A child can sit through every Mathematics lesson and still arrive at Year 2 without a stable sense of number.
They may count accurately but not trust that a rearranged set has the same quantity. They may know number facts one day and rebuild them from one every time the next. They may add by counting all, confuse the structure of “difference,” fail to see five and three as parts of eight, or lose the relation between concrete objects, diagrams and symbols. More whole-class exposure does not always repair these gaps because the lesson keeps moving while the missing structure remains upstream.
Targeted numeracy intervention creates a different learning environment. A small group receives additional time on the precise number relationships that are not yet secure. The adult can watch every strategy, interrupt inefficient counting, use mathematical language deliberately, vary representations and give many more opportunities to explain and practise than a whole class can usually provide.
The Education Endowment Foundation’s second effectiveness trial of 1stClass@Number 1, completed in September 2026, is unusually strong current evidence for this mechanism. Year 2 pupils received 30-minute small-group sessions three times a week for ten weeks, delivered by trained teaching assistants alongside normal Mathematics lessons. Pupils made the equivalent of two additional months’ progress on average, including pupils eligible for free school meals, with a high-security rating. EEF’s broader small-group tuition evidence similarly emphasises targeting specific needs.
The practical question is not “Does extra Maths help?” It is: what makes additional small-group number teaching more powerful than simply giving struggling learners more of the same work?
The 50-second answer
Targeted numeracy intervention is additional, time-limited Mathematics teaching for learners whose current number knowledge is limiting progress. The intervention is usually delivered in a small group so the adult can diagnose strategy use, model efficiently, give high response density, correct misconceptions immediately and adapt examples to the learner’s current state.
The mechanism is precision plus extra supported practice. The child receives more time, but time alone is not the treatment. Effective intervention identifies the bottleneck—number composition, place value, counting strategy, additive reasoning, comparison, mathematical language or another foundational relation—and teaches that relation through carefully sequenced examples, representations, talk and practice.
The 2026 1stClass@Number 1 trial gives strong evidence for one structured programme: groups of four, trained teaching assistants, 30-minute sessions three times a week for ten weeks, aligned with the ordinary curriculum. The correct generalisation is not that every small maths group produces two months’ progress. It is that targeted, structured, well-trained small-group intervention can create meaningful additional progress when it repairs a genuine bottleneck and reconnects to classroom Mathematics.
1. Small-group time should solve a visibility problem
Whole-class teaching limits how closely one adult can inspect each child’s strategy. A small group increases the number of times the teacher can watch, question and respond before misconceptions harden into habits.
Concrete example. Four pupils solve the same addition using counters; the adult notices one child recounts the entire set while another uses known number bonds.
Boundary and caveat. Small groups do not guarantee better diagnosis. If all pupils complete worksheets silently, the visibility advantage disappears.
2. Diagnosis should name the mathematical relation
Broad labels such as ‘weak at maths’ do not tell the adult what to teach. Intervention starts by identifying the relation that is insecure: magnitude, part-whole structure, place value, additive composition, multiplicative thinking or mathematical language.
Concrete example. Two pupils score similarly on a test; one cannot compose ten flexibly while the other misreads comparison language. They need different teaching.
Boundary and caveat. A short diagnostic is a hypothesis, not a complete profile. Watch how the learner responds during teaching.
3. Counting strategy matters
Children can produce correct answers with strategies that are too slow or fragile for later Mathematics. Intervention can help learners move from counting all toward more efficient derived facts and part-whole reasoning.
Concrete example. For 7+5, a pupil counts twelve objects from one. The adult models making ten: 7+3+2.
Boundary and caveat. Efficiency should follow understanding. Do not force a faster strategy before the child sees why it works.
4. Cardinality needs to remain stable across arrangements
Young learners may recount when objects are moved because they do not fully trust that quantity remains constant. Structured tasks can separate number from visual arrangement.
Concrete example. Six counters are spread apart, pushed together and rearranged while the child predicts whether the total changes before recounting.
Boundary and caveat. Recounting is not always a misconception; it can be a reasonable checking strategy. Look for repeated dependence.
5. Subitising can support number structure
Quick recognition of small quantities can reduce counting load and help children see numbers as composed patterns rather than serial lists.
