The Tutor Handbook Vol No.0113 | The Training Need Gate — How a Tutor Coach Separates a Tutor-Skill Gap From a Material or System Problem Before Assigning More Training

The Tutor Handbook · Volume 0113 · Series ID THB-0113

The lesson keeps ending badly.

The tutor rushes the final ten minutes. The learner leaves without a clear independent next move. The same pattern appears again the following week. A coach watches and says, “We need more training.”

Perhaps.

But “more training” is not yet a diagnosis.

The tutor may lack a teachable skill: planning the session backwards, closing with retrieval, setting a next-step handover, or protecting time for independent work. In that case, training may be the right response.

Or the materials may be too long for the allotted session. The timetable may make a proper close impossible. The programme may require tutors to record too much administration during teaching. The lesson template may contain four competing goals. The tutor may understand the closing routine perfectly but be carrying a group whose needs cannot fit the planned sequence.

Training a tutor to compensate for a system problem can produce a strange result: the person becomes more skilful at surviving a badly designed system while the system itself remains unchanged.

This article owns the decision that comes before training. How does a tutor coach decide whether an observed problem is actually a training need?

The direct answer

Treat training as one possible intervention, not the default explanation.

Start with the performance requirement: what should the tutor know, decide or do differently in a real session? Then gather evidence from more than one source where possible: live observation, tutor explanation, learner work, learner experience, materials, scheduling conditions and repeated patterns. Ask whether the problem persists when the environment supports the expected behaviour. Ask whether the tutor can perform the skill in a simpler or rehearsed condition. Ask whether several tutors fail at the same point using the same materials. Ask whether the tutor knows what to do but cannot execute it under the current constraints.

Only after those questions should a coach decide that training is the right lever.

This is not an argument against tutor development. It is an argument for aiming development at the mechanism that can actually change.

Why “training need” is often used too quickly

Training is attractive because it is actionable. A weak session appears. A workshop can be scheduled. A checklist can be issued. A module can be completed. Attendance can be recorded.

That makes training administratively visible.

System design is often less convenient. Fixing materials, workload, scheduling, unclear expectations or role boundaries can require organisational change.

So there is a predictable bias: when performance is weak, improve the person.

Sometimes that is correct. Sometimes it is the wrong level of analysis.

The National Student Support Accelerator’s Training Development Principles explicitly recommend using student academic data, student experience data and observations of tutor skills to identify additional training needs. The same guidance also says training should be adjusted to tutor selection criteria, content knowledge and ongoing observations, and that pre-service training alone is insufficient. This is strong research-informed programme guidance for the present problem because it treats training as responsive to evidence rather than as a fixed annual event.

However, the NSSA Rationale and Usage Guide also acknowledges an important evidence limit: relatively few studies directly isolate tutor training, so some recommendations draw from the teacher professional-development literature. That transfer should remain visible. A private three-student tutorial in Singapore is not automatically equivalent to a US school-based tutoring programme or a teacher-PD study.

Define the performance before diagnosing the gap

A training need cannot be identified from a vague complaint.

“The tutor needs better questioning.” “The tutor needs stronger classroom management.” “The tutor needs to engage students.” “The tutor needs better explanations.” Each statement is too broad.

A coach should define an observable professional performance.

For example: after a partly correct learner answer, the tutor should ask a follow-up question that preserves the learner’s reasoning rather than supplying the missing step. Before introducing a new task, the tutor should establish the learning target and what a successful response must demonstrate. During three-student independent work, the tutor should sample each learner’s thinking without allowing the fastest learner to dominate the evidence. At the end of the session, the tutor should leave each learner with a clear next move and an independence condition.

Now the coach can ask whether the tutor lacks the knowledge, discrimination, decision rule, practice or fluency needed to perform that move.

Without this specificity, “training” easily becomes a general response to dissatisfaction.

The five-way differential

When a tutor performance problem appears, separate at least five plausible causes.

