Bukit Timah Math tuition in a three-student small group is useful when the smaller class changes the teaching, not merely the seating plan. Parents comparing small-group Mathematics tuition, Secondary 1 to Secondary 4 support, E-Math, Additional Mathematics, IP or IB lessons need to look beyond the phrase personalised attention. What matters is whether the tutor can see each learner’s first attempt, identify the particular decision that needs teaching and check that the student can use the explanation independently afterwards.

A well-designed 3-pax Mathematics class in Bukit Timah combines individual observation with carefully managed peer discussion. One student may draw a diagram, another form an equation and a third notice that an apparently sensible answer violates a condition. Those differences can become useful teaching material. They can also conceal dependence if the quickest student supplies every method before the others have thought. Small-group Maths tuition therefore needs a deliberate sequence: individual thinking, justified teaching, purposeful comparison and a fresh individual check.

This guide explains why small-group Math tuition can work, when three-student classes fit, and how parents can judge the actual learning. It includes original worked examples in number, algebra, ratio, graphs, geometry, probability and appropriate A-Math connections. The central idea is simple: a small group earns its value when it makes each student’s Mathematics more visible and progressively gives the important decisions back to that student. Three seats alone cannot guarantee a distinction, a fixed grade improvement or an ideal match.

eduKate Singapore’s established small-group format is a maximum of three students with standard 1.5-hour lessons, subject to curriculum fit and availability. This page explains the format; current fees, teacher arrangements and available times must be confirmed directly. Alicia, Tricia and Kai Kai are fictional learners in the examples, not identifiable pupils or testimonials. Every proposed task and lesson sequence is an original teaching illustration, not an official assessment or a measured claim about results.

Your 50-second route through the guide

A small class should make it easier to see where the student needs help, but the help still has to be appropriate. Start with the teaching mechanism and what research does and does not show. To see the Mathematics, open the ratio lesson, equation lesson or probability lesson. For a family decision, use group compatibility and the parent observation guide. You need one relevant next action, not a requirement to finish the whole article.

Route 1: My child gets lost in a larger class

Look for an independent beginning, a visible point of uncertainty and a tutor response that fits that uncertainty. Read the chapters on observation and signed arithmetic. A smaller room is helpful only when the learner no longer disappears inside a shared explanation.

Route 2: My child follows friends instead of choosing a method

Use the private-start and peer-discussion chapters. A written first step before discussion and a changed individual question afterwards can separate useful collaboration from borrowing another student’s recognition.

Route 3: My child is strong and needs more than routine practice

Read the quadratic, geometry and proof comparisons. Challenge can mean explaining a condition, constructing an example or choosing between valid routes. It does not always mean importing material from a different syllabus.

Route 4: We are comparing a small group with individual tuition

Begin with compatibility, independent working intervals and the parent observation guide. A learner needing continuous explanation or a specialised course sequence may need another arrangement. No tuition is also a reasonable outcome when existing support is sufficient.

Route 5: We need to know whether the lessons are working

Examine fresh work completed with less help, not only corrected answers. Use the transfer and parent-review sections, and keep the student’s course, task difficulty, prior exposure and time conditions visible when comparing results.

Open the main contents

1. Three is a teaching condition · 2. Research boundaries · 3. The private start · 4. What the tutor observes · 5. Compatible, not identical · 6. Course boundaries · 7. Signed numbers · 8. Fractions · 9. Ratios and changing totals · 10. Percentage bases · 11. Equations and equality · 12. Graphs and fixed costs · 13. Similarity and scale · 14. Probability and sample spaces · 15. Quadratic representations · 16. A-Math connections · 17. Purposeful peer discussion · 18. A possible 90-minute lesson · 19. The independent return · 20. What parents can inspect.

For choosing a tutor rather than understanding the class format, use How to Choose the Right Mathematical Support. For the wider library, return to the Bukit Timah Mathematics Master Gateway. This article remains focused on the work that a three-student group makes possible and the conditions that can stop that work from happening.

1. Three is a teaching condition, not a mathematical guarantee

Imagine three students copying a complete solution from a board. The tutor speaks clearly, the pages fill up and everyone reaches the same answer. The class is small, but the central decisions may still belong entirely to the tutor. Now imagine the same three students first writing how they would begin. Their different starts reveal that one misread the quantity, one chose the wrong relationship and one made a sign error. The headcount is unchanged. The information available for teaching is not.

The useful mechanism is a sequence of opportunities. A small group can make it easier to inspect an attempt, ask a precise question, hear the explanation and return later to see whether it was used. These opportunities are not outcomes by themselves. They have to lead to an appropriate intervention. A tutor who notices a wrong equation but responds only with another worksheet has observed without yet teaching the missing relationship. A tutor who explains everything again may obscure what the learner already understood.

Peer comparison adds something different from individual attention. When Alicia uses a bar representation and Tricia uses an equation, the class can examine what each represents and why both reach the same result. Kai Kai may notice that the two methods express the same equality. This is useful only when the connection is made explicit. Simply displaying several solutions can become another performance by confident students while the quieter learner still cannot decide which route to use independently.

The format also creates a practical requirement: each student needs something worthwhile to do when the tutor attends to another learner. That independent interval should be designed, not left to chance. It might contain a fresh application, a worked example to explain or a short task already within reach. Giving an unprepared student an unrelated difficult problem is not independence training. Giving a secure student endless copies of a familiar exercise is not meaningful challenge. The tutor must match the interval to the learner.

Three students do not receive thirty minutes of identical individual tuition each simply because a lesson lasts ninety minutes. Shared explanation, individual observation and peer reasoning can serve several learners simultaneously, while different support needs consume different amounts of attention. Dividing time by headcount therefore does not describe learning quality. The better question is what mathematical decision each learner had an opportunity to make, how the tutor responded and what a later unaided attempt showed.

A small group earns its place between private teaching and a larger class when both sides of that design work. Individual thinking remains visible, and shared thinking adds useful comparison. Either can fail. The group can become three disconnected private lessons with long waiting periods, or it can become a miniature lecture in which the individual disappears. Parents should inspect the work and the teaching response rather than assume that a small number settles the educational question.

2. Use tutoring research without turning averages into promises

The Education Endowment Foundation’s small-group tuition review defines its category as one educator working with two to five pupils. It emphasises targeting particular learning needs and linking support to classroom content. It also warns that teaching quality and group composition matter. Its average findings do not establish that every three-student Mathematics class is effective, that three is always better than two or four, or that a particular child will make a specified gain.

The National Student Support Accelerator’s design brief describes high-impact tutoring as a package including small ratios, consistent tutors, curriculum alignment, progress monitoring and frequent sessions. Its description includes at least three sessions a week over a sustained period. A once-weekly 90-minute private lesson does not become the same intervention merely because it shares a three-student ratio. Frequency, duration, setting and implementation must remain part of the comparison.

The 2024 research synthesis by Nickow, Oreopoulos and Quan examines experimental tutoring evidence across varied programmes and contexts. It supports taking tutoring seriously as an educational intervention, while also studying differences among implementations. It is not an evaluation of this eduKate class. The defensible use of such research is to ask better design questions, not to borrow an average effect and present it as a local success rate.

There is a useful distinction between a supported principle and a tested programme. A principle might suggest attending to the learner’s current gaps or using fresh checks. A local programme then has to show that it implements those ideas coherently. Its own evidence should include what was taught, what support was used and what the student subsequently did. An attractive label such as evidence-based does not eliminate that obligation. The classroom examples in this guide are proposed applications, not experiments establishing a causal effect.

Be especially careful with claims measured in months, percentages or examination grades. A research average compares particular groups under specified study conditions. It cannot simply be converted into the number of lessons a family should buy or the number of marks one learner will gain. Even a genuine local improvement can have several contributing causes, including school teaching and independent work. A trustworthy discussion can recognise improvement without claiming that one component alone caused all of it.

For the parent, the practical result is not cynicism about research. It is a better way to use it. Ask whether the teaching targets an observed need, whether the group fits the actual course, whether the same tutor can maintain continuity and whether independent work is reviewed. Then inspect the child’s work over time. Research can inform the design; current learner evidence should guide the next teaching decision. Neither needs to be exaggerated for the other to matter.

3. Protect the first attempt before anyone announces the method

The first unaided attempt is one of the most informative parts of a small-group lesson. Once a classmate announces that a question uses simultaneous equations, everyone else’s work has changed conditions. A learner who then solves accurately may have demonstrated excellent execution but not independent recognition. That is useful information, provided the tutor does not confuse the two. A brief private start preserves the decision that would otherwise be supplied invisibly through discussion.

A private start need not mean complete silence for a long examination. Ask each student to write the unknown, one relationship and a possible route. For a suitable geometry problem, a labelled sketch may be enough. For an expression, the learner might identify the main operation or factor. The task should be small enough that the tutor can inspect it promptly. Its purpose is to reveal entry into the Mathematics, not make students struggle through unfamiliar content before any teaching is allowed.

Suppose the question says that two numbers have sum 23 and difference 7. Alicia writes a + b = 23 and a − b = 7. Tricia reasons that removing the difference leaves two equal smaller numbers, so the smaller is 8 and the larger 15. Kai Kai writes 23 − 7 = 16 and stops. The class now has useful differences to discuss. Kai Kai may need to see why sixteen represents two copies of the smaller number, not merely be given the remaining division.

The tutor should distinguish productive uncertainty from absence of a prerequisite. A learner drawing two bars and adjusting their lengths is doing relevant work. A learner copying the sentence repeatedly without identifying the quantities may need a prompt. There is no universal period of silence that proves good teaching. The tutor’s judgment should follow what the student is doing and what information is missing. Direct explanation remains appropriate when the relationship has not yet been understood.

After discussion, use a changed individual task. A new sum and difference can check the immediate relationship; a perimeter-and-difference context can add a representation demand later. Record what help was used. A copied correction answers a different question from a fresh unaided solution. Both belong in learning, but only the second checks whether the entry decision has transferred under those conditions. This simple separation keeps the group from producing a misleading appearance of equal readiness.

The private start also gives every learner something to contribute. A student does not have to win a race to the complete answer in order to have a mathematical idea worth examining. The tutor can select a useful representation, a sensible partial step or an instructive error. Participation becomes connected to reasoning rather than speed. Over time, the student should become more willing to commit to an inspectable first step because that step is treated as information, not as a verdict on ability.

4. Observe the decision sequence, not only the final answer

A wrong final answer can sit several steps downstream from its cause. The tutor should inspect what was asked, how the learner represented it, which method was chosen, how the transformations were executed and what the final quantity was taken to mean. These are practical reading questions, not a clinical diagnostic framework. They help prevent one broad response, such as more practice, from being applied to mathematically different problems.

For example, a student asked for the area of a rectangle with sides x and x + 3 may form x(x + 3) correctly but expand it as x² + 3. Another may add the sides because they have confused area with perimeter. A third may form and solve the equation correctly but retain a negative geometric length. All may lose the final mark, yet the next explanation should be different. The first needs multiplication structure, the second the measured quantity and the third interpretation of admissible values.

Written working is useful because it leaves a trace. Verbal explanation can clarify an ambiguous step, but should not replace every written attempt. A fluent spoken account after the tutor’s explanation may not show what the learner could originally construct. Likewise, a terse correct solution can conceal excellent understanding or a lucky guess. Compare the work with a short question and a fresh example. No single signal needs to carry the whole judgment.

Preserve correct decisions explicitly. If the learner formed the model correctly and only mishandled a coefficient, say so. Otherwise, correction can accidentally teach the student to distrust a valid method. A local repair should remain local when the evidence supports it. When the first equation is wrong, explain why later accurate calculations are solving another problem. The scope of the error determines how much work should be rebuilt.

The tutor should also notice assistance. A topic heading, a visible worked example or a peer’s comment may supply the recognition step. A small hint can restore a forgotten relationship. That does not make the resulting solution worthless. It tells the tutor which capability is available with that support and which decision still needs an independent check. A useful record names the help rather than merely classifying the whole answer as right or wrong.

The outcome of observation is a next task. The tutor might isolate signed multiplication, compare two representations, revisit an old method or use a timed section to inspect repeated checking. A label without a changed task has limited teaching value. A small group should make it easier to gather this evidence and act on it during the same lesson, while leaving enough room for the learner to attempt the relevant decision again without immediate rescue.

5. A compatible group shares useful work, not an identical profile

Students do not need identical marks, personalities or school names to learn together. They need enough common mathematical content for shared teaching to be useful, and enough capacity for independent work that individual attention can move among them. A group of three with incompatible courses can be less coherent than a larger well-matched class. Placement should therefore consider the course, current sequence, support demand and the kind of work the student needs now.

One useful group might share linear equations while differing in how much support is needed. Alicia may need to retain signed brackets, Tricia may need to explain equality and Kai Kai may need to form the equation from words. A common example can reveal all three issues, followed by differentiated tasks. The shared topic gives the lesson coherence. The different follow-ups make the support appropriate. Neither requires pretending that all learners are at the same stage.

Another grouping may be less suitable. One student needs continuous teaching of fractions, another is learning trigonometric identities and another needs a complete timed paper. The tutor may be able to coordinate part of their work, but should not assume that a low headcount makes every combination practical. Sustained waiting or repeated unrelated explanations are signs to review the arrangement. That is a fit decision, not a judgment that a child is difficult or unsuitable for learning.

The ability to work independently for short intervals should be assessed in context. A student may manage a familiar consolidation task alone but need substantial help with a new concept. That is normal variation in task readiness, not a permanent independence label. The tutor can prepare an appropriate task for the interval while another learner receives explanation. The key question is whether the task is within reach and whether the tutor returns to inspect what happened.

Review compatibility as school demands change. A group that shares algebra this month may diverge later because schools teach topics in different orders. A learner may develop a new need for sustained prerequisite repair, or become ready for a more independent format. Placement should not be treated as a one-time decision that remains correct indefinitely. The family should understand what evidence would trigger a review and how the student’s current work would be carried into another arrangement.

