How Self-Explanation Improves Mathematics | A Worked-Example Workbook for Reasoning, Error Repair and Independent Problem Solving

Self-explanation improves mathematics when a learner explains why a step is allowed, what its quantities mean and when the method would stop working. In this worked-example workbook, mathematical reasoning becomes visible through fractions, ratios, percentages, algebra, geometry, probability and selected Additional Mathematics problems. The purpose is better maths problem solving, not longer answers: a useful explanation identifies the relationship that makes the calculation valid, then helps the learner solve a different problem without copying the example.

For primary and secondary students, learning mathematics can become a frustrating gap between following a teacher and working independently. A correct model answer may look obvious until its numbers, diagram or wording change. The exercises here address that gap with self-explanation prompts, fully worked examples, common mathematical mistakes, corrected reasoning and new questions with answer discussions. Parents and teachers can select the relevant level rather than taking a younger learner through the entire workbook.

Self-explanation is a promising learning strategy, but asking for explanations is not automatically better than providing clear instruction. A meta-analysis by Bisra and colleagues found benefits across varied learning conditions; mathematics research also shows that the design of the example and prompt matters. This workbook therefore uses a deliberately narrow principle: explain the important decision, check its accuracy, and then remove the support. Its original exercises are teaching proposals, not a programme whose effectiveness has been experimentally established.

Your 50-second route through this workbook

Following examples but getting stuck alone? Begin with the explanation test, then choose one worked example and attempt its changed question without looking back. Making repeated number mistakes? Use fractions, percentages or rates. Losing algebra marks? Start at brackets and negative signs and continue to equations, inequalities and domain restrictions. Knowing formulas but choosing the wrong one? Use the geometry, probability and representation chapters. Teaching or supervising? Read the evidence caution before adding more explanation prompts.

The smallest useful routine is: attempt one question, identify the first uncertain decision, explain that decision in ordinary language, check it against a trustworthy model, and solve one altered question. Stop explaining steps that are already secure. Return later to find out whether the idea remains available without the explanation prompt.

Open the workbook contents

1. What an explanation must accomplish · 2. Evidence and limits · 3. Equality · 4. Fraction units · 5. Multiplying and dividing fractions · 6. Ratios · 7. Percentages · 8. Rates and units

9. Average speed · 10. Distribution and negative signs · 11. Linear equations · 12. Simultaneous equations · 13. Inequalities · 14. Quadratics and lost solutions · 15. Algebraic fractions and domains · 16. Functions and inverses

17. Indices and logarithms · 18. Area, perimeter and scale · 19. Similarity · 20. Pythagoras · 21. Coordinates and gradients · 22. Averages and data · 23. Probability · 24. Differentiation

25. Optimisation · 26. Representing word problems · 27. Repairing an explanation · 28. A mixed transfer assessment · 29. Teaching without overprompting · 30. A seven-day practice design · 31. Supporting different learners · 32. Questions, teaching guide and next routes

All named learners, classroom exchanges, exercise situations and sample responses below are fictional teaching examples. A sample mistake is not a description of a particular child. The numerical questions were written for this workbook; they are not copied examination questions or official mark schemes. The chapters move from elementary quantities to advanced topics, and some later sections require algebra or calculus that a younger learner has not yet studied.

1. The explanation test: why this step, not merely what happened?

Consider two answers to the same question: why did you divide both sides of 4x = 28 by four? The first says, “Because four goes to the other side.” The second says, “Both sides represent equal quantities. Dividing each by the same nonzero number preserves their equality, and four equal groups of x become one group.” Both learners may write x = 7. Their explanations nevertheless reveal different resources for the next problem. The second explanation names a relationship that also works when the coefficient is a fraction or the variable appears in a less familiar position.

A useful self-explanation answers one of three questions. What does this quantity represent? What mathematical property permits this transformation? What condition must hold for the method to apply? These questions are narrower than “Explain everything.” They invite a learner to identify the load-bearing decision. In a percentage question, that decision may be the reference amount. In a triangle question, it may be the right angle. In an equation, it may be the operation that preserves the set of possible solutions. Naming every arithmetic action can leave the decisive condition entirely unmentioned.

Try this diagnostic without coaching. A learner writes 3(x + 2) = 18, then x + 2 = 6, then x = 4. Ask why the first transformation is valid. “I divided by three” describes the action. “The whole left side contains three copies of x + 2, so dividing both equal sides by three leaves one copy” explains its structure. Now show 3x + 2 = 18. The learner should not divide only 3x and 18 while leaving the two untouched. A familiar-looking coefficient does not license changing selected pieces of an equation.

The contrast also shows why a correct answer is insufficient evidence of a secure explanation. In the first question, a memorised sequence can happen to work. In the second, the same sequence produces a false equation if applied selectively. Ask the learner to point to the complete expression affected by the operation. A box around 3(x + 2), followed by a box around the whole right side, may communicate more clearly than a paragraph of polished prose. The target is a correct mathematical relationship, not a performance of academic vocabulary.

For an independent check, use 5(y − 1) = 30 and 5y − 1 = 30. The answers are y = 7 and y = 31/5 respectively. In the first, dividing both sides by five gives y − 1 = 6. In the second, adding one gives 5y = 31. Dividing the complete second equation by five is also valid, but it gives y − 1/5 = 6, not y − 1 = 6. Two methods can both be legal; the explanation must account for every term affected by the chosen operation.

Use the student’s response to choose the next action. An accurate but slow explanation may need practice. An explanation that names the operation but not its scope may need a contrasting example. An explanation containing a false rule needs correction before more independent practice. Silence may mean uncertainty, limited language for describing a known idea, or simply an unfamiliar request. Offer a diagram or a choice between two explanations before concluding that the mathematics itself is absent. This is an informal teaching decision, not a validated diagnostic test.

Finally, ask a different question later without saying, “Remember to explain.” Does the learner still respect the structure? The best outcome is not permanent dependence on being asked why. It is a learner who notices a risky transformation and checks it spontaneously. This workbook develops that applied habit. For the broader learning mechanism, use eduKate’s existing How Self-Explanation Works; the present article supplies the mathematical cases in which the habit can be practised and inspected.

2. What the evidence supports—and what it does not

The phrase “self-explanation works” is too broad to design a lesson. A learner can explain a correct example, justify an incorrect solution, repeat the teacher’s sentence, infer a missing relationship or produce an elaborate misconception. These activities are not interchangeable. Rittle-Johnson, Loehr and Durkin’s mathematics-focused meta-analysis examines self-explanation alongside instructional design. Its relevance here is the need to specify what students are explaining and what support makes that explanation mathematically useful, rather than treating any extra talk as an improvement.

There is also a caution that should change practice. Barbieri and colleagues’ 2023 meta-analysis of worked examples in mathematics reported an overall worked-example benefit, while the presence of self-explanation prompts was associated with smaller effects in its moderator analysis. That comparison does not establish that every prompt harms every learner. It does mean that adding prompts cannot be assumed to improve an already helpful example. More instructional activity can impose costs as well as supply information.

Imagine a learner who has not understood what a denominator measures. Asking for a lengthy account of common denominators may invite guessing. The more appropriate first move is a clear model of equal-sized parts. Once the learner can identify the unit, a brief explanation can check whether that model has been understood. By contrast, asking an experienced student to explain every routine multiplication may interrupt productive work. The proposed rule in this workbook is therefore selective: request an explanation where a decision is uncertain, consequential or commonly misapplied, and reduce the request when that decision becomes secure.

Worked-example programmes provide useful design illustrations without making this workbook a copy of them. The SERP Institute’s MathByExample materials distinguish explaining why from merely reporting what a solution does. Their teacher guidance also recognises that students may need support expressing explanations. The practical lesson is to accept mathematically adequate everyday language. “These are thirds of the same whole” can be a more informative response than an impressive sentence that never identifies the unit.

This article is not claiming that its exact questions, order or seven-day schedule have been tested in a trial. They are an original instructional design informed by the literature and by the mathematical relationships made explicit in the examples. Any local evaluation should examine what learners can do later, not merely whether they complete the pages. A student who writes better explanations but still selects the wrong method on an unlabelled question has made one kind of progress while leaving another problem unresolved. Record both outcomes rather than allowing the first to conceal the second.

Choose a comparison that is fair enough to guide teaching. Keep an initial independent answer, offer the selected example and prompt, then use a new question requiring the same relationship. Return after a delay with no reminder. The new question should not be harder in every other respect: otherwise a vocabulary burden or complicated arithmetic may hide the concept being tested. Conversely, simply changing 12 to 14 can leave all the original cues intact. Change one important surface feature while preserving the principle, and explain what your small classroom check can and cannot establish.

There are clear stopping conditions. Stop an explanation exercise when the student is rehearsing a false rule without correction, when the language demand overwhelms the mathematical objective, or when the same secure step is being described repeatedly without new information. Supply instruction, simplify the representation, or move to independent practice. The official IES guide on organising instruction and study provides a broader evidence-informed context for examples, questions and review. No single strategy should replace the teacher’s responsibility to inspect the learner’s actual response.

3. Equality: protect the relationship, not the appearance of the line

An equals sign says that the expressions on its two sides have the same value. It is not merely a signal that the answer comes next. This distinction matters even before formal algebra. In 8 + 4 = □ + 5, the blank is seven because both sides must total twelve. A learner who enters twelve in the blank may be treating the equals sign as an instruction to calculate what precedes it. Rather than repeating “That is wrong,” ask the learner to calculate the entire right side after the proposed answer has been inserted.

The incorrect right side becomes 12 + 5 = 17. The mismatch is visible without a lecture. Now ask, “What must the box and five make together?” The phrase “make twelve together” supplies the needed relationship. The calculation 12 − 5 = 7 follows from that relationship rather than from a positional rule about boxes. For another representation, place twelve counters on one side of an imaginary balance and a covered group plus five counters on the other. Removing five from both sides leaves seven to match the covered group.

The same idea protects later algebra. From 2x + 5 = 17, subtracting five from both complete sides gives 2x = 12. Dividing both sides by two gives x = 6. These are equivalent equations because each step preserves precisely the same possible value of x. The language of “moving five across” may be convenient shorthand for someone who understands the transformation, but it should not become the reason the transformation is believed. Ask what happened to both sides, including the term that disappeared when five was subtracted from itself.

A useful wrong example is 2x + 5 = 17, followed by x + 5 = 8.5. The student has divided selected terms by two while leaving five unchanged. Dividing the whole equation is legal, but the left side must become x + 2.5. There is no requirement that the most familiar method be the only accepted method. What matters is a complete transformation. A student who chooses the less convenient but valid division route should be asked to finish it accurately, not told that mathematics permits only one sequence.

Try three transfer questions: 15 = 9 + □; 7 + □ = 10 + 4; and 3a + 6 = 24. The missing values are six, seven and a = 6. The first reverses the usual position of the numerical answer. The second places operations on both sides. The third replaces a blank with a letter and groups several equal copies. Ask what is unchanged across the tasks. The answer is not their layout but their demand that two complete expressions represent the same quantity.

Now introduce a statement with more than one equals sign: 6 + 3 = 9 + 4 = 13. This chain is false because 6 + 3 equals nine, while 9 + 4 equals thirteen. A learner may intend to record two successive calculations. Separate them as 6 + 3 = 9 and 9 + 4 = 13, or write a sentence describing the next action. The explanation should distinguish a sequence in time from equality of value. Neat notation is not cosmetic here; the notation makes a mathematical claim.

For a final independent check, show 4b − 3 = 21 and ask for two legal first steps. Adding three gives 4b = 24. Dividing everything by four gives b − 3/4 = 21/4. Both lead to b = 6. The learner need not choose both in ordinary work, but recognising both reveals the underlying principle. A strong explanation can be one sentence: “I changed both complete sides in the same reversible way, so the value that makes them equal remains the same.” Later chapters will inspect cases where an apparently similar operation is not reversible without additional conditions.

4. Adding fractions: explain the unit before adding the count

Three apples and two apples make five apples because the counted unit is the same. Fraction addition has the same requirement, although the unit is a part of a whole rather than a whole object. One third and one sixth cannot be combined by adding the numerators while leaving their units unexplained. A third is twice the size of a sixth when both refer to the same whole. Rewrite one third as two sixths; two sixths plus one sixth make three sixths, which is one half.

For 2/3 + 1/4, use twelfths as a shared unit. Each third contains four twelfths, so two thirds contain eight twelfths. Each quarter contains three twelfths. The sum is eleven twelfths. The familiar multiplication of numerator and denominator is not an arbitrary ritual: multiplying both by the same positive integer partitions each existing part into smaller equal pieces without changing the quantity represented. The denominator changes the name and size of the part; the numerator changes the number of such parts correspondingly.

Inspect the wrong answer 3/7. The learner may have added two to one and three to four. Ask what kind of part a seventh would represent and why adding thirds to quarters should suddenly create sevenths. A rough size check also exposes the problem: two thirds is already larger than one half, so adding a positive quarter cannot produce a result smaller than one half. The correction therefore has two supports, a unit explanation and an estimate. Either can warn the learner before a detailed calculation is complete.

Do not silently assume that every stated fraction uses the same whole. Half of a small cake and a quarter of a much larger cake cannot be combined as three quarters of one cake without defining a common size. If the small cake weighs 400 grams and the large cake weighs 800 grams, the portions weigh 200 grams each, for a total of 400 grams. Fractions describe relationships to reference wholes. A diagram with equal-looking bars can mislead when the real wholes differ. Ask “A fraction of what?” before operating on the symbols.

Practise with 3/4 − 1/6. Twelfths give nine twelfths minus two twelfths, or seven twelfths. Explain why the answer should be less than three quarters but greater than one half: subtracting one sixth from three quarters leaves approximately 0.583. A decimal is not necessary to justify the exact fraction, but a benchmark comparison can help inspect its plausibility. Now try 5/8 + 1/4. Rewriting one quarter as two eighths gives seven eighths. The denominator need not be the product of the original denominators; any suitable common unit works.

