ROUTE 03 · FOR TUTORS

Tutors: the Handbook is the technical layer

The Tutor System explains the public purpose. The Tutor Handbook handles the professional decisions.

That distinction matters. A parent should be able to understand the learning route without needing to operate every diagnostic gate. A student should be able to understand what is expected without carrying the tutor’s full decision architecture. The tutor, however, needs enough technical depth to distinguish different causes that produce the same visible result and enough discipline to avoid overclaiming from weak evidence.

Three connected tutor layers

The first lesson is an information problem before it is a coverage problem

A first lesson can easily become a performance by the tutor. There is a syllabus to cover, a parent wants to know whether the tutor is good, the student may be nervous and everyone wants visible progress. That pressure can make the tutor teach too much too soon.

A more useful first lesson reduces uncertainty. What can the learner begin without help? Which prerequisites are available? What happens when the surface changes? Where does the learner ask for help? What kind of prompt is enough? Which school demand is actually causing the current concern? Which hypotheses survive contact with the work?

The Tutor Handbook follows that opening sequence through The First Lesson, The First Month and The First Review. The sequence matters because early tutoring should turn impressions into a revisable route rather than a permanent label.

Progress: marks matter, but one number cannot explain itself

School marks are part of the real environment the learner has to navigate. They should not be dismissed. But the same mark can be produced by different mechanisms, and the same learning improvement can take time to appear in a headline score.

A tutor therefore needs two disciplines at once. First, take outcomes seriously: examination performance, school work and the learner’s ability to meet actual curriculum demands matter. Second, avoid reading more into one result than it can support. A score may be affected by topic coverage, paper difficulty, timing, question selection, task cues, help during preparation, marking differences, fatigue or a single high-weight component.

The stronger review asks what changed in the learner and whether that change survives. Can the learner retrieve after a delay? Choose the method without a chapter heading? Explain why it works? Transfer it into a changed question? Recover after an error? Perform with less prompting? Convert the capability into marks under representative conditions?

For this reason, How We Know Learning Has Really Held sits between teaching and the decision to declare a repair finished. The Learning Runtime then supplies a stricter evidence discipline when the score and the apparent learning do not agree.

Subject knowledge still owns the subject

A tutoring system can organise diagnosis, evidence, support and review. It cannot replace the intellectual structure of English, Mathematics, Science or another discipline.

If a learner cannot distinguish evidence from inference in English comprehension, the repair eventually belongs inside English. If the learner does not understand conservation of energy, the scientific mechanism belongs inside Science. If an algebraic transformation is invalid, Mathematics owns the mathematical reason. The Tutor System helps decide where to look, what evidence to collect and how to manage support. Then the subject must teach the subject properly.

When the three people disagree

Parents, students and tutors do not always read the same situation in the same way. A parent may see falling marks. The learner may say the work is becoming clearer. The tutor may see a repaired foundation but unstable timed performance. A school comment may point somewhere else again.

The solution is not to average those accounts into one vague judgement. Each source may be describing a different part of the system. The useful move is to make the disagreement testable.

• Parent: describe what has changed at home or across school results.
• Student: describe what now feels easier, harder or still uncertain.
• Tutor: identify the capability being taught and the evidence used to judge it.
• Work sample: test whether the capability appears under a suitable independent condition.
• Next decision: change one meaningful variable, collect a fresh sample and update the route.

The Tutor Handbook’s Parent Review, Student Review, Learning Contract, Evidence Conference and Priority Arbitration Gate exist because tutoring is a relationship as well as an instructional process. Good coordination makes disagreement informative instead of political.

Boundaries: tutoring should know what it does not own

A tutor can observe learning, teach a subject, design practice, give feedback, organise a route and communicate educational evidence. Those responsibilities do not turn ordinary tutoring into medical, psychological, legal or school-authority practice.

A tutor should not infer a clinical condition from ordinary lesson behaviour, make promises about examination outcomes, rewrite assessed work so that the submission becomes the tutor’s product, or use one ambiguous performance as a permanent label. When the remaining issue belongs to another professional, institution or authority, the responsible tutoring move is a handoff.

The Tutor Handbook’s Scope Boundary owns that decision in technical detail.

Frequently asked questions

How a family enters the Tutor System

Most families do not begin with a clean educational diagnosis. They begin with an event.

A test comes home lower than expected. Homework that once took thirty minutes now takes two hours. A child says, “I understand in class but I cannot do the questions.” A teacher writes that answers need more evidence. Mathematics feels fine until problem sums become multi-step. Science facts are remembered but application questions collapse. A previously confident learner reaches Secondary school and suddenly seems slow. Another child is scoring well but every week depends on a parent sitting beside them.

Those events matter, but they are not yet the learning problem. They are the entrance.

Step 1 · Preserve the event before explaining it

Keep the paper. Keep the question. Keep the teacher comment. If the difficulty happens during homework, note what the task was and roughly where the learner stopped. If the problem is repeated, collect two or three examples rather than relying on the memory of an argument at the kitchen table.

