The Tutor Handbook · Volume 0180 · Series ID THB-0180
The Tutor Handbook: Complete Series Index
A lesson can drown in everything the learner once knew
A tutor opens a lesson with review. Five minutes becomes twelve. Twelve becomes twenty-five. There is last week’s algebra, the vocabulary list from August, a graph-reading weakness from the previous term, three formulas that may matter next month, an old error that once cost marks, and a parent request to “make sure she doesn’t forget anything”. Every item has a defensible reason to be there. Together they consume the lesson.
The opposite failure is easier to miss. The tutor moves relentlessly forward. A topic is taught, practised, checked, then left behind. The learner looks strong while the work is fresh. Weeks later, a prerequisite disappears just as a harder topic begins to depend on it. The tutor then has to repair under pressure what could have been kept alive with a smaller earlier return.
Both failures come from the same mistaken idea: that “review” is a single activity rather than a selection problem. The tutor cannot revisit everything equally. They need to decide what deserves to come back, when, in what form, for how long, and what the result will change.
This is the Cumulative-Review Selection Gate. It asks how a tutor chooses a small portfolio of past knowledge and skills to revisit so review protects future learning without turning every lesson into a museum of old work.
Quick answer
Choose review items because they are likely to matter again, because their stability is uncertain, because they are prerequisites for imminent work, or because the cost of unnoticed forgetting is high. Keep review short enough that current learning still has room. Use review as evidence: a successful return can reduce special monitoring; a weak return can trigger a narrower check, reteach or repair. Do not include an item merely because it was taught before, because it appears on a checklist, or because the tutor is anxious about forgetting.
A good cumulative review is selective, spaced, mixed enough to reveal retrieval and method choice, and responsive to what the learner actually shows. It is not a second syllabus running in parallel with the first.
Why apex guidance makes this gap visible
The Australian Education Research Organisation’s current Revisit and Review practice guide, updated in May 2026, recommends regular opportunities to return to prior learning. Its related classroom examples emphasise regular, short, appropriately paced reviews, high participation and careful selection of knowledge and skills that will maximise learning progress. AERO’s examples are school-classroom guidance, not proof of one optimum routine for three-student tuition. They nevertheless expose an important design question: if review is supposed to be regular and short, something must decide what gets included and what stays out.
The same guidance warns indirectly against review taking over. AERO’s practice-guide materials note that daily review should not occupy such a large share of a lesson that new learning is displaced. This matters even more in tuition, where the learner may have only one or two sessions a week and the tutor is balancing school alignment, current repair and independent practice.
The advanced problem is therefore not “Should tutors review?” It is “How should a tutor allocate scarce review space across a growing history of learned material?”
Review has at least four different jobs
The first job is reactivation. Old knowledge is needed now, and bringing it back makes the current lesson intelligible. A fraction operation before algebraic fractions, a sentence-function distinction before editing, or a force concept before a new mechanics problem may belong here.
The second is maintenance. A capability was previously established and matters enough to keep alive. The existing Maintenance volume owns the question of preserving a confirmed gain without reopening full repair. This article addresses the portfolio problem around it: when there are many possible gains to maintain, which ones should occupy today’s limited review window?
The third job is monitoring. The tutor does not know whether an old capability has held. A small fresh return provides information. This is not the same as reteaching. If the learner succeeds independently, the item may leave the active review queue for longer. If the learner fails in a meaningful way, the tutor decides whether to sample again, connect the missing prerequisite, or reopen diagnosis.
The fourth is integration. Past knowledge is mixed with newer knowledge so the learner has to choose, connect and apply rather than respond to a single obvious cue. Integration can prepare the learner for real curriculum and examination conditions, but it should arrive only when the component skills are sufficiently stable to make the mixture interpretable.
These jobs can use similar-looking questions while serving different purposes. Declaring the job changes what a result means.
The four reasons an old item earns review space
An old item earns scarce review space most strongly when one or more of four conditions applies. First, it is a dependency: upcoming learning will fail if this knowledge cannot be retrieved or used. Second, it is fragile: recent evidence suggests the capability may not survive delay, variation or reduced support. Third, it is high-cost: unnoticed loss would create a large downstream problem relative to the small cost of checking it. Fourth, it is low-opportunity: ordinary schoolwork is unlikely to bring it back naturally soon, so tuition may be one of the few opportunities to keep it accessible.
These conditions help solve a common puzzle. Two equally important topics do not necessarily need equal review. If one recurs naturally every week in current work while the other will disappear for two months and then return as a prerequisite, the second may deserve more deliberate review even if both were learned equally well.
Composite case: Beatrice’s review list has become a second curriculum
This case is fictional and constructed for teaching. Beatrice is a Secondary learner whose tutor keeps an impressive review spreadsheet. Every topic is colour coded. Items move from weekly to fortnightly to monthly review. Nothing is ever deleted because “we might need it again”. After six months, the first thirty minutes of a ninety-minute lesson are consumed by old material.