A second transfer task changes the format. A container is three fifths full. An additional amount equal to one tenth of the container’s capacity is poured in, with no spillage. It becomes seven tenths full. The phrase “of the container’s capacity” is important because it fixes the whole. If instead the extra amount were one tenth of the liquid already present, the calculation would differ: one tenth of three fifths is three fiftieths, and the new amount would be thirty-three fiftieths of capacity. The words determine the quantity model.

Ask the learner to finish this explanation in their own form: “I can add the numerators after I have…” A useful completion is “expressed the amounts in equal-sized parts of the same whole.” “Made the bottoms the same” describes an appearance but can conceal the reason. Accept a carefully labelled strip diagram instead of a sentence. Then remove the diagram and give one new fraction problem later. If the learner only remembers which numbers to multiply, return to the shared unit; if the unit is secure, spend the next practice period on independent accuracy rather than another long explanation.

5. Multiplying and dividing fractions: identify what the operation asks

The statement “multiplication makes numbers bigger” is not a reliable rule. Multiplying a positive quantity by a factor between zero and one selects a fraction of it, making it smaller. Two thirds of three quarters of a metre is one half of a metre. The calculation is (2/3) × (3/4) = 6/12 = 1/2. A rectangle divided into twelve equal small parts can represent the overlapping selection. The operation is not asking for two thirds plus three quarters; it is asking for a specified portion of an existing portion.

Cancellation should also have a reason. In (2/3) × (3/4), the factors of three in numerator and denominator can be divided by three, leaving 2/4. This is division of the entire product’s numerator and denominator by the same nonzero factor. It is not permission to cross out matching digits wherever they appear. The contrast with 2/3 + 3/4 matters: the expression contains a sum, not one combined product. A cancellation rule detached from structure can produce a correct answer in one layout and a false answer in the next.

For division, ask what is being counted. Three quarters of a litre divided into servings of one eighth of a litre gives six servings: (3/4) ÷ (1/8) = 6. Express three quarters as six eighths. There are six one-eighth units in that amount. This explanation makes sense before the reciprocal algorithm appears. It also shows why division can produce a number larger than the starting numerical value. The answer counts small units; it is not a smaller amount of liquid. Keeping the answer’s unit visible prevents a misleading size rule.

Now consider (2/3) ÷ (4/5). Let q be the quotient. Then q × 4/5 = 2/3. Multiplying both sides by 5/4 gives q = (2/3) × (5/4) = 5/6. The reciprocal appears because 4/5 multiplied by 5/4 equals one. This is the structural reason for “invert and multiply.” It does not apply to division by zero, which has no multiplicative inverse. A short explanation of the inverse relationship is more durable than treating the reversal of numerator and denominator as a mysterious command.

Compare two word problems using the same numbers. “Find three quarters of eight” gives six. “How many three-quarter units fit into eight?” gives 32/3, or ten and two thirds. The first selects an amount; the second measures an amount in units of size three quarters. A learner who sees a fraction and a whole number but chooses an operation before interpreting the relationship may calculate accurately and answer the wrong question. Ask for a rough verbal prediction before arithmetic: are we taking a part, combining amounts, sharing equally or counting how many units fit?

Use these independent questions. First, find five sixths of three fifths; the answer is one half. Second, divide seven eighths by seven sixteenths; the answer is two, because seven eighths equals fourteen sixteenths. Third, a ribbon of length 9/4 metres is cut into pieces of length 3/8 metre; exactly six pieces are possible. Fourth, a recipe uses 3/4 cup of flour for one batch. Five half-sized batches use (5/2) × (3/4) = 15/8 cups. Require units in the ribbon and recipe answers.

A useful error discussion begins with the statement “I divided by a fraction, so the answer must be smaller.” Ask the learner to test it with 1 ÷ 1/2. Two halves fit into one, so the proposed rule fails. Replace it with a conditional statement: dividing a positive number by a positive number below one increases its numerical value; dividing by a number above one decreases it. The conditions matter. The goal is not to memorise a more complicated slogan, but to understand the unit-counting relationship well enough to predict a sensible answer before executing the algorithm.

6. Ratios: distinguish the relationship from the total

In a box, the ratio of red counters to blue counters is 2:3. This says that each two red parts correspond to three blue parts. It does not say that two thirds of all counters are red. The total contains five ratio parts, so red counters make up two fifths of the total. If there are forty counters altogether, each ratio part represents eight counters. There are sixteen red and twenty-four blue. The calculation depends on identifying whether a given quantity describes one category, another category or their combined total.

The most useful explanation may therefore precede the division. Ask, “Why are you dividing forty by five?” A structural answer is, “The forty counters correspond to the combined two-plus-three ratio parts.” “Because you add ratios” is too imprecise; some ratio questions do not give a combined total. If the problem instead says that there are forty red counters, forty corresponds to two parts, each part is twenty, and there are sixty blue. The same ratio is present, but the mapping from real quantity to ratio parts has changed.

A labelled bar model can expose the mapping. Draw two equal boxes for red and three equal boxes for blue. Place the given number beneath the exact group it measures. The location of forty is now mathematically consequential. The drawing is not merely decorative support for a calculation already chosen. It prevents the learner from applying a memorised “total divided by sum” routine to a number that is not a total. Once the mapping is secure, the learner may use equations or mental scaling instead; the diagram need not become a permanent requirement.

Ratios remain unchanged under multiplication by a common positive factor, not under addition of a common amount. Starting with two red and three blue counters, adding one of each gives 3:4, not 2:3. Multiplying both counts by two gives 4:6, which preserves the relationship. The difference is visible in fractions: 2/3 and 4/6 are equal, while 3/4 is different. Ask the learner to explain whether a real action scales both categories or changes them by a fixed amount. This distinction becomes important in mixture and age problems.

Consider a changed-state problem. A box begins with red:blue = 2:3 and contains thirty counters. There are twelve red and eighteen blue. Six red counters are added, giving eighteen red and eighteen blue, so the new ratio is 1:1. It would be wrong to add six directly to the ratio number two and report 8:3; the original ratio part represented six counters, not one. The ratio numbers describe a scale-free relationship. An actual addition must be interpreted at the real scale before the new ratio is simplified.

For independent practice, use three variants. A:B = 3:5 and the total is sixty-four: A = 24, B = 40. A:B = 3:5 and B exceeds A by eighteen: the difference of two parts represents eighteen, so A = 27 and B = 45. A:B = 3:5 and A is thirty: A’s three parts represent thirty, so B = 50. The same ratio and familiar numbers cannot decide the operation. The learner must first say which collection of parts corresponds to the given amount.

Finish with a boundary question: what would you need to know to turn 3:5 into actual counts? Some real-scale information is required, such as the total, one category or a difference. The ratio alone permits many possibilities: three and five, six and ten, thirty and fifty. This is a good place to distinguish insufficient information from inability. A mathematically competent answer can be “There is not enough information for a unique count,” supported by two valid examples. Explanations should protect that judgement rather than train students to manufacture a number whenever a question contains numerals.

7. Percentages: name the reference amount before using the rate

A percentage is a proportion expressed per hundred, and its meaning depends on the reference quantity. Twenty per cent of eighty is sixteen because 0.20 × 80 = 16. The question becomes less routine when the reference amount is unstated or changes between steps. A jacket costs eighty dollars and is discounted by twenty per cent. Its sale price is sixty-four dollars. Returning from sixty-four to eighty is an increase of sixteen dollars, but that increase is twenty-five per cent of sixty-four, not twenty per cent.

Ask why the two percentages differ when the dollar difference is the same. The answer is that the denominators differ: 16/80 = 0.20, while 16/64 = 0.25. The calculation is a comparison to a base, not an intrinsic property of sixteen dollars. A learner who says “The discount and increase cancel” may be using equal-looking percentage labels without attending to their reference quantities. Write the base beside every percentage during early practice. Once that habit is secure, the notation can become more compact.

Reverse percentages are another test of the reference amount. A sale price of seventy-two dollars follows a twenty per cent discount. Seventy-two represents eighty per cent of the original price, so the original is 72/0.8 = 90 dollars. Adding twenty per cent of seventy-two gives 86.40 dollars, which does not restore the original base. The self-explanation should identify what fraction of the unknown original remains. A simple equation, 0.8P = 72, makes the relationship explicit and avoids the misleading idea that reversing a percentage change means applying the opposite sign to the same percentage.

Successive changes multiply their factors. An amount rises by ten per cent and then falls by ten per cent. Starting from one hundred gives one hundred and ten, followed by ninety-nine. Algebraically, the final amount is 1.1 × 0.9 = 0.99 times the original. The net decrease is one per cent. A concrete starting amount makes the changing bases easy to see, but the factor calculation shows that the conclusion is not peculiar to one hundred. Ask the learner to explain why the second ten per cent is calculated from the enlarged amount.

Distinguish percentage change from percentage-point change. In a fictional class survey, the share choosing one option rises from forty per cent to fifty per cent. That is an increase of ten percentage points. Relative to the original forty per cent share, it is a twenty-five per cent increase. Neither wording is automatically preferable; they answer different questions. The explanation must state whether the comparison subtracts two percentages or expresses the difference relative to the initial percentage. This distinction is useful when reading charts as well as when solving school arithmetic.

Try these questions without the model. A price rises from sixty to seventy-five dollars: the increase is twenty-five per cent. A quantity is reduced by thirty per cent to forty-nine: the original is seventy. A value increases by twenty per cent and then by twenty-five per cent: its final factor is 1.2 × 1.25 = 1.5, a fifty per cent increase. A test score rises from twelve to fifteen out of twenty: the score increases by three marks, by fifteen percentage points of the maximum, and by twenty-five per cent relative to the original twelve marks.

A good final explanation is short: “I divided the change by the original amount because the question asks for change relative to that starting amount.” For a reverse problem, it becomes: “The stated result is this fraction of the unknown original, so I divide by the remaining factor.” Keep those two questions separate during correction. A learner who knows the arithmetic but selects the wrong base needs representation practice, not another page of multiplying decimals. For additional quantitative context, continue through the eduKate mathematics system rather than treating percentages as an isolated trick.

8. Rates and units: let the units help choose the calculation

A rate compares quantities of different kinds. If a printer produces ninety pages in three minutes at a constant rate, its rate is thirty pages per minute. Dividing ninety pages by three minutes produces the required unit, pages/minute. If the printer operates for eight minutes at that rate, multiplication gives 240 pages. The minutes cancel in 30 pages/minute × 8 minutes. This cancellation is not merely a notation trick; it shows that the calculation converts a duration into the corresponding output under the stated constant-rate assumption.

Units can reject an implausible operation before numerical work begins. Dividing thirty pages per minute by eight minutes produces pages per minute squared, not a number of pages. The calculation could have meaning in a different context, but it does not answer this output question. Ask the learner what unit the answer must have and then inspect whether the planned operation produces it. This method does not replace understanding the situation: two formulas can have the same dimensions while describing different relationships. It is a useful check, not an infallible method generator.

For speed, convert time consistently. A cyclist travels twelve kilometres in forty minutes. Forty minutes is two thirds of an hour, so the speed is 12 ÷ (2/3) = 18 kilometres per hour. Dividing twelve by forty gives 0.3 kilometres per minute, which is also correct if the unit is retained. Multiplying by sixty then gives eighteen kilometres per hour. The wrong answer “0.3 kilometres per hour” is not an arithmetic failure; it is a mismatch between the quantity calculated and the label attached to it.

A unit-price comparison offers a different application. One pack contains 750 grams and costs six dollars. Another contains 1.2 kilograms and costs nine dollars. Express both masses in kilograms. The first costs 6/0.75 = 8 dollars per kilogram; the second costs 9/1.2 = 7.50 dollars per kilogram. On price per kilogram alone, the second is lower. That mathematical conclusion does not establish that it is the better purchase for every person: quantity needed, waste and product differences may matter. Separate the computed criterion from a broader real-world decision.

Rates can also describe work completed per unit time. A tap fills a tank in six hours at a constant rate, so it fills one sixth of a tank per hour. A second tap fills the same tank in three hours, or one third per hour. Together they fill one half of a tank per hour and need two hours, assuming their rates remain unchanged and no water leaves. Adding the completion times, or averaging six and three, does not combine the rates. The explanation should use the shared whole: one tank.

Now change an assumption. If a drain removes one twelfth of a tank per hour while both taps operate, the net rate is 1/6 + 1/3 − 1/12 = 5/12 of a tank per hour. Filling an empty tank takes 12/5 hours, or two hours and twenty-four minutes. The negative contribution describes outflow, not a negative duration. This example is useful because it requires the learner to preserve the meaning of the unit while combining several processes. A diagram of arrows into and out of the same tank may clarify the model before fractions are calculated.

For transfer, calculate the speed of a runner covering 1.5 kilometres in six minutes: fifteen kilometres per hour. Find the cost per litre when 1.5 litres costs 4.80 dollars: 3.20 dollars per litre. Finally, a machine makes forty-five items in fifteen minutes; at a constant rate, it makes 126 items in forty-two minutes. In each answer, require the learner to identify the rate, the quantity sought and the assumption that allows scaling. The habit to keep is simple: define the unit before calculating, carry it through the operation, and check that the final label matches the question.

9. Average speed: combine journeys, not just the speed labels

A traveller covers sixty kilometres at thirty kilometres per hour and another sixty kilometres at sixty kilometres per hour. The arithmetic mean of the two speeds is forty-five, but the average speed for the whole journey is forty kilometres per hour. The first part takes two hours and the second takes one hour. Total distance is 120 kilometres and total time is three hours, giving 120/3 = 40. The slower section occupies more of the travelling time, so the two speed labels cannot be given equal weight simply because there are two sections.

The decisive explanation is the definition of the required quantity: average speed over a journey is total distance divided by total elapsed time for that journey. Begin there rather than selecting a special formula. A student can know how to calculate each section correctly but combine the wrong things at the end. Ask, “What imaginary constant speed would cover this total distance in this total time?” Forty kilometres per hour answers that question. Forty-five would cover 135 kilometres in three hours, which does not match the journey described.

Now compare a different journey: one hour at thirty kilometres per hour and one hour at sixty kilometres per hour. The traveller covers thirty plus sixty, or ninety kilometres, in two hours. The average is forty-five. The arithmetic mean works here because the speeds apply for equal durations. This contrast is more useful than saying “Never average speeds.” Sometimes the average of the labels is correct, but its correctness depends on the weighting. The explanation must name the condition that made a shortcut legitimate.

