A length is recorded as:
7.4 cm, correct to the nearest 0.1 cm.
Is the true length exactly 7.4 cm?
Not necessarily.
It could have been 7.36 cm.
Or 7.44 cm.
Both round to 7.4 cm to the nearest tenth.
A rounded measurement is a compressed interval. Bounds recover the range of exact values that could have produced the reported rounded value.
The quick answer: move half a rounding unit in each direction
If 7.4 cm is correct to the nearest 0.1 cm:
half of 0.1 cm is 0.05 cm.
Lower bound:
7.4−0.05 = 7.35 cm.
Upper boundary:
7.4+0.05 = 7.45 cm.
Error interval:
7.35 ≤ L < 7.45.
The lower bound is included.
The upper boundary is not, because 7.45 would normally round to 7.5 to the nearest tenth.
Why the interval is half-open
Values from 7.35 upward round to 7.4 until the halfway point to 7.5 is reached.
At 7.45, ordinary half-up school rounding sends the value to 7.5.
So the possible original values include 7.35 but exclude 7.45.
The upper bound is a boundary value, not necessarily a value the original measurement could actually have taken under the stated rounding rule.
Nearest whole number
A mass is reported as 84 g to the nearest whole gram.
For a continuous measurement, the mathematical interval would be:
83.5 ≤ x < 84.5.
But if the quantity is a count of people, only integer values are physically possible.
Context matters.
Bounds are strongest when the learner distinguishes continuous measurement from discrete counting.
Nearest ten
A number is 3,470 correct to the nearest 10.
Rounding unit = 10.
Half-unit = 5.
So:
3,465 ≤ x < 3,475.
Nearest hundred
5,200 correct to nearest 100 gives:
half-unit = 50.
5,150 ≤ x < 5,250.
Decimal-place bounds
3.82 correct to 2 decimal places.
Rounding unit = 0.01.
Half-unit = 0.005.
3.815 ≤ x < 3.825.
Significant-figure bounds depend on the place of the last retained digit
4,700 correct to 2 significant figures.
The last significant digit is in the hundreds place.
So the rounding unit is 100.
Half-unit = 50.
4,650 ≤ x < 4,750.
The phrase “2 significant figures” does not itself tell us a fixed decimal increment.
We must first locate the second significant digit.
Small-number significant-figure bounds
0.037 correct to 2 significant figures.
The second significant digit is the 7 in the thousandths place.
Rounding unit = 0.001.
Half-unit = 0.0005.
0.0365 ≤ x < 0.0375.
Absolute error
If 7.4 cm is measured to the nearest 0.1 cm, the maximum rounding error is at most 0.05 cm in magnitude.
We can describe the reported value as having a half-unit error tolerance around it.
This connects bounds to measurement uncertainty.
Bounds for addition
Suppose:
2.4 ≤ A < 2.5
and:
5.1 ≤ B < 5.2.
For A+B:
smallest occurs at lower+lower:
2.4+5.1=7.5.
Largest boundary occurs at upper+upper:
2.5+5.2=7.7.
So:
7.5 ≤ A+B < 7.7.
Bounds for subtraction require opposite endpoints
For A−B:
smallest difference uses smallest A and largest B.
Largest difference uses largest A and smallest B.
Using the same intervals:
lower boundary:
2.4−5.2 = −2.8.
upper boundary:
2.5−5.1 = −2.6.
The endpoint logic matters more than a memorised “lower minus lower” rule.
Positive multiplication bounds
If all quantities are positive:
minimum product uses lower×lower.
maximum product boundary uses upper×upper.
Example:
A length is 8.2 cm to nearest 0.1 cm:
8.15 ≤ L < 8.25.
Width is 3.6 cm to nearest 0.1 cm:
3.55 ≤ W < 3.65.
Area lower bound:
8.15×3.55 = 28.9325 cm².
Area upper boundary:
8.25×3.65 = 30.1125 cm².
The actual area lies between these values.
Bounds propagate measurement uncertainty through a formula instead of pretending rounded inputs were exact.