There are legitimate alternatives. Some learners need individual teaching for a period, a different group, a school consultation or no additional tuition. A useful support system permits these outcomes. The purpose of assessing fit is not to find a justification for filling an available seat. It is to decide whether the proposed teaching conditions can address the learner’s actual mathematical need while keeping the wider workload manageable.

6. Keep G1, G2, G3, A-Math, IP and IB boundaries clear

The student’s school year, subject level and examination route are different pieces of information. A Secondary 3 learner may be taking Mathematics and Additional Mathematics under different subject-level arrangements. An IP learner may follow a school-specific sequence. An IB learner needs the actual programme and course identified. A small group should not replace these distinctions with one worksheet bank labelled advanced Mathematics. The course defines required scope; the student’s work defines the support needed within it.

MOE’s Full Subject-Based Banding information explains the move to subject-level learning, and SEAB states that the SEC begins in 2027. For that examination year, Mathematics is listed as G1 K110, G2 K210 and G3 K310. The latter two listings separately identify Additional Mathematics as K232 and K341. Use the document for the actual examination year.

A shared prerequisite can be taught across routes without implying that the full syllabuses are the same. Several students may benefit from seeing why a signed coefficient distributes across a bracket. Their later application tasks can then differ in scope and complexity. Label extension as extension. A learner should not be told that compulsory preparation is deficient because they have not learned a topic belonging to another course. Equally, a demanding course label should not make a necessary foundational explanation embarrassing.

For IP, begin with the school’s actual material and intended qualification. Programme architecture is not a simple ladder placed above G3. As one concrete example, ACS Independent describes an IP leading to the IB Diploma. That example should not be generalised to every IP school. The tutor needs to know the learner’s present sequence and requirements rather than infer them from a broad label or a neighbouring school’s practice.

For the IB Diploma, the IB Mathematics overview distinguishes Analysis and Approaches from Applications and Interpretation, with Standard and Higher Level courses. The IB also describes a curriculum transition to first teaching in 2027 and first assessment in 2029. That future edition should not be silently substituted for a student’s current assessed course. School-provided course information remains essential for choosing topics, technology use and appropriate assessment preparation.

The examples that follow sample mathematical teaching decisions; they are not a promise that every task belongs to every level. Select only material appropriate to what the learner has been taught. For detailed route guidance, use the existing Bukit Timah Mathematics pathways page. A well-run small group keeps both ambitions intact: serious reasoning for every learner, and accurate boundaries around the actual course.

7. Signed numbers: three wrong answers need not mean one misconception

Give the group the expression −4 − 3(2 − 5). Ask for an individual attempt before discussion. The correct result is 5: the inner bracket is −3, multiplication gives −9, and subtracting −9 from −4 adds nine. The task is short, but it contains different uses of the minus sign. One belongs to a negative number, another indicates subtraction and another appears within a bracket. Reading those roles carefully matters before applying a remembered slogan about two negatives.

Alicia calculates 2 − 5 as 3. Her first difficulty is subtraction order inside the bracket. A number line or a change-from-two interpretation can clarify the direction. Tricia calculates the bracket correctly but writes −4 − 9. Her error concerns the signed product or the subtraction of that product. Kai Kai reaches 5 after hearing someone say two negatives make a positive but cannot explain which two operations they mean. The same task calls for different questions.

Use close numerical contrasts rather than a long collection of arbitrary signed calculations. Compare −4 − (−9), −4 + (−9) and (−4)(−9). Their results are 5, −13 and 36. The visual presence of negative signs does not make the operations interchangeable. Ask the learners to name the operation before calculating. A multiplication rule cannot be applied indiscriminately to addition. This is the relationship the lesson should establish, not merely a list of correct numerical outputs.

Now connect the same idea to algebra: simplify 5 − 2(3 − x). The result is 2x − 1. A learner can first distribute the signed coefficient, writing 5 − 6 + 2x. Substituting x = 4 into the original gives 7, matching the simplified expression. That one input is a useful slip check, not a proof of an identity for every x. Distribution establishes the identity. The group can discuss what each kind of evidence contributes.

The peer task is to identify the first invalid line in an invented solution, not to identify which classmate is the weakest. Present the work anonymously or as a tutor-created example. One learner explains the numerical bracket, another the signed distribution, and the third chooses a check. Rotate those roles in later tasks. A student should not become permanently responsible for errors or permanently exempt from explanation because they usually finish first.

The fresh individual check is 7 − 3(2 − a), which simplifies to 1 + 3a. Ask each learner to explain the sign of the variable term. Later, place the same operation inside a line equation or a suitable derivative. The purpose is a signed transformation that survives a change of context. If it is secure in the short expression but fails in the larger one, the tutor has found a new teaching boundary rather than proof that the initial lesson achieved nothing.

8. Fractions: compare the unit, the operation and the answer’s size

Ask the group to find 2/3 + 3/4. A common denominator of twelve gives 8/12 + 9/12 = 17/12. Before calculating, the learners can predict that the sum exceeds one because two thirds and three quarters are each greater than one half. This rough bound is not the exact answer, but it can reject 5/7. A useful small-group discussion connects the size estimate, the equivalent fractions and the final improper fraction rather than treating them as separate tricks.

Alicia writes 5/7 by adding numerators and denominators. Ask what one seventh would measure relative to thirds and quarters. Tricia finds 17/12 correctly but changes it to 12/17 because she expects a fraction to be less than one. Kai Kai can repeat the common-denominator procedure but cannot explain why the numerator of two thirds becomes eight. Each learner needs a different connection: common units, admissible magnitude or equivalence under multiplying numerator and denominator by the same non-zero number.

A diagram can help when the same whole is divided consistently. Split a rectangle into twelve equal parts, then represent eight and nine twelfths. Do not use two diagrams with different-sized wholes and silently compare their shaded fractions as though the units matched. The mathematical object includes the reference whole. This observation becomes important in word problems where a fraction of one person’s collection is compared with a fraction of another’s. Equal fractions need not mean equal quantities when the wholes differ.

Move to 3/4 ÷ 1/8. The result is six because six eighths fit into three quarters. This interpretation explains the reciprocal procedure in this case. Contrast 3/4 × 1/8, which is 3/32 and asks for an eighth of three quarters. The presence of the same numbers does not determine the operation. Ask the learners to explain the question each calculation answers before using symbols. A familiar algorithm should remain tied to a relationship.

For students ready for algebraic fractions, compare 2/x + 3/x with 2/x + 3/y, keeping denominators non-zero. The first combines as 5/x. The second is (2y + 3x)/(xy). The common-unit idea remains, while the symbols make the conditions more visible. A learner who writes 5/(x + y) may be extending the earlier numerator-and-denominator addition error. The tutor can use the numerical case to expose the same invalid structure before reconnecting it to algebra.

End with a fresh task chosen for each learner’s current stage. One explains why 5/6 + 1/3 exceeds one; another solves a fraction equation; another constructs an incorrect simplification and disproves it with an allowed input. The group shares a central relationship while the independent demands differ. This is more purposeful than giving everyone the same long fraction sheet and calling the resulting headcount personalised teaching.

9. Ratio: the quantities matter more than a remembered number of parts

In an original problem, red and blue counters are in the ratio 3:5, with 64 counters altogether. There are eight equal ratio parts, so each part represents eight counters. Red is 24 and blue is 40. A learner can solve this through a bar representation, a unitary method or equations. The useful discussion is not which method is fashionable. It is how each represents the same relationship and total.

Alicia multiplies 64 by 3/5, treating the total as the blue quantity. Tricia obtains both quantities but labels 40 as red. Kai Kai knows to divide by eight only after the tutor says add the ratio parts. The first error concerns the reference quantity, the second order and the third independent interpretation. Asking all three to do more ratios would be a weak response unless the new tasks isolate these different decisions.

Now add sixteen red counters. The quantities become 40 and 40, so the new ratio is 1:1. The original ratio parts cannot simply have sixteen added to the first number because three was not the number of red counters. Ask what each symbol measures. This is an opportunity for the group to distinguish a ratio from an actual count. A change in quantities has to be represented at the quantity level or through a correctly defined scale factor.

Reverse the problem. Red and blue start in the ratio 3:5. Adding sixteen red counters makes the quantities equal. Let the initial quantities be 3k and 5k. Then 3k + 16 = 5k, so k = 8 and the original counts are 24 and 40. The relationship is familiar, but the total is no longer supplied. A student who understood the first task only as divide the total by eight needs help seeing the more general representation.

Introduce a contrast where counters are transferred rather than added from outside. Starting from 24 red and 40 blue, reclassifying or replacing eight blue counters as red would give 32 of each while keeping the total 64. Adding sixteen red from outside makes the total 80. The final ratio can be the same while the conservation condition differs. The tutor should make the situation explicit and avoid treating add, transfer and remove as interchangeable keywords.

The shared discussion can compare the bar model with 3k + 16 = 5k. Ask the algebra learner to identify the sixteen on the diagram and the diagram learner to explain the two-part difference. Then give a fresh individual version, such as a 2:7 ratio that becomes 1:2 after six are added to the smaller quantity. The equation 2k + 6 = 7k/2 gives k = 4, so the original quantities are 8 and 28. The check confirms 14:28 = 1:2.

10. Percentages: identify the base before comparing changes

An item marked at 80 units of currency is discounted by 25 percent. The reduction is 20 and the new price is 60. Now ask what percentage increase takes 60 back to 80. The increase of 20 is measured against 60, so it is one third, or 33⅓ percent. The numerical change is the same size as the earlier reduction, but the reference base has changed. This is a useful small-group lesson because several superficially plausible answers reveal different reasoning.

Alicia says 25 percent because she remembers the discount. Tricia calculates 20/80 correctly for the first change and reuses the same denominator for the second. Kai Kai writes 80/60 and obtains a multiplier but does not know how it relates to a percentage increase. The tutor should ask each learner to complete the sentence twenty is what fraction of which starting amount. The missing word is mathematical information, not a minor detail in the story.

Represent the changes multiplicatively. The discount multiplies by 0.75. Reversing it requires multiplication by 1/0.75 = 4/3, not by 1.25. This connects percentages to inverse operations. A ratio representation of old:new = 4:3 gives another route. Ask the learners to show how the multiplier, the ratio and the percentage statement express the same relationship. The class comparison should build connections rather than present three disconnected recipes.

Use a changed example: a quantity rises by 10 percent and then falls by 10 percent. Starting from 200 gives 220, then 198. In general, the multiplier is 1.1 × 0.9 = 0.99. The second ten percent is based on the increased amount. A student who subtracts the percentages before identifying their bases is combining unlike reference quantities. The original example is about arithmetic relationships, not a prediction about real prices or investment returns.

A deeper extension lets the percentage change be p, written as a decimal with 0 < p < 1. Increasing and then decreasing by that fraction gives multiplier (1 + p)(1 − p) = 1 − p². The connection to a difference of squares can be explained after the numerical meaning is secure. Do not introduce the parameter merely to make the task look advanced. It is useful because it shows what remains true when the chosen percentage changes.

The independent check asks for the original amount when a 20 percent reduction leaves 72. The equation 0.8x = 72 gives x = 90. Another learner might explain why adding 20 percent of 72 does not recover the original. Both tasks target the base. A later return can use a different context and wording, while keeping current financial products, taxes or policy rates out of the example unless they have been separately verified.

11. Equations: equality should survive every transformation

Solve 3(2x − 5) = 4x + 7. Expanding gives 6x − 15 = 4x + 7, so 2x = 22 and x = 11. Substitution makes both original sides 51. The task contains distribution, operations on both sides and a final check. The group should be able to explain why those operations preserve the set of solutions, not merely narrate symbols moving across an equals sign and changing signs by themselves.

Alicia writes 6x − 5 in the first expansion. Tricia expands correctly but changes −15 to −15 again when isolating terms, obtaining an incorrect constant. Kai Kai follows the tutor’s procedure but cannot say why subtracting 4x from both sides is allowed. The errors are not all the same. The tutor can use a numerical balance analogy, explicit distribution or a carefully labelled intermediate equation depending on what the learner’s attempt reveals.

A close contrast is 3(2x − 5) = 6x − 15. Here every real x satisfies the equation because the two expressions are identical. Another is 3(2x − 5) = 6x − 14, which has no solution. A learner who always expects one numerical value may keep manipulating after the variable disappears. The group discussion should distinguish identity, inconsistency and a single solution. The question is what equality permits, not whether the worksheet contains enough algebra to produce an x.

For learners ready for a quadratic contrast, solve x² = 7x. Dividing by x without considering zero loses one root. Factoring gives x(x − 7) = 0, so x = 0 or 7. The earlier operations of adding the same expression to both sides were reversible without excluding an input. Division by a possibly zero expression has an additional condition. A tutor should explain that difference rather than teach that moving or dividing symbols is always safe.

The peer activity can present two solutions: one correct but long, another shorter but containing an invalid division. Ask which steps are justified and whether any answer is lost. This directs attention to reasoning rather than handwriting or speed. An unexpected valid method should be accepted. A tidy method should not receive automatic trust. The learner is learning how to judge an argument, a capability that later supports algebraic proof and equation solving under unfamiliar notation.

A fresh individual task is 5(x − 2) = 3x + 8, giving x = 9. Another task asks the learner to construct an equation with no solution. For instance, 2x + 1 = 2x + 4 works. Constructing an example can reveal whether the student understands the condition more deeply than a repeated solve-and-check routine. Choose the demand according to the learner’s stage while keeping the shared theme of equality intact.

12. Graphs: a fixed charge prevents direct proportion

Use a hypothetical printing model C = 12 + 0.4n, where C is cost in currency units and n is the number of pages in one order. The fixed charge is 12 and each page adds 0.4. At n = 20, the cost is 20. At n = 40, it is 28, not 40. Doubling the page count does not double the whole cost because the fixed component is charged once. No actual business price is being quoted here.