A stop changes the calculation if the question includes it in the journey time. Suppose a cyclist rides twelve kilometres in forty minutes, rests for twenty minutes and then rides eighteen kilometres in one hour. Total distance is thirty kilometres. Total elapsed time is two hours, so average speed including the stop is fifteen kilometres per hour. Average speed while moving excludes the twenty-minute stop: thirty kilometres divided by one hour and forty minutes gives eighteen kilometres per hour. Neither answer can be judged without reading which period the question asks about.

Build a small table with distance, speed and time for each segment. Fill the two known quantities before deriving the third. In the sixty-kilometre example, the two distances are equal, but the two times are not. The table makes that asymmetry visible. A common wrong explanation is “There are two speeds, so divide by two.” Ask what each entry being averaged represents and whether each represents an equal amount of the weighting quantity. That question also prepares the learner for weighted means in the data chapter.

For a new problem, a car travels ninety kilometres at sixty kilometres per hour and sixty kilometres at forty kilometres per hour. Each section takes one and a half hours, so the average speed is 150/3 = 50 kilometres per hour. Notice that the distances are unequal while the durations are equal. The arithmetic mean of sixty and forty works, but not because the journey has equal-distance halves. A strong learner should be able to explain why the shortcut happens to work and still derive the result from distance and time.

A harder transfer question asks for an unknown speed. A cyclist travels ten kilometres out at twenty kilometres per hour and ten kilometres back at an unknown speed. The average for the complete ride is sixteen kilometres per hour. Total time must be 20/16 = 1.25 hours. The outward ride takes 0.5 hour, leaving 0.75 hour for the return. The return speed is 10/0.75 = 40/3 kilometres per hour. Setting the arithmetic mean of twenty and the unknown speed equal to sixteen would give twelve, which does not reproduce the required total time.

Finish by asking what can be checked before exact arithmetic. If both sections involve positive travel at thirty and sixty kilometres per hour with no stops, the overall average must lie between those speeds. That bound does not distinguish forty from forty-five, so it cannot replace the definition. It can reject ninety or fifteen in the original example. Good checking combines several levels: a broad bound, a correctly represented journey and a calculation that reproduces the given totals. The explanation should use whichever level can catch the learner’s actual error most directly.

10. Brackets and negative signs: identify the expression being multiplied

The expression 3(2x − 5) represents three copies of the entire bracket. Expanding gives 6x − 15 because both terms are multiplied by three. The incorrect expansion 6x − 5 changes only the first term. Ask the learner to test both expressions at x = 4. The original is 3(8 − 5) = 9; the correct expansion is 24 − 15 = 9; the incorrect expansion is 24 − 5 = 19. A substitution can disprove an asserted identity, although one successful substitution alone cannot prove that two expressions are identical for every x.

Negative multipliers require the same complete scope. For −2(3x − 4), distribute negative two to obtain −6x + 8. The second product is positive because negative two multiplied by negative four equals positive eight. A learner saying “Change all the signs” may produce the correct result here but remain unclear about the multiplication. Contrast it with 2(3x − 4), where the signs do not all change. The sign pattern follows the actual multiplier; it is not a property of brackets in general.

Subtraction of a bracket is multiplication of that bracket by negative one. In 5x − (2x − 7), rewrite the subtraction as 5x + (−1)(2x − 7). Expansion gives 5x − 2x + 7 = 3x + 7. The explanation “I am subtracting both contributions in the bracket” can help, but it must handle the negative contribution accurately. Subtracting negative seven adds seven. Avoid teaching a visual trick based only on the sign nearest the bracket; ask which complete expression is being subtracted.

Nested expressions offer a useful next step without requiring harder arithmetic. Simplify 4 − 2(3 − x). First expand the product: 2(3 − x) = 6 − 2x. Subtracting that product from four gives 4 − 6 + 2x = 2x − 2. Alternatively, treat the multiplier as negative two immediately: 4 + (−2)(3 − x) = 4 − 6 + 2x. Both routes are valid. Ask the learner to mark whether the minus sign belongs to an outer subtraction or has already been incorporated into the multiplier, so it is not applied twice.

Distinguish a negative number being squared from the negative of a square. At x = 3, −x² = −9, whereas (−x)² = 9. The exponent applies to the base indicated by the notation. In the first expression, x is squared and the result is negated. In the second, the whole negative quantity is squared. Writing a substituted value with brackets, such as −(3²) or (−3)², can make the difference visible. This is another scope question: what exactly is the operation acting on?

Try four independent simplifications. First, 4(2a + 3) − 5a = 3a + 12. Second, 7y − 3(2y − 4) = y + 12. Third, 2 − (5 − 3b) = 3b − 3. Fourth, −3(2c − 5) + 2(c + 1) = −4c + 17. For one answer, require a substitution check with a nonzero value that will not make the disputed term vanish. For example, c = 2 gives nine in both the original fourth expression and its simplified form.

An explanation can be concise and still expose the complete reasoning: “Negative three multiplies both terms, giving negative six c and positive fifteen; the second bracket contributes two c and two.” Compare that with “I opened the brackets.” The latter names a classroom action but does not reveal whether the learner knows what must be preserved. During correction, ask for the products before combining like terms. This separates a distribution error from an error in collecting terms, making the next practice question more precisely targeted.

Do not demand repeated narration once the structure is secure. A student accurately expanding several varied expressions may benefit more from a mixed set in which expansion is sometimes unnecessary. For example, solving 5(x − 2) = 35 by dividing first is shorter than expanding. The ability to distribute does not imply that distribution is always the best first move. Independent mathematical control includes deciding whether a transformation helps the present goal. Ask why the learner chose that move, not only whether they can perform it.

11. Linear equations: preserve solutions while simplifying the demand

Solve 4x + 7 = 2x + 19. Subtracting 2x from both sides gives 2x + 7 = 19. Subtracting seven gives 2x = 12, and dividing by two gives x = 6. The important explanation is not that letters must move left and numbers must move right. That arrangement is a convenient choice. Subtracting 4x instead would also be valid and would lead to an equation with a negative coefficient. The goal is to preserve the solutions while producing a simpler equation to inspect.

A substitution check returns to the original statement: 4(6) + 7 = 31 and 2(6) + 19 = 31. This verifies that six is a solution. The sequence of reversible operations establishes that no other solution has been discarded along the route. These two kinds of justification do different jobs. Checking a candidate protects against execution errors; understanding equivalence protects the reasoning that produces the candidate. Encourage the learner to use both without turning every routine linear equation into a formal proof.

Brackets can make a familiar equation look more complex. Solve 3(2x − 1) = 4x + 9. Expanding the left gives 6x − 3 = 4x + 9, then 2x = 12 and x = 6. Ask why subtracting 4x is useful: it leaves the variable in one combined term. The response “Because that is the next step” is less informative than identifying the objective. A learner who can name the objective can evaluate a different valid route rather than relying on one memorised sequence.

Not every linear-looking equation has one solution. For 3(x + 2) = 3x + 6, expansion gives 3x + 6 = 3x + 6, which is true for every real x. Removing matching terms leaves 0 = 0. That does not mean x = 0; it means the equation imposes no further restriction on x. Test x = 0 and x = 5 to make the distinction concrete. Both satisfy the original, and the identity explains why all real values do so.

Now compare 3(x + 2) = 3x + 8. Expansion gives 3x + 6 = 3x + 8, leading to 6 = 8 or 0 = 2. No real x can make that true. The variable disappearing is therefore not enough to decide the result. It can leave a true statement, implying all permitted values, or a false statement, implying none. A student trained to force every exercise into “x equals a number” may invent a value rather than interpret the resulting statement.

Use a comparison set: 5x + 4 = 2x + 19 has x = 5; 5(x + 1) = 5x + 5 holds for every real x; 5(x + 1) = 5x + 6 has no solution. Ask the learner to classify the outcome before writing a final sentence. These questions share almost all their visual features, which makes them useful for testing the meaning of the equation rather than the ability to follow a familiar layout. Require the conclusion to describe the solution set accurately.

A word problem adds one further responsibility. “A taxi charge consists of a fixed four dollars plus three dollars per unit of distance; the total is nineteen dollars. Find the distance in those units.” The equation 4 + 3d = 19 gives d = 5. The final answer must return to distance, not simply state a naked five. If the resulting value were negative in a context that permits only nonnegative distance, the algebra might be valid while the model or stated data would need inspection. Mathematical solutions and contextual admissibility are related but distinct checks.

For delayed practice, ask the learner to create one equation with a unique solution, one with no solution and one with every real number as a solution. Example answers are 2z + 1 = 9, 2z + 1 = 2z + 3 and 2(z + 1) = 2z + 2. Creating examples is useful here because it requires control of the structure. Do not assess the task by how complicated the equations look. Assess whether the learner can justify the three different outcomes and recognise them when someone else changes the surface wording.

12. Simultaneous equations: one answer must satisfy both constraints

Suppose x + y = 11 and 2x + y = 15. The first equation permits many pairs: one and ten, two and nine, and so on. The second equation supplies another constraint. Subtracting the first equation from the second gives x = 4; substituting into x + y = 11 gives y = 7. The pair must satisfy both original equations. It is not enough for each value to look plausible separately. Simultaneous equations ask for the overlap between requirements.

Why can the equations be subtracted? For a pair satisfying both, the left-hand quantities equal their respective right-hand quantities. The difference between the left sides must therefore equal the difference between the right sides. In this case, the y terms cancel because they have equal coefficients and opposite contributions after subtraction. The explanation should describe that relationship, not say that y disappears because the method is called elimination. Naming a procedure is not the same as explaining what makes its operation valid.

When coefficients do not initially match, scale a complete equation. Solve 2x + 3y = 17 and 3x − 2y = 6. Multiply the first by two to get 4x + 6y = 34, and the second by three to get 9x − 6y = 18. Adding gives 13x = 52, so x = 4. Substitution gives 8 + 3y = 17, so y = 3. Every term, including the constant on the right, must be multiplied. A selective multiplication changes the constraint rather than producing an equivalent one.

Substitution offers another route. From x + y = 11, write y = 11 − x. Replacing y in the second equation gives 2x + (11 − x) = 15, so x = 4. Ask why the replacement is allowed: the first equation states that y and 11 − x have the same value for any acceptable solution. A good explanation identifies the expression being substituted and places brackets around it when necessary. In more complicated examples, omitting those brackets can turn a valid idea into an incorrect expansion.

Parallel and coincident lines provide a geometric interpretation. The equations x + y = 5 and 2x + 2y = 10 describe the same line; every pair on that line satisfies both. The equations x + y = 5 and 2x + 2y = 12 describe distinct parallel lines; no pair satisfies both. The elimination results, 0 = 0 and 0 = 2, match these interpretations. Seeing the diagram and the algebra agree can make the solution-set distinction less dependent on memorised labels.

In a fictional stationery problem, two notebooks and three pens cost seventeen dollars, while three notebooks and two pens cost eighteen dollars. Let n and p be the respective prices. The equations are 2n + 3p = 17 and 3n + 2p = 18. Multiplying the first by three and the second by two gives 6n + 9p = 51 and 6n + 4p = 36. Subtraction gives p = 3, then n = 4. Explain that the multiplication creates two imaginary purchases containing equal numbers of notebooks, allowing their notebook costs to cancel.

A useful transfer set is x + 2y = 13 with 2x − y = 1, giving x = 3 and y = 5; and 4a + b = 9 with b = 1 − 2a, giving a = 4 and b = −7. The second pair includes a negative result, which is acceptable in pure algebra. If a and b represent prices, that result would require reconsidering the context. Ask for substitution into both original equations, not only the last simplified equation. The original constraints are the final standard.

When a learner struggles, separate three possible difficulties: constructing the equations, selecting elimination or substitution, and executing the arithmetic. A correct explanation of elimination does not repair an equation built from the wrong quantities. Conversely, a well-modelled situation can fail because a negative sign is mishandled. Ask the learner to identify the first line at which their work ceases to match the intended relationship. This keeps correction focused on the earliest divergence rather than covering the page with comments about every later consequence of the same initial error.

13. Inequalities: explain why multiplying by a negative reverses order

The inequality 2 < 5 states an order. Multiplying both numbers by negative one gives −2 and −5, and their order is −2 > −5. On a number line, negation reflects positions across zero, reversing left and right. This is the reason an inequality sign reverses when both sides are multiplied or divided by a negative number. It is not a special punishment for moving a minus sign. The operation changes order in a predictable way while preserving the set of values that satisfy the transformed statement.

Solve −3x < 12. Dividing both sides by negative three gives x > −4. Check x = 0: the original becomes 0 < 12, which is true. Check x = −5: the original becomes 15 < 12, which is false. These substitutions support the direction of the answer. A learner who reports x < −4 has selected values that do not generally satisfy the original. The quick test with zero is especially helpful because it exposes the direction without elaborate arithmetic.

Adding or subtracting the same quantity from both sides does not reverse order. From x − 7 < 2, adding seven gives x < 9. The rule is tied to multiplication or division by a negative, not to the mere presence of subtraction. Contrast x − 7 < 2 with −7x < 2. The first gives x < 9; the second gives x > −2/7. Ask the learner to describe the actual operation rather than search the line for a minus sign and apply a visual trigger.

A compound inequality requires preserving both bounds. For 2 < 3x + 5 ≤ 14, subtract five throughout to obtain −3 < 3x ≤ 9, then divide throughout by three to obtain −1 < x ≤ 3. The left endpoint is excluded and the right endpoint included. Test the endpoints: x = −1 makes the first strict inequality false, whereas x = 3 satisfies the second equality. Open and closed endpoint symbols express these differences; they are not interchangeable drawing conventions.

Do not multiply or divide an inequality by an expression of unknown sign without handling the cases. For example, x² > x cannot safely become x > 1 by dividing by x unless x is known positive. Rearranging gives x(x − 1) > 0. The product is positive when both factors are negative, x < 0, or both are positive, x > 1. Dividing immediately would lose the negative solution interval and would be undefined at zero. The condition attached to a transformation is part of the mathematics.

