BTT Mathematics / Primary Mathematics Learning Hub / Working Mathematically

Primary Mathematics: Mathematical Modelling, Assumptions, Context and Validation | Worked Learning Guide

A mathematical model is useful when it captures enough of a real situation to answer the question—without pretending that the real world has become as simple as the equation.

Primary Mathematics is full of models. A bar stands for an amount. A map scale stands for a distance relationship. A table stands for repeated measurements. A speed calculation treats a journey as distance divided by time. A budget problem turns prices and limits into arithmetic. A rectangle can model a floor, but only if the shape and measurements actually match the floor well enough for the intended estimate.

Modelling is the disciplined movement between context and mathematics. The learner decides what quantities matter, which details can be ignored, what assumptions are being made, which relationship links the quantities, how to calculate, and whether the result still makes sense when returned to the original situation.

This guide develops that cycle explicitly. It is not about making Primary Mathematics artificially advanced. It is about exposing the reasoning that already sits inside word problems, measurement, rate, percentage, scale, data and estimation.

The modelling cycle · Choose quantities · State assumptions · Build relationships · Validate · Model limits · 24 questions · Worked solutions

1. The Primary modelling cycle

1. Context: What real situation is being described?
2. Question: What quantity must be found or decided?
3. Quantities: Which measurements, counts, rates or limits matter?
4. Assumptions: What are we treating as fixed, equal or negligible?
5. Relationship: Which mathematical structure links the quantities?
6. Calculation: Carry out the mathematics.
7. Interpretation: Return the numerical result to the context.
8. Validation: Check whether the result is plausible and satisfies the original conditions.

The cycle can move backwards. If the result is impossible, revisit the units, assumptions, data or relationship. A model is not sacred merely because the arithmetic was correct.

2. Start by naming the quantities

Question: A bus travels 180 km in 3 hours. Find its average speed.

Relevant quantities:

• distance = 180 km,
• time = 3 h,
• required quantity = average speed.

Relationship: speed = distance ÷ time.

Calculation: 180÷3=60.

Interpretation: average speed = 60 km/h.

The answer is not “60 km”. The model’s output has a new unit because it compares distance with time.

Irrelevant detail

If the bus is blue and carries 24 passengers, those facts do not matter to the average-speed calculation unless another part of the question uses them.

Modelling includes knowing what to leave out.

3. A representation is part of the model

A bar model can represent a quantity relationship without resembling the real objects. A map is not physically the territory. A table is not the experiment. The representation compresses selected information.

Worked example: comparison model

A has 18 more counters than B. Together they have 74.

Represent B as one unknown bar and A as the same bar plus an 18-unit segment. The model preserves the total and difference while discarding colour, container and arrangement.

2B+18=74, so B=28 and A=46.

4. Assumptions make hidden conditions visible

Suppose a recipe uses 250 ml of juice for each serving. To predict the amount for 8 servings by simple multiplication, we assume the same recipe ratio is used for every serving and there is no fixed extra quantity independent of serving count.

Under that model: 8×250 ml=2000 ml=2 L.

If the real recipe includes a fixed 500 ml base plus 150 ml per serving, direct proportion would be wrong. The model must match the stated structure.

Worked example: transport timing

A walker covers 6 km at an average of 4 km/h. Modelled travel time is 6÷4=1.5 h.

This model assumes the quoted 4 km/h is an average appropriate for the route. It does not claim the walker moves at exactly 4 km/h every second.

5. Exact models and approximate models

Some Primary Mathematics models are exact within the stated conditions. If 8 identical notebooks cost exactly $3 each, total cost is exactly $24.

Other models are approximations. Estimating paint for a wall may treat the wall as a rectangle and ignore a tiny fitting. Estimating crowd size may use average density rather than count every individual.

The answer should match the model’s precision.

Worked example: rectangular floor estimate

A room is approximately 5.8 m by4.1 m. Estimate area by rounding to6 m by4 m: about24 m².

Exact multiplication of the measured dimensions gives23.78 m², but the measurements themselves may already be approximate. Writing many decimal places does not make the physical estimate more certain.

6. Unit choice can simplify the model

A tank contains 2.4 L and receives 650 ml. Convert before combining:

2.4 L=2400 ml. New amount=2400+650=3050 ml=3.05 L.

Adding 2.4+650 without unit conversion creates a meaningless numerical combination.

