A resit is not simply the same course repeated one more time.
The learner already has history with the subject. They have sat the examination, received a result, formed beliefs about what that result says about them, and often experienced months or years of partial success, avoidance, rushed revision or inconsistent attendance. When tutoring begins, the tutor is not working with a blank slate. They are working with knowledge, habits, emotions and timing constraints that have already interacted once badly enough to produce failure or an insufficient grade.
That makes resit tutoring a distinctive educational problem.
The Education Endowment Foundation’s Tutor Trust pilot, completed in September 2026, tested a post-16 one-to-one tutoring and coaching programme for learners resitting GCSE Mathematics or English. Learners and staff reported improvements in attitudes, confidence, motivation, participation and persistence. Among the 117 learners for whom grades were available, 26 percent achieved grade 4 or above, higher than the national resit pass rate. But the pilot had no control group, the sample was limited, and attendance was a serious implementation problem: 31 percent withdrew before tutoring began and only 24 percent completed the minimum 12 hours. The programme was not judged ready for a randomised trial in its current form.
That evidence points to the deeper mechanism. A resit programme cannot help a learner who does not reliably receive the instruction, and instruction itself must be targeted at the reasons the first attempt failed.
The 50-second answer
Resit tutoring works when it treats the second attempt as a diagnosis-and-repair problem rather than a replay. The tutor identifies whether failure came from missing knowledge, weak retrieval, method selection, reading demands, timing, attendance, confidence, exam strategy or some combination. The tuition then concentrates practice, feedback and coaching around those bottlenecks while remaining explicitly linked to the qualification the learner must retake.
One-to-one tuition has a strong broader evidence base and EEF estimates an average positive effect across studies, especially when tutoring is targeted, structured and linked to normal learning. But the 2026 Tutor Trust resit pilot shows why post-16 implementation is different. Retention and attendance were difficult; only a minority completed the intended minimum dose. Positive attitudes and reported confidence are promising, but they do not establish causal attainment effects.
The practical rule is that dose precedes impact. The programme must first make attendance possible and worthwhile. Then tutoring has to repair the actual failure mechanism, not merely add more hours of the same teaching the learner already survived once.
1. Begin with the marked history, not the syllabus page
A resit learner has already produced evidence: past papers, school reports, mock exams, coursework, attendance records and remembered failure points. The fastest route is to inspect that evidence before deciding what to reteach.
A concrete example: a Mathematics tutor sees that algebra scores are acceptable but multi-step word problems collapse after the first modelling decision.
The caveat: Old marks can be noisy or incomplete. Diagnosis should use fresh tasks too.
2. Separate knowledge gaps from performance gaps
Some learners genuinely do not know the material. Others know it untimed but cannot retrieve, select or execute under exam conditions. These require different tutoring.
A concrete example: one learner cannot factorise quadratics at all; another can factorise but chooses the quadratic formula inefficiently and runs out of time.
The caveat: Treating every low score as missing knowledge leads to unnecessary reteaching.
3. Attendance is the first implementation variable
Tutoring effects depend on exposure. The Tutor Trust pilot showed severe attrition before and during delivery, making attendance a central mechanism rather than an administrative side issue.
A concrete example: a programme contacts learners quickly after absence and offers timetable options that avoid clashes with work placements.
The caveat: Attendance barriers may include transport, employment, caring responsibilities and motivation. They should be understood, not moralised.
4. A partnership lead can protect the programme
The pilot found implementation was stronger where a proactive setting lead monitored attendance and engagement. Somebody needs responsibility for making the tutoring system real.
A concrete example: a college coordinator checks missed sessions, communicates timetable changes and keeps tutors connected to subject staff.
The caveat: Coordination should not become bureaucracy that consumes the tutor’s teaching time.
5. The tutor should know the resit specification
Targeted support depends on the exact demands of the qualification: content, command words, calculator rules, writing formats, mark allocation and timing.
A concrete example: an English tutor distinguishes reading analysis from transactional writing demands rather than teaching generic ‘better English.’