Alicia doubles the first total. Tricia calculates the formula correctly but says the gradient is twelve. Kai Kai draws a line through the origin because he expects every relationship between two quantities to begin there. Each mistake calls for a distinct connection: separating fixed and variable components, interpreting slope and reading the intercept. A small class is useful when the tutor can distinguish those needs before assigning another graph worksheet.

Construct a short table and connect it to the equation. The values at n = 0, 10, 20 and 30 are 12, 16, 20 and 24. Each increase of ten pages adds four units. The graph’s vertical intercept is the fixed cost, while its rate of change is 0.4 per page. Ask a learner using the table to locate both quantities in the formula and the graph. A representation becomes useful when the connection is explicit.

Now compare another hypothetical model D = 4 + 0.6n. Setting C = D gives 12 + 0.4n = 4 + 0.6n, so n = 40 and the shared cost is 28. Below forty pages, D is lower; above forty, C is lower. The decision depends on the page count. A tutor can ask the learner to explain that conclusion from the equations, a difference expression or the crossing of two lines rather than make a blanket claim that one option is always cheaper.

Interpret the model’s domain. Pages are non-negative whole numbers in this example. A graph may be drawn as a continuous line to make its structure clear, but an order of 3.7 pages is not a literal allowed count. The equation also assumes the stated pricing rule applies throughout the range considered. A mathematical model should not be extended beyond its supplied conditions without explanation. This is a reasoning habit that can be taught without importing real commercial claims.

The fresh individual task supplies two data points and asks for a fixed-plus-variable equation. If ten units cost eighteen and twenty units cost twenty-six, the rate is 0.8 and the fixed component is ten. The model is C = 10 + 0.8n. Ask what each parameter means before using it. A correct pair of numbers is not a complete interpretation if the learner reverses their roles or treats transformed coordinates as original quantities in later work.

13. Similarity: a length factor is not an area factor

Two similar triangles have corresponding side lengths in the ratio 3:5. If the smaller triangle has area 27 square units, the larger has area 27 × (5/3)² = 75 square units. The area factor is the square of the length factor. The lesson should show why: corresponding base and perpendicular height each scale by 5/3, so their product scales by 25/9. The factor of one half in the triangle area formula remains on both sides.

Alicia multiplies the area by 5/3 and obtains 45. Tricia knows to square a factor but chooses 3/5, calculating in the wrong direction. Kai Kai pairs sides by their position in the drawing rather than corresponding vertices. The tutor should not label all three as weak in similarity. One needs dimensional scaling, another a clear source-to-target direction, and the third correspondence. A shared problem can support different repairs while preserving a coherent class discussion.

Rotate one triangle without changing its labelled relationships. The corresponding side does not become a different side merely because it is no longer horizontal. Ask the learners to state an angle or vertex correspondence before writing a ratio. If similarity itself has not been established or given, that is an earlier obligation. A drawing that looks similar is not enough. The tutor should distinguish using a known relationship from proving that it applies in the first place.

Compare the rectangle 3 by 4 with the rectangle 6 by 8. The side factor is two, the perimeter factor is two and the area factor is four. Then compare 3 by 4 with 6 by 4. The area doubles, but the rectangles are not similar. This contrast prevents the student from assuming that any increase in one dimension justifies a similarity rule. The relationship requires proportional change in all corresponding lengths, not merely a familiar pair of numbers.

For an appropriate extension, use similar solids with side ratio 2:3. Volumes scale as 8:27 because three length dimensions contribute. The tutor should not force this extension on a learner still establishing the area relationship. Instead, ask whether the existing explanation can be generalised when another dimension is involved. Depth comes from reasoning about the product of scales, not from memorising isolated instructions to square here and cube there.

A fresh individual check gives an area ratio 16:49 and asks for the corresponding positive length ratio. The answer is 4:7. Reversing the direction tests whether the learner understands the relationship rather than simply repeats a forward multiplication. Another learner can construct two non-similar figures with equal areas. These tasks give the group a shared topic and distinct useful challenges without turning one student’s strength into a permanent role as the other students’ unpaid tutor.

14. Probability: list the outcomes that are actually equally likely

An original bag problem contains three red counters and two blue counters. Two counters are drawn without replacement. The probability of two red counters is (3/5)(2/4) = 3/10. The probability of exactly one red is (3/5)(2/4) + (2/5)(3/4) = 3/5. These results depend on the sampling rule. The second draw is taken from four remaining counters, and exactly one red can occur in either order.

Alicia uses 3/5 for both red draws, as though replacement occurred. Tricia counts red-then-blue but omits blue-then-red. Kai Kai lists the colour outcomes RR, RB, BR and BB and assumes they have equal probabilities. The visible answer alone does not establish which misconception is present. A small-group discussion can compare a labelled tree, sequential reasoning and a list of individual counters, making the sampling process explicit.

Label the counters R1, R2, R3, B1 and B2. There are twenty ordered pairs of distinct counters when every first choice and every remaining second choice are equally likely. Twelve pairs contain exactly one red: three red choices followed by two blue choices, and two blue choices followed by three red choices. This gives 12/20 = 3/5. The labelled outcomes explain why the four colour patterns are not equally likely even though they are easy to list.

Now change only the replacement rule. With replacement, the chance of two red counters becomes (3/5)² = 9/25. Exactly one red becomes 2(3/5)(2/5) = 12/25. Ask what has changed in the physical process and which line of the calculation represents it. The task is not harder because of larger numbers; it requires the learner to attend to a condition. This is the kind of close contrast that can make group discussion mathematically useful.

Use a total-probability check. Without replacement, the probability of RR is 3/10, BB is (2/5)(1/4) = 1/10 and exactly one of each is 3/5. Their sum is one. This checks that the three categories form a complete, non-overlapping partition of the possible colour outcomes. It does not prove every number independently, but a sum different from one exposes an inconsistency. The learner should know why the categories are exhaustive before using the check.

The individual exit task changes the counter counts or asks for at least one blue. The complement approach may be shorter than listing every favourable sequence. Let students compare valid methods and explain why their events are the same. A learner who chooses a method independently and accounts for both order and replacement has demonstrated more than one who copies the class tree. Preserve that distinction when reporting what the lesson established.

15. Quadratics: compare representations for the question being asked

Let f(x) = x² − 8x + 7. Its factor form is (x − 1)(x − 7), and its completed-square form is (x − 4)² − 9. Ask three different questions: where is f zero, what is its minimum over the real numbers, and what is f(0)? Each form makes some information particularly visible. The roots are one and seven, the minimum is minus nine at x = 4, and the vertical intercept is seven. The learner should choose a representation for a purpose.

Alicia finds the roots quickly but calls seven the maximum because it is the largest number she calculated. Tricia completes the square correctly but cannot explain why the square creates a lower bound. Kai Kai can carry out either procedure after it is named but hesitates when the questions are mixed. The group should not receive an identical additional factorisation drill. The current needs are interpretation, justification and selection.

Place the forms side by side and ask what remains unchanged. They describe the same function for every real x. Expanding either transformed form recovers the original. Substituting a chosen input can catch a mistake, though agreement at one input is not a proof of equivalence. This distinction matters because students sometimes trust a transformed expression after one favourable numerical test. Algebraic identities justify the representation; selected values provide practical checks against slips.

Change the target to f(x) ≤ 0. The factor form identifies boundary roots one and seven, while the upward-opening graph or a sign analysis shows 1 ≤ x ≤ 7. Knowing the roots is necessary but not sufficient to identify the interval. A learner who lists only the roots has answered the equality problem rather than the inequality. The tutor can compare the two requests and ask what values between the roots do in the original expression.

For a deeper task, restrict the domain to 0 ≤ x ≤ 2 and ask for the minimum. The unrestricted minimising input x = 4 is not allowed. The function decreases on the given interval, so the minimum is f(2) = −5. The completed square remains useful because x = 2 is the allowed value closest to four. This tests whether the learner carries the question’s conditions into the interpretation rather than automatically reports the number outside a square.

The IES algebra guide recommends purposeful comparison of strategies and attention to algebraic structure, with different evidence ratings across recommendations. The classroom comparison here is a proposed application of those ideas. Its success is checked through a fresh individual quadratic with a changed target, not established merely because several methods appeared on the board.

16. A-Math: connect a derivative to the information the question needs

Use this example only after the relevant calculus has been taught. For y = x³ − x + 2 at x = 2, the curve point is (2, 8) and the derivative is 3x² − 1, giving tangent gradient 11. The tangent is y − 8 = 11(x − 2), or y = 11x − 14. The calculation brings together a function value, a derivative value and a straight-line equation. Each plays a different role.

Alicia differentiates correctly but substitutes eleven as the point’s vertical coordinate. Tricia finds the point and gradient but writes y + 8 = 11(x − 2). Kai Kai knows how to write a line once the point and gradient are supplied, but does not know how the derivative provides the gradient. A complete answer supplied by the tutor can hide all three distinctions. An individual attempt reveals whether the repair concerns interpretation, signed substitution or the calculus connection.

Ask what f(2) and f′(2) mean. The first locates the curve at the input; the second describes the local tangent gradient. A sketch with separate labels can help. Then verify that the proposed line passes through (2, 8). This check catches an incorrect intercept but does not independently establish the derivative. The learner should know the reach of a check. Passing through the required point is one obligation, not a substitute for every earlier justification.

Reverse the task: find the points where the tangent gradient is eleven. Solving 3x² − 1 = 11 gives x = ±2. The corresponding points are (2, 8) and (−2, −4). There are two relevant tangents. The learner must solve the derivative condition, then return to the original function for coordinates. This changed direction of information is a useful transfer task. It tests more than repeating the standard sequence from a supplied input.

Peer comparison can focus on which information is still missing at each stage. One student explains why differentiation alone does not locate the point; another explains why solving the derivative equation can produce more than one input; another checks the final line equations. These roles should rotate. The tutor remains responsible for mathematical correctness and should not allow a confident peer explanation to become an unchecked source of a new misconception.

For the full final-year lesson process, use Inside a Secondary 4 A-Math Lesson. For how help changes across a programme, read What to Expect from Bukit Timah A-Math Tuition. The small-group principle here is to keep each learner’s particular connection visible before and after the shared explanation.

17. Peer discussion should have a precise mathematical job

A lively class is not automatically a thinking class. Students can talk extensively while repeating answers, guessing methods or waiting for the confident learner to decide. The tutor should give discussion a specific job: compare two routes, identify a first invalid line, justify a condition or construct a counterexample. These tasks make contributions inspectable. They also give the quieter student a way to participate that does not depend on being the first to announce a complete solution.

Compare routes only when the learners have enough understanding to make the comparison meaningful. An initial explanation of a new method may be needed before anyone can evaluate alternatives. Showing three sophisticated solutions to a novice can create confusion rather than flexibility. The tutor should choose a contrast at the right level: perhaps a diagram and one equation expressing the same relation, rather than several algebraic techniques with unfamiliar notation.

Require reasons, but keep them proportionate. A useful explanation might be that the denominator changes because a counter was not replaced, or that the negative coefficient makes the quadratic open downwards. It need not become a speech. Ask a listener to restate the relationship and then apply it to a small change. This checks whether the discussion conveyed a usable idea rather than only created an impression of engagement.

Use errors without turning a learner into the example of failure. A tutor-created incorrect solution can contain the same mathematical issue without publicly identifying who made it. Ask where the solution stops being equivalent to the original and what should be preserved. This approach focuses attention on the reasoning. It does not require pretending that wrong answers are correct or avoiding direct correction; it keeps the correction tied to a mathematical decision rather than a person’s status.

Watch for agreement without understanding. Three students saying yes to a peer explanation do not provide the same evidence as three fresh individual applications. The tutor should include a short independent check after discussion. One learner may have understood the procedure, another the reason, and another only the answer. That difference is not a reason to avoid collaboration. It is a reason to distinguish what collaboration helped teach from what each learner can now do alone.

The group should gradually develop mathematical listening. A student can ask which whole a fraction refers to, whether the interval includes an endpoint or why a line is perpendicular. These questions can improve a solution without requiring the listener to have completed it first. The ultimate value of discussion is not that students become dependent on one another’s approval. It is that they acquire questions they can later ask of their own work.

18. A possible 90-minute group lesson, with flexible allocation

Consider a group working on linear models. The tutor has seen that one learner confuses fixed and variable components, another calculates gradients inaccurately and a third can use a formula but cannot construct it. The lesson’s shared question is how a constant rate and a starting amount appear in words, tables and graphs. The minute ranges below are proposed planning choices, not a prescribed schedule or research-validated allocation for every child.

During the opening ten minutes, each learner attempts a short fresh model and one earlier skill without a demonstration. The tutor also checks the current school topic and any returned work. This is not a long test. It provides enough evidence to confirm or revise the planned focus. If the supposed graph difficulty is actually an arithmetic gap, the tutor should notice before spending the whole session on a polished explanation of intercepts.

The next twenty minutes connect a hypothetical cost rule to a table and graph. Each representation is labelled clearly. The tutor asks why doubling the input does not double the output when a fixed charge is present. The class compares a proportional model with a fixed-plus-variable model. This shared teaching establishes a relationship everyone can use, while the tutor listens for the particular interpretation each learner still finds uncertain.

In the following twenty minutes, tasks differ. One student identifies intercepts from descriptions; another calculates gradients from clean data; the third forms equations from a short context. The tutor moves between them, but every independent interval has a defined job. A student finishing early can explain a counterexample to a false claim rather than receive unnecessary copies. A student unable to begin receives teaching rather than being told that silence itself is productive struggle.

The next fifteen minutes reunite the group around two competing hypothetical models. They identify the intersection and discuss which output is larger in different input ranges. One learner uses algebra, another a graph and another a difference table. The tutor connects the routes and checks the assumptions. The shared comparison should make the relationship clearer, not become a contest to decide which student has the best method in every situation.

