Education: Standards for streaming and setting across mainstream secondary education

Ministerial submission

To: The Secretary of State for Education

Issue

1. The government has no statutory position on the use of streaming or setting in mainstream secondary schools. Schools — including academies and free schools, which are not bound by the National Curriculum — are free to organise teaching as they choose. In practice, subject setting is already the dominant model: 97% of secondary pupils in England are educated in schools that use ability grouping for at least some subjects, and setting for mathematics is near-universal. Despite this prevalence, there are no national standards for how sets are formed, staffed, reviewed, or monitored for equity. This submission concerns whether that absence of regulation is appropriate.

Background

2. Setting and streaming are distinct practices. Streaming places pupils in fixed ability groups that apply across all or most subjects; it is relatively uncommon in English secondary schools and is more strongly contraindicated by the research evidence. Setting places pupils in ability groups within individual subjects, allowing a pupil to be in a higher set for mathematics and a lower set for English; it is the predominant practice in English secondary schools for core subjects and is the primary subject of this submission. Both are well-established features of English secondary education and have been endorsed as matters of school autonomy by successive governments. The previous Labour government of 1997-2010 explicitly encouraged setting, with Tony Blair describing it as the expectation for schools seeking to raise standards; the current government has not taken a position on the practice.

3. An analogous stratification operates at the level of school choice in parts of England and lies entirely outside the scope of any within-school grouping policy. In cities where secondary school density supports genuine parental choice — London and Leeds are examples — oversubscription criteria, inspection ratings, and the proximity rules of popular schools effectively concentrate higher-attaining and more advantaged pupils in higher-rated institutions. This school-choice effect is unavailable in smaller towns and rural areas, where a single school or a small number of local schools serves the whole population irrespective of attainment. England retains 163 selective grammar schools in 36 local authority areas, providing an explicit institutional expression of the same principle. Periodic calls to expand the grammar school sector draw on the same rationale as within-school setting applied at institutional level; successive governments have declined to expand selective admissions, but the question recurs as a strand of the wider ability grouping debate.

4. The Education Endowment Foundation (EEF) Teaching and Learning Toolkit rates setting and streaming as having zero months’ average progress impact across all pupils. For lower-attaining pupils the effect is negative, with an estimated loss of one to two months’ progress per year compared to similar pupils taught in mixed-ability groups. For higher-attaining pupils the evidence of benefit is minimal. The EEF rates the evidence security as very limited — based on 58 studies, many dated, few of which are randomised controlled trials — and cautions that this reflects the difficulty of conducting robust research in this area rather than confidence in the null finding.

5. Research led by Professor Becky Francis at University College London, based on a study of 140 secondary schools, identifies three structural problems with current setting practice. First, approximately 31% of pupils are misallocated to sets above or below their assessed attainment level, with disadvantaged pupils and some ethnic minority groups disproportionately placed in lower sets. Second, lower sets are systematically taught by less qualified and less experienced teachers, creating compounding disadvantage. Third, set placement tends to be self-reinforcing: pupils placed in lower sets receive less challenging instruction and are less likely to move upward. The Francis intervention testing best-practice setting — with careful allocation, regular review, and teacher parity — produced no measurable improvement in outcomes and suffered from very low school uptake, suggesting the model is difficult to implement at scale.

6. Your government has committed to raising attainment and reducing the gap between disadvantaged pupils and their peers. This submission examines whether the current absence of any policy or quality framework for grouping practice is consistent with that commitment, and what intervention — if any — is warranted.

The General Case

7. The case for government intervention in grouping practice rests on three grounds:

• (a) the absence of quality standards for setting has produced widely variable and frequently inequitable practice; the misallocation rate and teacher quality disparity identified by the Francis research represent systemic failures that school autonomy has not corrected;
• (b) the scale of ability grouping in England — exceptionally high by international standards, with 97% of pupils in grouped settings against an OECD average of 37% — means that any systematic problems with the practice affect the overwhelming majority of secondary pupils;
• (c) the government’s attainment gap commitments are difficult to reconcile with a practice that the best available evidence associates with widening disparities between higher and lower-attaining pupils.

8. The case against intervention also rests on substantial grounds:

• (a) school autonomy in pedagogical organisation is well-established and valued by heads and governors; government mandates on grouping would represent an extension of central control into classroom practice with limited evidence of benefit;
• (b) the EEF evidence base is rated very limited; the research does not provide sufficient confidence to justify statutory requirements in either direction;
• (c) setting for mathematics is already near-universal and is strongly preferred by many parents and teachers; any policy that appeared to discourage or restrict the practice would be politically contentious.

Options

Option A: Preserve school autonomy

9. Under this option, the current position is maintained: schools determine their own grouping practices without statutory obligation or official guidance. No new Ofsted inspection criteria are introduced for grouping arrangements.

10. For:

• (a) preserves a well-established tradition of professional autonomy in pedagogical organisation;
• (b) avoids legislating on a matter where the evidence base is genuinely limited and contested;
• (c) allows schools to respond to their own pupil cohorts and parental expectations, which vary substantially across different communities.