Knowledge gap. The tutor does not know something necessary: subject content, curriculum expectation, assessment rule, instructional principle or procedure.
Skill gap. The tutor understands the idea but cannot yet execute it reliably in live teaching.
Judgement gap. The tutor can execute several moves but does not recognise when each is appropriate.
Material or system problem. The expected performance is made difficult or impossible by lesson design, resources, workload, group composition, timing, technology or programme rules.
Implementation or priority problem. The tutor knows the expected move and can perform it, but it is not being used consistently because another priority is winning, the expectation is unclear, or feedback has not been translated into a routine.

These categories can overlap. They are not a validated taxonomy and should not become a pseudo-scientific score. Their value is diagnostic: each points toward a different improvement response.

A knowledge gap may need study or content training. A skill gap may need modelling, rehearsal, observation and feedback. A judgement gap may need contrasting cases and decision practice. A system problem may need redesign. A priority or implementation problem may need clearer expectations, protected time, reminders, feedback or accountability rather than another workshop.

The coach should not prescribe training until the likely cause is narrow enough that training could plausibly change it.

Can the tutor explain the move?

One useful probe is explanation.

Ask the tutor to explain what they were trying to do, what they noticed, which alternatives they considered and why they chose the observed move.

This is not a test of eloquence. A tutor may teach well without sophisticated professional language. The question is whether the underlying decision is available for inspection.

Suppose a tutor gives the answer after three seconds of silence. A coach might assume the tutor lacks wait-time skill.

Then the tutor explains: “I knew I should wait. I cut in because the other two learners had finished, one was starting to distract the group and we had four minutes left.”

That answer does not prove the interruption was correct. But it changes the diagnosis. The issue may involve orchestration, timing or session design rather than ignorance of wait time.

Now compare another tutor who says: “If they do not answer immediately, I tell them so they do not lose confidence.” That may reveal a belief or judgement issue about struggle, confidence and support.

Same observed behaviour. Different improvement need.

Can the tutor do it in rehearsal?

The next probe is controlled performance.

If a tutor can describe a move but not execute it in a short rehearsal, there is stronger evidence for a skill gap.

Ask the tutor to practise responding to a partly correct answer; giving one piece of feedback without rewriting the response; opening a lesson with a learning target; ending with a handover; or redirecting the fastest student while preserving another learner’s thinking time.

Rehearsal reduces live complexity. If the tutor still cannot perform the move, training and coached practice become more plausible.

If the tutor performs it well in rehearsal but not in the real lesson, do not conclude that training failed. Ask what changes under live conditions: timing, emotional pressure, multiple learners, unfamiliar content, material sequence, behaviour, technology or competing programme requirements.

This is the distinction between competence in a simplified condition and performance in the real one.

The Rehearsal-to-Live Gate owns the transfer from rehearsal into real teaching. This article stops earlier. It asks whether rehearsal and training are the right intervention in the first place.

Look for cross-tutor patterns

One of the strongest clues that a problem may not be individual is repetition across several tutors.

If one tutor repeatedly fails to complete a lesson sequence, investigate the tutor. If six tutors using the same resource all abandon the same final activity, investigate the resource. If several tutors misunderstand the same instruction, investigate the instruction. If every tutor spends the first fifteen minutes fixing a prerequisite that the lesson assumes is already secure, investigate the curriculum dependency.

This does not prove a system cause. Tutors can share the same training gap. But the pattern changes the probability and should stop a coach from individualising the problem too quickly.

A useful question is: if a highly capable tutor entered this exact system tomorrow, what would still be difficult?

If the answer includes impossible timing, contradictory goals, inaccessible materials or missing information, professional learning alone is unlikely to solve the whole problem.

A fictional composite case: the rushed ending

This example is fictional and constructed. It does not describe a real employee or customer.

A coach observes Denise, a tutor in a three-student group, for three sessions. Each lesson begins well. The explanations are clear, the students attempt work and feedback is specific. But the final ten minutes collapse. Denise speeds through corrections, tells students what to finish at home and does not obtain a fresh independent attempt.

The first hypothesis is a coaching need: Denise must learn a better lesson close.

The coach does not stop there.

Denise can explain the intended closing routine accurately. In rehearsal she can run it. She knows the purpose: independent retrieval, next-step ownership and a clean handover.