This simple habit protects the next conversation from becoming a contest between labels. “Careless”, “weak”, “lazy”, “not confident”, “bad at comprehension” and “does not understand Math” may contain part of the truth, but each is too large to design a lesson from. The work gives everyone something smaller and more honest to inspect.

Step 2 · Ask what the learner can already carry

Before listing weaknesses, locate capability. Can the learner explain the concept verbally? Can they perform the first step? Can they solve a familiar version? Can they retrieve the relevant fact? Can they identify what the question is asking? Can they notice that an answer is implausible? Can they correct the error after it is pointed out?

This matters because tutoring should begin at the first useful weak link, not at the beginning of the entire subject. A student who can set up the algebra but repeatedly loses the sign does not necessarily need the whole chapter retaught. A student who can state a scientific concept but cannot use it in an unfamiliar situation may need transfer work rather than more notes. A student who understands an English passage but writes vague answers may need answer precision rather than another reading-comprehension worksheet.

Step 3 · Separate the symptom from the job

The same visible symptom can imply several tutoring jobs. A low score can come from missing knowledge, weak retrieval, misreading, poor method selection, slow execution, incomplete checking, examination pacing or a mismatch between what was practised and what was tested. A blank page can come from not knowing the content, not knowing how to start, overload, fear of being wrong, unclear task instructions or a habit of waiting for someone else to supply the first move.

The Tutor System therefore moves from symptom to job. What is the smallest educational job that, if solved, would materially improve the learner’s next attempt?

Step 4 · Decide whether tutoring is the right owner

Some needs belong naturally in tutoring: subject explanation, prerequisite repair, practice, feedback, examination preparation, route design and support with becoming more independent. Other needs may require coordination with school, a change in schedule, a different resource or a qualified professional outside ordinary tuition. The presence of a learning difficulty does not automatically mean the answer is “more tuition”.

A useful system is allowed to say that the tutor is not the right owner. That is not failure. It is scope control.

Step 5 · Match the support to the present condition

When tutoring is appropriate, the next question is not simply “Which tutor has the strongest profile?” It is “What must the tutor actually be able to do for this learner now?” A clear explainer can be exactly right for a contained conceptual gap. A diagnostic tutor becomes more important when several plausible causes compete. A route designer is useful when the learner has multiple interacting gaps or a transition ahead. A performance coach becomes useful when knowledge exists but examination output is unstable.

The tutorial structure matters too. Some learners benefit from close one-to-one diagnosis. Some benefit from a well-designed group in which they must attempt, explain, compare and wait without receiving continuous rescue. Some are sufficiently independent that a larger class can provide useful coverage and exposure. Format should serve the learning job.

Step 6 · Define the first review before the first month disappears

Before tuition becomes a routine simply because it has started, decide what should be clearer after several lessons. Perhaps the tutor should know whether the main issue is retrieval or application. Perhaps the learner should be beginning questions with less prompting. Perhaps one recurring error should be less frequent. Perhaps the student should be able to complete a fresh problem after the worked example has been removed.

The first review does not require a dramatic grade jump. It requires better information and a justified next move. Continue if the route is working. Change it if the evidence points elsewhere. Reduce support if the learner is already carrying more. Stop a particular intervention if its job is complete.

Is tutoring the right tool?

Tutoring is powerful precisely because it can respond to an individual learner or a small group. That responsiveness can also tempt families to use tutoring as the answer to every educational difficulty. A stronger system asks what kind of problem is actually present.

Tutoring is a strong fit when the job is teachable and observable

• A prerequisite concept is missing and later work depends on it.
• The learner understands explanations but cannot yet retrieve or apply the idea independently.
• The student needs carefully designed practice rather than more random volume.
• A recurring error needs diagnosis before more work is added.
• The learner knows the content but cannot convert it into timed examination performance.
• A school transition has changed the level of abstraction, pace or independent organisation required.
• The learner needs feedback on real attempts and opportunities to act on that feedback.
• The learner needs support with planning, checking, recovery and more precise help-seeking inside a subject.

Tutoring may not be the first tool when the main problem is elsewhere

If a learner is exhausted because the week is overloaded, adding another class may worsen the environment in which learning has to happen. If the issue is a current school administrative decision, the school is the owner of that decision. If a concern involves health, mental health, developmental assessment, safeguarding or another professional domain, an ordinary tutor should not convert educational observation into a clinical conclusion. If the student already understands the work but has no time to sleep, another worksheet is not the repair.

Sometimes the correct tutoring decision is coordination: keep the academic work focused while another part of the learner’s environment is handled by the appropriate person. Sometimes it is dosage reduction. Sometimes it is a short targeted intervention rather than open-ended tuition.

Ask what tutoring is expected to change

“Improve English” is too large. “Increase marks” is an outcome, not yet a tutoring mechanism. A useful tuition goal names a capability that can be taught and checked.