The system looks rigorous. It is also beginning to harm the current route. New work is rushed. Beatrice learns that the beginning of each lesson is a long sequence of familiar questions, so she becomes fast at retrieving the review format rather than at choosing among methods under less predictable conditions. The tutor sees high review accuracy and interprets it as evidence that the review system is working.
The better move is not to abandon review. It is to apply a selection gate. Topics that recur naturally in current school work can leave the special queue unless evidence says otherwise. Stable low-cost skills can move to longer intervals. High-value dependencies for the next unit can temporarily move closer. One old fragility can receive a fresh, unfamiliar probe rather than five rehearsed examples. The review window shrinks, but its information value rises.
A queue needs an exit rule
Review systems grow because adding an item is easy and removing one feels risky. The tutor remembers the pain of the original weakness and thinks, “What if it comes back?” Without an exit rule, the queue becomes a permanent archive of past anxiety.
An item can leave frequent review when repeated evidence shows that it survives the conditions that matter and ordinary curriculum use is likely to exercise it enough. Leaving the active queue does not mean deleting the history. It means the capability no longer deserves scarce special attention. The Retirement volume addresses that transition for known fragilities. In cumulative review, the practical implication is simple: if nothing can graduate, review cannot remain selective.
Do not review only what is easy to ask
Another distortion appears when the tutor selects items by convenience. Vocabulary definitions, formula recall and short computations fit neatly into five-minute warm-ups. Multi-step planning, interpretation, explanation quality and method selection are harder to sample quickly, so they disappear from review even when they matter more.
A review window should not become a test of whatever can fit on a flashcard. Sometimes one richer item is more useful than ten rapid ones. If the target is choosing a method, present a problem that does not announce the method. If the target is explanation, ask for the causal connection rather than the keyword. If the target is writing control, use a short paragraph that requires the learner to revise meaning rather than circle an isolated grammar answer.
The duration can stay short while the evidence becomes more representative.
Recent learning and old learning should not have identical schedules
Freshly taught content often needs closer return because the learner has had little opportunity to stabilise it. Older, well-established content may need only occasional sampling. Yet age alone is not enough. A recently taught low-dependency enrichment point may deserve less review than an older algebraic manipulation that supports half the next term.
The tutor should therefore combine time since learning with dependency, evidence of stability and natural recurrence. A calendar can remind the tutor to look again, but it should not decide that every topic follows the same forgetting timetable.
What to do with an unexpected failure
A cumulative review is valuable precisely because it can surprise the tutor. The learner who was reliable three weeks ago may now hesitate. But one weak return does not automatically mean the topic has “been forgotten” or requires full reteaching.
First identify what failed. Was the fact unavailable, or was the learner unable to recognise when it applied? Did the representation change? Was the item more complex than the original target? Was there a copying or language issue? Was the previous success heavily scaffolded? A second small probe can often distinguish these possibilities better than immediately reopening the whole chapter.
If the capability genuinely weakened, the tutor can move it from maintenance back into a more active route. The existing Reopening and Second Repair volumes cover that larger transition. The review itself should remain an early sensor, not become a hidden repair session every time something goes wrong.
Cumulative review in a three-learner tutorial
Three learners create a special selection problem because their review needs will never be identical. If the tutor builds the whole review around the weakest historical item of each learner, the group can fragment into three private lessons. If the tutor gives everyone the same review, much of the time may be irrelevant.
A practical solution is to separate shared dependencies from individual maintenance. One short common task can sample a capability everybody will need for today’s lesson. Then each learner may receive one brief individual return tied to their own evidence. The tutor can rotate those individual items rather than forcing all differences into every lesson.
This is not mechanical equality. One learner may need no individual maintenance that day. Another may need a single prerequisite check before the group begins. A third may need a delayed transfer item because the tutor is deciding whether a previous repair can finally leave active monitoring.
The review item should sometimes change form
If every return uses the same wording, layout and sequence, the learner can become good at recognising the review rather than using the knowledge. Variation should be meaningful but controlled. The tutor can change the numbers, context, representation or neighbouring distractors while keeping the underlying target stable enough to interpret.
The Example-Variation Gate explains how to vary examples without changing the target into another skill. In cumulative review, that principle protects against a false sense of stability from repeated familiar forms.
Do not use review to punish forgetting
A learner who cannot retrieve an old idea has provided information, not committed an offence. Review becomes counterproductive when every miss produces frustration, public comparison or a lecture about “we already did this”. The tutor’s job is to find the smallest useful interpretation and next action.
This matters because review deliberately creates conditions in which imperfect retrieval can become visible. If the learner learns that a miss will trigger embarrassment, they may guess, copy a peer, avoid responding or over-rehearse before tuition. That damages the evidence the review was meant to provide.