A further fifteen minutes provides fresh individual application, with an appropriate time window only where the Mathematics is sufficiently stable. The final ten minutes review what was independent, what assistance remained and what will return later. A brief maintenance task can protect an older skill. The lesson ends with a next action for each learner. Its quality is judged by the resulting work, not by exact obedience to this example clock.

19. The learning should return after the shared explanation is no longer available

Immediate success after discussion is valuable, but it does not answer every question about learning. The student may still be relying on a recently heard phrase, a visible diagram or memory of the answer. A later fresh task asks whether the relationship remains available after attention has moved elsewhere. The return should be manageable and purposeful. It is not an invitation to test every historical topic every week.

The IES guide on organising instruction and study recommends spacing learning, revisiting content through quizzes and connecting representations, with different evidence ratings for its recommendations. These ideas support planning returns rather than relying only on same-day fluency. They do not prescribe this class’s exact timetable or prove that one weekly lesson reproduces the effects of a different research programme.

For the ratio lesson, a later task might reverse which quantity is supplied or change addition into transfer. For the graph lesson, it might supply a table rather than an equation. For similarity, it might give the area ratio and request the length ratio. Each variation tests a specific connection. Changing every feature at once can make failure difficult to interpret. Keeping the intended relationship visible allows the tutor to locate the next difficulty accurately.

Record assistance honestly. If the learner opens notes during practice, that can be a sensible learning step, but the result is not an unaided check. Preserve the first attempt and use another fresh item later. A corrected page with unknown support conditions can mislead the tutor into withdrawing help too early. A short note describing the exact cue needed is usually more useful than a long account of how frustrating the homework felt.

When a return fails, ask what remains understood. A student may explain the concept but not retrieve the first step, or execute a method but fail to recognise it in a new form. Another may encounter an entirely new prerequisite. The response should follow that evidence. Repeating the complete original explanation unchanged is not always the best next action. The return exists partly to tell the tutor which part of the learning has and has not travelled.

As independence improves, the learner can help choose returns. Ask which earlier method has not been attempted without help recently and why it matters to current work. Compare that judgment with a fresh task. The tutor remains available to correct inaccurate self-assessment. The goal is not a child who never needs teaching again, but a learner who increasingly recognises what to practise, how to check it and when a specific request for help is appropriate.

20. Parents can inspect learning without monitoring every minute

A parent does not need to become another Mathematics tutor to judge whether a small-group programme is coherent. Ask for the current learning priority, the work that supports it, the teaching response and the next independent check. These four pieces of information can be concise. They are more useful than a catalogue of worksheets completed or a promise that confidence has improved without any account of what the learner can now do differently.

A good update might explain that the learner can calculate a scale factor but applies it directly to area; the lesson connected both length dimensions to the squared area factor; a fresh example was solved independently; the next return reverses the ratio. The family can understand that progression without reviewing every algebraic line. The report should also name uncertainty. One fresh success does not establish that every geometry problem is now secure.

Ask whether the first attempt was visible before peer discussion. A complete class solution may be evidence of good collaborative teaching but not of each learner’s independent recognition. Ask what happened after the explanation. Did the student solve a changed question, explain a condition or catch an inconsistency? The answers need not use technical language. They should describe actions the student actually performed and what support remained.

Compare marks cautiously. Different school papers vary in topic mix and difficulty, and practice sets vary in familiarity and assistance. A higher percentage can include genuine learning without proving that the class format alone caused the change. A stable percentage can conceal a repaired skill and a new topic difficulty. The tutor should inspect the script, not either dismiss the total or treat it as a complete explanation.

Review practical fit as well. Can the learner sustain appropriate independent work while the tutor helps another student? Does the group follow a compatible course sequence? Is homework proportionate to school demands? Are current fees and absence arrangements clear? These conditions matter to the family’s decision, but should be confirmed directly rather than inferred from a general article. A programme can be well designed in principle and still be the wrong arrangement for a particular learner at a particular time.

The useful long-term direction is less unnecessary rescue. The learner increasingly starts a problem, states a relationship, checks a condition and asks a precise question when stuck. That does not eliminate mistakes or guarantee an examination outcome. It makes mathematical responsibility more visible inside the student. A three-student class justifies itself when its individual attention and shared discussion both contribute to that direction.

The small-group casebook: make the difference in thinking visible

The following classroom cases take the format beyond a general promise of attention. Each begins with a mathematical relationship, identifies different possible interpretations and shows how shared teaching can lead back to individual work. Alicia, Tricia and Kai Kai remain fictional learners. Their roles change across the examples; none represents a permanent ability type. Select tasks appropriate to the student’s actual course and taught material. The cases illustrate teaching choices, not measured results or an official assessment sequence.

Open the mathematical casebook contents

21. Weighted means · 22. Average speed · 23. Simultaneous equations · 24. Inequalities · 25. What a diagram establishes · 26. Mean, median and spread · 27. Sequences and general rules · 28. The same answer, different causes · 29. One model across three representations · 30. Units and dimensions.

21. Weighted means: two averages cannot always be averaged equally

One fictional class has twelve students with a mean quiz score of 70. Another has eighteen students with a mean score of 80 on the same quiz. What is the mean for all thirty students? The first class contributes 12 × 70 = 840 score points, and the second contributes 18 × 80 = 1,440. The combined total is 2,280, so the combined mean is 76. The answer is not the unweighted average of seventy and eighty because the two means represent different numbers of students.

Ask each learner to write the quantity represented by seventy before calculating. Alicia writes the total for the first class as seventy, confusing a per-student average with a group total. Tricia knows that a total is needed but calculates (70 + 80)/2 because two classes are mentioned. Kai Kai obtains seventy-six using a remembered weighted-mean formula, but cannot explain why eighteen appears beside eighty. The correct final answer in his case does not remove the need to inspect the relationship. The tutor needs to distinguish execution from meaning.

A useful shared representation is a two-row table with number of students, mean score and total score. The final row adds counts and totals, not means. This makes the denominator visible: the resulting mean is per student, not per class. Ask the group what a simple average of the two class means would describe. It treats each class as one equally weighted observation. That can be a legitimate statistic for another purpose, but it does not answer the question about the average individual student’s score across these unequal class sizes.

Use a limiting example to expose the difference. Suppose one group has a single score of zero and another has nine scores of one hundred. The average across all ten scores is ninety, not fifty. The small example makes the weighting effect difficult to hide behind a formula. Then return to the original class problem. A smaller example is useful because it preserves the exact obligation: reconstruct the total associated with each average before combining observations.

The group can compare two valid calculations. One adds the reconstructed totals. Another starts from seventy as a baseline: all thirty students at seventy would contribute 2,100, and the eighteen students in the second class contribute an additional ten points each, adding 180. The result is again 2,280/30 = 76. The second route explains why the combined mean is closer to eighty than to seventy. It is not required as a replacement method; it offers a structural check on the first.

Now change only the group sizes so both contain fifteen students. The ordinary average of the two means is then correct. Ask why the shortcut becomes valid. The lesson should not create a new prohibition that means can never be averaged. It should teach the condition under which equal weighting matches the target population. A student who can state that condition has learned something more transferable than a rule to reject one familiar-looking formula.

For the individual return, give eight observations with mean six and twelve observations with mean eleven. The combined mean is (48 + 132)/20 = 9. Another task asks for the missing group mean when the overall mean and both group sizes are supplied. For example, ten observations with mean eight and five more observations have overall mean ten; the five additional observations must total seventy and have mean fourteen. The tutor can identify whether the learner reconstructs totals in both directions without a prompt.

The parent-facing evidence is specific. The learner now distinguishes an average from the total it represents and can combine unequal groups under stated conditions. This does not establish that every statistics question is secure. A later problem may add missing data, a different unit or a misleading graph. The small-group benefit is that several interpretations were compared without losing the responsibility to check each student’s new individual calculation.

22. Average speed: the denominator is elapsed time, not the number of speeds

A cyclist travels eighteen kilometres at twelve kilometres per hour and then eighteen kilometres at eighteen kilometres per hour. The first stage takes 1.5 hours and the second one hour. Total distance is thirty-six kilometres and total time is 2.5 hours, so average speed is 14.4 kilometres per hour. The arithmetic mean of twelve and eighteen is fifteen, but the cyclist spends unequal amounts of time at those speeds. Equal distances do not supply equal time weights.

Alicia averages the two speeds immediately. Tricia forms total distance correctly but divides it by two because there are two stages. Kai Kai calculates both journey times but adds the speeds before dividing. The tutor should ask each learner to write the units of the proposed result. Kilometres divided by hours has the required speed unit. Dividing kilometres by the number of stages produces distance per stage, a different quantity. Units cannot prove every calculation, but they can expose a mismatched denominator.

The shared table should contain distance, speed and time for each stage. Ask learners to fill the missing column before combining anything. A table is helpful only when its headings have meaning. Do not let it become a memorised layout that the student fills without knowing why d/v gives time. A brief numerical question, such as how long twelve kilometres takes at twelve kilometres per hour, can reconnect the formula to its rate interpretation.

Now change the problem to one hour at twelve kilometres per hour followed by one hour at eighteen kilometres per hour. The distances are twelve and eighteen kilometres, giving thirty kilometres in two hours and average speed fifteen. The same two speeds now have the arithmetic mean because the durations are equal. This near contrast gives the lesson a precise question: which quantity is held equal? A blanket instruction never to average speeds would be as misleading as always averaging them.

Add a thirty-minute rest between the original equal-distance stages and ask for the average speed for the entire elapsed journey. Total time becomes three hours and the average is twelve kilometres per hour. If the question instead asks only about moving time, the denominator remains 2.5 hours. The wording decides the measurement. A tutor should make that distinction explicit rather than silently choose one interpretation and call the other answer careless.

A stronger learner can derive the equal-distance formula. For distance d travelled at positive speeds u and v, total time is d/u + d/v. Dividing 2d by that time gives 2uv/(u + v). This derivation should follow the meaning, not replace it with another rule to memorise. The cancellation of d also explains why the average is independent of the common distance, provided the assumptions remain the same. The denominator requires positive speeds in this physical model.

The peer discussion can compare a time-weighted mean with total-distance-over-total-time. They are two expressions of the same rate calculation. Ask one learner to explain why the slower stage receives more time weight in the equal-distance journey, and another to identify when that reasoning would change. The tutor remains responsible for correcting the model. A confident peer should not turn a shortcut from one version into a universal statement about all journeys.

The independent task gives ten kilometres at five kilometres per hour and twenty kilometres at ten kilometres per hour. Both stages take two hours, so average speed is thirty divided by four, or 7.5 kilometres per hour. A second task can use unequal durations or include a rest. The intended evidence is not familiarity with a cyclist story. It is the student’s choice of a denominator that represents the requested period and a total that measures the requested distance.

23. Simultaneous equations: compare routes without confusing alternatives with inconsistency

In a hypothetical stationery order, two notebooks and three pens cost thirteen units, while three notebooks and two pens cost twelve units. Let n and p be the respective unit prices. The equations are 2n + 3p = 13 and 3n + 2p = 12. Eliminating n gives 5p = 15, so p = 3 and n = 2. Substituting into both original equations verifies the two conditions. The prices are invented for this mathematical example, not a current retail quotation.

Alicia eliminates correctly. Tricia subtracts one equation from the other first and obtains n − p = −1, so n = p − 1; substitution then gives the same solution. Kai Kai begins by adding the equations, obtaining 5n + 5p = 25 and therefore n + p = 5. His route is also productive, provided he combines it with another independent relationship. The tutor should not discard a valid start merely because it differs from the prepared model solution.

The useful group comparison asks what each operation achieves. Multiplying equations creates equal coefficients for elimination. Subtracting exposes a simple difference. Adding exposes a simple sum. The original pair happens to support all three efficiently. The student should learn why a route is convenient here, not that addition or subtraction is always the best first move. Later coefficient choices can change the practical cost without changing the validity of the methods.

Inspect a common invalid operation separately. If a student multiplies 2n + 3p = 13 by three but writes 6n + 9p = 13, the right side was not transformed. The next explanation concerns equality, not the choice between elimination and substitution. Another student may form the equations with notebook and pen coefficients reversed. That error belongs to representation. Both can produce wrong final prices, but a small class should allow the tutor to identify the earliest relevant discrepancy before choosing practice.

Now compare 2n + 3p = 13 with 4n + 6p = 26. The second is a multiple of the first and adds no new independent condition. There are infinitely many real pairs satisfying them, although a specific contextual domain could narrow possibilities. Replace twenty-six with twenty-seven and the equations become inconsistent. This contrast teaches why two written equations do not automatically determine one pair of values. The relationships must supply independent, compatible information.

A graphical interpretation uses two lines. Distinct non-parallel lines intersect at one point, coincident lines share all their points, and distinct parallel lines do not intersect. The diagram can make the solution types visible, while the equations establish their exact relationships. Ask learners to connect a zero-equals-zero elimination result with coincident lines, and a contradiction with no intersection. Do not treat the disappearing variable as an instruction to invent a numerical answer.

The independent check uses a new pair, such as 3a + b = 11 and a + 2b = 12, giving a = 2 and b = 5. Learners may choose different routes, but each should preserve equality and verify both starting conditions. A stronger student can construct two equations with the same chosen solution but different coefficients. Another can explain why one condition alone is insufficient. These tasks share a mathematical core while giving different learners meaningful responsibility.

A useful lesson note records the distinction between forming the equations, selecting a route and executing it. The learner who needed a prompt to define the prices should not be reported as independently modelling simply because their elimination was flawless afterwards. The learner who chose an unexpected valid route should not be called confused. The group makes alternatives visible; the tutor’s role is to keep the reasons and assistance conditions accurate.

24. Inequalities: a sign reversal should have a reason

Solve 5 − 2x > 11. Subtracting five gives −2x > 6. Dividing by negative two reverses the inequality, giving x < −3. Test x = −4: the original left side is thirteen, which is greater than eleven. Test x = 0: five is not greater than eleven. These checks help reject the opposite interval. They do not replace the general reasoning, but they make the consequence of a missed reversal visible.