The caveat: Specification familiarity should not reduce the course to exam tricks. Foundational subject capability still matters.
6. Confidence repair needs evidence
Learners who have failed once may say ‘I am just bad at Maths’ or ‘English is not for me.’ Reassurance alone is weak. Confidence grows more credibly when the learner experiences specific, repeatable success.
A concrete example: the tutor tracks that the learner can now solve six ratio problems independently after previously needing prompts.
The caveat: Celebrating easy work can feel patronising. Success should represent genuine capability growth.
7. Coaching can support persistence
The Tutor Trust model trained tutors in coaching strategies around relationships, confidence and motivation. Coaching can help learners plan, reflect and persist alongside subject instruction.
A concrete example: a tutor asks the learner to identify the exact point they usually stop on a long question and builds a restart routine.
The caveat: Coaching cannot replace missing subject teaching. Motivation without knowledge still produces wrong answers.
8. One-to-one time should increase response density
The advantage of tutoring is not intimacy by itself; it is the ability to make the learner think, answer and receive feedback repeatedly.
A concrete example: in thirty minutes, the learner solves, explains or retrieves dozens of targeted items rather than watching a long explanation.
The caveat: Tutors can overtalk. One-to-one teaching easily becomes a private lecture unless response opportunities are designed.
9. The tutor should target the bottleneck, not the favourite topic
Learners often prefer practising what already feels comfortable. Diagnosis should decide the focus.
A concrete example: a student wants to revise percentages because they are familiar, but the evidence shows weak algebraic rearrangement is causing larger mark losses.
The caveat: Relentless focus on weakness can damage motivation. Mix repair with visible strengths and progress.
10. Link tutoring to ordinary teaching
EEF’s broader one-to-one tuition evidence suggests tutoring is more effective when additional to and explicitly connected with normal lessons.
A concrete example: the tutor knows what the learner is currently doing in college and reinforces the same notation and core methods.
The caveat: Tutoring should not merely duplicate class. It should add intensity, diagnosis and practice the main setting cannot provide.
11. Short regular sessions can outperform sporadic marathons
Broader tutoring evidence suggests regular sessions over a defined period are often effective because practice accumulates and feedback loops stay short.
A concrete example: two or three focused weekly sessions maintain momentum better than one long session every few weeks.
The caveat: Post-16 timetables may make ideal frequency impossible. Feasibility is part of design.
12. Retrieval should begin before notes appear
Resit learners often overuse rereading because it feels safe. Tutoring can make retrieval the default evidence of what is actually available.
A concrete example: the learner attempts formula recall or a paragraph plan from memory, then checks and corrects.
The caveat: Unsupported retrieval can be demoralising when knowledge is absent. Teach first when necessary.
13. Error logs should describe mechanisms
A list of wrong questions is less useful than a classification of why they went wrong.
A concrete example: errors are tagged as knowledge, command word, method selection, arithmetic, evidence use, timing or checking.
The caveat: Too many categories become administrative. Keep the taxonomy small enough to guide teaching.
14. Method selection deserves direct practice
Examinations often require choosing among known methods. Learners may fail because they cannot recognise which tool fits.
A concrete example: the tutor shows five problems and asks for method choice without completing them.
The caveat: Method labels without execution can create false confidence. Follow selection with actual solving.
15. Reading load can hide subject knowledge
In both Mathematics and English resits, dense wording can block performance. Tutors should distinguish language difficulty from conceptual difficulty.
A concrete example: a learner can solve a ratio relation once the question is paraphrased, showing that interpretation is part of the bottleneck.
The caveat: Over-paraphrasing can remove the exact reading demand the exam requires. Support should fade.
16. Timing should be added after accuracy stabilises
Learners who failed under time pressure often want immediate timed practice. If the method is unstable, the clock can train rushed errors.
A concrete example: first secure a valid solution route, then add a reasonable time target and compare decisions.
The caveat: Some timing problems only appear under pressure, so timed diagnostics still have a role.
