The Tutor Handbook · Volume 0077 · Series ID THB-0077
The Tutor Handbook: Complete series index.
Three learners are sitting around one table. One needs the tutor now. One needs the tutor to stay away for another two minutes. One needs a question, not an explanation.
This is the central difficulty of small-group tutoring. The tutor is not merely teaching three people at once. The tutor is continuously deciding who should think, who should speak, who should attempt, who should observe, who should receive feedback, who should wait, and when a shared explanation is genuinely shared rather than merely convenient.
A weak small group can look impressively busy. Everyone has a worksheet. The tutor moves quickly. Questions are answered. Corrections happen. Yet one learner receives most of the cognitive work, another receives most of the tutor’s speech, and the third completes tasks whose errors nobody has had time to inspect.
A strong small group does not solve this by dividing a ninety-minute lesson into thirty equal minutes. Equal time can be unfair to the learning job. The learner who needs a two-minute discriminating question should not receive ten minutes of unnecessary explanation merely because another learner needed ten minutes. The learner who needs twenty minutes of independent practice should not be interrupted every five minutes to keep attention visibly balanced.
Three-Learner Orchestration is the tutor’s real-time allocation of attention, questioning, explanation, independent attempt, peer visibility, feedback and observation so that each learner receives the learning condition they need without one learner’s need becoming the class rule for everyone else.
The broader canonical page How 3-Pax Small-Group Tuition Works owns the general model of attention, comparison and independent thinking in a three-student class. This Tutor Handbook volume owns a narrower professional decision: how the tutor actually orchestrates the room moment by moment when the three learners are not in the same state.
Quick answer
Do not equalise minutes. Equalise access to the learning operations that matter.
Every learner needs enough opportunity to think before answers are supplied, enough independent attempt to make their own state visible, enough observation for the tutor to see what happened, and enough feedback to improve the next attempt. The amount and timing of those ingredients can differ. Fairness means the tutor can justify the difference by the learner’s current job rather than habit, preference, confidence, speed or volume of speech.
The tutor should know, at any point in the lesson, which learner is in direct instruction, which learner is in guided practice, which learner is in independent work, what evidence is expected to return, and what event will cause the tutor to switch attention. This is not a rigid rotation. It is a controlled distribution of cognitive opportunity.
1. A small group is not three simultaneous one-to-one lessons
If the tutor tries to run three one-to-one lessons in parallel, the room fragments. Each learner waits while the tutor completes a private explanation for someone else. The tutor becomes a scarce travelling resource. Every transition costs context. The group offers little advantage beyond physical proximity.
A small group becomes educationally different when some learning conditions can be shared without erasing individual diagnosis. A common explanation may serve two learners. A peer’s wrong method may become a comparison case for the other two. One learner can work independently while the tutor observes another. A shared question can produce three different responses that reveal three different models.
The EEF’s current small-group tuition synthesis defines small-group tuition as one educator working with a small number of pupils and reports positive average effects, while emphasising targeting support to specific needs. It does not establish a universal optimal three-student routine. The orchestration described here is a practical eduKate framework for preserving that targeting inside a shared lesson.
2. Fairness is not equal tutor contact
Imagine that Alicia is practising algebra method selection independently. She needs uninterrupted attempts because tutor presence itself can become a cue. Beatrice is reading a difficult paragraph and needs one question that helps her distinguish relevant evidence from a general sentence. Ciara has misunderstood a new Science mechanism and needs a concise explanation before she can practise meaningfully.
If the tutor distributes ten minutes to each learner in turn, Alicia is interrupted unnecessarily, Beatrice may receive nine minutes of excessive help after her question is resolved, and Ciara may wait too long for information she genuinely lacks.
A fairer distribution could be two minutes with Beatrice, eight minutes explaining and checking Ciara, then a return to Alicia’s completed independent set. Alicia may have received the least direct tutor talk and the most valuable learning condition.
This is why the tutor should distinguish attention received from learning opportunity received.
3. Every learner needs a visible current state
Orchestration fails when the tutor only knows what the learner is doing, not why the learner is doing it.
