From raw list to frequency table
Raw data is honest but unreadable: thirty numbers in the order they happened tell you almost nothing at a glance. A frequency table fixes that. You list each possible value once, tally how often it occurs, and write the total as its frequency. The table holds exactly the same information as the raw list — just organised.
Example 1The shoe sizes of 20 students are: 36, 37, 37, 38, 36, 39, 38, 37, 36, 38, 40, 37, 38, 39, 37, 36, 38, 37, 39, 38. Construct a frequency table.
| Shoe size | Tally | Frequency |
|---|---|---|
| 36 | |||| | 4 |
| 37 | 6 | |
| 38 | 6 | |
| 39 | ||| | 3 |
| 40 | | | 1 |
Grouping continuous data: class intervals
A frequency table with one row per value only works when there are few different values. Continuous data is the opposite: measure 30 sprint times to 0.01 s and you will probably get 30 different values. The fix is to group the data into class intervals and count how many values land in each interval.
Where does a boundary value go?
A filled circle means the boundary is included; an open circle means it is excluded. The value t = 20 fails t < 20 but satisfies 20 ≤ t, so it belongs to the second interval. No value can ever belong to two intervals.
Class interval builder
Set a lower bound, a class width and a number of classes — then edit the data list and watch each value drop into its correct class. Try putting a value exactly on a boundary, e.g. 20.
Example 2The times, in seconds, taken by 30 students to solve a puzzle are: 34, 18, 42, 27, 51, 33, 22, 45, 38, 29, 16, 40, 31, 24, 47, 35, 20, 43, 28, 55, 37, 26, 49, 32, 19, 41, 30, 23, 52, 36. Construct a grouped frequency table using classes 10 ≤ t < 20, 20 ≤ t < 30, …
| Time t (seconds) | Tally | Frequency |
|---|---|---|
| 10 ≤ t < 20 | ||| | 3 |
| 20 ≤ t < 30 | 8 | |
| 30 ≤ t < 40 | 9 | |
| 40 ≤ t < 50 | 7 | |
| 50 ≤ t < 60 | ||| | 3 |
Two-way tables
A two-way table organises data by two variables at once — one along the rows, one along the columns. Each inner cell is a joint count (“boys who chose cereal”), and the Total row and Total column summarise each variable on its own.
Exam questions usually give you a table with several cells missing, plus a couple of facts in words. The technique never changes: find a row or column with exactly one gap — that gap is forced — fill it, and repeat until the table is complete. Then re-add every row and column.
Example 3120 students were asked what they eat for breakfast. 64 are boys. 50 students in total eat cereal. 21 boys eat toast and 18 boys eat nothing. 13 girls eat nothing. Complete the two-way table.
| Cereal | Toast | Nothing | Total | |
|---|---|---|---|---|
| Boys | 25 | 21 | 18 | 64 |
| Girls | 25 | 18 | 13 | 56 |
| Total | 50 | 39 | 31 | 120 |
Stem-and-leaf diagrams
A stem-and-leaf diagram is a clever halfway house: it groups the data like a frequency table, but keeps every actual value visible. Each number is split into a stem (the first digit or digits) and a leaf (the final digit). The leaves in each row are written in order, and a key tells the reader how to rebuild a value.
Nine quiz scores: 12, 15, 17, 21, 24, 24, 28, 31, 33
| Stem | Leaf |
|---|---|
| 1 | 2 5 7 |
| 2 | 1 4 4 8 |
| 3 | 1 3 |
Key: 2 | 4 means 24. The diagram is really a bar chart lying on its side — the longest row of leaves is the modal group — but unlike a bar chart you can still read every value: the median of these 9 values is the 5th, counting through the leaves in order: 12, 15, 17, 21, 24. The mode is 24 (the repeated leaf).
Back-to-back stem-and-leaf
To compare two data sets, put a second set of leaves on the left of the same stems, read right-to-left. The result is two histograms sharing a spine — ideal for “compare the two classes” questions, which Example 6 tackles in full.
Drawing conclusions — and knowing their limits
Organising data is not the end of the cycle: the syllabus expects you to read, interpret and draw inferences from tables, to compare two data sets, and to recognise when a conclusion is not justified.
Comparing two data sets
Always compare two things, in context:
- an average (usually the median): “Class B’s median of 34 is higher than Class A’s 31, so Class B scored better on average.”
- a measure of spread (usually the range): “Class B’s range of 30 is smaller than Class A’s 33, so Class B’s scores are more consistent.”
A number on its own scores nothing — the comparison word (higher, smaller, more consistent) and the context earn the marks.
When a conclusion is not safe
- Small sample — 10 students cannot represent a whole school.