Worked example: money

$3.75 can be modelled as375 cents when sharing or integer arithmetic is easier. If five equal items cost $18.75, then1875 cents÷5=375 cents=$3.75 each.

The representation changes; the monetary quantity does not.

7. Choose the relationship, not the story keyword

The word “more” can mean different structures.

• “A has 5 more than B” → A=B+5.
• “A is 5 times B” → A=5B.
• “A is 5% more than B” → A=1.05B.

A model fails if it uses the wrong relationship even when the arithmetic is flawless.

Worked example: percentage base

A price rises from $80 by 25%.

Increase=25% of80=$20. New price=$100.

Do not calculate 25% of the new price when the stated increase refers to the original price.

8. Proportional models need a zero-compatible structure

If notebooks cost $4 each with no fixed fee, cost is directly proportional to number of notebooks. Zero notebooks cost zero dollars; doubling quantity doubles cost.

If a delivery service charges $5 plus $2 per kilometre, total cost is not directly proportional to distance. At zero kilometres the model still gives $5.

Worked example

For 7 km, total cost=5+2×7=$19.

For 14 km, total cost=5+2×14=$33, not $38. Doubling distance does not double total because the fixed fee is counted once.

9. Piecewise rules require the learner to locate the correct interval

Suppose a parking model charges $3 for the first hour and $2 for each additional started hour.

A stay of 2 h10 min uses the first-hour charge plus two additional started hours? No. The first hour covers hour1. The interval from1 h to2 h is the first additional started hour. The partial third hour is the second additional started hour.

Total=$3+$2+$2=$7.

The rule changes after the first hour. A single constant multiplication would not model the tariff.

10. Discrete and continuous quantities behave differently

53 pupils travelling in vans of capacity8 require ceil(53÷8)=7 vans. The arithmetic quotient is6 remainder5, but six vans are insufficient.

A liquid volume of53 L divided equally among8 containers can produce6.625 L per container if fractions are allowed.

The context determines whether the answer must be a whole count or can be fractional.

11. Rounding direction belongs to the model

If 53 people need vans of capacity8, round up because every person needs a place.

If $53 is available and each ticket costs $8, the maximum whole number of tickets is6; round down because a seventh ticket would exceed the budget.

“Always round up” and “always round down” are both bad modelling rules. Interpret the remainder.

12. Geometry models need shape assumptions

To model a garden as a rectangle, the length-width description must be adequate for the question. An irregular garden with curved boundaries cannot be measured exactly by one rectangular formula unless the task explicitly asks for an approximation.

Worked example: path around rectangle

A rectangular field is60 m by40 m. One lap of the boundary is2(60+40)=200 m.

If the actual path cuts across rounded corners, the rectangle perimeter becomes an approximation rather than an exact route length.

13. Data models summarize and therefore hide detail

Two datasets can share the same average:

A: 10,10,10,10.

B: 4,8,12,16.

Both average10, but their spreads differ.

Using the average as “the typical value” may be useful, but it does not preserve every feature of the data.

14. Sampling creates model limits

If 20 pupils in one class are surveyed about a school-wide preference, their responses describe that sample. Generalising to the entire school requires an assumption that the sample is sufficiently representative for the purpose.

Primary learners need not use formal statistical theory to understand the principle: a small or biased sample can produce a misleading model of a larger group.

15. Scale drawings preserve selected relationships

At scale1 cm:5 m, a 7 cm drawing represents35 m.

The scale model preserves proportional length. It does not necessarily preserve texture, colour, height or other real-world properties.

Worked example: area scaling

If every length doubles, area becomes four times as large, not twice. A 2 cm by3 cm rectangle has area6 cm². Enlarging both dimensions by2 gives4 cm by6 cm, area24 cm².

Length scale and area scale are different model outputs.

16. Build a model from a messy context

A class trip has a fixed bus fee of $120 plus $8 per student. There are 27 students.

Relevant quantities: fixed fee120, per-student fee8, student count27.

Model: total=120+8n.

For n=27: total=120+216=$336.

If a second bus becomes necessary above30 students, the simple model would need revision for larger n.