“Alicia is on Question 7” is activity information. “Alicia is completing a mixed set without method cues so I can observe selection errors” is an operating state. “Beatrice is reading page three” is activity. “Beatrice is choosing the most direct textual evidence before she writes an explanation” is a learning state.
A tutor can orchestrate well only when each learner’s current state contains four things:
- the learning job;
- the level of allowed support;
- the evidence expected to return;
- the event that should trigger tutor re-entry.
Volume 0066, The Configuration Snapshot, owns the wider current-route record. In a live three-learner room, the tutor needs a lightweight mental or written version of that information for each learner.
4. Direct instruction should be shared only when the prerequisite is shared
A shared explanation saves time only when the learners genuinely need the same explanation.
Suppose all three students are learning a new algebraic transformation and all three lack the same prerequisite. A concise shared model can be efficient and coherent. But if Alicia already understands the transformation and needs method selection while Beatrice and Ciara need conceptual explanation, forcing Alicia through the same exposition spends her attention on information she already owns.
Shared teaching should therefore pass a simple test: is the missing knowledge or decision actually common?
Grouping by age, textbook page or lesson time is not enough. The National Student Support Accelerator’s session-content guidance recommends purposeful small-group formation based on content needs and targeted learning goals. That programme-level guidance does not prescribe private-tuition grouping, but the underlying idea matters: common delivery should follow a common learning need.
5. Independent work is not the tutor’s waiting room
When the tutor turns to another learner, independent work should continue the learner’s route rather than merely occupy time.
Busywork is easy to assign. Good independent work has a specific function: retrieve, select, execute, explain, compare, revise, transfer, or produce evidence under reduced support. The learner should know the finish condition and what to do if genuinely stuck.
This is particularly important in a three-person class because independent time can be substantial. If each learner spends forty minutes of a ninety-minute session away from direct tutor attention, that forty minutes must still be designed teaching time. Otherwise the economics of small-group tuition are achieved by quietly converting much of the lesson into supervised worksheet completion.
The National Student Support Accelerator’s session-structure guidance explicitly includes purposeful independent practice and formative assessment inside a consistent session structure. The practical lesson is not to copy its template mechanically, but to recognise independent practice as a deliberate component rather than leftover time.
6. Tutor re-entry needs a trigger
If learners must repeatedly call, “Teacher, I’m done,” orchestration depends on interruption. If the tutor constantly checks every learner, independence is broken by surveillance.
A better design gives each learner a re-entry trigger. Examples include:
- after three independent attempts;
- after the learner has marked where uncertainty began;
- after choosing between two methods and recording the reason;
- after completing a short retrieval set;
- when the same error repeats twice;
- when the learner reaches a designated checkpoint;
- when the learner has tried the agreed help-seeking step and remains blocked.
The trigger lets the learner continue without the tutor while preserving a reliable return point.
7. Questioning is an allocation of thinking, not airtime
A small group can appear participatory because many questions are asked. Yet the same confident learner may answer first, the fastest learner may dominate explanations, and the quietest learner may spend the lesson listening to correct reasoning produced by others.
The tutor should therefore decide who the question is for before allowing the room to answer it.
Sometimes all three should think silently before anyone speaks. Sometimes each should write a first step. Sometimes one learner should explain while the other two compare the explanation with their own. Sometimes the tutor should deliberately ask the learner whose model is least visible. Sometimes a question should remain private because another learner’s answer would destroy the diagnostic value.
AERO’s current Monitor Progress examples emphasise varied methods, active participation and frequent checking for understanding. Its Vary Practice resource also highlights varying participation routines and encouraging students to question and explain their thinking. These classroom examples do not establish one ideal tutoring sequence, but they reinforce the principle that participation design matters for what thinking becomes visible.
8. Protect private thinking before public explanation
Peer explanations can be powerful, but they can also erase evidence.
If Alicia immediately explains the correct method aloud, Beatrice and Ciara may recognise the route without ever showing whether they could have selected it themselves. The group gains information and loses diagnostic clarity.
When independent evidence matters, use a simple sequence: private think or attempt first, then shared explanation. This protects each learner’s first model before the group improves it.