Concrete example. Dot patterns for five, six and seven are flashed briefly; pupils describe how they saw ‘five and two’ or ‘three and three.’
Boundary and caveat. Subitising practice should connect to arithmetic relationships, not become a speed competition.
6. Part-whole composition is a major foundation
Addition and subtraction become more flexible when children understand that a whole can be decomposed and recomposed in multiple ways.
Concrete example. Using two colours of cubes, pupils make eight as 5+3, 6+2 and 4+4, then discuss what stays the same.
Boundary and caveat. Manipulatives are a representation, not the final goal. Move toward diagrams and symbols as understanding stabilises.
7. Place value requires unitising
Understanding 23 as two tens and three ones involves treating ten ones as a new unit. Learners who see only separate digits can struggle with later calculation.
Concrete example. Pupils bundle ten sticks, exchange ten ones for one ten, and connect the action to the written number.
Boundary and caveat. Concrete materials can become ritual if exchange rules are followed without understanding the invariant quantity.
8. Comparison language can hide numerical understanding
Words such as more, fewer, difference, greater than and less than carry relationships that can block problem solving even when computation is available.
Concrete example. A pupil can calculate 12−8 but misinterprets ‘How many more?’ The intervention contrasts several comparison sentence forms.
Boundary and caveat. Keyword rules are dangerous. Teach the relationship, not a word-to-operation shortcut.
9. Representation should change deliberately
Moving among objects, number lines, ten frames, bar models and symbols helps learners see the same relation in different forms. The adult should make correspondences explicit.
Concrete example. A ten-frame showing 8+4 is connected to a number line jump and then to an equation.
Boundary and caveat. Too many new representations at once add cognitive load. Stabilise one before using variation to test transfer.
10. Mathematical talk makes strategy visible
Small groups create space for every learner to explain how they know. Language reveals whether the child sees structure or only performs actions.
Concrete example. Each pupil finishes the stem, ‘I knew there were nine because…’ using a different representation.
Boundary and caveat. Oral fluency can mask shallow understanding, and expressive-language needs can mask strong mathematical thought. Accept gesture and representation as evidence too.
11. Adult questioning should diagnose, not interrogate
Useful questions expose relations: What changed? What stayed the same? How else could you make it? Which is greater and how do you know?
Concrete example. After a child makes seven from five and two, the adult asks for another decomposition without adding more explanation.
Boundary and caveat. Constant questioning can slow practice and feel like testing. Use questions at decision points.
12. Immediate feedback prevents repeated error
The small-group setting allows the adult to correct a misconception before the learner practises it ten more times.
Concrete example. A child writes 42 for four tens and two ones in a task that actually represents two tens and four ones; the adult returns to the units and asks for another attempt.
Boundary and caveat. Not every hesitation needs interruption. Protect enough flow for independent thinking.
13. A second attempt turns feedback into learning
The adult’s explanation is not the endpoint. The child needs to act on the correction and then solve a near problem independently.
Concrete example. After correcting a comparison error, the pupil solves a fresh example with different quantities without the model visible.
Boundary and caveat. Near examples show immediate uptake, not long-term retention. Revisit later.
14. Response density is a major small-group advantage
Four pupils can each produce many more mathematical responses in thirty minutes than in a whole-class queue. That matters when the responses are meaningful and corrected.
Concrete example. Each learner solves, explains, points, predicts or constructs multiple examples across one session.
Boundary and caveat. High response count is not enough if tasks are trivial. Density and quality must coexist.
15. Training shapes teaching quality
The 1stClass@Number 1 trial included substantial TA training and structured resources. This matters because intervention often places significant instructional responsibility on teaching assistants.
Concrete example. TAs learn how to elicit strategies, use mathematical language and respond to errors rather than simply supervise practice.
Boundary and caveat. Training attendance does not guarantee execution. Coaching and observation can help when the method is complex.
16. Teaching assistants need mathematical knowledge for teaching
Knowing how to calculate is different from recognising why a child counts inefficiently or how a representation can reveal structure. Intervention staff need knowledge of likely developmental routes and misconceptions.
Concrete example. A TA recognises that a correct answer produced by counting all still signals an important instructional target.
