Four containers hold 3, 5, 8 and 12 litres of water.
The usual rule says:
3 + 5 + 8 + 12 = 28.
28 ÷ 4 = 7.
The average is 7 litres.
Correct.
But the rule hides the idea.
The mean is the value every item would have if the total were redistributed equally while the total stayed unchanged.
Imagine pouring water from the fuller containers into the emptier ones until all four contain the same amount.
No water is created.
No water disappears.
The total remains 28 litres.
When redistributed equally across four containers, each must hold 7 litres.
This equal-redistribution interpretation gives the formula meaning:
average = total value ÷ number of data values.
In Singapore’s current Primary Mathematics syllabus, average is part of Primary 6. The updated October 2025 MOE syllabus explicitly connects average with total value and number of data. That relationship is more useful than memorising “add and divide” because it can be rearranged to solve several different problem types.
The three quantities that define an average problem
For an ordinary arithmetic mean, three quantities are connected:
- average;
- total value;
- number of data values.
The core relationship is:
total = average × number of data values.
Therefore:
- average = total ÷ number;
- total = average × number;
- number = total ÷ average, when the context makes that quotient meaningful.
Many difficult average questions are simply these relationships hidden inside a story.
Why equal redistribution is stronger than a recipe
Consider the values:
4, 4, 4, 16.
Total = 28.
Average = 7.
No original value equals 7.
So the average is not necessarily “the most common value”, “the middle value” or “a value already in the data”.
It is a balancing level.
The 16 contributes 9 units above 7.
Each 4 is 3 units below 7.
Three deficits of 3 exactly absorb the surplus of 9.
At the mean, total positive deviation above the balancing level equals total negative deviation below it.
Primary learners do not need formal deviation notation, but the balancing image gives them a durable model.
Worked example: finding the average
Five scores are:
12, 17, 19, 21, 26.
Total:
12 + 17 + 19 + 21 + 26 = 95.
Number of scores = 5.
Average:
95 ÷ 5 = 19.
Interpretation:
If the total score of 95 were shared equally among five positions, each would receive 19.
Worked example: finding a total from an average
The average mass of 8 parcels is 3.5 kg.
Total mass:
3.5 × 8 = 28 kg.
This direction is often overlooked because learners associate average only with division.
But once average is understood as equal redistribution, the reverse relationship is natural:
eight equal shares of 3.5 kg contain 28 kg altogether.
Worked example: finding a missing value
Four numbers have an average of 18.
Three numbers are 11, 16 and 24.
Find the fourth.
Required total:
18 × 4 = 72.
Known total:
11 + 16 + 24 = 51.
Missing value:
72 − 51 = 21.
The key move is not “add and divide”. It is reconstructing the total implied by the average.
When one new value changes the average
A set of 4 values has average 12.
Total = 4×12 = 48.
Add a new value of 22.
New total = 70.
New number of values = 5.
New average = 70 ÷ 5 = 14.
The average rose because the new value, 22, was above the old average of 12.
This gives a powerful qualitative rule:
- add a value above the current mean → mean rises;
- add a value below the current mean → mean falls;
- add a value equal to the current mean → mean stays unchanged.
Removing a value works in the opposite direction
A set has average 20.
If a value greater than 20 is removed, the remaining average tends to fall.
If a value less than 20 is removed, the remaining average tends to rise.
This reasoning can predict direction before any exact calculation.
Average can be found by balancing around a convenient centre
Consider:
48, 51, 53, 47, 51.
Instead of adding directly, use 50 as a reference.
- 48 is −2 from 50;
- 51 is +1;
- 53 is +3;
- 47 is −3;
- 51 is +1.
Total deviation:
−2 + 1 + 3 − 3 + 1 = 0.
So 50 is exactly the mean.
This is a useful mental strategy because averages are balancing points.
Average is not the same as median or mode
For data:
2, 3, 3, 4, 18
Mean:
30 ÷ 5 = 6.
Median = 3.
Mode = 3.
The large value 18 pulls the mean upward.
This does not make the mean wrong.
It tells us that the mean uses every value and is therefore sensitive to unusually large or small observations.
Outliers can make an average unrepresentative
Suppose five monthly incomes in a tiny example are:
$2,000, $2,100, $2,200, $2,300, $20,000.
The mean is much higher than four of the five values.
That mean is mathematically correct but may not describe a “typical” member well.