• Distinguish stated evidence from inference in comprehension.
• Retrieve multiplication facts quickly enough that problem solving is not overloaded.
• Select the right algebraic method in a mixed set without a topic label.
• Explain a Science relationship using cause, mechanism and outcome.
• Plan an argumentative paragraph before drafting.
• Complete a timed section while preserving enough time to check high-risk errors.
• Begin homework without waiting for a parent to identify the first step.

When the goal becomes this specific, tutoring can be evaluated. The tutor can choose a task that exposes the capability, teach the missing structure, create practice, reduce help and test whether the learner can now carry the job.

Choosing a tutor: move from biography to function

Qualifications, subject knowledge, experience and safeguarding matter. They are necessary parts of selection. But a biography cannot by itself tell you how a tutor will respond when a learner gets a familiar question wrong for an unfamiliar reason.

A stronger selection process moves through four layers: credibility → subject fit → tutoring function → working fit.

1 · Credibility

Confirm what is relevant to the role: identity, appropriate qualifications or subject competence, experience with the stage and examination where required, clear terms, professional communication and transparent arrangements. Where a programme makes specific claims, ask what those claims mean and what evidence supports them.

Credentials matter because tutoring is educational work. They should not be converted into a guarantee. A distinguished academic profile does not automatically produce strong diagnosis, feedback, group orchestration or the ability to help a novice think.

2 · Subject fit

A tutor must understand the subject at the level required. For examination courses, that includes the relevant assessment demands and the difference between subject knowledge and examination technique. An English tutor may need to diagnose reading, vocabulary, syntax, answer scope and writing organisation. An A-Math tutor may need to see algebraic structure, prerequisite gaps and the sequence in which later topics depend on earlier ones. “Good at teaching” cannot substitute for subject knowledge.

3 · Tutoring function

Ask what the tutor does when the answer is wrong.

• Do they immediately explain?
• Do they ask the learner to show the last point that was secure?
• Do they distinguish a knowledge gap from a method-selection problem?
• Can they design a small diagnostic task rather than adding a full worksheet?
• Do they know how to move from teaching to independent verification?
• Can they explain how feedback will change the next attempt?
• Do they know when not to help?

These questions reveal function. A Homework Helper, Explainer, Drill Builder, Diagnostic Tutor, Route Designer, Performance Coach and Learning Architect can all be useful. The error is treating them as the same job.

4 · Working fit

Fit is not a personality contest. Warmth can matter. Trust can matter. So can clarity, boundaries, intellectual challenge and the learner’s willingness to make thinking visible. The relevant question is whether the relationship supports the educational job.

A student may like a tutor who rescues every difficult moment, yet become more dependent. Another student may initially find a tutor demanding because the tutor waits for an attempt, but later become more capable. Conversely, a tutor can mistake pressure for rigour and make the learner less willing to expose uncertainty. Fit must be read through the work, not only through whether everyone leaves smiling.

Questions worth asking before committing

• How do you use the first lesson?
• What information would you like to see from school or previous work?
• How do you decide what to teach first?
• How do you tell whether a learner understands rather than recognises?
• How much independent attempt time is built into a lesson?
• What happens when the same mistake keeps returning?
• How is homework chosen, and what happens if the learner cannot do it independently?
• How do you review progress with the student and parent?
• When would you change the teaching route?
• When would you recommend less tuition, a different tutor or another kind of support?

Signals to examine carefully

A single signal is not a verdict. But certain patterns deserve questions: promises of guaranteed results; the same material sequence for every learner despite different needs; very heavy homework that cannot be completed independently; constant tutor talk with little learner production; every error being explained as “carelessness”; all progress being reported through volume completed rather than capability demonstrated; or a tutor who cannot say what would make them change their mind about the current diagnosis.

The goal is not to find a flawless person. It is to find a professional relationship in which evidence can change the plan.

Continue with the Tutor Classification Model when you need the full function map, then return to Tutor–Student Fit once teaching has begun.

When tuition stalls: diagnose the tutoring system before adding more

When progress feels slow, the instinct is often to increase something: more lessons, more homework, more past papers, more reminders, more detailed notes, more pressure. Sometimes additional time is exactly what a learner needs. Sometimes it amplifies the wrong intervention.

A stalled result should trigger a differential diagnosis of the tutoring system itself.

Possibility 1 · The original diagnosis was wrong

The tutor may be treating a surface symptom. A learner who loses marks in fractions may actually have unstable multiplication facts. A student who writes short comprehension answers may understand the text but misread the question’s required scope. A learner who seems weak in Chemistry calculations may be blocked by algebraic manipulation. If the first weak link was misidentified, conscientious practice can make the wrong skill stronger while the original failure remains.

Possibility 2 · The diagnosis was right but the intervention was not run faithfully

A good plan can fail because it was never really implemented. Homework was repeatedly displaced by urgent school work. The learner missed sessions. The tutor planned mixed practice but kept returning to familiar blocked sets. The parent supplied too much help during the independent work. The student used an answer key before attempting. Before discarding a route, check whether the route actually occurred.