11. Against:

• (a) the evidence of systematic inequity in current setting practice — misallocation, teacher quality disparities, demographic stratification — constitutes a strong case for quality standards even if the evidence on attainment impact is ambiguous;
• (b) the absence of any government position on a practice affecting 97% of secondary pupils is increasingly anomalous given the government’s stated commitment to closing the attainment gap;
• (c) school autonomy has not produced convergence on better practice over the decade since the Francis research began; the market for quality improvement in this area has not worked.

Option B: Evidence-based guidance with revised Ofsted inspection criteria

12. Under this option, the Department for Education would publish non-statutory guidance on best practice in ability grouping, drawing on the EEF and Francis research. The guidance would cover placement processes, set review intervals, teacher allocation across sets, and demographic monitoring. Ofsted’s inspection framework would be updated to require inspectors to report on grouping arrangements and their equity implications as part of the quality of education judgement. No schools would be required to set or to cease setting, but those whose practice is inequitable would be expected to improve.

13. For:

• (a) addresses the quality and equity problems in current setting practice without mandating any particular organisational model;
• (b) uses the existing Ofsted accountability mechanism rather than new legislation;
• (c) is consistent with the government’s broader commitment to evidence-based education improvement and could be implemented quickly with limited cost.

14. Against:

• (a) non-statutory guidance has limited traction with schools, particularly academies and free schools, which are not bound by DfE guidance in the same way as maintained schools;
• (b) updating Ofsted criteria without a clear evidence base for what good practice looks like risks creating compliance activity — schools gaming inspection criteria — rather than genuine improvement;
• (c) the Francis research’s own best-practice intervention produced no measurable improvement at scale, suggesting that guidance and inspection pressure may be insufficient to change entrenched practice.

Option C: Statutory requirement for subject setting in core subjects with quality safeguards

15. Under this option, all mainstream secondary schools — including academies and free schools — would be required by statute to operate subject-specific setting for mathematics, English, and science, subject to statutory quality standards. The standards would require: set allocation based on assessed attainment with defined review intervals (at minimum annually); teacher qualification parity across sets; demographic monitoring with reporting obligations; and parental access to information about set placement decisions and the process for requesting review. A new inspection framework element would assess compliance.

16. For:

• (a) formalises and regulates a practice that is already nearly universal, extending its consistent application to the minority of schools that do not currently set and imposing quality standards on those that do;
• (b) the statutory quality safeguards directly address the three structural problems identified by the Francis research — misallocation, teacher quality disparity, and demographic stratification — providing a framework that voluntary guidance has not delivered;
• (c) parental transparency requirements and a formal review mechanism address the due-process deficit in current practice, where set allocation is often opaque to families.

17. Against:

• (a) the EEF evidence does not support a conclusion that well-implemented setting improves overall attainment; statutory requirements could embed a near-universal practice more deeply at a time when the evidence base calls its effectiveness into question;
• (b) binding academies and free schools through primary legislation on pedagogical organisation represents a significant encroachment on their statutory autonomy and would face resistance from academy trusts;
• (c) whole-school streaming — a more extreme and more strongly counterindicated form of grouping — is not meaningfully addressed by a framework focused on subject setting; the statutory regime could inadvertently endorse streaming by omission.

Resource and Cost Implications

18. Option A involves no cost. Option B would require DfE guidance development and Ofsted framework revision, both modest costs manageable within existing budgets. Option C would require primary legislation, an implementation period for schools to develop compliant systems, and a new Ofsted inspection element; the DfE would also need to develop the demographic monitoring framework and reporting obligations. Costs are not estimated here but Option C carries significant implementation complexity, particularly in extending the regime to the academy sector.

Legal and Devolution Considerations

19. Education is substantially devolved: Scotland, Wales, and Northern Ireland each have their own curriculum and inspection frameworks, and this submission concerns England only. Within England, the position of academies and free schools is the principal legislative complication: they operate under funding agreements with the Secretary of State rather than under local authority oversight, and their autonomy from the National Curriculum is a statutory feature. Any Option C requirement would need to be implemented through a combination of primary legislation and academy funding agreement revision; the latter would be legally complex and potentially contested.

20. No human rights concerns are identified: the right to education under Article 2 Protocol 1 ECHR does not prescribe any particular teaching method or grouping arrangement. The demographic monitoring obligations under Option C would require careful design to comply with the UK GDPR and the Equality Act 2010, as they would involve processing special category data (ethnicity) for regulatory purposes; this is achievable but requires explicit legal basis.

Requested direction

21. Ministers are invited to determine whether the current absence of any policy or quality framework for ability grouping in secondary schools is consistent with the government’s attainment gap commitments; and, if not, to indicate whether non-statutory guidance and revised Ofsted criteria (Option B) or a statutory quality framework for subject setting (Option C) should be developed for further consideration.