The current lesson still has a claim on time
The Learning Budget makes the broader point that learner time and attention are finite. Cumulative review must pay rent inside that budget. Its value is not measured by the number of old topics touched. It is measured by how much future learning, retention and route clarity the selected returns protect.
When a review item repeatedly consumes time without changing any decision, ask why it is still there. Perhaps the learner is already stable and the item can move farther apart. Perhaps the question is too easy to reveal anything. Perhaps the capability recurs naturally elsewhere. Perhaps the tutor is maintaining the routine because removing it feels unsafe. These are different problems, and adding more review does not solve them.
Composite case: Denise fails the old item but succeeds in the new lesson
This case is fictional. Denise’s tutor opens with a percentage problem from six weeks ago. Denise hesitates and makes an arithmetic error. The tutor is ready to move percentages back into repair. Ten minutes later, the new lesson involves comparing percentage changes in a data context. Denise chooses the correct base, explains the comparison accurately and completes the arithmetic without help.
The two pieces of evidence conflict only if the tutor treats the opening item as a verdict. Perhaps the first error was a slip. Perhaps the context later activated the relevant relationship. Perhaps the isolated warm-up used a representation Denise had not seen recently. The correct move is a narrower follow-up, not an automatic retreat through the unit.
A review queue is useful when it creates questions like this. It becomes harmful when its results are interpreted more strongly than the task supports.
A selection rule the tutor can actually use
At the end of a lesson, do not simply add every corrected error to future review. Ask: Will this capability matter again? Is its stability uncertain? Is ordinary upcoming work likely to exercise it anyway? What would a future miss cost? What is the smallest later check that would tell me something useful?
Before the next lesson, select only the few items with the highest current value. A shared prerequisite may come first. One recently repaired fragility may deserve a delayed return. One older high-cost dependency may be sampled because it has been absent from natural practice. The rest can wait.
After the review, act on the evidence. Stable items move out or farther apart. Ambiguous misses receive a discriminating follow-up. Genuine weakness receives the appropriate repair. Review is a queue with movement, not a permanent collection.
A worked selection example across one month
Consider a fictional learner, Emily, who is beginning a month of Secondary Mathematics work on algebraic fractions and simultaneous equations. Her tutor has twenty-seven plausible review items from the previous term. A completeness-first system would sample all of them on rotation. A selection-first system starts with dependencies. Integer sign control, common factors, fraction operations and linear-equation manipulation directly support the coming work. These four deserve closer attention than, for example, an older geometry construction that is currently stable and will not be used this month.
The tutor then checks fragility. Sign control was repaired recently and has held twice, but not after a long delay. It earns one fresh item. Fraction operations are stable in routine form but have not been seen inside algebraic expressions, so the tutor chooses one transfer item rather than five arithmetic examples. Linear equations recur naturally in the new unit, so they may not need a separate warm-up at all; current work will provide plenty of evidence. Common factoring is both a dependency and relatively low-frequency in Emily’s school homework, so it receives a short deliberate return.
That first lesson may therefore use three carefully chosen review prompts, not twenty-seven. In week two, the sign item is successful under changed numbers and no cue. It moves farther out. Factoring is weak in a new representation, so it stays active. In week three, algebraic fractions themselves provide natural retrieval of fraction operations; no separate arithmetic review is needed. By week four, the queue looks different because the learner’s evidence changed. The timetable follows the route rather than controlling it.
Review should distinguish retrieval from reconstruction
Sometimes the learner cannot produce an old answer immediately but can rebuild it from more fundamental knowledge. That is different from having no route at all. A Science learner may not recall an exact phrasing but can reconstruct the mechanism from conserved quantities and the conditions in the question. An English learner may not remember a named term but can identify the sentence relationship accurately from meaning. A Mathematics learner may forget a shortcut but derive the result from a secure representation.
The tutor should decide whether the target requires fast retrieval, reconstructive reasoning, or both. If rapid recall is genuinely load-bearing for later work, reconstruction alone may signal that more fluency is useful. If the curriculum requires reasoning rather than a memorised label, a slower reconstruction may be entirely acceptable. Review becomes misleading when every hesitation is treated as forgetting and every quick answer as mastery.
Review before examinations is not the same as cumulative review during learning
Near an examination, the portfolio may broaden because the learner must be ready for a wider range of content under less predictable selection. That does not mean the tutor should run examination-style broad review throughout the year. During ordinary learning, review can be more selective and dependency-led. Examination preparation has a different job: reintegration across a larger assessed domain, method selection under uncertainty, timing and reliable execution. Those performance demands belong partly to the existing Examination Craft owner.
This boundary matters because “the exam covers everything” can otherwise become an excuse for perpetual everything-review. A topic’s future assessability is one factor, not the only factor. The tutor should still ask whether today’s review of that topic will change readiness more than the competing use of the same time.