A good data reader therefore asks two questions:
- Is the calculation correct?
- Is the mean a useful summary for the purpose?
Mathematics includes both computation and interpretation.
Average of averages can be dangerous
Class A has 10 students with average score 70.
Class B has 30 students with average score 80.
Is the combined average:
(70 + 80) ÷ 2 = 75?
No, because the groups are different sizes.
Class A total = 10×70 = 700.
Class B total = 30×80 = 2,400.
Combined total = 3,100.
Combined students = 40.
Combined average = 3,100 ÷ 40 = 77.5.
The larger group must contribute more weight to the combined mean.
This is an early encounter with weighted averaging.
Worked example: changing average after an extra item
A student has an average of 16 points over 5 games.
Total points = 16×5 = 80.
In the sixth game the student scores 28.
New total = 108.
New average = 108 ÷ 6 = 18.
Because 28 exceeds the old average, the mean rises.
Worked example: what score is needed?
A learner wants an average of 75 across 4 tests.
The first three scores are 68, 72 and 79.
Required total:
75×4 = 300.
Current total:
68 + 72 + 79 = 219.
Required fourth score:
300 − 219 = 81.
The average is acting as a total constraint.
Averages can hide variation
Two groups can have the same mean and very different data.
Group A:
9, 10, 10, 10, 11.
Group B:
0, 0, 10, 20, 30.
Both have average 10.
But their spread is completely different.
The mean tells us about centre.
It does not tell the whole shape of the data.
This prepares learners for later measures of spread such as range and standard deviation.
Common misconception 1: average must be one of the original values
3, 5, 8 and 12 average to 7, even though 7 is not in the original data.
Repair: interpret the mean as equal redistribution rather than selection.
Common misconception 2: divide by the largest value
The denominator is the number of data values, not the largest observation.
Repair: count how many equal shares the total is being redistributed into.
Common misconception 3: every “average” in everyday language is the arithmetic mean
People sometimes use “average” loosely to mean ordinary, typical or middle.
In a mathematics question, identify whether mean, median or another statistic is intended.
Common misconception 4: average two averages directly
This works only when the groups have equal size.
Repair: reconstruct each group’s total before combining.
Common misconception 5: the mean describes everyone
A mean summarises a group. It need not match any individual member.
Repair: distinguish the statistic from the observations it summarises.
A diagnostic ladder for average
- Can the learner explain average as equal redistribution?
- Can the learner find average from total and count?
- Can the learner find total from average and count?
- Can the learner find a missing value from a required average?
- Can the learner predict whether adding a new value raises or lowers the mean?
- Can the learner reconstruct a combined average from groups of different sizes?
- Can the learner distinguish mean from median and mode?
- Can the learner notice when an outlier makes the mean less representative?
- Can the learner compare two data sets with the same mean but different spread?
- Can the learner explain what the mean does and does not tell us?
A five-minute home investigation
Use five small cups and 30 counters.
Distribute the counters unevenly, for example:
2, 4, 5, 8 and 11.
Now move counters between cups until each cup contains the same number.
Each ends with 6.
Then ask:
- What stayed unchanged?
- What changed?
- Why is 6 the mean?
- Could the same mean come from a different original distribution?
The physical redistribution makes the conservation of total visible.
What parents should listen for
- “The average is what each would have if the total were shared equally.”
- “Average 18 across four values means the total must be 72.”
- “The new score is above the old mean, so the mean should rise.”
- “I cannot average these two class averages directly because the classes have different sizes.”
- “The mean is correct, but the outlier makes it less representative of most values.”
How this fits Singapore Primary 6 Mathematics
The updated October 2025 MOE Primary Mathematics syllabus places average of a set of data in Primary 6. The official content states the mean as total value divided by number of data and develops the relationship among average, total value and number of data.
This article stays centred on that mathematical core while extending the reasoning into missing-value, changing-average and combined-group problems that help learners transfer the relationship rather than memorise one direction of the formula.
The deeper lesson: average is conservation plus equality
An average calculation has two hidden ideas.
First, conserve the total.
Second, redistribute it equally.
That is why average can be used forward and backward.
It can summarise data.
It can reconstruct totals.
It can reveal missing values.
It can predict the effect of adding or removing observations.
“Add and divide” is the procedure. Equal redistribution while preserving the total is the mathematics.