For a simpler contextual problem, a bag may weigh no more than eight kilograms. It already contains 2.5 kilograms of equipment, and each identical book weighs 0.6 kilogram. If n is the number of books, 2.5 + 0.6n ≤ 8 gives n ≤ 55/6. Since n must be a nonnegative whole number, at most nine books can be added. Reporting 9.166 books ignores the quantity’s domain. Ten books would weigh 8.5 kilograms in total and exceed the stated limit.

Try 5 − 2t ≥ 13, giving t ≤ −4; 4u + 1 < 17, giving u < 4; and −6 ≤ 2v − 4 < 8, giving −1 ≤ v < 6. For each, select one value inside the proposed solution set and one outside it, then check the original. These checks do not replace a valid derivation, but they are a practical defence against reversed signs and misplaced endpoints. Ask the learner to explain the sign reversal only in the first question, where division by a negative actually occurs.

The teaching aim is conditional control. A memorised sentence such as “Change the sign with negatives” is dangerously incomplete. A stronger account is “Multiplying or dividing both sides by a negative reverses their order; adding a negative to both sides shifts them equally and preserves order.” The learner may demonstrate this with a number line rather than recite it. Once that distinction is secure, mixed practice should require deciding whether a reversal is necessary, because that decision is what an unlabelled question will demand.

14. Quadratics: do not lose a solution while simplifying

Solve x² = 5x. Dividing by x appears to give x = 5, but x = 0 also satisfies the original equation. The division excluded a value for which the divisor was zero. A safer route is x² − 5x = 0, then x(x − 5) = 0. A product of two real numbers is zero only when at least one factor is zero, so x = 0 or x = 5. The crucial explanation concerns a condition on division, not merely a preference for factorisation.

The zero-product property applies when the product equals zero. From (x − 2)(x + 3) = 0, the solutions are x = 2 or x = −3. From (x − 2)(x + 3) = 4, it is not valid to set either factor equal to four or to zero. Expand and rearrange: x² + x − 6 = 4, so x² + x − 10 = 0. The solutions are (−1 ± √41)/2. The right-hand side is therefore part of the method’s condition, not a detail that can be ignored while noticing brackets.

Consider x² − 7x + 12 = 0. Factorising gives (x − 3)(x − 4) = 0. Explain why these factors reconstruct the original: multiplying gives x² − 4x − 3x + 12, whose middle terms combine to −7x. The solutions are three and four. A learner who remembers “Find two numbers that add and multiply” should still be able to connect that search to expansion. Otherwise the rule may be misapplied when the leading coefficient differs from one.

For 2x² + 7x + 3 = 0, the factorisation is (2x + 1)(x + 3) = 0, giving x = −1/2 or x = −3. The cross-products are six x and one x, which combine to seven x. Ask the learner to inspect the middle coefficient instead of checking only the first and last terms. Several plausible-looking pairs of brackets can produce the correct leading and constant terms but the wrong middle term. Expansion is an independent verification of the factorisation, not a redundant extra step when uncertainty remains.

Completing the square offers another explanation of a quadratic’s structure. For x² + 6x + 5 = 0, rewrite x² + 6x as (x + 3)² − 9. The equation becomes (x + 3)² − 4 = 0, so (x + 3)² = 4. Both x + 3 = 2 and x + 3 = −2 are possible, giving x = −1 or x = −5. The ± is necessary because two real numbers can share the same positive square. It is not a decorative symbol added because quadratic questions usually have two answers.

Not every quadratic has two distinct real roots. The equation (x − 2)² = 0 has one repeated root, x = 2. The equation (x − 2)² = −1 has no real solution because a real square cannot be negative. When working over the real numbers, that domain must be stated or understood. A learner should not force a negative quantity through a real square-root calculation and then invent a real value. The structure of the equation can establish the outcome before any formula is used.

For practice, solve x² = 9x, giving zero or nine; x² − x − 12 = 0, giving four or negative three; and 3x² − 12 = 0, giving two or negative two. Then inspect the proposed answer x = 3 for x² = 9. It is correct but incomplete because negative three also works. Ask the learner to distinguish an incorrect candidate from an incomplete solution set. Both lose marks in many assessment settings, but they require different repairs and reveal different weaknesses.

A contextual quadratic adds a final filter. If a rectangle’s side lengths are x and x + 2 and its area is fifteen, x(x + 2) = 15 gives x = 3 or x = −5. The negative value solves the algebraic equation but cannot be a side length in this model, so the dimensions are three and five. Do not discard a negative solution merely because it looks unusual; discard it because the defined quantity requires a positive length. State the condition, solve the mathematics and then return to the context.

15. Algebraic fractions: a simpler expression can retain an old restriction

The expression (x² − 9)/(x − 3) factorises to [(x − 3)(x + 3)]/(x − 3). Cancelling the common nonzero factor gives x + 3, but only for x ≠ 3. The original denominator is zero at three, so the original expression is undefined there. The simplified formula can be evaluated at three, yielding six, but that does not repair the original undefined value. An explanation that says only “The matching terms cancel” misses the condition under which cancellation was permitted.

Think of simplification as preserving values on the original domain, not automatically creating a new domain. For x = 5, the original expression is sixteen divided by two, or eight, matching x + 3. For x = 3, the original is zero divided by zero, which is not a defined real quotient. A graph of the original relation would follow the line y = x + 3 with the point at x = 3 missing. This is a useful connection between algebraic manipulation and the meaning of a function’s permitted inputs.

Cancellation concerns factors, not any symbols repeated above and below a fraction bar. In (x + 3)/x, it is not valid to cancel x and obtain three. The numerator is a sum. It can be separated as x/x + 3/x = 1 + 3/x for x ≠ 0, but the remaining term still depends on x. At x = 3, the original equals two, not three. A numerical counterexample can reject the proposed simplification; the factor-versus-sum explanation identifies the reason it failed.

Equations with denominators require restrictions before clearing fractions. Solve 1/(x − 1) = 2. First require x ≠ 1. Multiplying by x − 1 gives 1 = 2x − 2, so x = 3/2, which satisfies the restriction. Substitution gives 1/(1/2) = 2. Recording the restriction may feel unnecessary in this simple case, but it prepares the learner for equations in which an algebraic transformation produces a forbidden candidate. The habit is to inspect where the original statement is defined.

For a revealing contrast, solve (x² − 1)/(x − 1) = 2. The restriction is x ≠ 1. Simplification gives x + 1 = 2, whose only candidate is x = 1. Because that candidate is forbidden in the original equation, there is no solution. A student who reports one has solved the simplified unrestricted equation rather than the actual problem. The restriction is not an optional footnote; it decides the answer. Ask the learner to show exactly which original denominator makes the candidate inadmissible.

Now solve (x² − 4)/(x − 2) = 5. The restriction is x ≠ 2, and simplification gives x + 2 = 5, so x = 3 is valid. Comparing this with the previous no-solution example isolates the role of the candidate check. The same simplification method is used, but one candidate lies outside the domain and the other does not. Do not teach students that a cancelled factor always means the final answer is forbidden. It means the corresponding input must be examined against the original restriction.

Practice three expressions: (a² − 16)/(a − 4) = a + 4 for a ≠ 4; (2b + 6)/(b + 3) = 2 for b ≠ −3; and (c² + c)/c = c + 1 for c ≠ 0. Then ask why (c² + c)/c can be simplified even though its numerator is a sum. The answer is that the entire numerator factorises as c(c + 1). The rule is not “Never cancel across a numerator containing a plus sign.” The rule is to identify a genuine common factor of the whole numerator and denominator.

This chapter is a good test of precision in explanations. Overly broad slogans can be as misleading as careless arithmetic. “Cancel identical things” is too permissive; “Never cancel when there is addition” is too restrictive. The useful statement names both structure and condition: “I divide the complete numerator and denominator by a common factor, provided that factor is nonzero on the permitted inputs.” A student may use simpler words, but those two ideas must remain. Link this work with Additional Mathematics when functions and domains become part of the learner’s course.

16. Functions and inverses: reversing a process requires a recoverable input

A function assigns one output to each permitted input. Let f(x) = 2x + 3. If the output is eleven, the input must be four. Reversing the process requires undoing the addition of three and then the multiplication by two: f⁻¹(y) = (y − 3)/2. The inverse notation does not mean the reciprocal 1/f(y). It names a function that recovers the original input from the output, on the appropriate domains. The order of the reverse operations follows the order in which the original process was built.

A concrete machine description can clarify the idea. The forward machine doubles an input and then adds three. Its reverse machine subtracts three and then halves the result. If the reverse machine halves first and then subtracts three, an output of eleven becomes 2.5 rather than four. The order matters because the operations do not generally commute. Ask the learner to verify the proposed inverse by composition: applying the reverse to 2x + 3 gives [(2x + 3) − 3]/2 = x.

Not every function is reversible over its entire stated domain. For f(x) = x² on all real numbers, both two and negative two produce four. The output alone cannot identify which input was used. There is no single-valued inverse function on that unrestricted real domain. Restricting the domain to x ≥ 0 makes the square function one-to-one, and the inverse is the nonnegative square root. Restricting instead to x ≤ 0 gives a different inverse branch, the negative square root. The chosen domain is what makes recovery unique.

Compare solving x² = 9 with evaluating √9. The equation has two real solutions, three and negative three. The principal square-root symbol √9 denotes the nonnegative value three. Confusing these tasks can lead to either missing a solution or attaching an unnecessary ± to a function value. Ask the learner whether the question requests all inputs producing an output or the value of a specifically defined function. The notation is compact, but it carries a distinction about the task.

For composition, let f(x) = x + 4 and g(x) = 3x. Then f(g(x)) = 3x + 4, while g(f(x)) = 3x + 12. The two compositions are different because the second multiplies the added four by three. A learner who writes both as 3x + 4 may be treating composition as unordered collection of operations. Draw the input passing through one labelled box and then the other. The diagram’s purpose is to preserve order; it should be checked against the algebra rather than treated as an unrelated illustration.

Domain restrictions can also arise inside a composition. If h(x) = 1/(x − 2), the input two is forbidden. If k(x) = x + 5, then h(k(x)) = 1/(x + 3), and x = −3 is forbidden because it would make k(x) equal two. The restriction is not necessarily the same numeral that appeared in the original function. It must be traced through the input transformation. This is a useful explanation prompt: “Which starting value would send the inner output to a forbidden input of the outer function?”

For independent practice, find the inverse of p(x) = 5x − 7; it is p⁻¹(x) = (x + 7)/5. For q(x) = (x + 1)/3, the inverse is q⁻¹(x) = 3x − 1. Then explain why r(x) = x² + 1 on all real numbers has no inverse function without a restriction. On x ≥ 0, its inverse is √(x − 1), defined for x ≥ 1. The output range of the original becomes the input domain of its inverse. Include those restrictions in the explanation, not merely in a memorised formula.

Finish with a nonnumerical analogy that can be checked rather than merely admired. Recording only a person’s birth month does not identify the person uniquely; several people can share the same output. A reversible record would require a rule and domain that permit unique recovery. The analogy illustrates many-to-one information loss, but it does not replace the mathematical definition. Return to the function and name two distinct permitted inputs with the same output. That concrete pair is enough to show why a proposed inverse cannot recover every original input uniquely.

17. Indices and logarithms: use the law that matches the operation

The law a³ × a² = a⁵ follows from counting factors: three copies of a multiplied by two more copies produce five copies. The operation between the powers is multiplication. It does not follow that a³ + a² equals a⁵. At a = 2, the sum is eight plus four, or twelve, while the fifth power is thirty-two. A learner who adds exponents whenever the bases match has noticed a visual feature while ignoring the operation that licenses the law.

A power of a power counts groups of factors. In (a³)², there are two groups, each containing three factors of a, so the result is a⁶. Compare a³ × a² with (a³)². The first adds exponents; the second multiplies them. Ask the learner to expand a small integer example into repeated factors before accepting the symbolic shortcut. The explanation should not be “Outside powers multiply” unless the learner can identify the whole base to which the outside exponent applies.

For nonzero a, the quotient a³/a³ equals one. To preserve the exponent law aᵐ/aⁿ = aᵐ⁻ⁿ, the corresponding zero power is a⁰ = 1. The nonzero condition matters because the quotient is undefined at a = 0. Negative exponents similarly represent reciprocals: a⁻² = 1/a² for nonzero a. The minus sign in an exponent does not make the value negative. For example, 2⁻³ = 1/8, whereas −2³ = −8. The position of the sign changes its mathematical role.

For positive real a, a¹ᐟ² denotes √a and a³ᐟ² denotes (√a)³. Using a = 16 gives 16³ᐟ² = 4³ = 64. Keeping a positive in this introductory real-number discussion avoids complications about negative bases and fractional powers. A learner should not generalise every exponent identity to every imaginable real expression without checking where the expressions are defined. Stating a domain is not excessive caution; it tells us what mathematical objects the law is describing.

A logarithm asks for an exponent. The statement log₂8 = 3 means 2³ = 8. For real logarithms with base b, require b > 0, b ≠ 1 and a positive argument. The product law log_b(uv) = log_b u + log_b v for positive u and v reflects multiplication of powers with the same base. It does not create a corresponding rule for log_b(u + v). For instance, log₂(4 + 4) = 3, while log₂4 + log₂4 = 4.

Solve log₂(x − 1) = 3. First require x − 1 > 0, so x > 1. Rewriting the logarithmic statement in exponential form gives x − 1 = 8, hence x = 9, which is permitted. Now consider log₂x + log₂(x − 2) = 3. The original requires x > 2. Combining logs gives log₂[x(x − 2)] = 3, so x² − 2x − 8 = 0. The candidates are four and negative two, but only four satisfies the original domain.

The discarded negative candidate illustrates why the combined product’s positivity is not enough to recover the original restrictions. At x = −2, the product x(x − 2) is positive, but the separate real logarithms log₂x and log₂(x − 2) are undefined. Replacing two expressions by one may hide conditions that were explicit earlier. Ask the learner to record the restrictions before combining. The reasoning resembles algebraic fractions: simplification preserves values where the original expression is defined, not wherever the new expression happens to be defined.