17. Validate by returning to the context

After calculation, ask:

• Are the units correct?
• Is the magnitude plausible?
• Does the result satisfy all limits?
• Must the answer be whole?
• Did the model use the intended base or rate?
• Does a reverse calculation recover the original data?

Worked example: budget

Budget $100. Four items cost $18.50 each and delivery is $12.

Total=4×18.50+12=74+12=$86.

Validation: below100 by14, so the budget is sufficient and $14 remains.

18. Validate with a boundary or extreme case

Model: cost=5+2d.

At d=0, model gives5. If the real service truly has a $5 fixed starting fee, this makes sense. If zero use should cost zero, the model is incompatible with the context.

Boundary cases expose hidden assumptions cheaply.

19. Validate with a second representation

Ratio3:5 total64 gives one unit8 and quantities24,40.

Check using fractions: A is3/8 of total, and3/8×64=24. B is5/8×64=40.

Agreement between independent representations strengthens confidence that the model and arithmetic align.

20. A mathematically correct model can still answer the wrong question

A family wants the total journey duration including a 30-minute stop. The learner correctly calculates driving time as3 h and reports3 h. The arithmetic is correct but the model omitted a required component.

Correct output:3.5 h total elapsed time.

Validation means returning to the exact wording of the original decision.

21. Model limits should be stated, not hidden

Examples:

• “Assuming the speed remains constant…”
• “Treating the room as a rectangle…”
• “Using the sample as representative…”
• “Ignoring the thickness of the border…”
• “Assuming all packets have equal mass…”

A model can be useful even when simplified. The important question is whether the simplification matters to the answer.

22. Improve a model only when the added detail changes the decision

Suppose a rough classroom estimate of floor area is24 m². Measuring every skirting-board notch to the nearest millimetre may add enormous effort without affecting a decision about how many 2 m² mats are needed.

More detail is not automatically a better model. Precision should match purpose.

23. Common modelling errors

• Using every number in the question simply because it is present.
• Ignoring units.
• Assuming direct proportion when there is a fixed fee or threshold.
• Rounding mechanically without interpreting the remainder.
• Treating an approximate physical model as exact.
• Using an average as though it describes every individual value.
• Forgetting the assumptions that made a formula valid.
• Checking arithmetic but never returning the result to the real context.

24. A six-question modelling audit

1. What decision or quantity does the problem actually ask for?
2. What information matters?
3. What am I assuming?
4. What mathematical relationship represents the situation?
5. What does the calculated number mean in context?
6. What observation, estimate or condition would make me revise the model?

25. Practice: 24 original modelling questions

Questions 1–8: Build the model

1. A bus travels180 km in3 h. Identify the quantities and find average speed.
2. A recipe uses250 ml per serving. Model the amount for8 servings.
3. A taxi-style service charges$5 plus$2 per km. Find cost for7 km and14 km.
4. A class trip costs$120 fixed plus$8 per student. Find total for27 students.
5. 53 pupils travel in vans holding8 each. How many vans are needed?
6. A student has$53 and tickets cost$8 each. What is the maximum number of whole tickets?
7. A room measures approximately5.8 m by4.1 m. Give a sensible area estimate before exact multiplication.
8. Four identical notebooks cost$18.75. Find unit price using cents if useful.

Questions 9–16: Assumptions and boundaries

9. Explain why cost=$4n is directly proportional when each item costs$4 with no fixed fee.
10. Explain why cost=5+2d is not directly proportional to distance.
11. A path is modelled as a60 m by40 m rectangle. Find perimeter and state one assumption.
12. At scale1 cm:5 m, what real distance does7 cm represent?
13. If both dimensions of a rectangle double, what happens to area?
14. Two datasets10,10,10,10 and4,8,12,16 share what summary and differ in what feature?
15. A survey of20 pupils from one class is used to describe the whole school. State the modelling assumption.
16. A parking rule charges$3 first hour and$2 per additional started hour. Find charge for2 h10 min.

Questions 17–24: Validate and improve

17. Budget$100. Four items cost$18.50 each plus$12 delivery. Is budget sufficient?
18. A learner calculates average speed60 but writes60 km. Identify the modelling/communication error.
19. A journey has3 h driving plus30 min stop. Learner reports3 h total elapsed time. What was omitted?
20. For model C=5+2d, evaluate d=0 and explain what this tests.
21. A ratio3:5 has total64. Solve, then validate using fractions of the total.
22. Give one example where a rectangular model is useful but approximate.
23. Give one example where extra precision would not change the practical decision.
24. Create a simple real-world Primary Mathematics model, state one assumption and describe one way to validate it.