After the private attempt, public reasoning becomes more valuable because learners can compare rather than merely receive. “I chose the same method for a different reason” or “I thought the diagram meant the opposite” gives the tutor richer information than a chorus of agreement after the strongest learner speaks.
9. Peer visibility is not peer teaching by default
One learner’s work can help another learner without turning the first learner into a substitute tutor.
The tutor might display two anonymous solution routes and ask all three to compare them. One learner can explain a decision while the others identify where their own reasoning diverged. Two learners can predict the next step in a third learner’s incomplete method. These routines use peer visibility as evidence and comparison.
The tutor should be careful when one learner repeatedly becomes the explainer for the group. Explanation can strengthen the speaker’s understanding, but it can also make the other learners dependent and burden the strong learner with unpaid instructional responsibility. The group is not well orchestrated if one student becomes an assistant teacher because the adult cannot attend to everyone.
10. Observation needs protected time
Tutors often speak because speaking feels like teaching. In a small group, some of the most valuable minutes involve the tutor saying almost nothing.
Watch Alicia choose a method before she writes. Watch Beatrice move between the question and the passage. Watch Ciara reconstruct a Science mechanism before comparing her wording with the model. Watch Denise recover after a difficult item. Watch Emily decide which study task enters the next block. Watch Faith decide whether she needs help.
These are learner operations that disappear when the tutor intervenes too early. Orchestration must therefore allocate not only tutor attention but tutor restraint.
11. Fast learners should not become invisible learners
A learner who completes work quickly can receive less observation precisely because the room seems to need the tutor elsewhere.
Speed may hide fragile checking, shallow explanation, over-reliance on familiar patterns or avoidance of challenge. A fast correct answer still deserves occasional inspection of method, transfer and reasoning.
The tutor does not need to slow the learner down artificially. Instead, schedule evidence touchpoints that make quality visible. Ask for a justification, a changed-condition problem, a comparison of two methods or an explanation of what would make the chosen method fail.
12. Slow learners should not receive permanent rescue
The opposite risk occurs when one learner struggles more visibly. The tutor stays close, prompts frequently and begins to carry the learner through every difficult step. That learner receives the most tutor time and the least independent evidence.
High support may be correct during new instruction or genuine repair. The orchestration question is whether the support has a boundary. Can the learner receive one explanation and then attempt? Can the tutor step away long enough to see the first independent move? Can another learner’s independent period provide the time needed for that observation?
Volume 0038, The Dose, owns how much help to give. Three-Learner Orchestration asks how the room is arranged so that the right dose can actually be delivered without the other learners losing their own routes.
13. A shared error can create a shared micro-lesson
Efficiency becomes educationally legitimate when the evidence genuinely converges.
Suppose all three learners confuse correlation with cause in different Science or data-reading contexts. The tutor notices the same reasoning failure across their independent attempts. A five-minute shared comparison may now be useful. Each learner brings a different case, the tutor names the common distinction, and all three return to their own tasks.
The micro-lesson emerged from common evidence rather than timetable convenience. That is the key difference.
14. One learner’s error should not automatically become everybody’s lesson
Alicia makes a sign error. The tutor stops the room and gives a five-minute reminder on sign control. Beatrice and Ciara listen politely even though neither has shown the problem.
This looks efficient because the explanation is delivered once. It may be inefficient because two learners spend time on a repair they did not need.
Before turning an individual error into a group explanation, ask whether the others have shown the same weakness, whether the concept is an imminent shared prerequisite, or whether the explanation is brief enough that the opportunity cost is genuinely small.
15. Worked case: three learners, one algebra lesson
This is a fictional composite case, not a report of real students.
Alicia is ready for mixed algebra method selection. Beatrice understands the methods but becomes slow because she writes every intermediate step. Ciara has just encountered simultaneous equations and still needs conceptual modelling.
The weak orchestration is to teach simultaneous equations to all three because Ciara needs it, then give everyone the same worksheet. Alicia is under-challenged, Beatrice practises a method that is not her bottleneck, and Ciara receives insufficient guided work once the tutor moves around the table.