For independent practice, simplify t⁴ × t³ to t⁷; (t⁴)³ to t¹²; and t⁴/t³ to t for t ≠ 0. Evaluate 3⁻² as 1/9 and 25³ᐟ² as 125. Solve log₃(2x + 1) = 2 to obtain x = 4. Finally, explain why log₁₀(x²) = 2 can be rewritten as x² = 100, giving x = ±10, while 2log₁₀x = 2 permits only x = 10. The expression log₁₀(x²) is defined for negative nonzero x, whereas log₁₀x is not.

18. Area and perimeter: the unit tells you which feature is being measured

A rectangle measuring eight centimetres by five centimetres has perimeter twenty-six centimetres and area forty square centimetres. Perimeter measures the length around its boundary; area measures the amount of two-dimensional space inside. The formulas 2(8 + 5) and 8 × 5 answer different questions. A learner who can recite both formulas may still choose incorrectly when a problem asks about fencing, edging, covering or painting. Ask what physical quantity the answer represents before asking which formula has been remembered.

The area formula can be explained by a grid. Eight unit squares fit along one row, and five such rows contain forty squares. The square-centimetre unit records that two perpendicular length dimensions have been combined. It is not enough to append a small two to the unit because the question is called area. If lengths are measured in centimetres and metres in the same calculation, the units must first be made consistent. A room five metres long and 300 centimetres wide has area fifteen square metres, not 1,500 square metres.

Equal perimeters do not guarantee equal areas. A rectangle with sides eight and two has perimeter twenty and area sixteen. A square with sides five and five also has perimeter twenty but area twenty-five. The same boundary length can enclose different amounts of space. Conversely, rectangles with sides six and four, and eight and three, both have area twenty-four but have perimeters twenty and twenty-two. These paired examples reject a false generalisation more directly than another isolated formula exercise.

Scaling both side lengths by a factor of three multiplies the perimeter by three and the area by nine. Starting with an eight-by-five rectangle gives a twenty-four-by-fifteen rectangle. The perimeter changes from twenty-six to seventy-eight, while the area changes from forty to 360. The area factor is nine because both independent length dimensions were multiplied by three. A drawing with a three-by-three arrangement of original-sized rectangles can show the same relationship. The explanation should identify how many dimensions are being scaled.

A composite-shape problem tests whether a learner chooses a partition that preserves the target. Suppose a ten-by-eight rectangle has a four-by-three rectangular corner removed. Its remaining area is eighty minus twelve, or sixty-eight square units. The perimeter is not found by subtracting the removed rectangle’s entire perimeter from the original perimeter. Some removed edges were on the outer boundary, while new interior-cut edges become part of the boundary. Trace the actual remaining outline. In this corner-cut configuration, the perimeter remains thirty-six units because the removed horizontal and vertical lengths are replaced by equal cut lengths.

That perimeter result depends on the location of the cut. A rectangular hole completely inside a larger rectangle adds an inner boundary if the task asks for the total boundary length of the remaining region. Removing a notch from the middle of one edge creates yet another outline. The area subtraction may look similar across these situations, but the perimeter reasoning changes. This is a useful self-explanation prompt: “Which segments are actually on the boundary after the change?” The learner should not infer a perimeter rule solely from the area calculation.

Try three transfer questions. A nine-by-four rectangle has area thirty-six and perimeter twenty-six. Doubling both sides gives area 144 and perimeter fifty-two. A square with area eighty-one square centimetres has side nine centimetres and perimeter thirty-six centimetres. A rectangle with area forty-eight square metres and width six metres has length eight metres and perimeter twenty-eight metres. Require the learner to distinguish taking a square root to recover a square’s side from dividing area by a known side of a rectangle.

The final check is representational. Ask the learner to write a short situation that would require perimeter and another that would require area for the same garden. Buying edging around the outside and covering the ground with turf are suitable examples. Then ask what additional information might matter in real purchasing, such as gaps, overlap or waste, without changing the idealised mathematical answer. A good explanation respects both levels: the clean model used in the exercise and the assumptions that would need inspection before applying it to a physical project.

19. Similarity: match corresponding features before forming a ratio

Similar shapes have matching angle structure and proportional corresponding lengths. The word “corresponding” carries much of the work. Suppose triangle ABC corresponds to triangle PQR in that order, and AB = 4 while PQ = 10. The length scale factor from ABC to PQR is 10/4 = 2.5. If BC = 6, the matching side QR is fifteen. Comparing AB with an unrelated side of the second triangle would create a ratio of numbers but not the scale factor the similarity relationship requires.

Rotation or reflection does not change which vertices correspond. A diagram can make a short side appear in a different position, so “left side matches left side” is not a reliable rule. Use the stated vertex correspondence or matching angles. If angle A matches angle P and angle B matches angle Q, then side AB matches side PQ. Ask the learner to write the paired vertices before calculating. This small preparation prevents a large class of ratio errors without demanding a long written explanation.

For triangles, establishing two equal corresponding angles is enough to establish similarity because the third angles then also match. Equal-looking pictures are not evidence. A triangle drawn nearly to scale may invite measurement, but the proof should use the given angles or valid geometric relationships. In an exercise with a line parallel to one side of a triangle, corresponding or alternate angles can establish the smaller triangle’s similarity to the larger one. State those angle relationships rather than concluding “They look the same.”

Consider triangle ABC with D on AB and E on AC, with DE parallel to BC. Suppose AD = 3, DB = 5 and DE = 6. The whole side AB is eight, so the small-to-large length ratio is 3:8. The corresponding large side BC is 6 × 8/3 = 16. The common error is to compare AD with DB, using 3:5, because both numbers appear next to the same drawn side. But DB is the remaining segment, not the full side corresponding to AD in the large triangle.

Ask the learner to outline each complete triangle separately. The small triangle is ADE; the large one is ABC. The side AD corresponds to AB, not DB. A correct self-explanation names the two objects being compared before naming the ratio. This mirrors the percentage-base problem: the arithmetic can be flawless while the reference object is wrong. When the learner cannot identify the complete objects, additional cross-multiplication practice will not address the first error.

Areas of similar shapes scale with the square of the length factor. If the length scale factor is 3/2, the area factor is 9/4. A smaller triangle of area twenty square units corresponds to a larger triangle of area forty-five. Reversing the question, an area ratio of 25:49 implies a positive length ratio of 5:7. The square root appears because two length dimensions contributed to area. It is not appropriate to apply the area ratio directly to a side length.

For independent practice, two similar triangles have corresponding sides five and twelve. A second side of the smaller triangle is seven, so its match is 84/5, or 16.8. Their area ratio is 25:144. In another pair, the larger area is sixty-three and the smaller area twenty-eight. The large-to-small area ratio is 9:4, so the length factor is 3/2. A corresponding smaller side of ten therefore becomes fifteen. Require the learner to label the direction of the ratio so an otherwise correct factor is not inverted.

A final boundary question asks whether two rectangles with one pair of proportional sides must be similar. They need not be: a two-by-three rectangle and a four-by-five rectangle share one doubled side, but their other sides do not share that factor. Similarity is a claim about the whole shape’s corresponding structure. In triangles, a valid criterion can establish that structure efficiently; outside those criteria, one convenient ratio is insufficient. The learner should explain what evidence makes the proportion applicable, not merely demonstrate that a proportion can be solved.

20. Pythagoras: locate the right angle and the hypotenuse

For a right-angled triangle with perpendicular side lengths a and b and hypotenuse c, a² + b² = c². The hypotenuse is opposite the right angle and is the longest side. If the perpendicular sides are six and eight, the hypotenuse is √(36 + 64) = 10. The condition that the triangle is right-angled is essential. Three side lengths appearing in a diagram do not automatically license the formula. Ask the learner to identify the right angle before choosing which length to square on the right-hand side.

If the hypotenuse is thirteen and one perpendicular side is five, the remaining perpendicular side b satisfies 5² + b² = 13². Thus b² = 169 − 25 = 144, giving b = 12 because a length is positive. Adding the squares of thirteen and five would produce a new hypotenuse for a different triangle. The operation depends on which side is unknown. Rather than teaching separate “add” and “subtract” tricks, start from the same relationship and isolate the required quantity.

Orientation should not alter the reasoning. Rotate the triangle so that the hypotenuse is not the side that looks horizontal or slanted in the familiar textbook arrangement. Its identity remains determined by the right angle. A learner relying on “the diagonal is c” may fail on a rotated or composite diagram. Ask them to point to the side opposite the right angle and explain why it must be the longest. Naming the geometric role is more reliable than memorising the side’s usual position on a page.

The converse can test whether a triangle with given positive side lengths is right-angled. For sides seven, twenty-four and twenty-five, the largest is twenty-five. Since 7² + 24² = 49 + 576 = 625 = 25², the triangle is right-angled. For sides six, eight and eleven, 36 + 64 does not equal 121, so it is not right-angled. The largest side must be the candidate hypotenuse. Do not test an arbitrary arrangement of the three numbers and infer geometry from that mismatch.

A rectangle’s diagonal offers a common application. In a nine-by-twelve rectangle, the diagonal and two adjacent sides form a right triangle, so the diagonal is fifteen. A ladder model may also produce a right triangle when the ground and wall are perpendicular, but a practical safety decision is not determined by that length calculation alone. In this workbook, treat such descriptions as idealised geometry, not instructions for placing equipment. State the modelling assumption that creates the right angle and keep the mathematical conclusion within that assumption.

In three dimensions, a cuboid with edges two, three and six has space diagonal seven. First find the base diagonal: its square is 2² + 3² = 13. The space diagonal then has square thirteen plus 6², or forty-nine. The result is seven. This works because the vertical edge is perpendicular to the base plane, creating a second right triangle. The familiar formula √(a² + b² + c²) compresses two justified applications; explaining those two triangles is useful when the three-dimensional picture feels mysterious.

For practice, find the hypotenuse when the perpendicular sides are nine and twelve: fifteen. Find the missing perpendicular side when the hypotenuse is seventeen and the other side eight: fifteen. Determine whether sides eight, fifteen and seventeen form a right triangle: yes, because sixty-four plus 225 equals 289. Determine whether a triangle with sides five, six and seven is right-angled: no. These tasks deliberately alternate calculation and condition checking so that recognising the formula’s applicability is part of the work.

A useful explanation sounds like, “This triangle is right-angled at B, so the opposite side AC is the hypotenuse; therefore AB² + BC² = AC².” That one sentence names the condition, correspondence and relationship. A less useful explanation merely says, “I used Pythagoras.” During later independent work, ask for the full sentence only when the choice of triangle is uncertain. When the geometry is clear, a correctly labelled equation may be enough. The goal is purposeful justification, not a fixed amount of writing attached to every formula.

21. Coordinates and gradients: distinguish position from change

The gradient of a nonvertical straight line is the change in vertical coordinate divided by the corresponding change in horizontal coordinate. Between A(2, 3) and B(6, 11), the vertical change is eight and the horizontal change four, giving gradient two. The coordinate eleven is not the gradient; it is the height of one point. A learner who divides eleven by six has measured from the origin, which is only appropriate if the relevant line passes through the origin and that point is being used with it.

Subtraction order must be consistent. Using (11 − 3)/(6 − 2) gives two. Reversing both differences, (3 − 11)/(2 − 6), also gives two. Reversing only one difference gives negative two and changes the direction incorrectly. Explain the quotient as one directed change divided by the matching directed change. A small arrow from A to B, or from B to A, can make the pairing visible. The choice of direction is free; mixing two directions in one ratio is not.

Once the gradient is known, a point determines the intercept. For the line through A(2, 3) with gradient two, y = 2x + c. Substitution gives 3 = 4 + c, so c = −1 and the equation is y = 2x − 1. Check B: 2(6) − 1 = 11. The intercept is the y-value when x = 0, not necessarily the first point shown in the question. Asking what c represents can prevent a learner from inserting a convenient coordinate without justification.

Horizontal and vertical lines expose the limits of the quotient. A horizontal line has zero vertical change and nonzero horizontal change, giving gradient zero. A vertical line has zero horizontal change, so its gradient is undefined rather than zero. The line x = 4 is vertical; y = 4 is horizontal. They look similar as equations but constrain different coordinates. Ask the learner to name two points on each line. For x = 4, points such as (4, 1) and (4, 7) make the vertical structure clear.

Units give a gradient its contextual meaning. If y is distance in metres and x is time in seconds, a gradient of two represents two metres per second on a straight distance-time graph. If y is cost in dollars and x is the number of items, the same numerical gradient represents two dollars per item. The number alone is not the full answer. A learner can carry out the coordinate calculation correctly while explaining the wrong real-world relationship. Require a sentence naming what changes when the horizontal quantity increases by one unit.

A decreasing line has a negative gradient, but that does not automatically mean the measured quantity itself is negative. A water-level graph can fall from eight centimetres to two centimetres while remaining above zero throughout. Between times zero and three minutes, the gradient is (2 − 8)/(3 − 0) = −2 centimetres per minute. The minus sign describes direction of change. It does not say that the water depth is negative two centimetres. Distinguishing state from rate is a central mathematical habit that returns in calculus.

For independent practice, find the gradient through (−1, 5) and (3, −3): negative two. Its equation is y = −2x + 3. Find the line with gradient three through (2, 1): y = 3x − 5. Find the gradient through (4, 2) and (4, 9): undefined. Finally, a tank contains twenty litres at minute two and eight litres at minute five, with a straight-line change between them. Its rate is −4 litres per minute over that interval. Include the interval; the data do not establish the rate outside it.

A useful extension asks what the straight-line assumption contributes. Two measured endpoints determine an average rate over the interval, but they do not establish that the rate was constant at every moment. A curved graph could share the same endpoints. In a pure coordinate question about a straight line, constant gradient is part of the object. In a data question, it may be a modelling assumption. Asking the learner to distinguish these contexts prevents a formula from carrying more meaning than the information supports.