Alicia remembers to reverse the symbol but cannot explain why. Tricia forgets the reversal because she treats the inequality like an equation. Kai Kai avoids dividing by a negative number by adding 2x to both sides, obtaining 5 > 11 + 2x, then −6 > 2x and −3 > x. His route is valid and leads to the same interval. The tutor should compare the paths without requiring every student to use the same arrangement of terms.

A numerical contrast can establish the reversal. Start from 2 < 5. Multiplying both by positive three gives 6 < 15; multiplying both by negative three gives −6 > −15. On a number line, multiplication by a negative number reverses orientation. The rule is about order under the operation, not a magical consequence of a minus sign appearing somewhere on the page. Adding a negative number to both sides does not reverse an inequality.

Use that distinction immediately. From x + 4 > 9, subtracting four gives x > 5 with the direction unchanged. From −3x > 9, division by negative three gives x < −3. A learner who reverses whenever a negative appears needs the operations separated. The small-group discussion can ask each student to construct an example where a negative number is involved but reversal is not required. Construction reveals whether the rule has a meaningful boundary.

For students ready for quadratic inequalities, consider (x − 2)(x − 5) < 0. The roots are two and five, and the product is negative between them, so 2 < x < 5. The answer is an interval, not only the boundary roots. A sign table, a graph or a product-sign argument can justify it. Ask what happens at the endpoints. The strict inequality excludes values making the product zero. Equality and inequality questions share algebra but make different demands.

Do not multiply a rational inequality by an expression of unknown sign as though its effect were fixed. Even where that topic is only an extension, a simple warning can connect to the established principle: the sign of the multiplier matters. Use a sign analysis or cases when appropriate to the course. The tutor should avoid introducing a sophisticated method before its prerequisites are understood, but should also avoid teaching an unrestricted shortcut that later becomes false.

The independent check is 7 − 3x ≤ 16, giving x ≥ −3. Ask the learner to represent the answer on a number line and test the boundary and one interior value. Another task asks for an inequality whose solution is x < 4. A possible answer is 3x < 12, though many others work. Accepting a valid constructed example helps the learner see a solution set as a mathematical object rather than merely the final symbol in a routine.

The group has done useful work when each learner can state which operation changes order and can apply that understanding in a fresh task. A student who remembers the spoken cue only while the tutor is nearby still needs a later return. The record should distinguish that stage. The lesson is not successful because everyone repeated reverse the sign together; it is successful when the relevant decision is made independently where the operation requires it.

25. Geometry: the drawing suggests a relationship, but the conditions establish it

Draw triangle ABC with D on AB and E on AC. Mark AD = DB and AE = EC. The midpoint relationships imply that DE is parallel to BC and half its length. A student may recognise a familiar theorem immediately. The useful teaching question is whether the learner knows which conditions allow it to be used and which conclusion follows. A horizontal-looking line in a diagram does not become parallel merely because it resembles the textbook picture.

Alicia uses the midpoint theorem correctly. Tricia begins by assuming DE is parallel to BC and then uses angle relationships to justify the same result, making the argument circular if parallelism was not given. Kai Kai identifies the midpoints but cannot connect the half-length ratios to similar triangles. The tutor should distinguish recall of a theorem from justification and from recognising its underlying relationship. These are different learning jobs within one geometry problem.

One proof route notes AD/AB = AE/AC = 1/2 and the common included angle at A. The appropriate side-angle-side similarity condition makes triangles ADE and ABC similar with the corresponding vertices in that order. Corresponding angle equality then supports the parallel conclusion, and the side ratio gives DE/BC = 1/2. The learner should label the correspondence rather than copy a chain of letters whose order is not understood.

Rotate the entire drawing and stretch its appearance while preserving the labelled relationships. The proof remains valid. Then change one midpoint condition: let AD/AB = 1/2 but AE/AC = 1/3. The previous similarity argument no longer applies. The class should explain which assumption was lost and why the conclusion is no longer guaranteed from those conditions alone. This is more informative than memorising another upright diagram with different side lengths.

A second case concerns a quadrilateral that looks like a rectangle. Give only one pair of parallel sides and one right angle. Ask whether all four angles must be right angles. They need not be. A suitable trapezium provides a counterexample. The point is not to teach distrust of diagrams; diagrams are valuable representations. It is to distinguish a drawn appearance from a stated or established property. The tutor should help the learner know what information may be used in the proof.

For peer work, assign a proof-reading job rather than a fastest-answer job. One student identifies the given information, another checks the reasons attached to each step and another tries to find an unstated assumption. Rotate those responsibilities. The learner who usually calculates quickly may discover that explaining why a statement follows is a different demand. The quieter learner may contribute a decisive question about a missing condition without completing every numerical step first.

A fresh individual task can ask which of several diagrams supplies sufficient information for a particular theorem. Another asks the student to finish a proof with one reason missing. Do not interpret a correct multiple-choice selection as complete proof-writing mastery. These tasks sample different stages. Follow with an appropriate independent explanation when that is the intended capability. The programme should make the distinction clear rather than use a short recognition task to claim a much broader result.

The parent update can describe a change in the learner’s use of evidence: the student now identifies what is given, avoids assuming the target conclusion and maintains correspondence when the diagram is rotated. Those observations are more useful than saying geometry confidence improved. They also identify a transferable reasoning habit, while recognising that each new geometric theorem still requires its own knowledge and conditions.

26. Statistics: the same centre can hide a different distribution

Compare two small numerical data sets: A = 4, 5, 6, 7, 8 and B = 0, 5, 6, 7, 12. Both have mean six and median six. Their ranges are four and twelve respectively. A learner who reports only the shared mean has not described the different spread. A learner who claims that B must have a higher mean because it contains twelve has not accounted for its zero. The complete set of observations matters.

Ask each learner to say what average is intended before calculating. In everyday speech the word can be loose; in a mathematical task the requested measure should be clear. Alicia computes the mean correctly but calls it the median. Tricia finds the middle value without first sorting a newly rearranged list. Kai Kai compares only the largest observations. These errors involve terminology, procedure and selection of evidence. Another long mean worksheet would not address all three equally well.

Use the two sets to ask whether a higher range proves that most values are more spread out. Range uses only the extremes. Other measures and the distribution itself provide different information. Within the learner’s taught scope, compare dot plots or ordered lists before introducing additional formulae. The lesson should not imply that one statistic tells the complete story. It should explain what a measure uses and what it ignores.

Now replace the eight in set A with twenty-eight. The mean becomes ten, while the median remains six. Ask why one measure moves and the other does not. The total changes by twenty, distributed over five observations in the mean calculation. The middle ordered observation stays six. This is a concrete explanation of different sensitivities, not a rule that the median is always better. The appropriate summary depends on what the question asks and the context being described.

A graphical case can use the same counts with two vertical-axis choices. If bars representing 48 and 50 begin visually at zero, their heights are similar. If the displayed axis starts at 47, the difference can look much larger. The numerical difference remains two. The tutor should ask the learner to read the axis labels and distinguish the data from the visual impression. No particular real organisation or chart is being accused; this is an original demonstration of how scale influences interpretation.

The peer discussion can compare two written conclusions. One says both groups have the same average, therefore their results are identical. Another says the means agree but the observed spreads differ. The second is more accurate for the original sets. Ask which additional information would be needed to make a broader claim about a population. A five-value classroom example should not be presented as strong evidence about a whole school or a teaching programme.

The independent task asks learners to construct five numbers with mean ten and median eight. One valid set is 4, 6, 8, 12, 20. Their total is fifty and the middle ordered value is eight. Another task asks for two different sets with the same mean but different ranges. Construction forces the learner to satisfy several conditions at once. It can provide purposeful challenge without importing an advanced statistical method that the course has not yet introduced.

The small-group value lies in making the interpretations discussable and then checking them individually. A learner might calculate fluently yet make an unjustified conclusion about the data. Another may reason sensibly but misuse a term. The tutor should preserve those distinctions in the next task and the parent report. Statistical literacy includes both the calculation and the claim the calculation can reasonably support.

27. Sequences: a pattern should become a rule that can be explained

Consider a row of adjoining squares made from matchsticks, with each new square sharing one side with the previous square. One square uses four sticks, two use seven and three use ten. For n squares in this specified construction, the number of sticks is 3n + 1. One route starts from four and adds three for each of the remaining n − 1 squares. Another counts n top sticks, n bottom sticks and n + 1 vertical sticks. Both explain the same rule.

Alicia sees the difference of three and writes 3n, forgetting the offset. Tricia writes 4n and double-counts shared sides. Kai Kai obtains 3n + 1 from a memorised linear-sequence formula but cannot identify where the extra one comes from in the figure. The tutor should use the physical structure to distinguish these needs. A correct algebraic rule can still benefit from an explanation that connects it to the object being counted.

Ask the class to test n = 1 and n = 2. These checks reject 3n and 4n for the given construction. They do not by themselves prove a general rule. The counting argument establishes why each additional square contributes three new sticks. This distinction matters when learners infer rules from a short numerical list without a construction. Many different formulas can match a finite set of values. The problem’s structural description helps justify the intended generalisation.

Now ask how many squares can be made with sixty-one sticks in this row pattern. Solving 3n + 1 = 61 gives n = 20. The unknown is no longer the number of sticks. The learner must reverse the relationship and interpret n as a positive whole-number count. With sixty sticks, an exact row of this form would not use all the sticks: the formula gives n = 59/3. The tutor should discuss whether the question asks for an exact construction or the maximum number with unused sticks allowed.

Change the construction to separate squares that do not share sides. The rule becomes 4n. This makes Tricia’s earlier expression correct for another situation. Such a contrast is useful because it shows that an error can be a valid model applied to the wrong structure. The tutor need not describe the learner as bad at sequences. The teaching job is to identify what is shared and what is counted more than once.

For a more demanding but related task, arrange n squares around a closed loop where the geometry is explicitly defined, or use a different repeated tile construction whose joins are clear. Do not invent a diagram rule from a vague description. Each new case should make the counting assumptions visible. A stronger learner can explain how a constant starting component and a repeated component produce a linear expression, while another learner continues connecting the concrete count to the first few terms.

The group can compare the two derivations 4 + 3(n − 1) and n + n + (n + 1). Expanding gives the same 3n + 1. Ask each learner to label every term with its counting meaning before simplifying. Algebra is then a way to show that two counting views agree. The discussion should not end at announcing they are the same formula; it should explain why they count the same objects without omission or duplication.

A fresh individual task might use a row of adjoining triangles with a clearly specified shared-side arrangement, or a numerical rule supplied in words. The learner states a formula, checks an initial case and solves a reverse question. The tutor records whether the structure was recognised without a prompt. The result is stronger evidence than a page of continued sequences where the student only needs to repeat the last observed difference.

28. The same answer can conceal different causes, including a lucky cancellation

Take the expression 2(x + 3) − x at x = 4. The correct value is ten. Alicia simplifies correctly to x + 6 and substitutes four. Tricia substitutes first, obtaining 2(7) − 4 = 10. Kai Kai writes an incorrect intermediate simplification, 2x + 3 − x = x + 3, but then mistakenly adds six during substitution and also reports ten. The final answers match. Only the working reveals that the third route contains errors that happen to cancel numerically.

This is why checking only final answers gives an incomplete picture. It can classify a sound alternative as suspicious because it looks different, or accept an invalid method because the number is right. A small group makes it more practical to ask each learner how the result was obtained. The tutor should inspect enough working to identify whether the relationships were preserved. The purpose is not to demand maximum writing on every simple task, but to make important decisions visible when diagnosing a current difficulty.

Now use a wrong shared answer. In solving 4(x − 2) = 2x + 10, the correct solution is x = 9. One learner obtains x = 6 after expanding the left side as 4x − 2. Another expands correctly but moves the constant incorrectly and reaches a different invalid equation that also yields six. A third copies six from a classmate and cannot begin independently. The tutor should not infer one common misconception from the matching final number.

Ask for the first line each learner can defend. The first case may need distribution explained as multiplication of every term. The second may need equality-preserving operations. The third may need an unaided representation or a foundational explanation before practice. These are materially different interventions. A generic instruction to check the answer would reveal the failure through substitution but would not by itself teach the missing relationship.

A useful peer exercise presents two invented solutions with the same result and asks whether both are valid. Students identify the justification for each transition rather than vote according to the answer key. Another exercise presents different-looking correct expressions, such as (x − 2)² and x² − 4x + 4, and asks for proof of equivalence. Together, the cases teach that matching answers do not establish matching reasoning and differing forms do not establish disagreement.

The tutor should not overinterpret a single slip as a stable misconception. Ask the learner to perform a close fresh example and explain the intended operation. A one-off copying mistake, a repeated invalid rule and dependence on a peer can look similar on one page but differ over several observations. Good diagnosis uses the current evidence provisionally. The next task is a way to test the explanation, not merely a punishment for the first error.

For parent reporting, avoid counting every eventually correct answer as independent success. A more informative statement is that the student selected the method independently but needed a correction at distribution, then completed a new example without that prompt. Another learner may have needed the method named but executed it accurately. Both records acknowledge useful learning and make the next check clear. They do not rank children by a hidden permanent trait.

The independent return should target the suspected cause without repeating the same values. Change the bracket coefficients, the variable’s position or the requested form. Observe whether the repaired operation survives. The small-group advantage is not that three students can all write the same correct answer together. It is that the tutor can identify what each answer actually demonstrates and use that distinction to choose a better next explanation.

29. One model, three representations: let students connect rather than merely compare

A fictional tank begins with twelve litres of water and receives water at a constant rate of three litres per minute, with no outflow during the period considered. Its volume is V = 12 + 3t. At zero, two, four and six minutes, the volumes are twelve, eighteen, twenty-four and thirty litres. The equation, table and graph describe the same model. The lesson’s purpose is to help learners move between them while preserving the meaning of the initial amount and rate.