17. Exam strategy should remain subject-specific
Generic advice such as ‘skip hard questions’ is incomplete. Good strategy depends on mark structure, dependencies and response format.
A concrete example: a Mathematics learner learns when to park a low-mark item; an English learner allocates time based on writing demands and planning needs.
The caveat: Strategy cannot compensate for insufficient knowledge.
18. Tutors need a progression map
Random worksheet practice creates activity without a coherent route. A tutor should know what prerequisites lead into the target topics.
A concrete example: weak percentage problems trace back to multiplicative reasoning, so practice repairs that relation before exam-style questions.
The caveat: Do not reopen every historical gap. Repair what is causally upstream of current failure.
19. Feedback should produce a second attempt
Correction becomes learning when the learner must act on it.
A concrete example: after one prompt, the learner closes the model and solves a near problem independently.
The caveat: If the tutor explains every correction fully, the learner can mistake following for mastery.
20. Attendance data should inform interpretation
A learner who completed four hours of a twelve-hour programme has not received the same intervention as a learner who completed the full dose.
A concrete example: progress records include actual tutoring hours alongside assessment outcomes.
The caveat: Dose-response patterns from one local cohort do not prove causation, but they improve interpretation.
21. Motivation problems may be rational
A learner who has failed repeatedly may see little probability of success or may prioritise work and family demands. Motivation is not simply an attitude defect.
A concrete example: the tutor makes the route concrete by showing which small set of skills would unlock a meaningful grade improvement.
The caveat: Optimism must remain realistic. Do not promise a grade that evidence cannot support.
22. Relationship quality can increase diagnostic honesty
Learners reveal more when they trust that mistakes will not be used to shame them.
A concrete example: a student admits they have been guessing on algebra questions instead of pretending the topic is understood.
The caveat: A good relationship should support challenge, not reduce standards.
23. Tutor training matters
EEF’s broader evidence notes the importance of experienced, trained and supported tutors, especially when non-teachers deliver programmes.
A concrete example: tutors learn a common correction routine, diagnostic framework and escalation route.
The caveat: Scripts can help consistency but should not prevent responsive teaching.
24. Coaching goals should be observable
‘Become more motivated’ is vague. Better goals describe actions under the learner’s control.
A concrete example: attend three consecutive sessions, complete two independent practice blocks, or use the restart routine after getting stuck.
The caveat: Behavioural goals should not ignore structural barriers such as timetable clashes.
25. Practice needs sufficient similarity before transfer
Early repair may use close examples so the learner can stabilise a method. Later practice should vary surfaces so the learner learns to recognise structure.
A concrete example: ratio practice moves from direct numerical questions to unfamiliar contexts and mixed problem sets.
The caveat: Variation too early can overwhelm; variation too late can produce brittle learning.
26. Mock tests should diagnose, not merely rank
A resit mock is valuable when errors become a repair map.
A concrete example: the tutor analyses where marks were lost by section and mechanism, then changes the next fortnight’s plan.
The caveat: Repeated full papers without targeted repair can rehearse the same failure pattern.
27. The programme needs an engagement strategy
The pilot showed learner retention was difficult. Programme design should therefore include communication, scheduling, reminders and visible value from the first sessions.
A concrete example: the first tutoring meeting produces a concrete diagnosis and one immediate success rather than spending the whole time on administration.
The caveat: Engagement tactics cannot become coercive or ignore learner autonomy.
28. Post-16 learners deserve adult treatment
Many resit students are balancing employment, vocational study and adult responsibilities. Tutoring should respect that context.
A concrete example: sessions begin on time, goals are transparent and practice is linked to the qualification’s real consequences.
The caveat: Respect does not mean lowering expectations or avoiding corrective feedback.
29. Success should include more than pass rate but not hide behind softer outcomes
Confidence, persistence and attendance matter because they can enable learning. They should be reported honestly alongside grades and skill evidence.
A concrete example: a programme tracks independent diagnostic scores, attendance and learner confidence separately.