- Unrepresentative sample — asking only the football team about sport.
- One occasion — one quiz does not prove one class is “better at maths”.
- Grouped data — the raw values are gone, so statements about individuals are guesses.
More worked examples
Example 4Using the completed breakfast table from Example 3, one of the 120 students is picked at random. Find the probability that the student (a) is a girl who eats toast, (b) eats nothing, given that the student is a boy.
Example 5The marks of 15 students in a test are: 23, 35, 41, 28, 36, 44, 31, 27, 39, 42, 25, 38, 38, 30, 47. Draw an ordered stem-and-leaf diagram, and find the mode, the median and the range.
3 | 5 6 1 9 8 8 0
4 | 1 4 2 7
3 | 0 1 5 6 8 8 9
4 | 1 2 4 7
Key: 3 | 5 means 35 marks
Example 6The back-to-back stem-and-leaf diagram shows quiz scores for two classes of 13 students. Compare the two classes, and explain why it would not be safe to conclude that Class B is better at mathematics.
Class A | Class B
8 6 4 1 | 2 | 2 5 7 9
8 6 5 3 1 | 3 | 2 4 5 7
5 2 | 4 | 1 3 4 8
Key: 2 | 3 | 2 means 32 for Class A and 32 for Class B
range(B) = 48 − 18 = 30
Your turn
Practice set
- Write down the four class intervals of width 5 that organise heights from 150 cm up to (but not including) 170 cm, using correct inequality notation for a height h.
- A table uses the classes 0–10, 10–20, 20–30 for the mass m kg of parcels. Explain the error, and rewrite the classes correctly.
- Using the grouped table from Example 2, write down (a) the modal class, (b) the number of students who took less than 30 seconds, (c) the fraction of students with times of 40 seconds or more.
- 90 students chose one activity. 48 are in Year 10; 25 students in total chose swimming; 19 Year 10 students chose swimming; 30 Year 11 students chose football; the rest chose athletics. Complete the two-way table (rows Year 10 / Year 11; columns Swimming / Football / Athletics) and find how many Year 10 students chose athletics, given that 26 Year 10 students chose football.
- Draw an ordered stem-and-leaf diagram for the data 5.2, 6.1, 4.8, 5.9, 6.4, 5.2, 4.5, 6.8, 5.7, 6.1, 5.2 (use stems 4, 5, 6 and a key), and find the mode and median.
- 150 ≤ h < 155, 155 ≤ h < 160, 160 ≤ h < 165, 165 ≤ h < 170 — each lower bound included, each upper bound excluded.
- The classes overlap: a parcel of exactly 10 kg fits in both 0–10 and 10–20. Correct version: 0 ≤ m < 10, 10 ≤ m < 20, 20 ≤ m < 30
- (a) 30 ≤ t < 40 (frequency 9) (b) 3 + 8 = 11 (c) (7 + 3)/30 = 10/30 = 1/3
- Year 10 athletics = 48 − 19 − 26 = 3. Full table: Year 11 swimming = 25 − 19 = 6; Year 11 total = 90 − 48 = 42; Year 11 athletics = 42 − 6 − 30 = 6; athletics total 3 + 6 = 9; check: 25 + 56 + 9 = 90 ✓ (football total 26 + 30 = 56). Answer: 3 students
- Stems 4, 5, 6 with leaves in order: 4 | 5 8; 5 | 2 2 2 7 9; 6 | 1 1 4 8; key 5 | 2 = 5.2. Count: 2 + 5 + 4 = 11 ✓. Mode = 5.2; median = 6th of 11 values = 5.7
Key points to remember
- Tally in groups of five, in one pass through the raw list, and always check that the frequencies total the number of data values.
- Continuous data is grouped with inequality class intervals: 10 ≤ t < 20 — lower bound in, upper bound out, no gaps, no overlaps.
- A boundary value like t = 20 always goes up into the class that starts at it.
- Grouping loses the raw values — you keep the shape of the data but lose the detail.
- In a two-way table, repeatedly fill the row or column with exactly one gap, then re-add everything; “given that” probability questions use the group total, not the grand total, as denominator.
- A stem-and-leaf diagram must be ordered and must have a key; read off mode, median (the (n+1)/2th leaf) and range directly.
- To compare two data sets: one average + one measure of spread, both in context — and remember a small, one-occasion sample cannot support a sweeping conclusion.
Where next?
🎥 Take the interactive module
Representing Data — the full teaching module for this topic, with real past-paper questions and mark schemes.
✏️ Practise exam questions
Statistics — 25 exam-style questions worth 99 marks, with full answers and the mark scheme.
Progress is saved on this device only and appears on the notes home page.