Alicia receives the verbal description, Tricia a table and Kai Kai a graph with labelled axes. Each first writes what they believe the rule is and identifies the starting volume. This is not a test of three unrelated problems. It is an opportunity to see which representation makes the relationship accessible to each learner and which conversion creates difficulty. The tutor should make the units and time interval explicit so that the representations are genuinely comparable.

The group then reconstructs the missing forms. The table learner explains why an increase of six litres over two minutes means three litres per minute. The graph learner identifies the intercept at twelve and the rise over a chosen time interval. The equation learner locates the same information in its constant and coefficient. Ask each student to point to the physical quantity represented by a symbol. Saying gradient or intercept is not enough if the learner cannot explain what it measures here.

Next ask when the tank contains forty-five litres. The equation 12 + 3t = 45 gives t = 11 minutes. A graph can estimate this input, while the equation gives an exact value within the idealised model. A table can be extended to show it. Discuss the difference between an exact algebraic solution and a reading limited by graph scale. The tutor should not mark a reasonable graphical estimate as a conceptual failure when the task permits estimation, nor accept an estimate where exactness is required.

Now introduce a tank capacity of forty-eight litres and state that inflow stops when the tank is full. The linear formula applies from t = 0 to t = 12. Continuing it to t = 20 would predict seventy-two litres, which violates the capacity condition under this model. A later piecewise description would hold volume at forty-eight after inflow stops. The learner does not need advanced notation to recognise that a formula has a range of physical applicability. The condition belongs to the model.

A changed case begins with thirty litres and drains at two litres per minute. The expression becomes V = 30 − 2t until the tank empties at fifteen minutes. Ask why the slope is negative but the volume is not allowed to continue into negative values. This separates a negative rate of change from an inadmissible quantity. A student who associates every minus sign with a wrong answer needs the roles of the quantities clarified.

The independent exit gives a different representation to each learner from the one they found easiest initially. Alicia might start from a graph, Tricia from words and Kai Kai from an equation. Each reconstructs one other form and solves a reverse question. The tutor records where help was needed. Success in the original familiar form should not be treated as proof of all conversions, but failure in a new form should not erase the relationship already understood.

This is a useful model for small-group differentiation because the shared object remains coherent. The students are not completing random tasks merely to occupy themselves. Their different representations contribute to one explanation, then their individual checks reveal whether those connections became usable. The tank story is incidental. The transferable capability is moving between a relationship, its symbols, its data and its graph while keeping units and conditions intact.

30. Units: a conversion changes the numerical measure, not the physical object

A rectangle is 2.4 metres long and 75 centimetres wide. Its area is 2.4 × 0.75 = 1.8 square metres. Converting both lengths to centimetres gives 240 × 75 = 18,000 square centimetres, the same physical area. The conversion between these area measures is ten thousand, not one hundred, because one square metre contains one hundred centimetres in each of two perpendicular directions.

Alicia multiplies 2.4 by 75 and labels the answer square metres, mixing length units. Tricia converts correctly to centimetres but divides the resulting area by one hundred, applying a length factor to area. Kai Kai obtains 1.8 using a calculator but cannot explain why the centimetre answer is much larger numerically. The tutor should identify the unit relationship before adding more multiplication practice. The arithmetic is not necessarily the main difficulty.

Use a square one metre on each side. Mark that each side is one hundred centimetres, so the count of one-centimetre squares is 100 × 100. This connects unit conversion to the earlier similarity lesson, where an area scale factor uses two length dimensions. The number changes because the unit used to measure the same area changes. A larger numerical value in smaller units does not mean that the object has physically grown.

Compare perimeter and area for the same rectangle. Perimeter is 2(2.4 + 0.75) = 6.3 metres or 630 centimetres. The conversion factor is one hundred for this length quantity. Area uses a different factor because it is measured in squared units. The learner should identify what is being measured before selecting a conversion. Memorising isolated factors without that distinction can produce correct answers in familiar worksheets and fail when the question changes quantity.

For a suitable extension, a box measures 30 centimetres by 20 centimetres by 15 centimetres. Its volume is 9,000 cubic centimetres, equal to nine litres using 1,000 cubic centimetres per litre. Converting all lengths to metres gives 0.3 × 0.2 × 0.15 = 0.009 cubic metres. The class can connect cubic centimetres, litres and cubic metres without implying that every dimensional conversion uses the same numerical factor.

Rates add another useful contrast. Seventy-two kilometres per hour equals twenty metres per second because the distance unit contributes a factor of one thousand and the time unit contributes a factor of 3,600. The ratio becomes 72,000/3,600. A shortcut such as divide by 3.6 is valid here, but its derivation explains its direction. A learner who multiplies instead may need to reconstruct the units rather than repeat the mnemonic more firmly.

The peer task is to inspect a solution with correct arithmetic and incorrect units. Ask what physical quantity the answer would represent as written. Another task gives two numerically different answers in different units and asks whether they agree. These exercises teach learners to read the full mathematical statement, including its unit. A number is not complete information when the question concerns a measured quantity.

The independent return can use mixed units in a new geometric context or reverse a speed conversion. The tutor should record whether the learner standardises units before calculating and whether the final unit matches the requested quantity. A small group is useful because one student’s area-factor explanation can illuminate another’s conversion error, but the final check still needs to show that each learner can perform the conversion without borrowing the explanation in real time.

From the worked example to a functioning small group

The mathematical cases show what can be taught. The next question is how the class protects time, attention and evidence so that those explanations become useful to three different learners. The following practical designs are proposals to adapt, not a claim that every lesson uses a fixed script. They retain the same standard throughout: each learner should have a worthwhile task, an appropriate opportunity for help and a clear way to show what they can subsequently do without that help.

Open the practical chapters and independent task bank

31. Allocating attention · 32. Rotating mathematical roles · 33. Reading the mathematical request · 34. Constructing questions · 35. Using a correction meeting · 36. Homework that returns information · 37. Different school sequences · 38. Timing without a speed contest · 39. Reading the review record · 40. When the format should change · 41. Student-directed practice · 42. Independent task bank · 43. A family-readable agreement.

31. Allocate attention according to the current task, not a permanent ranking

In a three-student lesson, equal care does not require identical tutor talk at every moment. A learner encountering a genuinely new relationship may need a complete explanation. Another may need a short question to retrieve an old method. A third may need the tutor not to intervene because an independent attempt is already progressing. The allocation should follow those current needs. It should not become a permanent arrangement in which one child is always taught, one is always checked and one is always left alone because of a previous score.

Suppose all three are working with quadratic forms. Alicia is independently solving a changed minimum problem. Tricia is attempting a direct example after a newly explained square-completion step. Kai Kai cannot yet identify why the target suggests a completed square. The tutor may spend a focused interval with Kai Kai while the other two have appropriate work. On returning, the tutor should inspect Tricia’s new transformation and Alicia’s interpretation of the domain. Attention moves, but the evidence from the independent intervals is not discarded.

The quality of the task given during that interval matters. Alicia might compare a minimum with a restricted-domain minimum, rather than receive ten copies she already controls. Tricia might explain one missing line in an example before trying a fresh one. These are different jobs linked to the same relationship. A small class does not become personalised simply because worksheets differ. The differences should correspond to observed learning needs and lead to a check the tutor can actually use.

There should also be a way to signal genuine blockage. A student can mark the last justified line and write the unresolved question while moving to a designated maintenance item where appropriate. This is not permission to avoid every difficult task. It prevents an entire independent interval from being spent repeatedly staring at a missing concept. The tutor later reads the marked point and decides whether a prompt, a different representation or direct teaching is needed.

Repeated continuous support demand is information about group fit. If the same learner cannot productively begin any available independent task because several prerequisites are missing, the class may need another arrangement or a different entry point. The response should not be to blame that learner for consuming attention. Nor should the other students’ learning be indefinitely displaced. A responsible review asks whether the teaching conditions can still serve all three current needs.

A parent can ask what the student did while the tutor attended elsewhere and what was learned from that work. The answer should be more specific than additional practice. It might be a fresh application, a return to an earlier skill or a comparison of two valid representations. The independent interval is part of the lesson, not empty time between explanations. Its value becomes visible when the tutor returns to the work and uses it to decide what happens next.

32. Rotate mathematical roles without fixing students into identities

A shared problem can give learners different responsibilities: explain the representation, inspect the transformations and choose a verification. These responsibilities make discussion more precise. They should not become permanent identities such as the clever one, the careful one and the weak one. A student who usually calculates quickly still needs to justify a model. A student who usually hesitates may notice an omitted condition. Rotating the work keeps the group focused on capabilities that can develop.

Take the equation (x − 2)(x + 5) = 18. Expanding and rearranging gives x² + 3x − 28 = 0, so (x + 7)(x − 4) = 0 and x = −7 or 4. One learner explains how the product becomes a quadratic equation, another verifies the factorisation and another substitutes the candidates. These are not three unrelated fragments assigned to avoid thinking about the whole problem. After discussion, each learner should understand the complete route and attempt a changed question individually.

Now place the same algebra inside a rectangle problem with side lengths x − 2 and x + 5 and area eighteen. The positive-length condition leaves only x = 4, giving sides two and nine. The verification role now has an additional obligation: return the candidate to the context, not only the algebraic equation. A learner who previously checked substitution can lead the discussion of admissibility. Another can explain why the negative root remains valid in the original pure equation but not in the geometric model.

Roles should remain small enough that the students do real Mathematics rather than manage a complicated classroom procedure. There is no need for elaborate job titles, scoring systems or a permanent hierarchy. The tutor can ask a focused question and change who answers next. The purpose is to distribute reasoning opportunities. A routine that consumes more attention than the problem itself should be simplified.

Do not make a confident learner responsible for continuously tutoring the others. Explaining can be a useful task, but that student also needs fresh challenge and feedback on their own work. Similarly, do not ask a struggling learner only to read the question while others make all the mathematical decisions. Give an appropriately sized responsibility that can be taught and checked. Participation should move the learner’s capability forward, not merely allow them to appear included.

The independent return reveals whether the shared role became transferable knowledge. Give a different quadratic context and ask everyone to identify candidates and restrictions. A learner may have understood their assigned fragment while missing another part of the chain. That is useful evidence, not a reason to abandon role-based discussion. The tutor can reconnect the missing stage. The class format works when participation and individual accountability support each other.

33. Read the mathematical request before treating the difficulty as calculation

A student can understand the individual words in a question and still misread their mathematical relationship. At least, at most, exactly, remaining, twice as much and increased by describe different conditions. A small-group tutor should inspect that interpretation before attributing every wrong answer to algebra. This is not a diagnosis of a language disorder or a claim about a learner’s general reading ability. It is a question about what the particular mathematical sentence requires.

Compare three requests about drawing counters: exactly one red, at least one red and no red. The first excludes two red counters; the second includes them; the third is the complement of at least one red. A learner may calculate probabilities accurately for the wrong event. Ask each student to list the colour sequences included before writing a fraction. The representation makes the interpretation inspectable and allows the tutor to correct the event definition before lengthy calculation.

In an inequality context, a maximum capacity of thirty means a count cannot exceed thirty. If each box holds six objects and there are n boxes with no other objects, 6n ≤ 30 gives n ≤ 5 for non-negative whole n. A student who writes 6n ≥ 30 has reversed the meaning of the condition, not necessarily mishandled inequality operations. The tutor should ask whether a proposed value of six boxes satisfies the real-world statement. Concrete testing can clarify the language before symbolic manipulation.

Question verbs also matter. Simplify an expression, solve an equation, show an identity and explain a conclusion require different outputs. Present the same expression in several requests and ask what would count as a complete response. A numerical value is not always required. A student who keeps trying to isolate x in an identity may be following the most familiar task pattern rather than reading the mathematical object. The group can compare these distinctions without adding more computational difficulty.

A useful peer task is to paraphrase a condition while preserving its mathematical meaning. Another learner checks whether the paraphrase changed the base, direction, range or target. For example, twenty percent of the remaining amount is not the same as twenty percent of the original amount. The tutor should verify the equivalence rather than accept a more conversational sentence simply because it sounds clearer. Clarity must preserve the relationship being described.

The independent check uses a fresh context with a similar logical condition. The learner identifies the event, interval or reference quantity before solving. Keep the wording appropriate to the course; making every sentence longer does not necessarily create a better transfer test. The parent update should name the issue precisely, such as distinguishing at least from exactly, rather than generalise from one mathematical reading difficulty to the child’s entire language ability.

34. Constructing a question can reveal whether a relationship is understood

After students can solve a family of problems, ask them to construct a new one meeting stated conditions. This changes the direction of reasoning. A learner no longer receives all the quantities and applies a method; they must choose quantities that make the intended relationship possible. The task should be bounded. Asking for any difficult question can produce unnecessary complexity, while asking for a linear equation with solution four gives a clear mathematical target.

For example, a student might construct 3x + 5 = 17. Another writes 2(x − 1) = 6. A third writes x² = 16, which has both four and negative four as real solutions. The third example contains the requested solution but not exactly that solution if uniqueness was required. The group can discuss the difference between includes four and has four as its only real solution. Construction has exposed a condition that ordinary calculation might not have revealed.

For a quadratic task, ask for a function with minimum value negative two attained at x = 3. The form a(x − 3)² − 2 works for any positive real a. Students can choose different positive coefficients and explain why each function meets the conditions. A negative coefficient gives a maximum instead. A zero coefficient produces a constant function with the minimum attained everywhere, not only at three. The role of each parameter becomes visible through the constructed examples.

A geometry version asks for two rectangles with equal perimeter but different areas. A 4 by 6 rectangle and a 3 by 7 rectangle both have perimeter twenty, with areas twenty-four and twenty-one. The examples disprove the claim that equal perimeter forces equal area. Ask the learner to verify both conditions rather than rely on a drawing. Another student can construct equal-area rectangles with different perimeters. The contrast makes the two measurements distinct.