Possibility 3 · Supported performance has been mistaken for independent learning

The learner appears fluent beside the tutor but loses the skill elsewhere. This can happen when the tutor chooses the method, reminds the learner of the first step, supplies key vocabulary, notices the error immediately or gives reassurance at every uncertain moment. None of those supports is automatically wrong. The mistake is using performance under those supports as evidence that the learner no longer needs them.

Possibility 4 · Practice is too narrow

A learner can become highly accurate within one question family and still fail when cues are removed. The textbook chapter names the topic. The worksheet groups identical structures. The tutor has just modelled the method. The same vocabulary repeats. In that environment, recognition can substitute for selection. Mixed practice and changed surfaces reveal whether the learner can identify the underlying structure independently.

Possibility 5 · The capability exists but does not survive the examination environment

Timed performance introduces new demands: pacing, question selection, switching between topics, managing partial uncertainty, checking under load and knowing when to move on. A student can genuinely understand the content and still need a separate performance layer. That is the job of the Performance Coach rather than a reason to reteach every topic.

Possibility 6 · The workload is consuming the learning

More material can create less useful attention. The student may be spending every free hour completing tuition work and school work, leaving little time for sleep, delayed retrieval, reflection or life outside study. The tutor should be able to prioritise. A learning budget asks which task deserves the next hour, not how to fill every available hour.

Possibility 7 · The tutor is part of the problem

Professional self-checking matters. The explanation may have been unclear. The example may have contained a hidden shortcut. The tutor may be asking a question whose wording creates unnecessary difficulty. Feedback may be too broad. A small group may be organised so that the fastest learner supplies the thinking for everyone else. A tutor who tests their own contribution before assigning every failure to the learner protects the quality of the diagnosis.

Possibility 8 · The tutoring structure no longer fits

A learner who once needed intensive Micro support may now benefit from Meso practice and peer calibration. Another learner may be disappearing inside a group and need a period of closer diagnosis. A strong student may be ready for broader exposure rather than continuous remediation. Fit can change because the learner changed.

Continue, change, reduce or stop

Tutoring often begins as a yes-or-no decision: have tuition or do not have tuition. Once support is underway, the more useful decisions are continuous.

Continue

Continue when the tutoring job still exists, the current method is producing useful evidence and the learner still benefits from the present level of support. Continuing should not mean repeating the same lesson shape indefinitely. The work can become more complex while the underlying route remains justified.

Change

Change when evidence says the route is targeting the wrong weak link, the tutor function is mismatched, the tutorial size is obscuring the learner, materials do not fit the curriculum demand or the relationship is preventing useful work from happening. Change can be small: one practice type, one feedback routine, one session frequency. It does not always require replacing the tutor.

Reduce

Reduce when the learner can carry more. That may mean fewer prompts inside the same lesson, less homework monitoring, a shorter review period, fewer sessions or movement from intensive repair into maintenance. Reduction is a success state when it reflects stronger independent capability.

Stop

Stop when the present tutoring job has been completed, when the remaining issue lies outside the tutor’s scope or when the cost of continuing exceeds the educational value. A learner can stop one form of tuition and still seek teaching later for a new stage or subject. Stop is not an ideological position. It is a legitimate system decision.

Return conditions should be clear

If support is reduced or stopped, agree what would justify returning. A new subject stage, repeated failure in a previously stable capability, a major examination transition or evidence that independent routines are no longer holding may justify another intervention. The return should respond to evidence rather than anxiety alone.

One-to-one, three students, small group or larger class?

Class size changes what a tutor can see and what a learner is required to do. It is not a quality score by itself.

One-to-one: maximum attention can be useful—and can become too much attention

One-to-one tuition can support close diagnosis, rapid adaptation and intensive repair. It can be valuable when a learner has hidden gaps, significant uncertainty or a need for concentrated feedback. It also creates a particular risk: the tutor can become available for every pause. If each hesitation immediately produces a cue, the learner may lose opportunities to begin, wait, check and recover independently.

The solution is not to withhold teaching. It is to design moments in which the learner owns the next move.

Three-student tutorials: visibility plus controlled independence

A three-student tutorial can give the tutor enough proximity to inspect individual work while creating natural intervals in which each learner must operate without constant tutor attention. One student can receive an explanation while another begins independently. One can explain a method while the others compare. The tutor can notice who waits, who checks, who dominates, who hides and who can transfer a method after hearing a peer.

This only works when the tutor orchestrates the group. Three students completing the same worksheet silently is not automatically superior to another format. The educational value comes from task design, grouping, individual accountability, questioning, observation and feedback.

Larger groups: useful when the learning job changes

Larger classes can efficiently provide structured coverage, exposure to many questions, polished explanations and a broader competitive or collaborative environment. They ask more independence from the learner because the tutor cannot continuously inspect every decision. For a strong or stable learner that may be entirely appropriate. For a learner with a hidden prerequisite gap, it may allow the problem to remain invisible.