The coach then inspects the materials. Each lesson contains approximately ninety-five minutes of planned activity inside a ninety-minute tutorial, before accounting for transitions or learner errors. The last task is labelled “extension” but is treated as compulsory in the tutor notes. A progress form must also be completed before the session ends.

The dominant problem is not that Denise lacks the closing skill. The system makes the closing routine the easiest part to sacrifice.

The improvement route therefore has two parts. The material sequence is reduced so the core lesson fits the real session. Denise receives brief coaching on a stopping rule: protect the final independent check even when earlier activity runs long.

Training remains useful, but it is no longer being asked to repair an impossible timetable.

When a problem really is a training need

The gate should not become an excuse to blame the system for every weak lesson.

Sometimes the evidence is straightforward. A tutor repeatedly converts learner errors into immediate explanations. The lesson materials allow plenty of time. Other tutors using the same materials preserve learner attempts. The tutor explains that giving the correct method immediately “saves time.” In rehearsal, the tutor struggles to generate diagnostic follow-up questions and frequently leaks the next step.

This pattern supports a real development need.

The coach can now specify it precisely: learn to distinguish error types, generate discriminating questions, preserve attempt ownership and practise the move until it can survive live timing.

The intervention can include a short conceptual explanation, models and non-models, contrasting cases, rehearsal, observation, specific feedback and a later live check. That is much stronger than assigning a general “questioning skills” module.

Training should have an exit condition

A common professional-development failure is that participation becomes the evidence.

The tutor attended. The tutor completed the module. The tutor passed the quiz. The tutor said the training was useful.

Those facts can be worth recording. They do not prove teaching changed.

A training need should therefore be commissioned with a performance receipt. What should become different in live teaching? Under what conditions? How soon should the change be visible? What evidence will distinguish improvement from ordinary session variation? What evidence would show that the original diagnosis was wrong?

For a questioning skill, the receipt might be that on several relevant learner errors, the tutor uses a follow-up question that preserves learner reasoning and obtains a fresh response before providing a fuller explanation. For session closure, the receipt might be that the tutor protects an independent final check and a learner-owned next action across several sessions, including one session that runs behind schedule.

The Coaching Receipt already owns the question of proving that professional feedback changed live teaching. The Training Need Gate supplies the upstream diagnosis so the receipt is attached to the right target.

Training content should match the gap

NSSA’s Training Topic Guidance lists broad areas such as tutoring basics, relationships, data and assessment, instruction and session structure, social-emotional support and content-specific knowledge. The depth should vary with tutor experience and existing knowledge.

That is important because training volume is not training quality.

A subject expert may not need another content lecture but may need practice turning expertise into explanations a novice can use. An experienced classroom teacher may need little help with general pedagogy but substantial orientation to a tutoring programme’s materials, data routines or one-to-three orchestration. A new tutor may require more explicit modelling of instructional moves. A tutor teaching outside recent subject experience may need content refresh before pedagogical rehearsal.

The gap defines the training job. The training catalogue should not define the diagnosis.

Do not confuse confidence with competence

Tutors can feel uncertain while teaching well. They can also feel confident while teaching badly.

Self-report is therefore useful but insufficient.

A tutor who says “I need more training” may be accurately identifying a gap. Or they may be responding to unclear expectations, difficult materials, a new group, one unusually hard lesson or lack of feedback. A tutor who says “I am fine” may still have a recurring skill gap visible in learner evidence.

The coach should treat confidence as one data source. Ask what the tutor can do, what they can explain, what happens in rehearsal, what appears in live teaching, what the learners experience and what conditions surround the performance.

The same principle applies to learner confidence: a feeling is real evidence about experience, but it is not a complete measure of capability.

Do not turn every variation into remediation

Good teaching is adaptive. Tutors will not all sound identical.

One tutor may use concise questions. Another may use longer verbal scaffolds. One may annotate visually. Another may use oral comparison. Variation is not automatically a defect.

Training should target differences that matter to the educational job.