22. Averages and data: reconstruct the total before combining groups

One group of ten students has a mean score of eighty, and another group of thirty students has a mean of sixty. The combined mean is not seventy. The first group’s total is 10 × 80 = 800; the second’s is 30 × 60 = 1,800. Together, forty students have a total of 2,600, giving a mean of sixty-five. Averaging the two group means equally gives each group the same weight even though one contains three times as many students. The correct weighting follows the number of observations represented.

The explanation should begin with the definition of the mean: total divided by count. Multiplying a mean by its count reconstructs the group’s total. Adding totals and counts then reconstructs the combined dataset’s mean. This method works whether the groups are equal in size or not. If both groups contained ten students, the combined mean would indeed be seventy. As with average speed, a shortcut can be valid under a specific weighting condition without being a universal rule.

Finding a missing value uses the same structure. Five numbers have a mean of twelve, so their total is sixty. Four of them are eight, eleven, thirteen and fourteen, whose sum is forty-six. The missing value is fourteen. A learner who subtracts the four-number mean from twelve is comparing averages instead of recovering totals. Ask, “What total must all five numbers have?” That question directs attention to the quantity that can be combined additively.

The median describes position in an ordered dataset rather than a redistribution of its total. For 2, 3, 4, 5 and 100, the median is four while the mean is 22.8. Neither is a computational error. The large final value influences the mean strongly, while the middle position remains four. A responsible interpretation names which summary is being used and why it is relevant. In a fictional spending dataset, the median may describe a typical middle observation more directly, while the mean remains essential when reconstructing the total spending.

For an even number of observations, the median is the mean of the two middle values after ordering. In 3, 7, 8, 10, 12 and 20, the middle values are eight and ten, so the median is nine. There need not be an actual observation equal to the median. A learner who selects the third value alone has transferred the odd-count procedure to an even-count dataset without checking the condition. Ask them to pair off one smallest and one largest value repeatedly until the middle structure is visible.

A frequency table is compressed data. If the values two, four and six occur three, two and five times respectively, the total is 2 × 3 + 4 × 2 + 6 × 5 = 44 across ten observations, giving mean 4.4. Adding two, four and six and dividing by three ignores the frequencies. Explain each product as the contribution of several repeated observations. The table has fewer rows than the dataset has members; the number of rows is not the sample size.

For practice, combine a group of eight with mean fifteen and a group of twelve with mean twenty: the combined mean is eighteen. Six values have mean nine, and five sum to forty-one: the missing value is thirteen. The ordered dataset 1, 2, 4, 8, 9 and 12 has median six and mean six, although those equal summaries arise by different calculations. Finally, values one, three and five with frequencies two, four and two have mean three. Require a total-and-count explanation for one weighted question.

Do not turn a summary into an unsupported cause. A fictional class using one study routine may have a higher mean than another class, but the averages alone do not show that the routine caused the difference. Prior knowledge, group composition and other differences may matter. Within this workbook, the arithmetic examples establish how summaries are calculated; they do not report real educational outcomes. A strong self-explanation includes the boundary of the conclusion: “This calculation describes the supplied data, but it does not by itself explain why those data differ.”

23. Probability: explain the sample space and what changes after a draw

If a fair six-sided die is rolled, the probability of an even result is three favourable outcomes out of six equally likely outcomes, or one half. The equal-likelihood condition matters. Counting categories alone does not always determine probability. A spinner divided into one large red region and one small blue region has two colours, but the colours need not be equally likely. Ask the learner why each counted outcome deserves the same weight before accepting the fraction favourable/total.

A bag contains three red counters and two blue counters, with each individual counter equally likely to be selected. The probability of red on the first draw is 3/5. If the selected counter is replaced and the contents mixed before another draw, the second draw has the same composition. The probability of two reds is (3/5) × (3/5) = 9/25. The explanation should connect replacement to the unchanged numbers of available counters, not simply say “Multiply because there are two events.”

Without replacement, two reds have probability (3/5) × (2/4) = 3/10. After a red first draw, only two red counters remain among four counters. The second fraction is conditional on the first outcome. A tree diagram can show separate second-stage compositions after red and after blue. Reusing 3/5 on every branch without replacement ignores the changed state. This is a probability version of a familiar modelling error: applying the original conditions after an action has altered them.

The probability of one red and one blue without replacement requires two possible orders. Red then blue has probability (3/5) × (2/4) = 3/10. Blue then red has probability (2/5) × (3/4) = 3/10. These mutually exclusive ordered paths add to 3/5. Counting only red then blue gives half the required probability. Ask whether the event description specifies order. “One of each” does not mean “red first.” The words define which paths belong in the event.

Complementary events can simplify “at least one” questions. With replacement, the probability of at least one blue in two draws is one minus the probability of no blue. No blue means two reds, whose probability is 9/25, so the answer is 16/25. This includes exactly one blue and two blues. It is not the same event as “exactly one blue,” whose probability is 2 × (3/5) × (2/5) = 12/25. The difference comes from including or excluding the two-blue outcome.

Distinguish mutually exclusive from independent. On one die roll, “odd” and “even” are mutually exclusive: they cannot both occur. They are not independent because knowing one has occurred rules out the other. On two separate independent fair die rolls, the first being even does not change the second’s chance of being even. Students can memorise both terms yet confuse their relationships. Ask for the conditional question: does learning that the first event happened change the probability of the second?

For independent practice, a bag has four green and three yellow counters. The chance of two green without replacement is (4/7) × (3/6) = 2/7. With replacement, it is 16/49. The chance of one of each without replacement is (4/7)(3/6) + (3/7)(4/6) = 4/7. A fair coin tossed three times has probability 7/8 of at least one head, because the complementary all-tail outcome has probability 1/8. Require an explanation of the event boundary in the last question.

A final check asks whether a probability should lie between zero and one. That basic bound can expose arithmetic mistakes, but an answer within the bound may still describe the wrong event. Use three checks together: the counted outcomes are appropriate, the conditional state is correct, and all qualifying paths have been included exactly once. A learner’s explanation should reveal the weakest of those layers. Adding more fraction multiplication will not repair an event that was misread before the calculation began.

24. Differentiation: a derivative describes change, not the original height

For learners who have begun calculus, self-explanation is especially useful at the boundary between an algebraic procedure and its meaning. If y = x², then dy/dx = 2x. At x = 3, the curve’s y-coordinate is nine and its gradient is six. These numbers answer different questions. A learner who substitutes three into the derivative and reports the point as (3, 6) has calculated a gradient but labelled it as a height. Ask which function supplies the coordinate and which supplies the rate of change.

The derivative can be motivated through a difference quotient. For f(x) = x², [f(x + h) − f(x)]/h = [(x + h)² − x²]/h = 2x + h for h ≠ 0. As h approaches zero, this approaches 2x. The cancellation is performed while h is nonzero; only afterward is the limiting behaviour considered. Substituting h = 0 at the start creates zero divided by zero and stops the calculation. The order of reasoning matters, connecting this chapter to the domain restrictions examined earlier.

To find the tangent to y = x² at x = 3, use the point (3, 9) and gradient six. The point-gradient equation is y − 9 = 6(x − 3), so y = 6x − 9. Checking the point gives nine on both expressions at x = 3. The tangent line and curve agree at that point and have the same instantaneous gradient there; they do not agree everywhere. A student should not replace the original curve by its tangent for a distant input without treating that as an approximation and considering its limitations.

Stationary points occur where the derivative is zero, but zero derivative does not automatically mean a maximum or minimum. For y = x³, the derivative is 3x², which is zero at x = 0. The curve continues increasing through zero rather than turning back. A sign check of the derivative on either side reveals positive values on both sides. Compare y = x², whose derivative changes from negative to positive at zero, producing a minimum. The condition identifies a candidate; additional reasoning classifies it.

The chain rule is another place where naming the operation is insufficient. For y = (3x + 1)², the derivative is 2(3x + 1) × 3 = 6(3x + 1). The outer square changes with its input, while that inner input changes three times as fast as x. Omitting the factor three treats the inner expression as if its derivative were one. Expanding first gives 9x² + 6x + 1, whose derivative is 18x + 6, matching the chain-rule result. Agreement between two valid routes is a useful check.

For a contextual example, suppose a model gives distance s(t) = t² + 2t metres at time t seconds, for t ≥ 0. Its velocity is ds/dt = 2t + 2 metres per second. At t = 3, distance is fifteen metres and velocity eight metres per second. The average velocity from t = 1 to t = 3 is [15 − 3]/2 = 6 metres per second. Instantaneous and interval-average rates are distinct quantities, even when the same function supplies both. Name the time point or interval in the answer.

Try three transfer questions. For y = 3x² − 4x + 1, the gradient at x = 2 is eight and the point is (2, 5), so the tangent is y = 8x − 11. For y = (2x − 5)³, the derivative is 6(2x − 5)². For y = x² − 6x + 11, the stationary point is at x = 3, with y = 2, and it is a minimum. Explain the classification using either the positive second derivative or the change in the first derivative’s sign.

When a learner can differentiate accurately but cannot interpret an answer, do not assign more mechanical derivative exercises alone. Ask them to label the original function, derivative, point, gradient, interval and unit in one complete problem. When interpretation is secure but rules are unreliable, isolate the algebraic technique. The explanation prompt should be matched to the actual difficulty. Calculus becomes more coherent when its symbols retain their relationships to change; it becomes less coherent when every new notation is treated as another unconnected instruction to memorise.

Units note for the preceding motion example: interpret s(t) as the coordinate along a straight track, measured from a fixed origin, with motion in the positive direction for t ≥ 0. Under that assumption its derivative is the stated velocity, and the distance travelled from the origin agrees with the positive coordinate. More generally, the derivative of cumulative distance is speed, while the derivative of signed position is velocity; a change of direction would require keeping that distinction explicit.

25. Optimisation: a stationary value must still answer the actual problem

A rectangle has perimeter forty metres. Let one side be x metres. The other side is 20 − x, so area A = x(20 − x) = 20x − x². The domain for a genuine rectangle is 0 < x < 20. Completing the square gives A = 100 − (x − 10)², whose largest value is one hundred when x = 10. The rectangle is then a square. The explanation identifies the quantity to maximise, the constraint connecting the sides, and the reason the square term cannot increase the area beyond one hundred.

A calculus route gives dA/dx = 20 − 2x. Setting this to zero gives x = 10, and the second derivative is −2, indicating a local maximum. In this quadratic model, the completed-square form also establishes the global maximum over the permitted interval. It is useful to distinguish the two claims. A stationary point is a candidate found from local change. A maximum for the entire problem must be compared with the permitted domain and any relevant boundaries or other candidates. The method should not stop merely because a derivative has been set to zero.

Change the physical constraint: thirty-six metres of fencing encloses three sides of a rectangle, with a straight wall forming the fourth side. Let x be the width perpendicular to the wall and y the length parallel to it. Then 2x + y = 36, not 2x + 2y = 36. Area is A = xy = x(36 − 2x), with 0 < x < 18. Differentiating gives 36 − 4x = 0, so x = 9 and y = 18. The maximum area is 162 square metres.

The wrong answer x = y = 9 may come from assuming that every maximum-area rectangle with a fencing limit must be a square. In the three-sided problem, the wall changes the cost of extending one direction relative to the other. The square conclusion from the previous problem depended on a different perimeter constraint. Ask the learner to explain the coefficient two in 2x + y = 36. It counts the two fenced widths. A correctly drawn boundary can repair the model before any derivative is taken.

Sometimes the largest or smallest permitted value occurs at a boundary rather than a stationary point. On the closed interval 1 ≤ x ≤ 4, the function f(x) = x² has minimum one and maximum sixteen. Its derivative 2x never vanishes inside that interval. A student who reports “No maximum or minimum” because no stationary point was found has ignored the domain. Compare this with the open interval 1 < x < 4, where the function approaches those boundary values but never attains them. Endpoint inclusion changes what can be claimed.

An integer constraint adds another layer. Suppose a model gives a score S(n) = 30n − 2n² for a whole-number choice n between zero and ten. The continuous vertex is at n = 7.5. Evaluate the adjacent permitted integers: S(7) = 112 and S(8) = 112. Both maximise the stated integer model. Rounding mechanically to eight would hide the equal alternative. The learner should explain why the mathematical variable is restricted to whole numbers and why neighbouring candidates must be checked rather than treating a decimal optimum as an automatically valid action.

For independent practice, a rectangle has perimeter twenty-eight units. Its maximum area is forty-nine at sides seven and seven. A three-sided enclosure has twenty-four units of fencing: widths six and parallel length twelve give maximum area seventy-two. For g(x) = (x − 2)² + 3 on 0 ≤ x ≤ 5, the minimum is three at x = 2 and the maximum twelve at x = 5. Require the learner to evaluate both endpoints in the last question and explain why the larger endpoint distance from two produces the larger square.

Before accepting an optimisation answer, ask for a sentence containing the target and the constraints: “This is the largest area among rectangles satisfying this fencing condition.” That statement is more precise than “This is the best shape.” A wider practical decision could involve access, materials, terrain or other factors not represented by the model. The exercise’s mathematical conclusion should remain exact without pretending to settle every real-world consideration. Self-explanation is strongest when it makes both the power and the boundary of the model visible.

26. Word problems: explain the representation before the procedure

A student can execute every algebraic step correctly while solving the wrong problem. Consider: “Mira has three times as many cards as Jon. After Mira gives Jon twelve cards, they have equal numbers. How many did Jon have at first?” Let Jon initially have x, so Mira has 3x. After the transfer, Jon has x + 12 and Mira has 3x − 12. Equality gives 3x − 12 = x + 12, hence 2x = 24 and x = 12. The total remains forty-eight throughout the transfer.

The tempting equation 3x − 12 = x accounts for Mira’s loss but ignores Jon’s gain. It gives x = 6, and a learner may present impeccable working from that point onward. Substitution into the story reveals the problem: initially Mira would have eighteen and Jon six; after the transfer, they would have six and eighteen, not equal amounts. The earliest wrong step was representation, not equation solving. Ask, “Where did the twelve cards go?” The answer must appear on both sides of the state change.

A before-and-after table can organise the quantities. The two rows are Mira and Jon. The two columns are before and after. Enter 3x and x before, then 3x − 12 and x + 12 after. Place the equality sign between the after quantities because that is when the story says they become equal. The table’s purpose is to preserve time and ownership. A bar model can serve the same purpose, but the learner must still identify which bars describe the original state and which describe the later state.