Boundary and caveat. Do not create a two-tier system where struggling learners receive less expert teaching. Training and teacher oversight matter.
17. Structured materials can reduce preparation burden
A well-designed programme supplies sequences, examples and resources so the adult’s limited preparation time goes into understanding pupils rather than inventing activities from scratch.
Concrete example. A handbook specifies the focus and progression while still allowing adaptation to the group’s responses.
Boundary and caveat. Scripts can become mechanical if adults follow them without understanding the mathematical purpose.
18. Adaptations should preserve the mechanism
Learners differ, so examples, pace and scaffolds may need adjustment. The programme’s central mathematical relations and response opportunities should remain intact.
Concrete example. A group spends longer on number bonds using different objects but keeps the planned part-whole reasoning.
Boundary and caveat. Adapting away the challenging relation can make sessions pleasant but ineffective.
19. Additional time should be truly additional
The tested programme added roughly fifteen hours of numeracy alongside ordinary lessons. Replacing core Mathematics with intervention changes the educational exposure and can reduce access to the wider curriculum.
Concrete example. Pupils receive three intervention sessions without missing the main daily Mathematics lesson.
Boundary and caveat. Timetabling is difficult. Schools should inspect what pupils miss when intervention is scheduled.
20. Curriculum alignment supports transfer
EEF’s implementation findings emphasised alignment with the national curriculum. The child should encounter compatible language, representations and methods in the intervention and main lesson.
Concrete example. The class and intervention both use the same language for tens, ones and equality.
Boundary and caveat. Alignment does not mean duplication. Intervention should add precision and intensity.
21. Teacher–TA communication matters
The class teacher knows current curriculum and broader pupil performance; the intervention adult sees detailed strategy use. Sharing concise information can improve both environments.
Concrete example. The TA reports that a pupil now composes ten securely but still struggles with comparison language; the teacher creates opportunities to use that language in class.
Boundary and caveat. Communication should be lightweight enough to sustain. Long reports can consume the preparation time the programme needs.
22. Progress monitoring should match the target
If intervention teaches number composition, use fresh composition and arithmetic tasks to see whether the skill improves. Broad term grades are too distant.
Concrete example. A pupil who once counted all now solves unfamiliar 8+7 problems using make-ten strategies with fewer prompts.
Boundary and caveat. A trained assessment format can inflate gains. Use fresh examples and classroom work too.
23. Transfer to ordinary Mathematics is the real goal
Smooth performance in the small group is only valuable if the learner uses the knowledge in class, homework and later topics.
Concrete example. The pupil applies part-whole reasoning during a new subtraction lesson without the intervention manipulatives present.
Boundary and caveat. Transfer can take time and may need explicit bridging. Do not assume it happens automatically.
24. Confidence should grow from capability
The trial reported improvements in pupils’ confidence and enjoyment, but the most durable confidence is attached to evidence that the learner can now do something independently.
Concrete example. A child says, ‘I don’t need to count them all anymore—I can make ten.’
Boundary and caveat. Easy praise can create short-term comfort without changing Mathematics. Keep success meaningful.
25. Group composition should support instruction
Groups of four are small enough for dense feedback but still require the adult to manage different states. Learners should share enough of the target need for one lesson sequence to fit.
Concrete example. Four pupils work on additive composition, though their exact strategies differ.
Boundary and caveat. Do not group solely by total score. Broad ‘low maths’ groups can contain incompatible needs.
26. Attendance affects dose
A ten-week intervention loses power if pupils repeatedly miss sessions. Record actual sessions received and respond when timetabling or absence reduces exposure.
Concrete example. A child receives 20 of the intended 30 sessions; outcome interpretation includes that reduced dose.
Boundary and caveat. Attendance data should not be used to blame families. It helps judge whether the programme had a fair chance.
27. Preparation time is part of implementation
EEF highlighted sufficient preparation time for TAs. Structured sessions still require materials, review of prior performance and understanding of the next learning step.
Concrete example. A TA spends ten minutes before the group reviewing yesterday’s errors and preparing the correct manipulatives.
Boundary and caveat. Unpaid or unrealistic preparation expectations can degrade quality over time.