The review queue can reveal curriculum dependencies
Over time, a well-run review queue teaches the tutor something about the curriculum itself. Some capabilities repeatedly support later work. Others are relatively self-contained. Some are stable once established; others need periodic reactivation because they occur in bursts. The tutor can use these patterns to improve future sequencing without turning them into fixed rules for every learner.
For example, if several learners repeatedly struggle when algebraic manipulation returns after a gap, the programme may need to inspect whether ordinary materials provide enough later recurrence. That is a programme-design question, not evidence that every learner has the same weakness. Conversely, if an item remains on every tutor’s review list but almost never changes a decision, it may be occupying space because of tradition rather than evidence.
The learner should gradually take over some selection
As learners become more capable, the tutor can involve them in proposing review items. The learner might say, “I can still do this method, but I keep forgetting which condition tells me to use it,” or “I have not seen this vocabulary in weeks and I am not sure I can use it in writing.” That information is useful, but it should not become the only sample. Learners may choose comfortable material or misjudge what matters.
A mature arrangement combines learner agency with tutor responsibility. The learner can nominate a concern and explain why it matters. The tutor can add a dependency the learner has not noticed. Over time, the student learns not merely to “revise old topics” but to ask which knowledge is load-bearing, uncertain or under-exercised. That is a valuable studying capability, even though this article remains a tutor-side decision guide rather than a general study-skills owner.
Failure modes
The completeness trap: the tutor tries to keep every past topic alive equally. The result is broad but shallow contact that crowds out current learning.
The comfort trap: review contains only items the learner can answer quickly, producing a pleasant accuracy score but little information about fragile or integrative knowledge.
The recency trap: the tutor reviews only last week’s content and calls the routine cumulative. Older dependencies quietly disappear because they are not on the immediate planning horizon.
The error-collection trap: every mistake becomes a permanent review item, even when the cause was a one-off slip or the skill already receives abundant natural practice.
The ritual trap: review continues because “we always start this way”, even after the questions stop affecting teaching decisions.
The support trap: the tutor supplies so many cues during review that the learner appears to remember. If the purpose is monitoring independent availability, answer-giving support invalidates the evidence. Legitimate access supports, however, should remain when they are not the target being tested.
Parents: what useful review looks like from home
Parents often worry when a child forgets something that was once learned. The right response is not necessarily to request more revision of everything. A better question is whether the forgotten idea is important, whether it is actually unstable, and whether it will matter for upcoming work.
A tutor should be able to explain why a particular old skill is returning. “This fraction operation is a dependency for algebra next week” is stronger than “We are revising because revision is good.” “I am sampling this writing move after a three-week gap to see whether the improvement held” is stronger than “We always do corrections.” Selective review should make the learning route clearer, not more crowded.
What the evidence can and cannot say
Research on retrieval, spacing and review supports returning to prior learning, but it does not provide one universal timetable for every school subject, learner or tuition format. AERO’s 2026 guidance is research-informed professional guidance developed for school teaching. It is useful for design principles—regular return, careful selection, short pacing, participation and using review to guide next instruction—but it does not validate a fixed five-minute or ten-minute prescription for Singapore tuition.
The tutor should therefore avoid pretending that a calendar is a scientific measurement of memory. The schedule is a planning aid. The learner’s later performance is the evidence that updates it.
One final discipline helps: review should preserve legitimate access conditions. If a learner normally uses an approved visual aid, enlarged text or another access support that does not supply the target answer, removing it merely to make review feel harder can change what is being tested. By contrast, a tutor hint that reveals the method may invalidate a check intended to test independent selection. The useful question is not whether all help disappears, but whether the support being removed is part of the capability the tutor is trying to verify.
Sources and further reading
The Australian Education Research Organisation’s Revisit and Review practice guide was published in November 2024 and updated 14 May 2026. Its Primary and Secondary practice video, updated 23 June 2026, illustrates regular short reviews, careful content selection and high participation. These resources provide the main research-informed teaching guidance used here.
AERO’s Monitor Progress guide, updated 14 May 2026, is relevant because review is not merely repetition; learner responses can inform additional instruction, guidance and feedback. The translation from school classrooms to a three-learner tuition setting in this article is a proposed professional application, not a directly tested tutoring protocol.
The point of review is not to keep the past busy
Cumulative review works best when it protects the future. It brings back the knowledge that current learning depends on, tests the fragilities that still matter, keeps high-value capabilities available and lets stable material move into ordinary life rather than permanent special monitoring.
The tutor’s skill is therefore not remembering to review. It is deciding what deserves another return.
When that decision is made well, review becomes smaller and more powerful at the same time. Old learning comes back because it has a job. New learning still has room to grow. And the learner gradually carries a curriculum that is connected rather than merely completed.
