← AS & A Level Mathematics

AS & A Level Mathematics · Probability & Statistics 1 · Chapter 1

Representation of data

  • Statistics begins with data: collecting, organising, and presenting facts to reveal patterns.
  • Discrete data are displayed using stem-and-leaf diagrams to preserve original values.
  • Continuous data require histograms with frequency density so column areas reflect frequencies accurately.
  • Cumulative frequency graphs allow quick estimation of medians, quartiles, percentiles, and box-and-whisker outlier fences.

Learning objectives

  1. Distinguish between qualitative (categorical) and quantitative data, and between discrete and continuous quantitative variables.
  2. Determine upper and lower class boundaries, class widths, and mid-points for rounded continuous data.
  3. Construct and interpret single and back-to-back stem-and-leaf diagrams with keys, including split-stem formats.

Book exercises: 1A to 1D, and End-of-chapter review 1.

1.1 Types of data

  • Qualitative (categorical) data: non-numerical descriptions or categories (e.g., eye colour, vehicle make, blood type).
  • Quantitative data: numerical measurements or counts. Quantitative data divide into two distinct categories:
    • Discrete quantitative: data that can only take specific, isolated values (usually whole numbers or counts, such as goals scored, shoe sizes 6, 6.5, 7, or coin values 1 cent, 5 cents). Gaps exist between possible values.
    • Continuous quantitative: data that can take any real value within a given range (e.g., height, race time, mass, temperature). Values are limited only by the accuracy of the measuring instrument.
Statistical Data Qualitative Words & categories Quantitative Numerical values Discrete Counted / specific values Continuous Measured / any value
Taxonomy of statistical data: Qualitative vs Quantitative (Discrete and Continuous).

1.2 Stem-and-leaf diagrams

  • Structure: best suited to small or moderate sets of discrete data (typically up to ~50 items).
  • Stem: contains the leading digits (e.g. tens or hundreds), written vertically in ascending order.
  • Leaves: contains the trailing single digit, written horizontally in ascending order away from the stem.
  • Equal spacing: every leaf digit must take up identical horizontal space, creating a natural bar-chart shape.
  • Crucial advantage: no raw data values are lost; every single original observation can be read directly.
  • Mandatory key: a key with units must always be written (e.g., Key: 5 | 3 = 53%).
Stem Leaves (ordered) 5 6 7 9 3 5 8 8 1 1 2 4 7 9 2 7 9 2 7 Key 5 | 3 = 53% Stem: 10s · Leaf: 1s Row length reflects frequency directly because digits are equally spaced in vertical columns.
Anatomy of a stem-and-leaf diagram: stem, leaves, alignment columns, and mandatory key.

Back-to-back stem-and-leaf diagrams

  • Purpose: compare two related distributions directly using a shared central stem.
  • Left-hand leaves: ascend from right to left (increasing outward from the center stem).
  • Right-hand leaves: ascend from left to right (increasing outward from the center stem).
Year 2016 (Rain days) Stem Year 2017 (Rain days) 0 0 1 5 8 9 1 2 2 3 4 6 7 8 8 9 1 0 1 2 3 6 7 1 3 2 0 Key: 5 | 0 | 6 represents 5 days in 2016 and 6 days in 2017
Back-to-back stem-and-leaf diagram: leaves on the left ascend leftwards; leaves on the right ascend rightwards.

📘 Worked Example 1.1 · Constructing a back-to-back stem-and-leaf diagram

The numbers of days on which rain fell in each month of 2016 and 2017 were recorded:

  • 2016: 17, 20, 13, 12, 10, 8, 0, 1, 5, 11, 16, 9.
  • 2017: 9, 13, 11, 8, 6, 3, 1, 2, 2, 4, 8, 7.
  1. Identify the minimum (0) and maximum (20) across both datasets to determine stems: 0, 1, 2.
  2. Group the 2016 values: stem 0 has 0, 1, 5, 8, 9; stem 1 has 10, 11, 12, 13, 16, 17; stem 2 has 20.
  3. Group the 2017 values: stem 0 has 1, 2, 2, 3, 4, 6, 7, 8, 8, 9; stem 1 has 11, 13; stem 2 is empty.
  4. Place 2016 leaves on the left ascending outward from right to left: 9 8 5 1 0.
  5. Place 2017 leaves on the right ascending outward from left to right: 1 2 2 3 4 6 7 8 8 9.
  6. Write a clear dual key: Key: 5 | 0 | 6 represents 5 days in 2016 and 6 days in 2017.
  7. Compare distributions: 2016 had rain on more days (total 122 days) compared to 2017 (total 74 days). Note: this compares frequency of rain days, not the volume of rain.

✏️ Try it · Data types & stem-and-leaf

Question 1: State whether the following variables are qualitative, discrete quantitative, or continuous quantitative: (a) Number of passengers on a bus; (b) Time taken to run 400 metres; (c) Grade awarded in an exam (A*, A, B, ...).