The question is not: did the tutor match the coach’s preferred style? The question is: did the learner receive a coherent opportunity to think, practise, receive useful feedback and become more independent relative to the target?

Standardisation is valuable where consistency protects quality, safety, valid assessment or programme coherence. It becomes harmful when surface conformity replaces functional evidence.

A coach should therefore distinguish required principles from optional style.

System fixes can be professional learning too

If the diagnosis points to materials or systems, the response does not have to exclude tutor learning.

Changing a lesson sequence can be paired with explaining why it changed. Clarifying a role can be paired with rehearsing the boundary. Reducing administrative load can be paired with teaching a better observation routine. Changing group composition can be paired with planning how the tutor will use the new grouping.

The distinction is not “training or system.” It is “what must change first, and which parts require human learning?”

That ordering matters. Otherwise the tutor may be trained on a behaviour the system still punishes.

The role of learner evidence

Learner outcomes can signal a problem, but they should not be used as a crude tutor score.

A weak result can arise from prior knowledge, attendance, task validity, school conditions, support, practice, assessment difficulty or many other factors. A strong result does not prove every teaching move was good.

The most useful learner data for a training diagnosis are often close to the target behaviour. If the target is tutor feedback, inspect whether learners can use the feedback on a fresh attempt. If the target is checking for understanding, inspect whether the tutor obtains responses from all learners and adapts instruction when misunderstanding appears. If the target is independence, inspect what happens when prompts are reduced.

Academic results remain important, but the inference from score to tutor skill must be cautious.

NSSA’s Training Development Principles use multiple data sources for precisely this reason: student academic performance, student experience and observations of tutor skills can jointly inform training needs.

Multiple sources do not need to agree perfectly. Their disagreement can be diagnostic.

The role of learner experience

A learner can tell you that the tutor moves too fast, explanations are hard to follow or questions feel impossible to answer in the available time. These reports matter.

But the coach should avoid converting satisfaction into instructional truth. A learner may dislike retrieval because it is effortful. A learner may enjoy a tutor who gives answers quickly. A learner may prefer easy work.

Experience data should answer experience questions. Performance evidence should answer performance questions.

The Learner Feedback Independence Check in the next volume deals with how to collect that feedback honestly. Here the key point is narrower: learner experience can help locate a professional-learning need, especially when it is specific, repeated and linked to observable teaching conditions.

When to use a second observer

Some performance questions are difficult to interpret from one observation.

The coach may have seen an unusual session. The tutor may behave differently because of being watched. The material may have been atypical. The group may have faced a one-off disruption.

Where the decision is consequential, a second observation, another coach, video review where appropriately authorised, or a later live sample can reduce overreaction.

This is not a requirement to collect endless evidence. The evidence should be proportionate to the decision.

A minor coaching suggestion may require little. A judgement that a tutor cannot perform a core function deserves stronger support.

The Observation Window already owns how much live teaching evidence to collect before feedback. The Training Need Gate uses that evidence to decide what kind of intervention follows.

A compact training-need record

Observed performance: what happened, without interpretation.
Expected performance: what should have happened.
Conditions: materials, group, time, task, technology, unusual events.
Tutor explanation: what the tutor intended and noticed.
Rehearsal evidence: can the move be performed outside live complexity?
Cross-tutor pattern: individual or shared?
Learner evidence: what happened to learner thinking, work or experience?
Working cause: knowledge, skill, judgement, system, implementation, or mixed.
Intervention: training, coaching, redesign, clarification, workload change, material change, or combination.
Receipt: what future live evidence will show improvement.

Again, this is not a validated instrument. It is a disciplined note that prevents the phrase “needs training” from ending the diagnosis prematurely.

Failure modes

The workshop reflex: every problem becomes a course.
The person-blame reflex: system friction is attributed to tutor weakness.
The system-blame reflex: genuine skill gaps are excused because conditions are imperfect.
The attendance receipt: completion of training is treated as improvement.
The generic-module failure: a broad course is assigned to a narrow gap.
The style-conformity failure: tutors are trained to imitate a coach rather than meet a functional standard.
The one-observation verdict: an atypical lesson becomes a stable trait judgement.
The outcome shortcut: one learner score is treated as proof of tutor quality.
The catalogue-first failure: available training determines the diagnosis.
The no-exit failure: the tutor remains “in training” without a defined live performance that would close the issue.