Now change only the event: “Mira has three times as many cards as Jon. Mira loses twelve cards, and they then have equal numbers.” The correct equation is now 3x − 12 = x, giving x = 6. The equation that was wrong in the transfer problem is right in the loss problem. This contrast prevents correction from becoming another rigid slogan. It is not always necessary to add twelve to Jon’s amount; it is necessary when the stated event gives him those cards. The language determines the model.

Comparison wording requires similar care. “A rope is five metres longer than another rope” states an additive difference. “A rope is five times as long as another rope” states a multiplicative relationship. If the shorter rope is x, the longer lengths are x + 5 and 5x respectively. A learner should be able to supply a concrete example: with x = 3, the lengths are eight and fifteen. The shared numeral five is not enough to choose the operation. Ask what stays constant when the shorter rope’s length changes: a difference or a ratio?

Insufficient information is also a valid result. “Two boxes contain thirty marbles altogether. How many are in the first box?” has no unique numerical answer without another constraint. Ten and twenty, or twelve and eighteen, both satisfy the total. The explanation should identify what additional information would determine the split, such as a difference or ratio. Learners should not be trained to believe that every collection of numbers must produce one number through a familiar operation. Recognising underdetermination is part of mathematical understanding.

For transfer, two friends initially have amounts in the ratio 4:1. The first gives the second nine tokens, and they become equal. Let the second initially have x. Then 4x − 9 = x + 9, giving x = 6; the first initially has twenty-four. If instead the first loses nine tokens, equality gives 4x − 9 = x, so x = 3. Ask the learner to explain why the first change reduces the difference by eighteen while the second reduces it by nine. The difference is affected by both sides of a transfer.

An effective representation explanation names the entities, quantities, units, time points and relation that matters. It need not retell every decorative detail in the story. “At the end, both amounts are equal; one decreases by nine and the other increases by nine” is enough to guide the equation. For more practice on this specific failure, use the existing problem-representation guide. Keep representation repair separate from arithmetic repair so the student understands why a perfectly calculated answer can still be wrong.

27. Repair an explanation without replacing the learner’s thinking

A correction should leave the learner with something specific to reconstruct. In a fictional tutoring exchange, Nia simplifies 4 − 3(2 − x) as 1(2 − x). The teacher could supply the correct expression immediately, but first asks what the original multiplication applies to. Nia points to the bracket. The teacher then asks whether the four is part of that product. It is not. This identifies the first structural mistake: Nia combined four and negative three as if they were like coefficients of the same bracket.

The teacher now models only the disputed structure: 4 + (−3)(2 − x). Nia computes the products, giving 4 − 6 + 3x = 3x − 2. The corrected answer is useful, but the learning check is not finished. Give 7 − 2(4 − y) without the first problem visible. The correct result is 2y − 1. Ask Nia to state why seven cannot be combined with negative two before distribution. Her explanation should refer to their different roles in the expression, not merely say that the teacher told her not to.

Compare another fictional learner, Arun, who expands the first expression correctly but writes 4 − 6 = 2. His first error is arithmetic with signed numbers, not scope of multiplication. Repeating the entire structural explanation would spend time on a secure part while leaving the actual slip insufficiently addressed. Ask Arun to locate four on a number line and move six units left, or use a signed-number fact he already understands. Then return to the expression. The same final wrong answer can arise from different first errors.

A repair record can contain four short fields: the original decision, the reason it failed, the corrected rule with its condition, and a new test. For a percentage error, the record might say: “I added twenty per cent of the sale price to reverse a discount. The discount was based on the original price, not the sale price. The sale price is eighty per cent of the original, so divide by 0.8. Test with a different discount and a different unknown.” This is more useful than copying the whole corrected solution without identifying the changed decision.

Do not turn the record into a catalogue of personal faults. Write “The equation omitted the recipient’s gain,” not “I am careless with word problems.” The first statement points to an inspectable mathematical feature; the second is broad and difficult to act on. A student may indeed make hurried slips, but the useful next action still needs to be concrete: underline the transfer, build the after-state quantities, or check the total. Describe the work precisely enough that the learner can recognise the same risk in a different task.

Feedback can be too complete. If every line is rewritten by the teacher and the learner merely copies it, the page may improve without showing that the student can generate the missing reasoning. Feedback can also be too sparse: a cross beside an unfamiliar concept may leave no route forward. Use a graduated response. Start with a question naming the relevant relationship. Add a contrasting example if needed. Supply a clear model when the knowledge is missing. Then ask for a fresh attempt rather than endless guessing.

A useful error-analysis exercise presents this incorrect solution: x² = 4x, so x = 4. Ask whether four is wrong, whether the answer is complete and which step needs a condition. Four is valid but incomplete; dividing by x excludes zero. The repaired solution is x(x − 4) = 0, giving zero or four. Now ask the learner to invent a parallel problem in which dividing by the variable loses a solution, such as y² = 7y. The creation task reveals whether the correction has become a general condition rather than a remembered answer.

For a final check, return to the mistake after intervening work. Do not announce its category. A learner who can identify the same structural risk independently has more control than one who succeeds only under the heading “Remember not to divide by zero.” This is the bridge from feedback to self-monitoring. The broader eduKate error-correction guide develops that process; here, the mathematics supplies specific evidence of whether the repair is holding.

28. A mixed transfer assessment: choose the principle without a topic label

The following assessment is an original practice set, not a standardised measure or an official examination paper. Select questions within the learner’s taught curriculum. Its purpose is to remove the helpful topic headings that accompany most examples in this workbook. For each selected question, ask for an answer and one sentence explaining the decisive relationship. Do not require equal explanation length for every item. A diagram, equation or counterexample may provide the most direct evidence. Keep the answer discussion closed until an independent attempt has been made.

Questions

Question A. A bottle is two thirds full. An amount equal to one sixth of its full capacity is added, with no spillage. What fraction of capacity is now filled? Question B. A price after a twenty-five per cent discount is fifty-four dollars. What was the original price? Question C. Two amounts are in the ratio 5:3 and differ by fourteen. Find both amounts. For each answer, identify what the given number or fraction refers to before calculating.

Question D. Simplify 6 − 2(4 − 3x). Question E. Solve 5 − 3y ≤ 17. Question F. Solve z² = 6z over the real numbers. Question G. Solve (t² − 16)/(t − 4) = 8, using the domain of the original expression. These questions test scope, order, solution completeness and restrictions. Do not assume that every item has exactly one numerical solution.

Question H. A traveller covers forty kilometres at twenty kilometres per hour and forty kilometres at forty kilometres per hour, without stopping. Find the average speed for the whole journey. Question I. A group of six has mean score twelve and a group of nine has mean score eighteen. Find the combined mean. Question J. A rectangle has sides three and seven. Both lengths are enlarged by factor four. Find the new perimeter and area, with appropriate units if the original lengths are in centimetres.

Question K. A right triangle has hypotenuse twenty-five and one perpendicular side seven. Find the other side. Question L. Two similar shapes have area ratio 16:81. A side of the smaller shape is twelve. Find the matching larger side. Question M. A bag has two orange and four purple counters. Two are drawn without replacement. Find the probability of two orange. Explain what changes after the first orange counter is removed.

Question N. Find the gradient and equation of the line through (1, 4) and (5, 12). Question O. For y = x² − 2x + 5, find the point and gradient at x = 3, then the tangent equation. Question P. Let f(x) = x² on the domain x ≤ 0. Write its inverse, including the inverse’s domain. These advanced questions are optional for learners who have not studied the relevant notation.

Open the answer discussion after attempting the selected questions

A: Five sixths. Both fractions use the bottle’s full capacity as the same whole, so two thirds becomes four sixths before adding one sixth. B: Seventy-two dollars. Fifty-four is seventy-five per cent of the original, so divide by 0.75. C: Thirty-five and twenty-one. The difference of two ratio parts equals fourteen, making each part seven. A learner dividing fourteen by eight has confused the difference with a combined total.

D: 6x − 2, because negative two multiplies the complete bracket, producing negative eight and positive six x. E: y ≥ −4. Subtract five to obtain −3y ≤ 12, then divide by negative three and reverse order. F: z = 0 or z = 6. Factor z(z − 6) = 0 rather than dividing by z and losing zero. G: No solution. The domain excludes t = 4, while simplification gives t + 4 = 8, whose only candidate is the excluded value four.

H: 80/3 kilometres per hour, or twenty-six and two thirds. The two sections take two hours and one hour, giving eighty kilometres in three hours. I: 15.6. The totals are seventy-two and 162, for 234 across fifteen observations. J: Perimeter eighty centimetres and area 336 square centimetres. The new sides are twelve and twenty-eight. Perimeter scales by four, while area scales by sixteen.

K: Twenty-four, because 25² − 7² = 576 and the positive square root is twenty-four. L: Twenty-seven. The length ratio is 4:9, so twelve in the smaller shape corresponds to twenty-seven in the larger. M: 1/15. The probability is (2/6) × (1/5), because one orange remains among five counters after the first orange is removed. Reusing 2/6 for the second probability would describe replacement, not the stated experiment.

N: Gradient two and equation y = 2x + 2. The vertical change eight is divided by horizontal change four, and substitution of (1, 4) gives the intercept two. O: The point is (3, 8), the gradient is four and the tangent is y = 4x − 4. The coordinate comes from the original function and the gradient from 2x − 2. P: f⁻¹(x) = −√x for x ≥ 0. The original domain selects the nonpositive input associated with each nonnegative squared output.

Interpret the responses by error type rather than only by the number correct. A and B inspect reference quantities. D inspects the scope of an operation. F and G distinguish a valid candidate from a complete, permitted solution set. H and I require combining totals with appropriate weights. A student may perform very differently across these categories. That pattern is more useful for planning than a broad label such as “weak at problem solving,” because each category points to a specific relationship that can be taught and retested.

For an explanation check, compare a correct numerical answer supported by a false rule with an incorrect answer following a valid setup and a small arithmetic slip. Neither should be ignored, but the repairs differ. The false rule needs a contrasting case that makes its failure visible. The execution slip needs a targeted check or brief practice of the operation. Do not reward a long explanation merely because it sounds confident. Ask whether its stated rule would remain valid if the numbers, order or domain changed.

Repeat a small selection later with new values and without the old headings. Keep the original conditions comparable, and record whether the student initiated the relevant check without prompting. This does not produce a causal experiment by itself: practice, feedback and familiarity all change across the period. It does provide useful evidence about the learner’s current independence. The assessment is successful when it tells teacher and student what to do next, even when the result reveals that a supposedly completed repair needs another pass.

29. Teaching without turning every solution into an essay

A self-explanation lesson can fail by asking too much language of too little mathematics. The learner spends the session composing sentences about steps that were already clear, while the unfamiliar decision remains hidden. Start with the purpose of the prompt. Are you checking a quantity’s meaning, a condition, a transformation or a method choice? Ask for that one thing. “Why is this the original amount?” is more targeted than “Explain your working.” The response should be long enough to reveal the relationship and no longer than the task requires.

A proposed twenty-minute lesson might begin with a two-minute independent attempt, followed by three minutes locating the first uncertain step. Spend five minutes on one correct model and a carefully chosen contrast. Use five minutes for the learner to complete a new problem with reduced support, and the final five for checking, a short repair note and a plan for later retrieval. These timings are examples of a workable allocation, not research-established optimal doses. A learner missing essential prerequisites may need more instruction and fewer independent questions at the outset.

In a small group, give each learner time to commit to an answer before discussion. Otherwise the quickest voice may supply a route that the others recognise but could not have generated. Ask different learners to explain different parts: one names the quantity, another identifies the legal transformation, and a third checks the answer against the original conditions. Then require a brief individual transfer question. Group discussion can produce a correct shared solution while leaving individual understanding uneven; the independent return makes that distinction visible.

Use sentence starters sparingly. “This step is allowed because…” and “The denominator represents…” can help a learner begin. A completed script containing the entire reason can also conceal whether the learner understands it. Fade support by moving from choosing between explanations, to completing an explanation, to producing one, and finally to solving without being explicitly asked to explain. The progression need not be linear for every topic. Restore a prompt when a new domain introduces a genuinely different condition.

Avoid the habit of asking why only after wrong answers. Students may learn that a request for explanation signals failure and immediately abandon a correct method. Ask occasionally about a correct answer, and make the purpose explicit: you are checking which relationship the learner used. Equally, do not treat every alternative route as a mistake. A valid but less efficient method may be a sound starting point. Compare efficiency after establishing correctness, and explain why a different route reduces work in this particular problem.

The teacher’s own explanation should show a decision rather than merely an expert performance. In a reverse-percentage problem, say, “The given amount is after the discount, so it represents the remaining fraction of the original.” Then write the equation. In a geometry problem, name the right-angle evidence before selecting Pythagoras. The model should expose the choice the novice is expected to make later. A flawless page of calculations that never reveals why a method was selected may leave the most important part of the task unexplained.

When choosing examples, vary the feature that tests the intended distinction. To teach equal-time versus equal-distance average speed, do not also introduce difficult decimals, unfamiliar units and a long narrative in the contrast. To test domain restrictions, include a candidate that is allowed and one that is forbidden while keeping the algebra manageable. Later, combine those demands. Early contrasts should make the condition visible; later mixed problems should test whether the learner can still find it when the page no longer announces what to notice.

Keep educational support distinct from a commercial claim. A small class may make it easier to hear individual explanations, but the useful outcome still depends on what is taught, checked and practised. Class size alone does not establish that a particular learner will improve. For families evaluating mathematics support, ask to see how initial work is diagnosed, how corrections lead to new attempts, and how later independent performance is checked. The existing mathematics-tuition explanation is the appropriate route for that separate decision; this workbook remains a free learning resource.

30. A seven-day practice design: make the return test part of the plan

The schedule below is an original planning example, not a claim that seven days is the optimum interval for every learner or mathematical topic. Its purpose is to prevent a common omission: the student receives an explanation, corrects one page and never checks whether the idea survives beyond that page. Choose one narrow relationship, such as the reference base in percentages or the nonzero condition on division. Do not attempt to repair fractions, algebra, geometry and calculus simultaneously in one short week.