28. Small-group numeracy should not become permanent tracking
Intervention is a route back into stronger participation, not a fixed identity. Review who still needs the support and what the exit criterion is.
Concrete example. A child leaves the group after repeated evidence of secure number foundations and returns to ordinary differentiation.
Boundary and caveat. Exiting too early can lead to relapse. Use repeated evidence, not one successful session.
29. Non-response requires rediagnosis
If the learner receives adequate dose and quality but progress remains limited, the school should reconsider the target, prerequisite profile and possible wider needs.
Concrete example. A pupil improves counting but not number comparison; the team checks language, magnitude understanding and appropriate support pathways.
Boundary and caveat. Teachers should not diagnose clinical conditions from intervention response alone.
30. Intervention can reveal core curriculum weaknesses
If large numbers of pupils need the same repair, the school should ask whether whole-class teaching or curriculum sequencing is generating the gap.
Concrete example. Half the cohort lacks secure part-whole knowledge, prompting leaders to strengthen earlier Mathematics teaching rather than expanding intervention indefinitely.
Boundary and caveat. Some cohorts genuinely enter with unusual gaps. System review and targeted support can both be appropriate.
31. The intervention should get simpler as competence grows
As learners internalise number relationships, prompts and concrete supports should fade. The child should carry more of the strategy and explanation.
Concrete example. The group moves from counters to diagrams, then to mental reasoning and symbolic tasks.
Boundary and caveat. Some representations remain useful even for experts. Fading means reducing dependence, not banning tools.
32. The strongest evidence is a changed mathematical habit
Progress is deeper than a higher post-test score when the child now sees number structure differently: decomposes flexibly, chooses efficient strategies and explains relationships.
Concrete example. A learner spontaneously says 14−6 is 14−4−2 because making ten is easier.
Boundary and caveat. Do not overclaim broad mathematical transformation from one intervention. Track the specific capabilities taught.
Worked cases
Case 1 — Counting all to making ten
A Year 2 pupil answers addition correctly but starts every problem from one. In the small group, the adult uses ten frames and number bonds to make ten visible, then varies the numbers and removes the frame. After several weeks, the child begins using make-ten strategies independently in class. The target was not ‘accuracy’—it was a more efficient number structure.
Case 2 — The word problem that is really language
A pupil calculates fluently when equations are shown but fails comparison stories. The intervention contrasts ‘how many more’, ‘difference’ and ‘fewer’ using objects and diagrams. The child then solves fresh word problems without keyword shortcuts. Targeted support prevented unnecessary arithmetic reteaching.
Case 3 — A trained TA changes the quality of practice
Before training, a teaching assistant mainly checks whether answers are correct. After structured professional learning, the TA listens for strategy, asks pupils to justify, models representations and requires second attempts after feedback. The same thirty minutes contains far more diagnostic and mathematical work.
Case 4 — When the whole curriculum needs repair
A school finds multiple intervention groups are reteaching the same place-value concept. Leaders audit the Year 1 sequence and discover insufficient exchange practice. Core teaching is strengthened while current pupils continue targeted repair. Intervention data becomes curriculum evidence.
Practical route for teachers and school leaders
Start with a precise diagnostic question: which number relation is preventing progress? Form groups around that need, protect additional time, use trained adults, and maximise pupil mathematical responses. Connect objects, diagrams, language and symbols deliberately. Feedback should lead to another learner attempt.
Record actual sessions received, the strategy the child uses, and whether the new relation appears in ordinary classwork. Review exit criteria on a planned date. If many pupils need the same repair, examine core curriculum as well as intervention capacity.
Practical route for learners
A small maths group is not a statement about what kind of mathematician you are. It is extra practice on a part of number that needs to become easier and more reliable.
Explain how you got the answer, not only what the answer is. Try the new method on a different problem. When a strategy starts to make sense, use it in ordinary class even when the intervention materials are not there.
Practical route for parents and families
Ask the school what specific Mathematics skill the intervention is targeting, how often sessions occur, who delivers them and how progress will be reviewed. ‘Extra maths’ is less informative than ‘building flexible number bonds and moving beyond counting all.’
At home, support simple number talk and games if the school suggests them, but do not turn family time into a second intervention programme. The core teaching and diagnosis remain school responsibilities.