Online and in-person are delivery conditions, not tutor classes

Online tutoring can offer strong access, shared documents, recorded resources and flexible scheduling. In-person tutoring can make paper-based work, physical diagrams, live monitoring and some forms of interaction easier. The relevant question is whether the delivery format allows the tutor to see the target learning process and whether the learner can participate effectively. A poor fit online does not prove online tutoring is poor; a distracted in-person lesson is not made strong by sharing a room.

EEF’s current reviews of one-to-one and small-group tuition report positive average effects for both and substantial variation between studies. That variation supports the Tutor System’s central design rule: teaching quality, targeting and fit matter more than treating class size as a universal ranking.

The three working agreements

Tuition becomes easier to interpret when the main responsibilities are explicit.

Parent ↔ Tutor

• The parent supplies relevant context without requiring the tutor to accept every interpretation.
• The tutor explains the active learning priority in language the parent can understand.
• Both distinguish current evidence from promises about future grades.
• Practical terms, attendance, homework expectations and communication boundaries are clear.
• Reviews focus on the learner’s work and the next decision.

Student ↔ Tutor

• The learner is allowed to be wrong without being allowed to disappear.
• The tutor can require a real attempt when an attempt is educationally appropriate.
• Help is available, but its size should fit the current purpose.
• The student can say when an explanation has not made sense.
• Both treat mistakes as evidence for the next move rather than as identity.

Parent ↔ Student

• The parent protects time, logistics and reasonable routines.
• The learner increasingly owns starting, packing materials, completing appropriate practice and reporting uncertainty.
• Parents help without quietly becoming the author of homework or the permanent homework supervisor.
• A result can be discussed seriously without becoming a statement about the child’s worth.

The Tutor Handbook’s Learning Contract formalises the same idea from the tutor side: support needs boundaries so responsibility can move rather than blur.

The tutor skill ladder in practical terms

The Tutor Classification Model becomes useful when each class is understood as a job, not a status badge. A tutor may operate several classes during one learner’s journey. The higher classes coordinate more uncertainty and more interacting variables; they are not automatically preferable for every task.

Class 0 · Homework Helper

The Homework Helper helps a learner organise, interpret and complete assigned work without silently becoming the person who does it. This can be valuable when the main problem is routine, task entry, missing instructions or the need for a reliable study structure.

The boundary is authorship. If the helper chooses every answer, rewrites the paragraph, solves each difficult problem or supplies the crucial reasoning, the homework may be finished while the educational evidence disappears. Class 0 work is strongest when it helps the learner understand the task, preserve school ownership and develop routines that require less external management over time.

Class 1 · Explainer

The Explainer makes a concept, instruction or method clearer. This is the tutor most families picture first: someone who can take a confusing idea and represent it differently until the learner can see the structure.

Explanation is powerful but easy to overvalue because the learner often feels improvement immediately. “That makes sense” is a useful moment, not the final receipt. The Class 1 job is complete only when explanation is followed by reconstruction and use. The tutor should eventually close the explanation and see what the learner can produce.

Class 2 · Drill Builder

The Drill Builder strengthens a capability that is basically correct but not yet reliable enough. Fluency matters because fragile basic operations consume attention that could be used for harder reasoning. The Drill Builder chooses the right amount, sequence and variation of practice.

Good drill is not mindless volume. It knows what is being stabilised. Early repetition can isolate a new operation; later practice should vary cues, mix alternatives and eventually require method selection. The Drill Builder also knows when continued repetition has stopped providing useful information.

Class 3 · Diagnostic Tutor

The Diagnostic Tutor becomes important when the visible mistake has several plausible causes. Instead of teaching the first explanation that comes to mind, the tutor designs small probes that distinguish between alternatives.

Consider a learner who cannot solve a percentage problem. Is the ratio structure missing? Is division unstable? Is the learner misreading “percentage increase”? Can the calculation be performed once the relationship is represented? Each answer points to a different intervention. Diagnosis reduces the amount of unnecessary teaching.

Class 3 should remain educational, not clinical. It diagnoses learning operations and subject prerequisites within the tutor’s scope. It does not turn lesson observations into medical or psychological labels.

Class 4 · Route Designer

The Route Designer decides what should happen first, next and later when more than one valid improvement is possible. A learner may need vocabulary repair, answer precision, examination timing and better study routines. Doing all four aggressively at once can make the learning system harder to operate.

Route design identifies dependencies. Which weakness blocks the others? Which capability can improve through normal school work once one prerequisite is repaired? Which issue is urgent because of an examination date? Which change should wait until the current route produces enough evidence? A good route makes the learner’s next action clearer.

Class 5 · Performance Coach

The Performance Coach works where capability exists but output under real conditions is unreliable. The coach handles timing, question selection, transition between tasks, recovery after a difficult item, checking strategy, answer scope and the difference between practice comfort and examination readiness.

Class 5 work should not hide missing knowledge under test-taking tactics. If the learner genuinely does not understand the content, return to teaching. Performance coaching is most useful when the knowledge base is sufficiently stable for the environment itself to become the main difficulty.