Day one: establish the starting point. Give two short questions requiring the target relationship, without hints. Keep the original answers. Ask for an explanation only after the learner has committed to a method. Record whether the problem is an absent concept, a false rule, an unhelpful representation or an execution slip. Use the evidence to select one chapter or worked example. The starting record should be brief enough to revisit; a photograph or retained page can preserve more useful information than a vague note saying “needs more practice.”

Day two: make the distinction visible. Work through one correct example and one contrast. For reverse percentages, compare “Find twenty per cent of eighty” with “Eighty is twenty per cent of what number?” The answers are sixteen and four hundred. Ask the learner to explain what eighty represents in each. Then use a new pair with different numbers. If the explanation remains inaccurate, provide a clearer model rather than adding five more pairs that rehearse the same confusion.

Day three: reduce support. Remove part of the worked solution. Leave the relationship statement visible but ask the learner to construct the equation and finish the arithmetic. Alternatively, provide the equation and ask what each term represents in the story. These two tasks inspect different directions of understanding. A student may move easily from words to symbols but struggle to interpret symbols, or the reverse. Choose the direction that was weak in the initial attempt instead of using the same worksheet format by default.

Day four: change the surface. Put the relationship into a different setting or representation. A percentage base can appear in prices, lengths, attendance proportions or a bar chart. A nonzero restriction can appear in an algebraic fraction or an equation where dividing by a variable loses a root. Keep the new context understandable. The purpose is to test whether the learner recognises the principle, not whether they can decode an unrelated vocabulary challenge that has accidentally been added to the problem.

Day five: mix the choice. Place the target question among two other familiar question types. Do not label the target chapter. The learner must decide which relationship matters rather than applying the method announced by a heading. Ask afterward what feature triggered the choice. “It looked like the example” may be a starting observation, but a stronger answer identifies the reference amount, equal-time condition, right angle or excluded denominator. The explanation should increasingly refer to structure rather than visual resemblance.

Day six: consolidate the correction record. Review the smallest useful statement of the repaired rule and its boundary. Add one example where the rule applies and one where an apparently similar move is not allowed. For cancellation, use a genuine common factor in one expression and an uncancelled sum in another. The learner should explain the contrast, not merely mark one tick and one cross. Keep this review short enough that it does not become another full reteaching session unless the evidence shows that reteaching is actually necessary.

Day seven: test independently. Use a new question with no model, no topic heading and no reminder of the repaired rule. Compare the work with day one. Check correctness, the first decision and whether the learner initiated a relevant check. A correct answer with the old false explanation is not a fully secure repair. An accurate representation with one arithmetic slip is different evidence from complete failure to recognise the method. Record the distinction and decide whether to move to maintenance, extend practice or return to instruction.

For a two-week extension, reduce the frequency of explicit explanation prompts while preserving occasional mixed questions. Add an unfamiliar representation only after the basic condition survives. A student who has mastered the reverse-percentage model might next interpret the same relationship in an equation with letters. A student who still cannot identify the original quantity should not be pushed into more complex algebra merely because the calendar has advanced. Progression follows evidence, not the number of days already spent.

The record can remain small: date, task, first decision, outcome, help needed and next return. “Help needed” should distinguish reading the question aloud, clarifying a word, supplying the mathematical relationship and performing the calculation. These supports have different implications for independence. The aim is not to withhold necessary support, but to know which part of the work the learner carried. A schedule becomes useful when it produces that clarity and turns the next session into a reasoned choice rather than another undirected page of practice.

31. Support different learners without confusing language fluency with mathematical understanding

A learner may understand a mathematical relationship before being able to describe it smoothly in English. Another may speak fluently while relying on an incorrect rule. The teaching task is to examine the mathematics rather than treating the polish of the explanation as a direct measure of understanding. Offer several response forms when appropriate: a labelled diagram, a numerical counterexample, a completed equation, a short oral explanation or a written sentence. The required relationship should remain the same even when the expressive route changes.

For a multilingual learner, it can help to discuss the idea in a familiar language before settling on the English terms used in the course. Keep the technical distinctions explicit: ratio and fraction, factor and term, area and perimeter, speed and distance. Do not assume that a translated word carries the same classroom associations automatically. Ask the learner to point to the mathematical object or construct an example. The goal is accurate connection between language and concept, not memorisation of an English sentence whose quantities remain unclear.

A younger learner may need concrete objects for an equality or fraction explanation. Three equal piles can make a coefficient meaningful; a strip divided into sixths can make a common unit visible. The object is useful only when its features are connected to the symbols. Ask what one pile represents, or which drawn part corresponds to the denominator. Otherwise the child may complete the practical activity while the symbolic relationship remains separate. Move between object, picture and notation with explicit correspondences.

An experienced learner may need fewer prompts and more strategically selected challenges. Ask for the condition behind a cancellation, a counterexample to an overbroad shortcut, or a comparison between two valid methods. Do not require a long explanation of every secure arithmetic step. For such a learner, the most useful question may be “What could make this elegant route invalid?” That question directs attention to domains, hidden assumptions and completeness rather than slowing all work to the pace of a first introduction.

When writing is physically effortful, separate the mathematical objective from the quantity of handwriting. An oral response or a brief typed explanation may allow the relationship to be checked without exhausting the learner on transcription. That choice does not mean abandoning written mathematical notation where it is part of the task. It means identifying whether the immediate goal is conceptual reasoning, written execution or sustained assessment performance. Supports should be chosen with the learner and relevant educators; this workbook is not a diagnostic or clinical guide.

A learner who becomes distressed by repeated public questioning may provide better evidence in a private written attempt or a quiet conversation. Avoid interpreting a pause as defiance or a wrong answer as a character flaw. At the same time, reassurance should not replace the needed mathematics. A calm response can be specific: “Your equation tracks the first amount correctly; now we need to show where the transferred counters went.” This acknowledges what is present and identifies the next repair without making a global judgement about the learner.

Parents can support the routine without pretending to be subject experts. Ask what the number represents, whether the units match, and how the answer can be checked against the original question. When the parent is unsure of a rule, consult the teacher or a reliable worked explanation rather than endorsing a confident guess. A calculator can verify arithmetic but may not reveal a wrong model. An automated explanation can also be incorrect or omit conditions. The standard remains the mathematical relationship, not the confidence of the person or tool presenting it.

Finally, recognise when self-explanation is not the next priority. A learner missing multiplication facts may need targeted fluency work alongside conceptual teaching. A learner unable to read the task may need language support. A learner who has not been taught a theorem needs instruction before being asked to justify its use independently. The strategy is not a substitute for those foundations. It is a way to make selected reasoning visible once there is enough knowledge to inspect, repair and eventually carry forward without support.

32. Questions, teaching guide and the next useful route

Does a correct answer prove that the student understands?

No single answer proves broad understanding. It may have come from a valid method, a memorised pattern, a lucky choice or a copied route. Ask for one decisive reason and then change an important feature of the task. In a ratio problem, change the given quantity from total to difference. In a quadratic, include a zero solution. The response to the change helps distinguish a useful principle from a procedure tied to one familiar appearance. Do not assume that every correct answer is shallow; use the check when the evidence is uncertain.

Should students explain every line?

Usually the better teaching choice is to explain selected decisions. Every-line narration can be helpful during a first introduction to a short procedure, but it can also add work that reveals little. Ask which step is most likely to hide the learner’s misunderstanding. A fraction problem may need one sentence about the shared unit; a domain problem may need one restriction; an optimisation problem may need a clear statement of the constraint. Reduce the prompt when that relationship is secure and test whether the learner can use it without being reminded.

What counts as a good explanation?

For this workbook, a good explanation is accurate, relevant and conditional where necessary. It identifies what the quantities mean or why the operation is valid. It does not need impressive language. “Both fractions refer to the same bottle capacity, and I have changed them into sixths” is good. “I used a sophisticated common-denominator strategy” is less informative if the learner cannot identify the whole. Judge the explanation by whether it would help produce a correct decision on a changed question.

What should happen when the explanation is wrong?

Do not ask the learner to repeat it more confidently. Locate the false relationship, provide a clear correction or contrast, and require a new attempt. A counterexample can expose an overbroad claim, but it may not supply the correct rule by itself. After showing that equal percentage increases and decreases do not cancel, explain the changing reference amounts. Then test a new pair of changes. The goal is a replacement model that can be used, not simply an admission that the old answer was wrong.

Are wrong worked examples always useful?

No. A wrong example without enough knowledge or corrective support can leave the learner uncertain about what should be believed. In this workbook, incorrect examples are explicitly labelled and followed by a mathematical account of the error. A novice may need a secure correct model first. Use wrong examples when they expose a particular misconception that the learner can inspect, not simply because finding mistakes appears more active than studying a correct solution. The evidence caution near the beginning of the article is important here.

Can a diagram replace a written explanation?

A diagram can provide strong evidence when its correspondences are explicit. In a ratio problem, the label showing that a difference of two parts equals fourteen may explain the calculation directly. An unlabeled sketch that merely resembles the story may not. Ask the learner to identify what each part represents and how the diagram justifies the chosen operation. For a formal assessment, follow the task’s stated requirements; this workbook’s flexible teaching responses are not a promise that every answer format is accepted in every examination.

How does this relate to showing working?

Working records the mathematical route, while explanation makes selected reasons for that route explicit. They overlap but are not identical. A complete equation with correctly labelled quantities may communicate the reasoning well. A long written paragraph without a valid equation may fail to solve the task. Use explanations to strengthen the decisions that produce clear working, not to replace working with commentary. In timed practice, the learner should know which notation efficiently preserves the relationship and which additional words are necessary for the question being asked.

Can this help with primary-school mathematics?

The earlier chapters are designed to make elementary relationships visible: equality, common fraction units, ratios, percentage bases and rates. Select only ideas that the learner has been taught. A primary learner does not need the calculus chapters to benefit from explaining why two quantities can be combined. Use concrete objects, diagrams and short sentences where helpful. The continuity across the workbook is the habit of preserving meaning, not an expectation that every child should study every topic immediately.

Can this help with Additional Mathematics?

The later chapters address conditions that become increasingly important in advanced work: nonzero divisors, logarithm domains, inverse-function restrictions, stationary-point classification and contextual optimisation. A learner may execute the symbolic rules rapidly while missing those conditions. Use the explanation prompt to identify the domain or reason a theorem applies, then return to independent mixed problems. The existing Additional Mathematics guide remains the broader subject route; this article is an applied reasoning workbook within that learning journey.

How should calculators or automated tools be used?

Use them for a defined checking role. A calculator can verify 72/0.8, but it cannot decide whether seventy-two represents the original or discounted price unless the situation has been modelled correctly. A symbolic tool may simplify a rational expression while presenting domain information in a form the learner overlooks. Ask the student to state the relationship and restrictions first, then compare the tool’s output. Follow the relevant school’s or assessment’s tool rules; this workbook does not prescribe permission to use a device in any particular examination.

How will a family know the practice is helping?

Look for specific changes in later independent work: the original quantity is identified, both sides of an equation are transformed completely, excluded values are checked, or a method is rejected when its condition is absent. Keep a small before-and-after sample. Improvement in a coached explanation is encouraging but different from improvement in an unprompted question. Do not demand that every sample improve monotonically or treat one result as proof of a universal method. Use the pattern to choose the next teaching action.

What is the most important habit to retain?

Before a consequential transformation, ask, “What makes this step legitimate here?” Sometimes the answer is a shared unit, sometimes an equality-preserving operation, sometimes a right angle, sometimes a domain condition. After solving, return to the original question and ask whether the result satisfies it. This pair of checks links meaning to procedure without requiring constant narration. The habit becomes valuable when it appears spontaneously on a new problem, not only when a worksheet heading tells the student to explain.

Teaching guide: turn one chapter into an independent capability

Choose one chapter by the first error in a real sample of work. Write the intended relationship in a single sentence before teaching. Select one correct example that makes that relationship visible, and one contrast that changes the relevant condition. Decide what response will count as evidence: identifying a whole, labelling a corresponding side, preserving all solutions, or checking a candidate’s domain. This preparation keeps the lesson focused on a mathematical decision rather than the vague ambition of making the student explain more.

During teaching, separate the student’s attempt from the model. Let the learner commit to a representation or next step, then compare it with the example. Ask what changed and why. Supply missing knowledge directly when necessary; do not turn uncertainty into a prolonged guessing contest. Follow the correction with a new problem whose surface features differ. When the learner succeeds, reduce the support. When the learner fails, identify whether the original concept, the new context or an execution demand caused the difficulty.

End with a small return plan. Record the condition to check, the next unprompted question and the evidence that would justify moving on. A useful lesson leaves the learner with a more accurate way to decide, not just another completed page. Its final product might be one reliable sentence—“That denominator cannot be zero”—combined with the ability to notice when the sentence matters. This is the central promise of the workbook as a teaching design: make the hidden mathematical decision visible, repair it accurately and give responsibility back to the learner.

Continue through the eduKate learning ecosystem

For the general strategy, continue to How Self-Explanation Works and the more specific self-explanation prompts guide. For the move from model to independent work, use example–problem pairs and worked-example fading. For wider subject navigation, use How Mathematics Works.

When the main difficulty is selecting or interpreting the task, use the problem-representation guide. When a corrected idea disappears on a new question, use transfer of learning. The preceding article in this improvement series, How Curiosity Improves Learning, addresses how questions begin an investigation; this workbook addresses how a learner justifies and checks the mathematical answer that follows.

Evidence used in designing the workbook

The research discussion draws on Bisra and colleagues on induced self-explanation, Rittle-Johnson, Loehr and Durkin on mathematics self-explanation, and Barbieri and colleagues on worked examples in mathematics. Instructional context also comes from the MathByExample research programme and the IES instruction-and-study practice guide. These sources are not endorsements of this article or evidence that its exact exercises have been tested.

Final checkpoint: explain the relationship that permits the step, preserve its conditions, verify the result against the original problem, and then solve again without the model. Mathematics becomes more independent when the learner can carry those decisions, not merely repeat the teacher’s sequence.