View solution

Step 1: (a) Number of passengers is counted in whole units: Discrete quantitative.

Step 2: (b) Time is measured on an unbroken scale: Continuous quantitative.

Step 3: (c) Letter grades are verbal categories: Qualitative (categorical).

Question 2: In a stem-and-leaf diagram with key 12 | 4 = 12.4 kg, write down the data value represented by stem 13 and leaf 0.

View solution

Step 1: Identify the place value from the key: the stem represents whole kilograms (12) and the leaf represents tenths of a kilogram (0.4 kg).

Step 2: Combine stem 13 and leaf 0: 13.0 kg.

Question 3: The scores of 11 students in a test are: 34, 45, 28, 37, 41, 34, 29, 48, 34, 42, 39. Find the modal score and the median score using a stem-and-leaf representation.

View solution

Step 1: Construct ordered stems: stem 2 has leaves 8, 9; stem 3 has 4, 4, 4, 7, 9; stem 4 has 1, 2, 5, 8.

Step 2: The most frequent leaf on stem 3 is 4 (appears three times). The mode is 34.

Step 3: With n = 11, the median is the 6th item: count 6 leaves from smallest: 28, 29, 34, 34, 34, 37. The median is 37.

Next: Histograms & frequency density →

1.3 Continuous data & histograms

  • Why histograms? Continuous data have no natural gaps. When continuous measurements are rounded, artificial gaps appear in tables (e.g. 10–14, 15–19). Histograms remove these gaps by plotting true continuous boundaries.
  • True class boundaries:
    • If data are recorded to the nearest whole unit, the class 10–14 has true boundaries 9.5 to 14.5. Class width = 14.5 − 9.5 = 5.
    • If data are recorded to 1 decimal place, the class 2.6–2.8 has true boundaries 2.55 to 2.85. Class width = 2.85 − 2.55 = 0.3.
  • The Golden Rule of Histograms:

    Column Area ∝ Class Frequency

  • When column widths are unequal, height cannot represent frequency! We define:

    Frequency Density =

    Class Frequency = Class Width × Frequency Density

Recorded: 146–150 Recorded: 151–155 Recorded: 156–160 145.5 150.5 155.5 160.5 Mid: 148 Mid: 153 Mid: 158
Continuous boundaries remove gaps between rounded classes. Class width = Upper boundary − Lower boundary.
1 2 3 4 5 Frequency density Area = 40 10 kg × 4 = 40 children Area = 60 20 kg × 3 = 60 children 40 50 60 70 Mass m (kg)
Worked Example 1.2: Histogram with unequal intervals. Column 2 has lower height (fd = 3 vs 4) but greater area (60 vs 40).

📘 Worked Example 1.2 · Histograms and Linear Interpolation

The masses, m kg, of 100 children are grouped into two classes: 40 ≤ m < 50 (frequency 40) and 50 ≤ m < 70 (frequency 60). Estimate the number of children with masses between 45 kg and 63 kg.

  1. Calculate class widths: Class 1 width = 50 − 40 = 10 kg; Class 2 width = 70 − 50 = 20 kg.
  2. Calculate frequency densities: fd₁ = 40 / 10 = 4 children/kg; fd₂ = 60 / 20 = 3 children/kg.
  3. Apply the linear interpolation assumption: assume children in each class are spread evenly across the interval.
  4. Find frequency between 45 and 50 kg: sub-width = 50 − 45 = 5 kg ⇒ Count = 5 × 4 = 20 children.
  5. Find frequency between 50 and 63 kg: sub-width = 63 − 50 = 13 kg ⇒ Count = 13 × 3 = 39 children.
  6. Add the two sub-interval estimates: 20 + 39 = 59 children.

🧮 Interactive Frequency Density & Sub-Interval Solver

Enter a class interval, its frequency, and a sub-interval to estimate frequency by linear interpolation.

✏️ Try it · Histograms & frequency density

Question 1: A class of exam marks given as 50–69 has frequency 40. The marks are rounded to the nearest integer. Find the true class boundaries, the class width, and the frequency density.

View solution

Step 1: True boundaries are 49.5 and 69.5.

Step 2: Class width = 69.5 − 49.5 = 20.

Step 3: Frequency density = frequency ÷ width = 40 ÷ 20 = 2.0.

Question 2: In a histogram, a class of width 4 cm is represented by a column of height 8 cm. A neighbouring class has width 6 cm and column height 5 cm. If the first class has frequency 64, find the frequency of the neighbouring class.

View solution

Step 1: Area of first column = 4 × 8 = 32 cm².

Step 2: Since area represents frequency, the scale factor is 64 ÷ 32 = 2 items per cm².

Step 3: Area of neighbouring column = 6 × 5 = 30 cm².