Positive division bounds
For positive A/B:
minimum quotient uses smallest numerator ÷ largest denominator.
maximum quotient uses largest numerator ÷ smallest denominator.
Why?
To make a positive fraction as small as possible, shrink the top and enlarge the bottom.
To make it as large as possible, do the opposite.
Worked example: speed from rounded distance and time
Distance = 120 km correct to nearest kilometre.
Time = 2.5 h correct to nearest 0.1 h.
Distance interval:
119.5 ≤ d < 120.5.
Time interval:
2.45 ≤ t < 2.55.
Speed = d/t.
Lower speed bound:
119.5/2.55 ≈ 46.86 km/h.
Upper speed boundary:
120.5/2.45 ≈ 49.18 km/h.
The computed speed from rounded centre values, 120/2.5=48 km/h, is only one value inside the possible range.
Why centre-value calculation can be misleading
If two measurements are rounded, calculating from the displayed values treats them as exact centres.
That can be acceptable when the problem asks for an estimate.
It is not enough when the problem asks for guaranteed bounds.
Maximum possible versus upper bound
This language needs care.
If an interval is x<7.45, then 7.45 is the upper bound but is not itself an attainable value under the interval.
There may be no single greatest possible real value; values can get arbitrarily close to 7.45 from below.
In school questions, “upper bound” is the safer mathematical phrase.
Bounds with discrete quantities need context
If a count is reported as 1,200 people to the nearest hundred, the continuous rounding interval is:
1,150 ≤ x < 1,250.
But people are counted in whole numbers.
So feasible actual counts are integer values from 1,150 to 1,249.
The mathematical interval and the physical domain should both be respected.
Bounds and engineering tolerance
Suppose a part must fit through an opening.
The rounded nominal dimensions may suggest a fit.
But engineers care about worst-case combinations:
- largest possible part;
- smallest possible opening.
Bounds turn the question from “Does the average-looking value fit?” into “Is the fit guaranteed under stated uncertainty?”
Bounds and percentage error
If a measured value m has maximum absolute rounding error e, a relative error scale can be estimated by:
e/m.
Percentage error scale:
(e/m)×100%.
This helps compare the significance of the same absolute uncertainty at different measurement scales.
An uncertainty of 0.05 cm is small relative to 100 cm and much larger relative to 0.2 cm.
Common misconception 1: lower and upper bounds are found by adding and subtracting the full rounding unit
Use half the rounding unit.
Common misconception 2: the upper bound is always included
Under ordinary nearest-value rounding, the upper boundary is typically excluded.
Common misconception 3: significant-figure bounds always use ±0.5
The half-unit depends on the place value of the last significant digit.
Common misconception 4: for subtraction use lower−lower and upper−upper
To minimise A−B, use low A and high B.
To maximise it, use high A and low B.
Common misconception 5: rounded measurements can be treated as exact in every context
If the question asks for guaranteed extremes or error intervals, the rounding uncertainty must be propagated.
A diagnostic ladder
- Can the learner identify the rounding unit?
- Can the learner find half the unit?
- Can the learner write lower and upper boundaries?
- Can the learner write the half-open interval correctly?
- Can the learner handle decimal-place bounds?
- Can the learner handle significant-figure bounds?
- Can the learner calculate bounds for sums?
- Can the learner choose correct endpoints for differences?
- Can the learner handle positive products and quotients?
- Can the learner preserve units and physical domain constraints?
- Can the learner distinguish a central estimate from a guaranteed bound?
How this fits Secondary Mathematics
Bounds connect rounding, significant figures, measurement, mensuration, rates and mathematical modelling. They make uncertainty explicit rather than hiding it behind rounded inputs.
Exact subject-level requirements should be checked against the current MOE/SEAB Mathematics syllabus for the learner’s cohort.
The deeper lesson: every rounded value has a hidden neighbourhood
Rounding simplifies communication by replacing many possible values with one representative number.
Bounds reverse that compression enough to recover the uncertainty that matters.
When a measurement is rounded, the mathematics should remember what the display forgot: the true value lives somewhere in an interval.