The caveat: Positive attitudes should not be used to imply attainment gains that have not been causally demonstrated.
30. The pilot pass-rate comparison is suggestive, not causal
In the Tutor Trust pilot, 26 percent of learners with available grades achieved grade 4 or above, but there was no control group and attendance was low.
A concrete example: treat the figure as evidence of promise that motivates better testing, not as a programme effect size.
The caveat: National-average comparisons cannot control for who enrolled, persisted or supplied grades.
31. Non-completion is part of programme effectiveness
If a theoretically strong tutoring design repeatedly loses learners before sufficient exposure, scalability is weak.
A concrete example: leaders redesign enrolment, scheduling and attendance support before running a larger trial.
The caveat: It is not enough to say the tutoring works ‘for those who attend’ if attendance is predictably difficult.
32. A resit plan should have milestones
The final exam is too distant to guide weekly decisions. Short milestones keep progress visible.
A concrete example: the learner aims to stabilise algebra accuracy by week four, then complete mixed timed sections by week eight.
The caveat: Milestones should adapt to evidence rather than become inflexible promises.
33. Exit from tutoring should mean increased independence
The goal is not indefinite tutor dependence. As capability grows, prompts should fade and self-study routines should take over.
A concrete example: the tutor moves from guided examples to independent mixed practice and self-marking with error analysis.
The caveat: With a fixed resit date, tutoring may end before full independence. The plan should still transfer as much control as possible.
34. The second attempt should not feel like a copy of the first
A meaningful resit programme changes something causal: instruction intensity, diagnosis, practice design, attendance support, feedback or strategy.
A concrete example: the learner can name what is different about this preparation cycle and why it addresses the previous failure.
The caveat: Changing only the tutor’s personality without changing the learning system may produce warmth but little repair.
Practical route for teachers
Start with evidence from the first attempt and a fresh diagnostic. Classify the main bottlenecks, then build a short progression that combines subject repair, retrieval, method selection and increasingly exam-authentic practice. Protect response density: the learner should do more thinking than the tutor.
Treat attendance as part of the intervention. Coordinate with the setting, review actual hours received, and respond quickly to disengagement. Track confidence, attendance and attainment separately so softer outcomes are valued without being mistaken for pass-rate effects.
Practical route for students
A resit is not proof that you are incapable. It is evidence that the first learning-and-performance system did not produce the required result. Your job is to find which part failed and change it.
Bring old papers if you have them. Track errors by cause. Practise from memory before looking. Ask your tutor why each task is in the plan. Most importantly, attend consistently: a good tutoring programme cannot work in hours you do not receive.
Practical route for parents and families
For older resit learners, support often means protecting logistics and dignity rather than supervising every homework task. Help with transport, schedule, quiet study time or reminders when welcomed. Ask whether tutoring is linked to a clear diagnosis and whether attendance is consistent.
Avoid turning every result into a judgement about character. Resit learners need honest standards and a credible route, not either panic or empty reassurance.
Common failure modes
- Repeating the original course without diagnosing why the first attempt failed.
- Treating low motivation as a character flaw while ignoring timetable and life barriers.
- Measuring planned tutoring hours instead of hours actually received.
- Using one-to-one time for long tutor explanations rather than learner responses.
- Reteaching favourite or easy topics while high-value bottlenecks remain.
- Adding time pressure before methods are accurate.
- Using full mock papers repeatedly without targeted repair.
- Reporting confidence gains as if they were causal pass-rate gains.
- Ignoring withdrawal and non-attendance when judging programme feasibility.
- Letting tutoring become a permanent crutch instead of transferring control to the learner.
Frequently asked questions
Is one-to-one tutoring effective in general?
EEF’s broader evidence base finds one-to-one tuition can be highly effective on average, especially when targeted, structured and linked to normal teaching. The exact effect varies by programme and context.
What did the 2026 Tutor Trust resit pilot show?