Class 6 · Learning Architect

The Learning Architect coordinates several interacting parts of the learner’s system: subject gaps, practice, school demands, study routines, tutoring dosage, tutor roles, performance, transitions, family constraints and the path towards greater independence.

Class 6 does not mean doing more. It often means removing unnecessary complexity. A Learning Architect should be able to operate a lower tutor class when that is all the learner needs. The highest-class error would be turning every learner into a grand project and making the tutor indispensable.

The architect’s strongest result is not a learner surrounded by an impressive support system. It is a learner who gradually internalises more of the planning, monitoring, checking and adaptation that the system once had to carry externally.

Micro, Meso and Macro: tutorial size as a teaching instrument

The eduKateSG control model uses Micro, Meso and Macro to describe different tutorial structures. These are not rigid universal categories. They are a way to ask what level of attention, interaction and independence the learner’s condition requires.

Micro

Micro tutoring is close enough to inspect the learner’s decisions at high resolution. One-to-one is the obvious example, though very small arrangements can also function in a Micro way for particular moments. Micro is useful when the tutor needs to see exactly where the chain breaks, adjust support quickly or rebuild a fragile prerequisite.

The risk is over-service. When a tutor can respond to every pause, the learner may receive less practice beginning, checking and recovering alone. Micro tutoring therefore needs deliberate independence windows.

Meso

Meso tutoring uses a small enough group for meaningful tutor knowledge of individual learners while preserving peer comparison, independent attempt time and varied interaction. eduKate’s three-student tutorials sit naturally in this zone.

The tutor can compare how three learners interpret the same question, ask one student to explain while the others test the explanation, branch temporarily for different weak links and observe what each learner does when tutor attention moves elsewhere. Meso works badly if grouping is poor or if one fast learner becomes the unofficial standard for everyone.

Macro

Macro tutoring can deliver broad coverage, structured revision, large question banks, strong lecturer explanations and efficient exposure to an examination landscape. It asks more self-regulation from the learner because individual state is less visible to the tutor.

Macro can be powerful for learners who already possess enough foundation and independence to process the material. It can fail quietly for learners who need close diagnosis because they may attend, copy and complete without the underlying weak link ever becoming visible.

Tutor skill class and tutorial size are separate decisions

A highly skilled tutor can teach a larger group. A one-to-one tutor can still operate only as a Homework Helper. A three-student class can contain Class 6 thinking if the tutor is genuinely coordinating diagnosis, route design, performance and independence. The structure creates opportunities; the tutor’s capability determines how those opportunities are used.

This is why the control formula combines student condition, tutorial size and tutor skill class instead of assuming any single variable decides quality.

Tutor professional practice: what happens behind the visible lesson

Strong tutoring includes work the learner may never see directly. The tutor chooses or adapts materials, anticipates likely misconceptions, reviews previous evidence, decides which question will reveal the next uncertainty, keeps some items fresh for later verification, records enough continuity to avoid resetting the learner and checks whether their own teaching move created the difficulty.

Professional practice also means knowing what not to record. A tutor needs enough educational information to continue the route, not an unnecessary archive of personal data. Notes should serve continuity, review and responsible communication.

Preparation is not prediction

A prepared tutor arrives with a coherent plan and materials. A responsive tutor is willing to change the plan when the learner’s actual work contradicts the assumptions behind it. The skill is to prepare enough structure without pre-deciding what the learner will need.

Tutor errors should be repairable

Tutors can make mistakes: a wrong explanation, an incorrect mark, an ambiguous question, a poor example or a misread learner state. Professional strength is not pretending those events never happen. It is identifying the error, correcting the record, explaining what changes and protecting trust without shifting the cost of the tutor’s mistake onto the learner.

The tutor’s knowledge has a boundary

A tutor should be able to say “I need to verify that” when a question falls outside reliable knowledge. Bluffing is especially dangerous in technical subjects because a fluent wrong explanation can be remembered very effectively. Search and AI can help verify, but the tutor remains responsible for checking the answer before teaching it.

Four worked journeys through the Tutor System

The Tutor System becomes easier to understand when we follow a learner from the visible problem to the next educational decision. The examples below are illustrative. They are not diagnoses of any particular child and they are deliberately simplified so that the routing logic is visible.

Journey 1 · Primary Mathematics: “careless” answers that are not really carelessness

A Primary learner is losing marks in multi-step problem sums. At home, the repeated description is “careless”. The child can often follow the tutor’s worked solution and can calculate accurately when the operation has already been chosen. Yet unfamiliar questions still break down.

The Tutor System does not begin by increasing the worksheet volume. It begins by preserving several failed questions and asking where the chain first becomes unstable. Can the learner identify what the quantities represent? Can the relationship be expressed before calculation? Does the learner confuse additive and multiplicative structures? Does the learner choose a plausible model when the wording changes? Can the arithmetic be completed once the relationship is set up?

Suppose the evidence shows that arithmetic is secure but representation is weak. The tutoring job is no longer “stop being careless”. It is to make quantitative relationships visible and to teach the learner a repeatable way to extract them from unfamiliar wording.