A stronger route begins with a two-minute shared retrieval check on an algebraic manipulation all three need. Alicia then receives a mixed set with no topic labels and is told the tutor will return after three method decisions. Beatrice receives a short timed section using familiar methods so the tutor can inspect whether step compression preserves accuracy. Ciara stays with the tutor for an eight-minute model of simultaneous equations followed by one guided attempt.
The tutor then returns to Alicia’s method choices, not merely her answers. While Alicia explains one selection, Ciara begins an independent second attempt and Beatrice completes her timing section. Beatrice’s work later shows that her speed improved but one skipped line created an error; the tutor gives a narrow correction. Ciara’s second attempt reveals that she can execute elimination but cannot yet explain why the operation preserves equality. The next shared moment is not automatic. Each learner has produced a different next job.
The lesson is coherent because the tutor always knows what each learner is carrying and what evidence should return.
16. Worked case: one passage, three English jobs
This is a fictional composite case.
All three learners read the same short passage. That does not mean they need the same comprehension task.
Beatrice is working on evidence precision. Emily is working on answer planning. Faith is working on independent task initiation. The tutor asks everyone to read privately. Beatrice must mark the one line that most directly supports an inference. Emily must outline a two-sentence response before drafting. Faith must begin without a task-restatement prompt and mark where she becomes uncertain.
The shared text reduces material complexity for the tutor while preserving three different learner jobs. After the first attempts, the tutor uses a short group discussion to compare how each person used the same passage. The discussion is useful precisely because their private work is already preserved.
17. Worked case: one learner needs silence
This is a fictional composite case.
Denise is preparing for longer Additional Mathematics sections. Her current job is to sustain accurate performance without pacing prompts. The tutor’s instinct is to check frequently because the work matters.
Frequent check-ins would change the condition. Denise needs a twenty-minute protected block in which the tutor does not announce time remaining, suggest question order or react visibly to errors. The other two learners need active teaching during exactly that interval.
Small-group orchestration creates the possibility of this independence. The tutor can work intensely with two learners while Denise performs under a more authentic condition. At the end of the block, the tutor has both instruction delivered elsewhere and a clean performance sample from Denise.
18. The tutor needs a rotation of evidence, not a rotation of people
A mechanical rotation says: learner A, then B, then C, ten minutes each.
An evidence rotation says: which learner’s current task will soon produce a decision-relevant receipt? Who is approaching a point where feedback will matter? Who requires information before practice can continue? Who should remain uninterrupted?
The tutor’s attention moves in response to those events. This is more demanding than a timer, but it is also more faithful to learning.
19. Use predictable routines to reduce orchestration overhead
Responsive teaching does not require an unpredictable classroom.
A consistent opening, visible learning objective, agreed help-seeking rule, familiar independent-work routine and clear checkpoint system reduce the amount of attention spent on logistics. The tutor can then adapt the intellectual work without constantly redesigning the room.
The National Student Support Accelerator’s session-structure guidance argues that consistent structure can improve predictability and allow tutors to focus more on instruction. That principle is especially useful in a three-person tutorial: stable routines create the operating space in which individualisation can happen without chaos.
20. Equal participation is not always equal speaking
Some learning jobs are visible in speech. Others are visible in writing, diagrams, method selection, annotation, silence, checking or revision.
A tutor should not force every learner to speak equally in every segment merely to make participation look balanced. Nor should quiet learners disappear behind written work for the whole session. Match participation mode to the capability being observed.
If the job is oral explanation, speech matters. If the job is independent problem representation, private writing may matter more. If the job is self-monitoring, the tutor may need the learner to articulate why they decided to seek help. Orchestration is about making the target operation visible in an appropriate form.
21. Feedback timing should follow the learning job
In a small group, the tutor cannot give immediate feedback to everyone at once. That is not always a defect.
For a learner practising a new and fragile procedure, quick correction may prevent repeated error. For a learner completing a diagnostic attempt, interruption may destroy evidence about what they can do alone. For a learner revising writing, a short delay may allow self-detection before tutor feedback arrives.
Volume 0072, The Feedback Receipt, owns what happens after feedback. Orchestration owns when and to whom the feedback is delivered so that it serves the present learning operation.
