’ eFSM, recognising the many reasons why a parent may not register their child.

The data on eFSM are not perceived by stakeholders to accurately reflect learners living in poverty, particularly due to narrow eligibility criteria that overlook many working families. Indeed, the fact that the criteria have not changed since 2019, despite increases in the national living wage and minimum earnings requirements for UC, means that fewer families in relative poverty are likely to be eligible. Equally, local authorities expressed concerns around families whose income sits on the boundary of the income threshold and whose eFSM status may fluctuate. This is particularly challenging considering the timing of PLASC data collection in January and the potential impact of seasonal work on parental income.

These limitations affect the accuracy with which the data can reflect the scale and nature of income-related disadvantage and their ability to serve as a meaningful proxy. It presents a particular challenge for calculating and distributing funding for demand-led fixed budgets, with the potential to misjudge the level of need or misallocate funds. Where learners living in poverty are missing from eFSM data, this also undermines its use in targeting support for learners.

Stakeholders also acknowledged that eFSM data, which is a proxy for income, does not capture the multidimensional nature of socio-economic disadvantage and overlooks other factors relevant to learner experience and may be instrumental in bringing about inequality in educational outcomes. This includes factors such as housing tenure and parental occupations and qualifications, and how these intersect and are compounded by other existing inequalities associated with factors such as ethnicity and ALN. It was suggested that relying on data that does not account for these factors may risk directing support erroneously or to an insufficient number of learners, and does not enable stakeholders to compare and respond to different needs fully.

The extent to which these limitations impact stakeholders’ use of eFSM data varies, with the biggest implication being for user confidence. Some local authorities acknowledge the problems but continue to use the data as the best available.

The introduction of UPFSM has raised concerns that parents will no longer register their primary school children as eligible for FSM if they are not required to do so to receive a free meal, which, in turn, may threaten the robustness of the data.

The research found that 16 of the 19 responding local authorities perceived that the introduction of UPFSM had impacted eFSM registrations. Of these, 10 authorities reported receiving fewer applications from parents than expected and presented several reasons for associating this with UPFSM.

Whilst the concerns relating to the impact of under-reporting of eFSM due to UPFSM are valid, any reduction in numbers cannot be disentangled from other confounding factors. For some local authorities, confusion about how Transitional Protection status is recorded appears to have had a greater temporary impact on the data than UPFSM. The exact effect of these matters on the level of registrations has also not yet been borne out or determined in official figures.

Given this uncertainty, there were mixed views among stakeholders regarding the implications of UPFSM and its potential effect on registrations. Some suggested it had not affected their use of the data, while others said it had further undermined their confidence in the data and reflected on the issues it might cause for funding.

Stakeholder views on potential solutions to the limitations and recent challenges surrounding eFSM data include improving validation processes, enhancing what is currently available by using averages rather than point-in-time figures, adopting ‘ever 6 FSM’ which identifies learners who have been eligible for eFSM at any point in the previous six years, or increasing the frequency of eFSM capture from which more accurate understanding of numbers eligible over a time period would be possible.

A key approach several stakeholders suggested would be to access benefits data from the Department for Work and Pensions to determine those who are eFSM and open up the potential for auto-enrolment (piloted in several locations in England).

A further solution to addressing eFSM’s inability to capture all learners living in poverty fully may be to increase the Universal Credit (UC) income threshold, introduced in 2019, so that it keeps pace with inflation. This would ensure learners do not become ineligible even if their household income does not improve.

Improvements and solutions may not fully meet stakeholders’ needs, and this appears to prompt many of them to draw on other forms of data available. The most common reason for doing so was to provide greater contextual information, allowing them to gain a better understanding of other aspects of disadvantage beyond income alone.

The most widely used alternative form of SED data by stakeholders is the Welsh Index of Multiple Deprivation (WIMD). It is used across all groups of stakeholders, including WG, local authorities, schools, and others, for their respective purposes, but with different approaches and motivations. There are, however, concerns that limits its utility, including that the data is outdated (an update is due in November 2025) and the fact that it is an area-based measure with a limited level of granularity.

Stakeholders also deploy a range of other household- or individual-level indicators, including Census data, Acorn/CACI data, and DWP data, whilst some users are actively testing correlations of these data with eFSM, such as for the Local Government Settlement. In a school improvement context, data on various aspects of SED from different sources are used, sometimes in bespoke composite measures to allocate targeted support to learners. However, for school improvement leads —and indeed for many schools — the role of practitioner judgment remained the decisive factor in identifying vulnerable learners.