26. Worked solutions

1. Distance180 km, time3 h, average speed60 km/h. 2. Assuming constant recipe ratio:8×250=2000 ml=2 L. 3. $19 and$33. Fixed fee prevents direct proportion. 4. $336.

5. 7 vans; six hold48 and five pupils remain. 6. 6 tickets; seven would cost$56. 7. About6×4=24 m²; exact product23.78 m². 8. $4.6875 each mathematically; if price is required to cents, the context must specify how currency rounding is handled rather than silently inventing a legal tender allocation.

9. Zero items cost zero and multiplying item count by a factor multiplies total cost by the same factor. 10. At zero distance cost is$5, so doubling distance does not double total cost. 11. Perimeter200 m; assume the route follows the rectangular boundary exactly. 12. 35 m.

13. Area becomes4 times as large. 14. Same average10; different spread/variation. 15. Assume the class sample is representative enough of the school for the intended conclusion. 16. $7.

17. Total=$86; budget sufficient; $14 remains. 18. Average speed needs km/h, not km. 19. The30-minute stop; correct total3.5 h. 20. It checks whether the fixed-fee interpretation at zero usage matches the real context.

21. One ratio unit8; quantities24 and40. Fraction check:3/8×64=24 and5/8×64=40. 22. Example: estimate an irregular room as a near rectangle for rough flooring quantity. 23. Example: estimating mat count may not require millimetre-level wall detail. 24. Answers vary; a valid response must define quantities, relationship, assumption and a plausible check.

27. Modelling laboratory: choose between two models

Service A charges$4 per kilometre. Service B charges$10 fixed plus$2 per kilometre.

Models: A=4d and B=10+2d.

At d=3: A=$12, B=$16, so A is cheaper.

At d=8: A=$32, B=$26, so B is cheaper.

Set equal to find the break-even distance:4d=10+2d, so2d=10 and d=5 km. At5 km both cost$20.

This is modelling because the equation is tied to a decision: which service costs less at a stated distance? The answer changes across the domain, so one sample journey cannot decide the whole comparison.

28. Parent and tutor guide

When a learner gets a word problem wrong, ask whether the failure happened before the calculation. Did the child identify the right quantities? Did a fixed fee get mistaken for a rate? Was the unit converted? Did the context require a whole-number answer?

Separate model error from arithmetic error. A correct multiplication inside the wrong model needs a different repair from a correct model with one multiplication slip.

Use estimation before exact calculation. If the result is far from the expected scale, return to inputs and units before repeating the arithmetic.

Delayed return

Three days later, present one proportional cost, one fixed-fee cost, one capacity problem and one approximate geometry estimate. Ask the learner to state the model and one assumption before calculating.

29. Mastery receipt

• I identify the quantity the context asks for.
• I separate relevant from irrelevant information.
• I state assumptions when they matter.
• I choose relationships rather than keywords.
• I distinguish direct proportion from fixed-fee and threshold models.
• I interpret discrete counts and remainders in context.
• I validate units, magnitude and conditions after calculating.
• I know a useful model may still have limits.

Sources and scope

Illustrative Mathematics — contextualising, representing and modelling a problem situation illustrates the movement between a context and mathematical objects, pictures or equations.

NSW Department of Education — Mathematics K–6 Effective Teaching Approaches describes problem solving as forming mental representations, applying mathematical relationships to familiar and unfamiliar problems and devising methods when needed.

For Singapore subject scope, use the MOE Primary Mathematics Syllabus P1–P6, updated October 2025 alongside the learner’s current school programme.

Continue the capability batch

• Problem Posing, Questioning and Creating Variations
• Reflection, Self-Monitoring and Metacognitive Checking
• Productive Struggle, Resourcefulness and Persistence
• BTT Primary Mathematics Learning Hub

The Quiet Return

A model is a bridge, not the destination. Build it carefully, cross it with mathematics, then turn around and ask whether the bridge still lands in the real problem you meant to solve.