It showed promising reported gains in confidence, motivation and persistence, and 26 percent of learners with available grades achieved grade 4 or above. But there was no control group, attendance was low, and the programme was not considered ready for an RCT in its current form.
Why did attendance matter so much?
Only 24 percent completed the minimum 12 hours and 31 percent withdrew before tutoring began. Without sufficient exposure, even good tutoring cannot plausibly produce its intended effect.
Should resit tutoring focus on exam technique?
Exam strategy matters, but it should sit on top of repaired subject knowledge, retrieval and method selection. Technique alone cannot compensate for missing capability.
How can a learner know tutoring is working?
Look for independent evidence: fewer prompts, stronger diagnostic scores, better method selection, improved timed performance and sustained attendance.
The final idea
A resit tutor is not being asked to make the learner relive the course more intensely. The job is to discover why the first attempt failed and build a different path through the same required destination.
That path is partly academic and partly operational. The learner has to show up. The tutor has to target the right bottleneck. Practice has to produce independent capability. Confidence has to grow from evidence rather than slogans. The programme has to survive timetables, work, fatigue and the history of previous failure.
The 2026 pilot is most useful because it makes implementation visible. Positive relationships and reported motivation are not enough if too few learners receive the intended dose.
The rule is: repair the failure mechanism, protect attendance, and make the second attempt genuinely different from the first.
Sources and further reading
- Education Endowment Foundation — Tutor Trust post-16 resit pilot
- Education Endowment Foundation — One to one tuition
- Education Endowment Foundation — Small group tuition
Continue exploring on eduKateSG
- How X Works Hub
- How Tutoring Works | The Full Picture
- How High Performance Diagnosis Works | Find the Constraint Before You Add More Effort
- How Mock Exams Fail | Why Simulation Without Fidelity, Diagnosis and Repair Creates False Readiness
A twelve-hour resit tutoring architecture
The Tutor Trust pilot used a minimum intended dose of twelve hours, but many learners did not receive it. A useful way to think about such a block is as four phases rather than twelve identical lessons.
Hours 1–3: diagnosis and quick wins
Inspect old evidence, run fresh diagnostic tasks and identify two or three high-value bottlenecks. Create one early success that is genuinely meaningful. The learner should leave knowing what failed previously and what the new plan is changing.
Do not spend all three hours retesting. Diagnosis must turn into teaching quickly.
Hours 4–6: core repair
Teach the prerequisites and methods that are causally upstream of repeated mark loss. Use concentrated response practice, immediate feedback and second attempts. Keep a small error log organised by mechanism.
Confidence during this phase should come from visible independence: fewer prompts, cleaner method selection and successful retrieval.
Hours 7–9: mixed transfer
Begin mixing topics, surfaces and command words. The learner now has to decide which knowledge applies. Add moderate timing pressure only after accuracy is stable.
This phase often reveals whether the early repair was robust or tied to familiar worksheet formats.
Hours 10–12: exam conversion
Use exam-authentic sections, pacing decisions, checking routines and realistic constraints. Analyse every remaining error by cause. Build a final independent revision plan for the period after tutoring.
The tutoring block should end with the learner owning more of the system than they did at the start.
An attendance rescue protocol
When a learner misses a session, respond quickly and without moralising. First identify the barrier: timetable clash, work, transport, caring responsibility, forgetting, low perceived value or avoidance after difficulty. Second, choose a proportionate response. Some barriers need rescheduling; some need clearer communication; some need a more compelling first-session experience.
Track consecutive attendance, not only total enrolment. The Tutor Trust pilot suggests that retention is not peripheral. A programme that cannot keep learners in the room has an implementation problem before it has an instructional one.
A resit error map
Use five broad categories: knowledge, interpretation, method selection, execution, regulation.
Knowledge means the fact, rule or concept is missing. Interpretation means the learner misreads language, data or command words. Method selection means they know techniques but choose the wrong one. Execution covers arithmetic, algebraic or writing-process errors after a valid method is chosen. Regulation includes timing, recovery after a difficult item and checking.