The tutor may first operate as an Explainer, using concrete or visual representations to show how the problem is structured. The tutor then becomes a Drill Builder, but the drill is not fifty near-identical questions. It is a carefully varied sequence in which wording changes while the underlying relationship remains stable. Later, mixed questions require the learner to decide which relationship is present without being told the topic.

The progress receipt is not simply a higher score on a worksheet completed immediately after teaching. It is a fresh question, a delay, a changed surface and eventually school evidence showing that the learner can represent the relationship without the tutor choosing the route.

If the learner can now model the problem but still loses marks through arithmetic slips under speed, the tutor changes job. The system does not insist that the first explanation must remain the explanation forever.

Journey 2 · Secondary English: the student understands the passage but still writes vague answers

A Secondary learner reads well and can discuss a passage intelligently. In tuition conversation, the student appears to understand the text. Yet comprehension answers are short, general and repeatedly lose marks. A generic response might be “do more comprehension”. The Tutor System asks what the answer is failing to do.

The tutor compares the question, the student’s oral explanation, the written answer and the marking demand. Perhaps the student can identify the relevant evidence orally but writes only the broad idea. Perhaps the student quotes without making the inferential link. Perhaps the answer contains the right thought but exceeds the question’s scope. Those are different language jobs.

Once the issue becomes language-specific, the route moves to the English owner rather than forcing the generic Tutor System to teach comprehension itself. The Singapore English Tuition Centre can handle question type, evidence, inference, answer scope, vocabulary precision and written formulation. The general Tutor Handbook remains useful for deciding how to diagnose, prompt, practise and verify the change.

The tutor might ask the learner to answer orally first, then identify which part of the oral reasoning failed to enter the written answer. A short sequence can contrast an answer that states evidence, one that states inference and one that connects both. The learner then writes a new answer without the model in view.

The parent report should not say merely “comprehension improving”. A stronger statement is: “The student now selects relevant evidence reliably in short inference questions. The remaining weakness is writing the inferential link precisely when two parts of the passage must be combined.” That claim is small enough to act on.

The next route is clear: preserve the reading strength, target answer construction, then verify that the skill transfers into a new passage and eventually a timed paper.

Journey 3 · Secondary A-Math: “I can do it in tuition but not in the test”

A Secondary learner can complete A-Math questions during tuition. The tutor explains clearly, the worked examples make sense and homework scores are high. School tests remain disappointing.

This is the classic point at which a tutoring system can make the wrong move. The tutor may reteach the topic from the beginning because the school result is low. The learner may practise even more questions of the same type. Both can increase familiarity without revealing why performance changes outside tuition.

The Tutor System asks for the support fingerprint. During tuition, is the topic known in advance? Is the relevant formula already on the board? Does the tutor say “this is a differentiation question” before the learner begins? Are questions blocked by chapter? Does the learner ask for confirmation after every line? Does homework use examples that closely resemble the lesson?

If so, the learner may possess execution skill without stable method selection. The next test should not be another explained example. Use a mixed set. Remove chapter labels. Include plausible alternatives. Ask the learner to name the relationship or method before computing. Record where the decision fails.

If selection becomes reliable untimed but breaks under time, the job moves again—towards Performance Coach work. Now pacing, task switching, recovery after a difficult question and checking strategy become visible. Full papers can be useful once the component skills are stable enough that the paper is testing coordination rather than merely exposing dozens of unresolved gaps at once.

The tutor’s success here may involve teaching less content and engineering better evidence. The learner’s success is the ability to recognise and execute the mathematics when the environment no longer tells them what to do.

Journey 4 · A high-performing student who does not need “repair”

Not every learner enters tuition because marks are falling. A strong learner may want deeper Mathematics, more sophisticated writing, broader General Paper knowledge, competition preparation or a faster route through material already mastered.

The Tutor System should not manufacture a weakness merely because tutoring is present. The first task is to identify the learner’s current ceiling under suitably demanding conditions. If ordinary school work is already secure, remediation is the wrong owner.

A Route Designer can build extension around genuine next demands: unfamiliar problem structures, proof, richer argument, broader reading, synthesis across sources, deeper scientific modelling or advanced examination control. The tutor can still use diagnosis, but the diagnostic question becomes “where does current capability stop being sufficient?” rather than “what is wrong?”

This distinction matters for motivation. A strong learner should not spend an entire year completing more of what they can already do simply because the tuition programme has a fixed volume. Extension should create new intellectual work.

The independence principle still applies. An advanced tutor should not become the learner’s source of every difficult idea. The tutor can open doors, model expert judgement, pose harder problems and provide calibration while leaving substantial room for the learner to read, investigate, attempt and form positions independently.

Journey 5 · Changing tutors without erasing the learner’s history

A tutor leaves, schedules change or a family decides that another tutor function is now required. The easiest transition is to start again: new assessment, new notes, new sequence, new vocabulary. For the student, that can mean paying the cost of adult transition.