Stakeholders expressed no strong views and offered limited feedback on specific additional data they would like to have access to beyond what is currently available. Rather than speculate on other forms of information and intelligence that might be useful, their focus was on maintaining and improving the data they have, using defined approaches and techniques to make the best of what is available.

Analysis of information gathered via proformas and interviews, however, provided insight into the kinds of data and their specific characteristics that different groups of stakeholders find most appealing to meet their needs, whether eFSM in its current or revised form or something else. This has been mapped to a series of tables in the technical annexe and should be consulted for a fuller understanding.

At a minimum, there is an expectation that data will be of high quality, have been validated to minimise errors, and involve transparency in collection methods. It should also be readily available and accessible for stakeholders who wish to use it. The understandability, timeliness, frequency, stability, comparability, and granularity of SED data were identified as important characteristics to consider in more detail. There is variation both within and across stakeholder groups in the extent to which they place importance on these and in their exact requirements for each.

All stakeholders feel it is important to continue to access information on income. They place high value on the eFSM proxy, particularly given its potential for improvement and enhancement, and on more direct access to data on benefits from the DWP. There is also a desire among some stakeholders for income data presented in gradients to enable an appreciation of different levels of poverty, and/or available over the longer term to provide insights into persistent poverty.

Stakeholders also have an interest in other contextual information on socio-economic disadvantage. This is to provide a better overall understanding of the factors influencing educational outcomes and the varied nature of learner needs, to deliver support in the most effective way possible. It includes insights on parental occupation and qualifications, housing tenure, and access to services, among others, with a view to differences in urban and rural experiences.

There is hesitancy amongst stakeholders towards the potential introduction of alternative or supplementary data, and several barriers and implications were raised. Concern is expressed around the limits of devolved power and WG’s ability to set its agenda and manage its own processes independently of the UK Government. In particular, there is awareness of obstacles to developing data-sharing agreements with DWP, which limit access to benefit data.

The most notable concern for stakeholders is related to resources and the potential burden involved in any change to the collection and recording process for local authorities and schools. There is also unease around ownership and availability of data, the ability of stakeholders to analyse it effectively, and risks associated with a lack of consistency in approach that may undermine user confidence.

In addition, a key challenge in overcoming the limitations of eFSM is the complexity of the solution required due to the volume of data, the number of users, and the range of purposes for which they use the data. Moreover, the collection and use of eFSM span a range of policy areas, and there could be considerable implications of a changing approach.

To facilitate any change in data use, stakeholders say there should be time and research dedicated to testing and demonstrating its appropriateness. They call for resources including guidance, training and leadership, as well as further consultation and engagement with stakeholders to ensure their needs are met.

Further consideration could be given to understanding the current and future impacts of challenges posed by Transitional Protection and the roll-out of UPFSM, including monitoring the validation and mitigation measures that have been put in place. Attention also needs to be given to the static earnings threshold for eligibility in light of inflationary pressures, and the implications it has for the representativeness of the data over time and its use for monitoring and intervention. Additional focus could be placed on exploring different options to address the shortcomings of eFSM data and improve its use, such as enhancements to the use of averages, the ever-6 eFSM, direct access to benefits data, or steps in line with recent developments in England to increase the eligibility threshold.

Many stakeholders require data with a variety of different characteristics, and there is interest in more contextual information that provides insights into other aspects of disadvantage than income alone. Any subsequent phases of the research could involve an assessment of how effectively different indicators and measures of socio-economic disadvantage, whether currently available or developed in a more bespoke fashion via data linking, meet stakeholder requirements.

In undertaking the next stages of research, consideration should be given to engaging with and consulting stakeholder groups. Given the varied requirements and the potential expectations of stakeholders that could emerge, consideration will also be needed for how their data needs can be accommodated within the scope of the resources available to deliver change.

Within the Welsh Government, several substantive projects and policy developments are in progress and may be affected by further research; these should be taken into account to minimise disruption. This includes an internal review of the PDG, a review of the local government funding formula, the development of approaches to monitoring the new Curriculum for Wales, and work related to the Welsh Benefits Charter.

Report author: Megan Clark, Oliver Allies, Huw Lloyd-Williams, Hilda Bernhardsson, Jakob Abekhon

Views expressed in this report are those of the researchers and not necessarily those of the Welsh Government.

For further information please contact:
Schools Research Branch
Knowledge and Analytical Services
Welsh Government
Cathays Park
Cardiff
CF10 3NQ

Email: SchoolsResearch@gov.wales

Social research number: 122/2025
Digital ISBN: 78-1-80633-758-3