The tutor should watch for a learner who creates a problem they cannot solve or whose conditions conflict. That is not automatically failure; it can become a useful discussion of consistency. However, a novice who is still learning the core procedure may need more direct examples before construction is productive. The task should be introduced when it can test a relationship already made accessible. Difficulty should have a reason rather than become an open-ended demand to be creative.

In a small group, peers can test a constructed problem without being told its intended answer. The author then explains the design. Follow with an individual construction or critique so the tutor knows what each learner controls. This gives stronger students depth without automatically accelerating the syllabus and gives all learners practice reading conditions. The useful outcome is a more deliberate grasp of what makes a mathematical question solvable and what a valid answer must satisfy.

35. A correction meeting should produce a new attempt, not a prettier archive

A small group can use returned work efficiently when errors are selected for a shared mathematical reason. The tutor need not discuss every student’s entire school paper publicly. Choose an appropriate common dependency or an invented example representing it. Preserve personal details and avoid turning individual scores into a classroom ranking. The purpose is to teach a relationship that appears in the work, then check how each learner uses it.

Suppose several scripts contain cancellation errors. Present (x + 3)/x with x ≠ 0 and compare it with 3x/x. The first is 1 + 3/x; the second is three. Matching x symbols do not justify cancelling across an addition. Ask the learners to factor the numerator where possible and identify the whole product being divided. A numerical substitution can reject an invalid simplification, while factor structure explains the valid operation. The correction should reveal that reason.

Then return to a more complex expression, such as (x² + 3x)/x for x ≠ 0. Factoring the numerator as x(x + 3) permits simplification to x + 3 on the original domain. This comparison helps the learner see why cancellation is now valid even though addition appeared before factorisation. The lesson should not replace an overused shortcut with a false rule that any numerator containing a plus sign can never be simplified.

The correction record can be small: the invalid operation, the valid replacement and a fresh task that tests it. A full copied solution may still be useful for studying the complete route, but should not obscure what the learner originally attempted. Keep the first work readable. Otherwise, the next review sees only corrected answers and cannot determine whether the same misconception is recurring or whether an entirely different difficulty has appeared.

A changed individual task might ask for simplification of (2x² − 8x)/(2x), with x ≠ 0, giving x − 4. Another asks whether (2x − 8)/(2x) also equals x − 4; it does not. The nearby contrast requires the learner to inspect factor structure again. The tutor should not supply the cancellation cue before the student has attempted it. The point is to see whether the correction has become a decision the learner can initiate.

Close the correction meeting by deciding where the skill will return later. It might reappear in a ratio expression, a trigonometric identity or a rational equation within the actual course. The learner should know which obligation travels across those contexts. A correction book becomes useful when it directs later practice. It becomes an archive when each polished solution is filed away without another opportunity to use the repaired relationship independently.

36. Homework should return information from three separate learners

After a shared lesson, homework need not be identical for all three students. It should remain connected to the taught relationship and manageable enough to review. One learner may need a direct fresh example, another a changed representation and another a mixed decision. The distinction should follow current evidence, not a permanent easy, medium and hard track. A student can need foundational practice in one topic and deeper reasoning in another.

Following a similarity lesson, Alicia might practise choosing the correct direction of a scale factor. Tricia might explain why area uses its square. Kai Kai might identify corresponding sides in rotated diagrams. Each task has a specific job. A later common problem can reconnect the decisions. The tutor should be able to explain why the assignments differ and what would justify reducing or changing them. Different worksheets alone are not evidence of intelligent differentiation.

Keep support conditions visible. A student may consult notes or receive a parent prompt during learning. That should be recorded neutrally. A corrected solution produced after seeing an example is not the same as an unaided first attempt, but it can still help learning. A fresh later item can test independent access. The class should not reward hidden assistance by treating every complete page as stronger evidence than an honest unfinished attempt with a precise question.

A bounded help-seeking note might say: I can identify the two corresponding sides, but I do not know whether to square the ratio because the question asks for perimeter. This tells the tutor exactly where to begin. A parent can encourage that specificity without teaching the answer. The learner should not need to diagnose every cause perfectly. They need enough language to preserve the uncertainty rather than erase it by copying a solution before the next lesson.

Review actual workload rather than page count alone. Two modelling questions can take longer than ten short substitutions. A student who repeatedly spends excessive time may need a smaller task, a clearer explanation or a different checking routine. Increasing volume without reading the cause can make the home problem larger while leaving the learning unchanged. School assignments may already provide a useful application, so duplicated tuition work should have an additional purpose.

The next lesson should use the returned homework. Inspect the original attempt, note the support and choose a fresh check where necessary. If the learner now controls the decision, move it into maintenance rather than repeating the same isolated repair indefinitely. If the dependency remains, change the teaching response. Homework is part of the small-group evidence cycle only when it informs what happens next, not merely when three completed sets are collected.

37. Different school sequences can coexist only when the shared purpose remains clear

Students from different schools may study the same broad course in different orders. A group can sometimes share useful prerequisite work while receiving different applications. It should not become an unexplained second curriculum that conflicts with every learner’s current demands. The tutor needs a concise account of what each school is teaching, what has already been learned and which assessment is approaching. That information helps distinguish current consolidation from preview, repair and optional extension.

For example, one school may be using linear equations in geometry while another is applying them to cost problems. Both learners can benefit from defining variables and preserving equality. The shared lesson can teach that relationship, while the independent tasks reconnect it to each student’s school work. This is coherent differentiation. By contrast, a learner who has not yet encountered a required new idea should not be expected to infer it merely because classmates have already learned it elsewhere.

Different valid methods should be connected rather than framed as a battle between school and tuition. A school may use elimination where a tutor demonstrates substitution. Explain why both preserve the relationships and which is convenient for the coefficients supplied. Follow explicit question instructions when a particular method or form is required. The student should learn to distinguish mathematical validity from a task-specific presentation requirement and from a teacher’s preference.

Use school feedback as evidence without assigning blame. A returned script may reveal an unfamiliar question form, a missing prerequisite or an execution error. It does not by itself prove that school teaching was inadequate or that the student did not try. The useful response is to locate the current decision and teach it. A small group should give the learner a more coherent understanding of the Mathematics, not a growing collection of conflicting adult narratives about who caused the score.

When the topic sequences diverge substantially, review the group. A limited common prerequisite may no longer justify a whole shared lesson if each learner needs sustained instruction elsewhere. The answer may be another group or another arrangement, subject to actual availability. The educational explanation should be specific. A low headcount does not remove the need for a coherent shared task, and the family should not be asked to accept repeated mismatch simply because the class was suitable earlier.

A useful continuity note contains the current school topic, the shared skill being taught and the independent application chosen for the learner. It should be short enough that it supports teaching rather than becomes an administrative project. The student should understand it as well. The purpose is to know why a task is present now and how it connects to the learner’s actual school environment, not to make every setting teach the same worksheet on the same day.

38. Timing should reveal decisions, not turn the class into a speed contest

Three students finishing at different times does not tell the tutor enough about their learning. One may be fluent, another may have omitted a condition, and another may be checking a correct answer repeatedly. A timed activity should investigate a specified behaviour in material the learner has been taught. It should not simply reward the first completed page. Speed has educational value when the Mathematics remains correct and the student can manage the task’s other demands.

Use a short mixed set and observe where time is spent. A long pause before the first equation points towards recognition or representation. Slow but valid transformations may indicate execution fluency. Repeated complete restarts may indicate difficulty trusting or inspecting intermediate work. Multiple identical checks may add little new information. These are different teaching questions. Asking all three learners to work faster ignores the location of the time cost.

For one learner, the useful change may be adding a line. If a signed bracket repeatedly causes a correction, writing the signed products explicitly can save time overall. For another, the change may be removing redundant copying while preserving essential equations. The same goal can therefore require opposite adjustments in working length. A universal demand for shorter solutions may remove exactly the evidence a student needs to stay accurate.

Practise recovery in a controlled way. Include an appropriately demanding item among more accessible questions and discuss whether the route is producing new valid information. A learner leaving the item should retain the last justified relationship and the unresolved target. On return, they begin there rather than reconstruct the entire problem. This is not advice to abandon any question that feels difficult. It is a reasoned decision about using limited time while preserving a route back.

Review timing privately enough that one student’s pace does not become another student’s benchmark. The intended comparison is with the learner’s own prior behaviour under reasonably comparable conditions. Freshness, topic mix and help matter. A faster repeat after seeing the solutions is not a clean test of general improvement. It may be useful rehearsal, but another fresh set is needed to see whether the targeted decision transfers.

End with one action to test next, such as writing the domain before accepting roots, retaining a return point or stopping after an adequate check. A timing session should produce a teachable change rather than only urgency. The tutor-designed practice window is not an official mark allocation or a prediction of examination speed. The student’s work should show whether the chosen adjustment improves useful performance without creating new avoidable errors.

39. A review record should describe the conditions of success

A small-group review should not reduce three learning processes to one class average. Each learner needs a current account of what was attempted, what help was used and what changed. The record can be concise. Its purpose is to choose the next teaching action, not to create an impressive dashboard. A chapter-completion tick says that material was covered; it does not establish that the learner can recognise and use it independently in unfamiliar work.

Consider two fictional records that both say six correct out of eight. In one, the tasks were fresh and unaided. In the other, four correct answers followed a method cue and two followed a visible worked example. Both sessions may contain useful learning. They do not provide the same evidence of independent recognition. The record should preserve the distinction rather than convert every completed correction into a mastery claim.

Use a short statement linking evidence to action. For example: the learner now calculates an area scale factor correctly, but still chooses the direction after a prompt; the next task will reverse the supplied and requested quantities without a cue. Another learner may independently choose the direction but need arithmetic repair. Their follow-ups should differ even if the same number of answers was correct. This is what individual visibility should make possible.

Do not infer a precise long-term trend from a few non-comparable marks. A new school paper may contain unfamiliar topics, while a tuition set may be heavily rehearsed. A higher mark can include real progress without proving that the class alone caused it. A stable mark can hide a repaired error and a new difficulty. Read the script and its conditions alongside the total. The record should improve judgment rather than create false statistical certainty.

Retire resolved repairs from the active list. If the learner has demonstrated a signed operation independently across changed and delayed tasks, maintain it through ordinary applications rather than keep it permanently at the top of a weakness report. Historical errors are not the child’s identity. A current plan should describe what matters now. The learner can participate by explaining which skill feels uncertain and comparing that self-assessment with a fresh attempt.

A review has served its purpose when it changes the next lesson’s first task, the support boundary or the group arrangement. If it produces only praise or criticism, add the next decision. If it is so detailed that nobody uses it, simplify it. A family should leave knowing what the learner can currently do with less help and what evidence will be sought next, rather than only hearing that the class is progressing well.

40. Sometimes the right response is to change the format

A three-student class is one teaching arrangement, not a universal destination. The current learner may need sustained individual explanation, a more compatible group, a short school consultation or no additional programme. These outcomes should remain available in the decision process. The goal is to match support to the mathematical need, not to justify the same format after every piece of evidence. A responsible class-fit review can conclude that the present arrangement should change.

Individual support may be appropriate when a learner cannot yet use any meaningful independent interval because several prerequisites are missing. The tutor might need to reconstruct number relationships, notation and equality together before a shared class becomes useful. This is not a claim that such a learner lacks potential. It is a judgment about the amount and continuity of teaching required now. A temporary arrangement can have a clear next review rather than become a permanent label.

A different group may be appropriate when the mathematical level is suitable but the topic sequence is not. A learner preparing a specialised school assessment should not repeatedly wait through unrelated explanations merely because everyone is the same age. The tutor should identify which requirements cannot be served coherently and discuss actual alternatives. Availability should be confirmed directly, not promised in an article or inferred from an empty seat.

Less support may be appropriate when the learner consistently starts fresh work, retrieves earlier methods, interprets conditions and uses school help effectively. Continuing tuition should have a clear learning purpose. More difficult worksheets are not automatically that purpose if they create an endless reason that the student can never be considered sufficiently independent. A strong learner can need new teaching later without needing the same intensity of support all the time.

Review evidence before changing solely because one lesson felt uncomfortable. A new concept can legitimately require more help; one poor paper can contain a narrow repair rather than a broad format failure. Conversely, repeated mismatch should not be ignored because the family has already invested time. The useful question is whether the present arrangement is changing the current difficulty in a way that survives independent work. That question is more precise than whether small groups are good or bad in general.

When a change is made, carry forward a concise account of secure knowledge, current needs, methods used and support still required. Do not make the student repeat an entire assessment history unnecessarily. The learner should understand why the arrangement is changing and what will be checked afterwards. Good support remains accountable to the work, including when that accountability leads to a different format or a reduction in help.

41. Let students gradually choose useful practice

A learner becoming more independent should increasingly understand why a task is chosen. This does not mean asking a teenager to design an entire curriculum alone. Start with a bounded choice among appropriate tasks. The student identifies one current uncertainty, one older method that needs a return or one changed application to test. The tutor reviews the choice against the work. Planning becomes another capability to teach, not an administrative responsibility suddenly handed over without support.

After a returned ratio task, Alicia might choose another routine total-and-ratio question because it feels comfortable. The tutor can ask whether that addresses the actual error, which concerned a transfer changing both quantities. A more useful choice may be a small transfer problem with easy numbers. Tricia may choose an excessively difficult extension when her current need is an area-unit conversion. The tutor helps align the practice with the observed decision rather than reward difficulty for its own sake.

Use the distinction between familiarity and availability. A student may feel that a topic is secure because the worked examples are recognisable. Ask for a fresh start without notes. If the method does not appear, the learner has useful information for planning a return. That result should not become a broad criticism. It helps calibrate the student’s judgment about what can currently be reconstructed and what still benefits from review.

A practical plan can remain small: one current repair, one delayed return and one mixed application, adjusted to the available time. The learner explains the job of each rather than only the number of questions. The tutor can identify duplication with school homework and remove it when no new learning purpose exists. Self-direction should make the workload more coherent, not create another elaborate schedule that competes with doing the Mathematics.