A continuity handoff is better. The outgoing route should identify the current learning priorities, what has already been tried, which gains appear stable, what support is still needed, what evidence remains ambiguous and which changes were made recently enough that their effects are not yet clear.

The new tutor does not have to inherit the previous tutor’s conclusions uncritically. They perform a Receiving Check: does the handoff match the work now in front of the learner? If yes, preserve continuity. If no, record the first justified change and why it was made.

This is the difference between continuity and conformity. The learner keeps useful history without being trapped inside an obsolete plan.

A quality standard for a connected tutoring ecosystem

A hub that connects parents, students and tutors needs more than a collection of links. The links need an organising principle. The standard below is the practical test for whether the Tutor System is doing useful work across the eduKate estate.

1 · The learner’s problem is described before it is labelled

Pages, tutors and parents should begin with observable work where possible. “Weak in English” can route to reading, vocabulary, syntax, writing organisation, inference, answer scope, oral communication, examination control or several interacting issues. The system becomes useful when a broad search phrase leads towards a more precise educational question.

2 · Discovery does not pretend to be diagnosis

A search page can help a family find a tutor for Primary 6 Mathematics, Secondary English, A-Math or GP. It should not imply that arriving on the page proves that tutoring is required or that a particular mechanism explains the learner’s difficulty. Discovery opens the door. Diagnosis requires learner evidence.

3 · The Tutor Handbook remains technical

The public wrapper should make the professional system understandable without diluting it. The Handbook can therefore remain rigorous: tutor classes, evidence gates, change control, support provenance, group orchestration, continuity, scope and intervention logic. Parents do not need every internal control in order to understand why a good tutor collects evidence before changing the route.

4 · Subject owners keep subject authority

The generic tutoring layer should not produce shallow copies of strong specialist material. English belongs with English owners; Mathematics with Mathematics; Science with Science. The Tutor System carries the cross-subject mechanics—diagnosis, support, feedback, verification and independence—then hands the learner to the domain that owns the content.

5 · Evidence is proportional to the claim

One correct answer can show that one answer was correct. It cannot by itself prove mastery, retention, transfer and examination readiness. One poor test can show that the learner performed poorly on that test. It does not automatically prove that all prior learning failed. The system should make claims at the size the evidence can support.

6 · Support is visible enough to interpret performance

As tutoring becomes increasingly digital, this standard becomes more important. A polished answer may contain learner knowledge, tutor prompts, parent correction, AI-generated phrasing and answer-key comparison. The work can still be educationally useful. It cannot be interpreted responsibly unless the relevant support condition is known.

7 · The system can change its mind

A diagnosis is a working hypothesis, not a permanent identity. A tutor fit that was excellent in February may be wrong by September because the learner’s condition changed. A small group can become appropriate after intensive repair. A programme can discover that the school has moved faster than expected. The ability to revise the route is a quality feature, not inconsistency.

8 · The learner receives capability back

This is the deepest test. If a tutoring system becomes more sophisticated every year while the learner becomes less able to plan, attempt, judge and recover alone, the system has accumulated support without completing the transfer. The best tutoring can be highly intensive at the right time and still aim towards a learner who carries more.

How the hubs should connect without cannibalising one another

The eduKate ecosystem now contains several kinds of owner. They should reinforce one another by owning different questions.

The Tutor System owns the relationship map

This page owns the public question: how do parents, students and tutors work together from first concern to increasing learner independence?

The Tutor Handbook owns professional tutoring decisions

The Handbook owns the tutor-side operating logic. Individual volumes answer technical decisions rather than competing with the wrapper’s public search intent.

Find-a-Tutor and programme routes own discovery

National, level, subject, examination and location pages help families find the relevant tutor or tuition route. Their job is to connect search intent to the correct service or decision guide, then hand off to the Tutor System when the family needs to understand fit and function.

Learning Runtime owns learner state and evidence mechanics

MindOS, Bolt and the Student Interface go deeper into what the learner knows, how confidently the evidence supports a conclusion and what task/help interface the learner is operating. That is a different level of the system from parent-facing tutor selection.

Subject hubs own disciplinary truth

When routing reaches a subject mechanism, the subject owner takes over. This prevents the Tutor System from becoming an encyclopedia that is wide but technically shallow.

A practical routing index

Use the question that most closely matches your present need.

What this hub deliberately does not claim

A comprehensive system should also make its limits visible.

• It does not claim that tuition is necessary for every learner.
• It does not claim that a particular tutor class, tutorial size or delivery format is universally superior.
• It does not guarantee examination outcomes.
• It does not turn educational observation into medical or psychological diagnosis.
• It does not treat a single study or average effect as a prediction for an individual learner.
• It does not treat every improvement as permanent until retention and transfer have been checked.
• It does not require parents or students to master the technical Tutor Handbook before using the system.
• It does not ask subject hubs to surrender their disciplinary ownership to one generic tuition page.

The architecture in one sentence

Find the person, understand the learner, match the tutoring function, teach the subject properly, verify the change, and return as much of the system as possible to the student.