Common failure modes
- Grouping pupils by a broad low score instead of a specific numeracy need.
- Using small-group time for silent worksheets rather than dense interaction.
- Assuming correct answers mean efficient or secure number understanding.
- Giving struggling pupils extra time but weaker instructional expertise.
- Using manipulatives without connecting them to mathematical relationships and symbols.
- Replacing core Mathematics lessons rather than adding targeted support.
- Allowing intervention language and methods to conflict with classroom teaching.
- Measuring confidence without checking mathematical transfer.
- Keeping children permanently in intervention groups after the target is secure.
- Expanding intervention indefinitely instead of strengthening core teaching when the same gap is widespread.
Frequently asked questions
What did the 2026 1stClass@Number 1 trial find?
Pupils receiving the programme made the equivalent of two additional months’ progress in Mathematics on average compared with controls. The result had a high security rating, and a similar average effect was found for pupils eligible for free school meals.
How was the programme delivered?
In the evaluated model, trained teaching assistants worked with groups of four pupils for 30 minutes, three times per week, over ten weeks, alongside normal Mathematics lessons.
Does any small-group maths intervention produce the same effect?
No. The trial tested a structured programme with training, resources, curriculum alignment and a specific dose. General small-group tuition evidence is positive on average, but effects depend on targeting and quality.
Why use teaching assistants?
Trained TAs can provide additional instructional capacity. The evidence does not imply that untrained adults or generic supervision are sufficient; training, preparation and teacher linkage matter.
When should intervention end?
When repeated fresh evidence shows the target skill is secure enough for the learner to progress with ordinary classroom differentiation, with continued light monitoring where needed.
The final idea
Targeted numeracy intervention works when it changes what the extra time is for.
The child is not simply given more Mathematics. They are given more visibility, more carefully chosen examples, more explanations, more corrective feedback and more chances to build the exact number relationship that ordinary lesson pace did not secure.
The 2026 evidence is strong for one structured programme and consistent with the broader logic of targeted small-group tuition. But the useful lesson is not a brand name or a magic group size. It is an instructional architecture: diagnose precisely, train the adult, add real dose, connect representations, require pupil thinking, align with class teaching, monitor transfer and stop when the support has done its job.
The governing rule is: repair the mathematical bottleneck, not the label ‘low attaining’—and make every extra minute produce more independent number sense than the learner had before.
Sources and further reading
- Education Endowment Foundation — 1stClass@Number 1 second trial, completed September 2026
- Education Endowment Foundation — 1stClass@Number 1 Promising Programme
- Education Endowment Foundation — Small group tuition
Continue exploring on eduKateSG
- How X Works Hub
- How Pre-Teaching Works | Why a Ten-Minute Head Start Can Change the Main Lesson
- How Guided Play Builds Early Numeracy | Why Games Need Mathematical Structure, Not Just Fun
- How Mathematics Works for Posting Group 1 (PG1) Students
A ten-week intervention architecture
A programme such as 1stClass@Number shows why small-group numeracy intervention benefits from a planned arc rather than thirty unrelated sessions. Weeks one and two establish baseline strategies and secure entry routines. Weeks three to six concentrate on the central number relationships and representations. Weeks seven to nine increase variation and problem solving so the learner must choose rather than copy a method. Week ten checks transfer and identifies what still needs ordinary classroom support.
The exact sequence depends on the programme and learner. The broader principle is cumulative design. Each session should inherit something from the previous one, and earlier ideas should reappear often enough that improvement becomes durable rather than session-specific.
A timetable full of sessions is not yet an intervention architecture. The content must accumulate.
A session structure that protects response density
Thirty minutes can disappear quickly in a group of four. A useful structure is: brief retrieval of earlier learning, one focused teaching episode, guided practice with every pupil responding, changed examples, and a final independent check.
This prevents the teaching assistant from turning the group into a mini whole-class lecture. Every learner should count, represent, explain, choose and correct repeatedly.
The final check matters because fluent group performance can be socially supported. A pupil may follow peers, echo language or infer the next move from repetition. One fresh individual item gives the adult better evidence about current control.