The parent-facing implication

Parents do not need to manage tutor professional development, but they can ask useful questions when a tutoring service says it is improving quality.

What exactly is expected to change? How will the organisation know the change reached live teaching? Does the service distinguish a tutor problem from a material or programme problem? Are tutors supported after initial training?

These questions are more informative than simply asking how many hours of training tutors receive.

Hours are inputs. The educational question is whether the right capability is built and whether it appears in practice.

The tutor-facing implication

A tutor should not experience the Training Need Gate as a defence against feedback. It is a way to make feedback fairer and more useful.

If the problem is a skill gap, the tutor deserves precise training, practice and follow-up rather than vague criticism. If the problem is materials, the tutor deserves materials that make good teaching possible. If the problem is workload or scheduling, the tutor deserves an operational solution. If the problem is judgement, the tutor deserves cases that help discriminate when different moves apply. If the problem is content knowledge, the tutor deserves time and resources to refresh it before being expected to teach confidently.

Good diagnosis raises the standard because it makes improvement more likely.

A delayed check matters

Immediate improvement after coaching can be misleading for the same reason immediate learner success can be misleading. The tutor has just heard the feedback. The model is fresh. The coach may still be nearby. The next lesson may contain unusually convenient opportunities to demonstrate the move.

A stronger check returns later.

Can the tutor still use the behaviour after several sessions? Can it survive a different learner, a harder task or a session that starts badly? Can the tutor explain when not to use it? Has the move become part of professional judgement rather than a short-lived imitation?

Changed-condition checks are especially important for judgement gaps. A tutor who has practised one questioning script may perform it well on the exact scenario used in coaching and still fail when the learner gives a different kind of error. Training has transferred only when the tutor recognises the underlying decision, not merely the surface wording.

This is also where a system diagnosis can be retested. If materials were shortened because the closing routine was repeatedly squeezed out, observe whether tutors now protect the close. If the problem remains after the system constraint is removed, a tutor-skill or priority issue becomes more plausible. If performance improves across tutors without additional training, the system hypothesis gains support.

Delayed checking therefore protects both sides of the judgement. It prevents coaches from declaring training successful too early, and it prevents organisations from blaming tutors for problems that disappear when the environment is repaired.

Research boundaries

The National Student Support Accelerator provides the most directly relevant public guidance used here. Its tutor-training materials recommend aligning training to tutor characteristics and needs, using observations and multiple forms of data, and providing ongoing in-service support. Its own rationale notes that the direct research base on tutor training is limited and draws in part from teacher professional development.

The Education Endowment Foundation’s teacher-professional-development systematic review and related resources provide a broader evidence base on mechanisms associated with effective teacher PD. Teacher PD is relevant by analogy and transfer, not identical to private tutoring.

The EEF School’s Guide to Implementation adds evidence-informed guidance on how context, behaviours and structured implementation influence whether an approach becomes real practice.

No source establishes the five-way differential in this article as a validated classification. It is a practical synthesis for tutoring judgement.

Sources and further reading

The final return

A training programme should begin one question earlier than training.

What is the problem?

Not who is the problem. Not which module is available. Not what workshop can be scheduled.

What performance is missing, under what conditions, and why?

Sometimes the answer will be clear: the tutor needs knowledge, rehearsal, feedback and practice. Sometimes the lesson template needs repair. Sometimes the timetable is impossible. Sometimes expectations are contradictory. Sometimes a tutor knows the move but has not yet learned to choose it under live complexity. Sometimes several causes are active together.

The Training Need Gate keeps professional learning honest by making training earn its place as the intervention.

When the diagnosis is right, training becomes more humane and more demanding at the same time. The tutor knows what to build. The coach knows what to observe. The programme knows what not to blame on training.

And the learner receives the part that matters: better teaching in the real lesson, not merely another certificate that says somebody attended a course.

Small Group Tutorials