22. The tutor should periodically audit invisible learners
Every tutor develops attentional habits. One learner asks more questions. One learner produces more dramatic errors. One learner is enjoyable to discuss ideas with. One learner rarely interrupts.
These social signals can distort instructional attention.
A simple weekly audit asks:
- Whose reasoning have I actually observed?
- Whose independent work have I reviewed rather than merely marked?
- Who receives the most explanations?
- Who receives the fewest questions?
- Who tends to answer for the group?
- Who is repeatedly interrupted during independent work?
- Whose support level has changed without the group routine changing with it?
The purpose is not to equalise a spreadsheet. It is to notice whether tutor habit rather than learner need is controlling the room.
23. Parent communication should explain differentiated attention
A parent can reasonably worry that a child in a small group receives less tutor time than another learner. The strongest answer is not “everyone gets equal attention.” That may be untrue and educationally undesirable.
A better explanation is concrete:
Your child is currently practising method selection without cues, so I deliberately leave longer uninterrupted blocks and then review the choices. Another student is learning new content and receives more direct explanation. If your child’s work shows a conceptual gap, the allocation changes immediately.
This makes differentiation visible without comparing children as better or worse.
24. Learners should understand the room well enough to self-operate
Strong orchestration does not require learners to wait passively for the tutor to direct every transition.
Each learner should increasingly know:
- what the current job is;
- what to attempt before asking for help;
- what counts as a checkpoint;
- what to record if stuck;
- what to do while the tutor is with another learner;
- how to resume after feedback;
- when a peer explanation is useful and when private thinking must come first.
The more learners can manage these transitions, the less tutor attention is consumed by traffic control and the more can be spent on actual teaching and observation.
25. A practical Three-Learner Orchestration card
- Learner A job: What operation is active?
- Learner B job: What operation is active?
- Learner C job: What operation is active?
- Shared need: Is any explanation or practice genuinely common?
- Private evidence: What should each learner attempt before peer answers are visible?
- Support level: What help is allowed for each learner?
- Independent block: Who needs protected time away from tutor cues?
- Re-entry trigger: What event brings the tutor back?
- Observation target: What should the tutor watch rather than explain?
- Feedback timing: Who needs immediate correction and who benefits from delayed review?
- Peer use: Where can comparison help without replacing independent thinking?
- Attention audit: Is tutor habit making anyone invisible?
26. Research foundation and limits
The research base supports the ingredients more strongly than this exact three-learner orchestration model. EEF’s small-group tuition evidence suggests positive average effects and emphasises targeting support to pupil needs. The National Student Support Accelerator’s Tutoring Quality Standards include purposeful grouping, structured instructional materials, research-based instructional practices, formative assessment and tutor consistency. Its session-structure resources include independent practice and formative assessment. AERO’s Monitor Progress and Vary Practice examples show varied participation and checking for understanding in classroom settings.
Those sources do not prove that three learners are universally optimal, that the specific routines above outperform other small-group designs, or that equal-time allocation is always inferior. Context, tutor expertise, learner needs, subject, age and task matter. The framework here is a disciplined way to turn the central promise of small-group tuition—visibility plus targeted support—into moment-by-moment tutor decisions.
27. What this volume does not own
This volume does not redefine the general evidence for small-group tuition, which remains owned by the broader 3-Pax Small-Group Tuition route and external evidence sources. It does not turn peers into tutors, prescribe a universal rotation schedule, or claim that every learner deserves equal minutes of direct adult interaction. It does not remove legitimate access support and it does not diagnose learners from participation style.
It owns one professional job: making sure a shared room still produces individual thinking, individual evidence and individual next steps.
Final principle
A three-learner tutorial is not fair because the tutor’s minutes can be divided evenly.
It is fair when every learner receives the learning conditions their current job requires, when differences in attention can be explained by evidence, and when independent work remains real learning rather than waiting for the adult to return.
Good orchestration is not equal tutor presence. It is the deliberate movement of teaching, silence, questions, attempts, observation and feedback so that all three learners keep thinking—even when the tutor is speaking to only one.