Diagnosis should identify the earliest unstable number relationship
A low Mathematics score is too broad for targeted numeracy teaching. The teacher or trained TA needs to know whether the learner’s first unstable relationship is counting sequence, cardinality, subitising, number composition, place value, additive reasoning or another prerequisite.
Consider a child who repeatedly struggles with two-digit addition. More addition worksheets may not help if the underlying problem is that ten ones and one ten are not yet represented flexibly. Repairing place value changes the problem upstream.
The diagnostic principle is the same one used across intervention: fix the first broken link that is causally limiting later performance, then check whether the downstream task improves.
Groups of four still need individual mathematics
A small group creates more opportunity, but one confident pupil can still carry the discussion. Adults should design simultaneous or rotating responses so every learner produces evidence.
Students can build their own representation, write an answer before discussion, hold up number cards simultaneously, or explain one step each. When one pupil answers aloud, another can be asked to paraphrase or test the claim with a different example.
The goal is not equal airtime for its own sake. It is enough independent mathematical action that the adult can see who understands which relationship.
Training changes what the teaching assistant can notice
The 1stClass@Number evaluation is useful because the intervention did not simply place pupils with an extra adult. Teaching assistants received structured training, materials and mathematical language. Staff reported growth in their own knowledge and confidence.
That matters. An adult needs enough subject knowledge to distinguish a counting error from a place-value error, recognise a useful representation, ask a question that exposes strategy, and decide when a child needs another example rather than the answer.
Extra staffing is not the mechanism by itself. Trained instructional attention is.
Adaptation should preserve the mathematical progression
Small groups invite responsiveness. A pupil may need more counters, a different representation or another day on one relation. Useful adaptation changes the route while preserving the programme’s core mathematical sequence and intended learning.
The danger is drift. If the adult replaces structured number work with generic homework help because it feels immediately useful, the intervention may lose the cumulative mechanism that justified the extra time.
Record important adaptations and ask whether they preserve the target relation. Flexibility is strongest when the adult knows what must remain invariant.
Link-teacher communication prevents a parallel curriculum
Targeted intervention should be additional to normal Mathematics, not a separate world with different terminology and methods. A link teacher can help align current classroom content, identify where transfer should appear and understand what the small group is revealing.
Communication can stay lightweight: current intervention target, current strength, one recurring misconception and one classroom bridge.
The purpose is not to make the class teacher reproduce the intervention. It is to ensure that when the learner returns to ordinary Mathematics, the repaired knowledge is recognised and used.
Confidence should be built from mathematical evidence
Small-group learners often arrive with a history of being slow, corrected publicly or grouped as weak. Intervention can improve confidence, but confidence repair is strongest when it follows real competence.
“I can now make ten in four ways without counting every object” gives a child a specific reason to update self-belief. “You are amazing at Maths” may feel positive and still provide no evidence.
Adults should notice strategy growth, successful correction and independence. Confidence then becomes an accurate prediction that effort has a usable route.
A parent-facing explanation of the group
Families benefit from a precise description: “Your child is receiving ten weeks of extra small-group Mathematics focused on number relationships that are making later work harder. It is additional to normal lessons. The group meets three times a week. We will review using fresh tasks and ordinary classroom work.”
This language protects dignity. The child is not assigned a permanent ability identity; they are receiving concentrated teaching for a current bottleneck.
Parents can support with ordinary number talk and games if they wish, but core intervention remains a school responsibility.
A local evaluation plan
A school using targeted numeracy intervention can monitor four layers. Delivery: did the planned sessions occur? Learning inside the target: did the specific number relation improve? Transfer: is ordinary class Mathematics changing? Sustainability: can staffing, training and preparation time be maintained?
Keep engagement and confidence as separate indicators. They matter, but they should not be used to imply attainment gains.
The evaluation does not need to mimic the EEF trial. It needs to give the school enough evidence to decide whether to continue, adapt, intensify or stop.
The exit should return the learner to ordinary Mathematics with more control
Targeted groups succeed when they make themselves less necessary. As the bottleneck stabilises, prompts and special representations can fade, and the learner should use the knowledge in whole-class work with fewer adult cues.
Exit should be based on more than one successful intervention session. Use fresh examples and classroom evidence.
Some pupils will need further support. The principle remains: the group is a repair mechanism, not a permanent destination.