Teach specific help-seeking alongside independent practice. The student can mark the last justified line, state what the question asks and explain which decision is unresolved. A precise question is compatible with independence. Guessing silently to avoid appearing dependent is not automatically better judgment. The programme should make it normal to distinguish a genuinely new concept from an earlier method that needs retrieval or a local error that can be checked.

The small group can compare practice choices without comparing worth. One learner explains why a counterexample task will test a condition; another why an old method needs a delayed return. The tutor then asks each student to carry out the chosen work individually. The plan earns its value through the resulting evidence. Over time, the learner should need fewer adult decisions about every page while remaining able to seek appropriate teaching for a new or persistent difficulty.

42. Independent task bank: try first, discuss afterwards

The following twelve original tasks can be used selectively after the relevant material has been taught. They are not a standardised assessment, an official paper or a basis for predicting a grade. A total score would conceal the different decisions being sampled. Preserve the learner’s first attempt, note any assistance and open the explanation afterwards. The tutor’s job is to identify which relationship should be taught or checked next, not to use a short bank to label the whole student.

Task 1: The signed coefficient belongs to the whole bracket

Simplify 8 − 3(2 − x), then evaluate the original expression and your simplified form at x = −1. Explain the sign of the x term. The first task tests distribution; the numerical comparison tests whether the chosen input exposes a discrepancy. Do not read a matching single value as proof that every proposed expression is equivalent for all inputs.

Solution and the decision to inspect

Distribution gives 8 − 6 + 3x = 2 + 3x. At x = −1, the original is 8 − 3(3) = −1, and the simplified form is also −1. The positive coefficient of x comes from multiplying negative three by negative x. If the learner writes 2 − 3x, inspect the signed multiplication rather than repeat every rule of algebra. If they simplify correctly but substitute −1 without brackets and change the sign, the current need is substitution control.

A later return can place the same operation inside an equation or a suitable derivative. The tutor should record whether the learner initiates the correct distribution without a prompt. A correct copy after explanation is useful learning, but a fresh unprompted transformation is the evidence needed before reducing that support.

Task 2: The whole determines the ratio fraction

Green and yellow counters are in the ratio 4:7, with fifty-five counters altogether. Find each number. Then add five green counters and state the new ratio. Explain why adding five to the number four in the original ratio is not the same operation as adding five counters to the collection.

Solution and the decision to inspect

Eleven ratio parts represent fifty-five counters, so one part is five. Green is twenty and yellow is thirty-five. After adding five green, the quantities are twenty-five and thirty-five, giving ratio 5:7. The number four represented four equal parts, not four actual counters. A learner who writes 9:7 has mixed a ratio description with a quantity change. A learner who computes fifty-five times 4/7 has treated the total as the yellow quantity.

The next task can reverse the information or change addition into transfer. Keep the quantities small enough that the model remains the main demand. The tutor is checking reference quantities and conservation, not only the ability to simplify a fraction at the end.

Task 3: Reversing a percentage requires the original base

A hypothetical price is reduced by fifteen percent and becomes 68 currency units. Find the original price and verify the reduction. Explain why adding fifteen percent of sixty-eight does not recover the original. This is an arithmetic example, not a current price, tax rate or financial recommendation.

Solution and the decision to inspect

The remaining amount is eighty-five percent of the original, so 0.85P = 68 and P = 80. Fifteen percent of eighty is twelve, and eighty minus twelve is sixty-eight. Adding fifteen percent of sixty-eight uses a different base and gives 78.2, not eighty. The relevant repair is identifying which amount the percentage refers to. A learner who knows how to multiply decimals may still need this relationship explained.

A changed check can give a percentage increase instead of a decrease. Ask the learner to write the multiplier before solving. The new task should test the base again without relying on the word discount as an automatic instruction to subtract.

Task 4: Two equations should preserve both conditions

Solve 2a + b = 13 and a − b = 2. Choose a route and explain what the first operation achieves. Check the result in both equations. Then decide whether doubling the first equation supplies a new independent condition or only another form of the same information.

Solution and the decision to inspect

Adding the equations gives 3a = 15, so a = 5 and b = 3. The first operation eliminates b because its coefficients are opposite. Both original equations hold: ten plus three is thirteen, and five minus three is two. Doubling the first equation gives 4a + 2b = 26, which is equivalent to it and does not add an independent restriction. A learner should not assume that more written equations always mean more information.

An alternative substitution route is valid. Evaluate it by its justification and execution rather than whether it matches the tutor’s preferred first step. A fresh pair with different coefficients can reveal whether method choice remains deliberate.

Task 5: A fixed component changes a doubling claim

A hypothetical model is C = 9 + 2n for non-negative whole n. Calculate C when n = 4 and when n = 8. Does doubling n double C? Explain the intercept and rate in a context you invent, keeping the quantities and their units consistent.

Solution and the decision to inspect

The outputs are seventeen and twenty-five, so the second is not double the first. The fixed component nine is included once in each case, while the variable component 2n doubles. A suitable invented context might charge nine units per order and two per item. The intercept represents the fixed component and the coefficient two represents the increase per item. The mathematical model does not establish that a real provider uses those charges.

A later comparison can use C = 2n, where doubling does double the output. The learner should identify the condition that makes proportional reasoning valid rather than attach a universal doubling rule to every straight-line relationship.

Task 6: Area ratios reverse through a square root

Two similar figures have areas in the ratio 25:64. A corresponding side in the smaller figure is fifteen centimetres. Find the side in the larger figure and explain why using the area ratio directly on the length would be inappropriate. State the correspondence before calculating.

Solution and the decision to inspect

The positive length ratio is 5:8, so the larger side is 15 × 8/5 = 24 centimetres. Area involves two length dimensions, so its ratio is the square of the corresponding length ratio. Multiplying fifteen by 64/25 would use the area factor on a length. A learner who chooses 5/8 instead may understand the square-root relationship but reverse the source and target. Those need different follow-ups.

The fresh return can supply a volume ratio for a suitable course, or rotate the figures while preserving labels. Choose one additional demand at a time so the tutor can distinguish dimensional reasoning from correspondence and arithmetic.

Task 7: An average must account for unequal group sizes

Six observations have mean twelve and nine observations have mean seventeen. Find the mean of all fifteen observations. Before calculating, predict whether the answer should be nearer twelve or seventeen and explain why. The prediction is a reasonableness check, not the exact calculation.

Solution and the decision to inspect

The totals are seventy-two and 153, giving 225 across fifteen observations and a combined mean of fifteen. It is closer to seventeen because more observations belong to that group. Averaging twelve and seventeen equally would give 14.5 and weight the two groups as if they contained equal numbers. The learner should reconstruct totals or use equivalent weights, with a denominator representing individual observations rather than the number of groups.

A reverse task can supply the combined mean and ask for a missing group mean. That checks whether the relationship is understood in both directions. A correct use of a memorised formula in the forward example alone does not establish that broader control.

Task 8: Exactly one success has two possible orders

A bag contains four white counters and two black counters. Two are drawn without replacement. Find the probability of exactly one black counter. Identify the favourable orders before multiplying probabilities, and explain why the denominator for the second draw differs from that for the first.

Solution and the decision to inspect

The orders are black then white and white then black. Their probabilities are (2/6)(4/5) and (4/6)(2/5), each 4/15, giving total 8/15. Only five counters remain for the second draw. A learner who includes one order only has omitted part of the event. A learner using denominator six again may have assumed replacement. A learner adding the original colour proportions without considering sequences has modelled another event.

The next task can change only the replacement rule or ask for at least one black. Keep the event definition visible. The purpose is to teach how the sampling process and request determine the calculation, not merely repeat a tree diagram after the tutor has already chosen every branch.

Task 9: A restricted domain can exclude the turning point

Find the minimum of x² − 6x + 13 for all real x. Then find its minimum when 0 ≤ x ≤ 1. Explain why the same expression gives different answers under the two domains. Use a completed square, a graph or another justified route appropriate to what has been taught.

Solution and the decision to inspect

The expression is (x − 3)² + 4, so its unrestricted minimum is four at x = 3. On the restricted interval, the allowed value closest to three is one, giving minimum eight. The turning point does not belong to that interval. A student who reports four twice may have completed the square correctly but ignored admissibility. A student whose transformed constant is wrong needs a different repair before domain interpretation can be judged fairly.

A changed task can use a negative leading coefficient and ask for a maximum. The learner should explain the sign and attainable input rather than memorise that the outside number always answers the requested extremum.

Task 10: The tangent needs both a location and a direction

For y = x² − 2x + 5, find the tangent at x = 3 after the relevant calculus has been taught. Write the curve point separately from the gradient. Explain what substitution into the final line checks and what it does not independently verify.

Solution and the decision to inspect

The point is (3, 8). The derivative is 2x − 2, so the gradient there is four. The tangent is y − 8 = 4(x − 3), or y = 4x − 4. Substituting x = 3 into that line gives eight and checks that it passes through the point. It does not independently establish that the derivative was calculated correctly. A learner using (3, 4) has confused the derivative output with the original function output.

A fresh reverse task can supply the gradient and ask for the point first. The tutor should inspect the information flow rather than assume that a correct routine tangent means every calculus-to-geometry connection is secure.

Task 11: Equal distances do not mean equal time weights

A traveller covers twelve kilometres at six kilometres per hour and another twelve kilometres at twelve kilometres per hour. Find the average speed over the journey, excluding any rest because none is stated. Then explain how the answer would change if the traveller spent one hour at each speed instead.

Solution and the decision to inspect

The equal-distance stages take two hours and one hour, so average speed is twenty-four kilometres divided by three hours, or eight kilometres per hour. With one hour at each speed, the distances are six and twelve kilometres, giving eighteen kilometres in two hours and average speed nine. The difference comes from the duration weights. A learner who averages the speeds in both cases has not identified which quantity is held equal.

The next check should ask for total distance and total time before a combined rate. That makes the denominator visible and prevents a formula from being applied without reference to the measurement the question actually requests.

Task 12: Construct an example meeting two conditions

Construct five non-negative numbers whose mean is eight and median is six. Show that both conditions hold. Then explain why specifying a mean alone would leave many possible data sets. This task tests construction and verification rather than a single routine calculation with all values already supplied.

One solution and the decision to inspect

One valid set is 2, 4, 6, 10, 18. The sum is forty, giving mean eight across five numbers, and the middle ordered value is six. Many other sets work. The total condition constrains the sum but does not determine every observation. The median adds an ordering condition. A learner should verify both rather than assume that matching the total automatically produces the required middle value.

A peer can offer a different valid set and compare its range. Follow with a fresh individual construction so the tutor knows whether the learner can satisfy the conditions without borrowing the class example. An open answer does not remove the need for mathematical checking; it makes the conditions especially important.

After using the selected tasks, do not stop at a total. Identify the first meaningful uncertainty in each learner’s work and decide what new attempt would clarify it. The same answer can reflect different reasoning, and different forms can be equally valid. Record the help supplied and preserve the original attempt. The bank is useful when it sharpens the next lesson, not when it becomes another set of marks detached from a teaching decision.

43. A family-readable agreement: what the small group is for now

The family should be able to describe the present purpose of the class in one clear sentence. The student may be learning to form equations independently, preserve algebraic conditions, interpret geometry or use known methods under time. That purpose can change. A broad ambition to improve Mathematics is a starting point, not a sufficient weekly plan. The class should connect the ambition to a current decision visible in the learner’s work.

State what evidence supports the priority. A returned paper, a fresh first attempt and a short explanation can together identify a useful next task. Do not turn one error into a permanent profile. Name secure knowledge too. A learner who forms an equation correctly but loses a sign should know that their model was valid. Preserving that strength prevents correction from becoming an instruction to distrust everything they did.

State how the shared and individual parts fit together. The group may compare representations, while each learner receives a different application or amount of support. The learner should understand why that is not a ranking of personal worth. The task differs because the current mathematical need differs. Later evidence can change both the assignment and the amount of help. A good group does not trap students in fixed roles based on the first lesson.

State what counts as a useful check. It may be a fresh task without the method being named, a delayed return after other work or an explanation of an original condition. Completing a corrected page is a learning stage, not the only standard. Help should be recorded without blame. The purpose is to know which decisions the student can now carry and which still belong partly to the explanation they have just received.

State how workload and fit will be reviewed. School demands, course sequence and the learner’s independent working intervals may change. A small group should remain educationally coherent and practically manageable. Current fees, times, absence arrangements and teacher availability require direct confirmation. A family can understand these practical conditions without being asked to accept a guaranteed grade or a claim that headcount alone proves value.

Finally, state the direction of travel: the learner should increasingly make the important mathematical decisions before someone else supplies them. This does not mean refusing help or never making an error. It means having more usable ways to begin, explain, check and recover. A three-student class is worthwhile when its close observation and shared reasoning both serve that independence, and when the programme is willing to change whenever the student’s work indicates a better next step.

Sources and the next reading route

Teaching and evidence: EEF small-group tuition review; National Student Support Accelerator design principles; Nickow, Oreopoulos and Quan’s tutoring research synthesis; IES instruction and study guide; IES algebra teaching guide. Research findings concern their studied settings, not guaranteed results for this class.

Course information: MOE Full SBB announcements; SEAB SEC overview; the official G1, G2 and G3 2027 syllabus listings; IB Diploma Mathematics. Always confirm the learner’s actual course and examination year.

Continue in the eduKate library: choosing mathematical support; the Mathematics programme guide; enrolment and class fit; Secondary 3 A-Math; Secondary 4 A-Math. For the bounded local specialist service, see Bukit Timah Tutor’s small-group Mathematics page.

Ask About Current Bukit Timah Mathematics Classes

Bring the student’s current course, recent original work and next assessment. Remove unnecessary personal identifiers from shared work. The first decision is whether a suitable class can address the observed learning need. Confirm current availability, fees and practical arrangements directly; this guide does not promise a place or a particular result.