A sixth category—attendance—belongs outside the paper but inside the programme. It explains whether the learner received enough teaching for the outcome to be interpretable.
When confidence repair is real
Confidence repair is real when the learner can point to evidence. “I can now start these questions without waiting.” “I finished the section in time twice.” “My last three paragraphs all explain the quotation after using it.” “I attended every session this fortnight.”
These statements are stronger than “I feel better at Maths” because they connect belief to capability.
The tutor’s job is not to manufacture optimism. It is to create conditions under which the learner has legitimate reasons to update their self-judgement.
A first-session script for resit tutoring
The first session should answer four questions for the learner.
What happened last time? Review evidence without shame. What are we changing? Name the few mechanisms the new plan will address. What can improve quickly? Create one credible early win. What will you need to do between sessions? Make attendance and independent practice expectations explicit.
This opening matters because resit learners often arrive with a global story—“I failed Maths”—rather than a usable diagnosis. The tutor’s first professional act is to turn that global story into a repairable map.
A final handover before the examination
Tutoring should end with a handover document owned by the learner, not the tutor. It can fit on one page: remaining high-risk topics, reliable methods, common error warnings, time checkpoints, recovery routine after a difficult question, and the final week’s practice schedule.
The learner should be able to explain every item on the page. If the tutor has to interpret it, control has not transferred.
This handover also protects against the emotional drop that can follow the last tutoring session. The programme ends, but the learner still has a compact operating plan for the days that remain.
A tutor dashboard that stays small
A resit tutor can track five lines after every session: attendance, target, independent success, recurring error, next move. This is enough to keep the programme responsive without turning tutoring into paperwork.
Attendance records whether the learner actually received the session. Target states the capability practised. Independent success records what the learner did without prompting at the end. Recurring error identifies the pattern that still matters. Next move decides what will change rather than merely what worksheet comes next.
Across several weeks, the dashboard reveals whether the learner is progressing, whether one bottleneck persists and whether tutoring dose is accumulating. It also makes conversations with the setting more concrete. Instead of “they are doing better,” the tutor can say that algebraic method selection has stabilised but timed reading interpretation remains weak.
When a resit learner should not simply receive more tutoring
More hours are not always the answer. If attendance is impossible because of timetable conflict, the programme design must change. If the learner has a substantial unaddressed language, sensory or learning need, appropriate school or specialist processes may be required. If the tutor repeatedly teaches the same content without independent improvement, the method should be reconsidered.
Tutoring is one support layer, not an explanation for every difficulty. The professional move is to recognise when the problem has moved outside the tutor’s scope and coordinate with the setting.
This boundary protects both learner and tutor. It prevents a supportive relationship from becoming an endless promise that more of the same will eventually work.
The learner’s weekly self-review
Once a week, the resit learner should answer four questions without the tutor supplying the language.
What can I now do that I could not do a week ago? Which error still costs me marks? What did I avoid or miss this week? What is the next action that would make the largest difference?
These questions keep tutoring from becoming something done to the learner. The goal is a growing internal model of performance.
A strong tutor can compare the learner’s self-review with the evidence. If the learner thinks timing is the problem but untimed accuracy is still weak, that mismatch becomes the next conversation. If the learner accurately identifies the bottleneck and proposes a sensible repair, self-regulation itself is improving.
Why the final grade should not erase the learning evidence
A resit result is important, but any single examination contains uncertainty. A learner can improve substantially and still miss a threshold, or scrape a pass while important weaknesses remain.
The programme should therefore preserve independent performance evidence alongside the final grade. That makes evaluation more honest and gives the learner a clearer account of what actually changed.
The qualification matters. So does the capability built while pursuing it.
One final resit rule
A second attempt deserves a different system. If attendance, diagnosis, practice and feedback remain unchanged, the programme is asking the learner to repeat the conditions that already failed. Tutoring earns its place when it changes those conditions in a way the learner can feel and demonstrate.
The tutor should therefore finish each week by asking not only whether more content was covered, but whether the learner now owns a stronger process.