Step 4: Frequency = 30 × 2 = 60 items.

Question 3: Three classes in a histogram have widths in ratio 1 : 2 : 5 and frequencies in ratio 4 : 6 : 5. Find the ratio of their frequency densities in simplest integer form.

View solution

Step 1: Frequency density fd = frequency ÷ width.

Step 2: Compute ratios: 4/1 : 6/2 : 5/5 = 4 : 3 : 1.

Step 3: In simplest integer form: 4 : 3 : 1.

Next: Cumulative frequency graphs →

1.4 Cumulative frequency graphs

  • Cumulative frequency (cf): the running total of frequencies of all values less than a given value.
  • Plotting at upper boundaries:
    • Points are always plotted at upper class boundaries because the total frequency up to the upper boundary is precisely known.
    • Plotting at class mid-points is a common error; mid-value totals are only estimates.
    • Always include the starting point: (L₁, 0) where L₁ is the lower boundary of the very first class.
  • Curve vs Polygon:
    • Cumulative frequency polygon: points connected with straight-line segments. Assumes even spread within classes, matching histogram linear interpolation exactly.
    • Cumulative frequency curve (ogive): smooth S-curve fitted through points. Useful for modelling underlying continuous populations.
20 40 60 80 Cumulative frequency (No. leaves) 0 2 4 6 8 10 12 Length l (cm) Longest 22 (cf = 58) Smooth curve (ogive) Straight-line polygon
Cumulative frequency polygon (dashed) vs smooth curve (solid) for 80 tree leaves (Worked Example 1.3).

🎮 Interactive Histogram & Frequency Density Simulator

Adjust the frequency sliders below. Observe how the column heights (frequency densities) adjust in real time to keep each column's area equal to its frequency.

Live Calculations & Working Steps

    ✏️ Try it · Cumulative frequency graphs

    Question 1: The masses of 80 objects are grouped into classes with upper boundaries 10, 20, 30, 40, 50 grams, and cumulative frequencies 8, 26, 58, 74, 80. State the coordinates of the five points to plot, and the starting point.

    View solution

    Step 1: Starting point is (L₁, 0) = (0, 0) assuming non-negative mass.

    Step 2: Upper boundary coordinates are: (10, 8), (20, 26), (30, 58), (40, 74), (50, 80).

    Question 2: From a cumulative frequency curve representing 200 examination candidates, the 60th percentile score is to be found. At what cumulative frequency should you read across from the vertical axis?

    View solution

    Step 1: 60th percentile corresponds to 60% of total candidates.

    Step 2: cf = 0.60 × 200 = 120.

    Question 3: A cumulative frequency polygon passes through (20, 45) and (30, 85). Estimate the number of items with values between 24 and 28.

    View solution

    Step 1: Gradient of segment = (85 − 45) / (30 − 20) = 40 / 10 = 4 items per unit.

    Step 2: At x = 24, cf = 45 + 4 × 4 = 61.

    Step 3: At x = 28, cf = 45 + 8 × 4 = 77.

    Step 4: Number between 24 and 28 is 77 − 61 = 16 items.

    Next: Quartiles & box plots →

    Quartiles, percentiles & five-number summary

    • Lower quartile (Q₁): the value 25% of the way through ordered data (cf = 0.25n).
    • Median (Q₂): the value 50% of the way through ordered data (cf = 0.50n).
    • Upper quartile (Q₃): the value 75% of the way through ordered data (cf = 0.75n).
    • Interquartile range (IQR): IQR = Q₃ − Q₁, measuring the spread of the middle 50% of observations. Unlike the full range, IQR is resistant to extreme outliers.
    • Five-number summary: Minimum, Q₁, Q₂, Q₃, Maximum.
    cf 75 50 25 Min Q₁ Q₂ (Median) Q₃ Max
    Direct alignment between a cumulative frequency curve and the resulting five-number summary box-and-whisker plot.

    The Cambridge 9709 1.5 × IQR Outlier Rule

    • Outlier definition: an extreme observation that lies unusually far from the main body of data.
    • Mathematical outlier fences:

      Lower fence = Q₁ − 1.5 × IQR

      Upper fence = Q₃ + 1.5 × IQR

    • Drawing outliers on a box plot:
      • Any data point lying below the lower fence or above the upper fence is marked as an individual outlier (using a small cross × or dot).
      • The whisker extends only to the most extreme data value inside the fence (not to the fence itself).
    Lower fence: Q₁ − 1.5×IQR Upper fence: Q₃ + 1.5×IQR Outlier Outlier IQR = Q₃ − Q₁
    Anatomy of a box plot with outlier fences at Q₁ − 1.5 × IQR and Q₃ + 1.5 × IQR.

    Assessing skewness from box plots

    • Symmetrical: Q₂ − Q₁ ≈ Q₃ − Q₂, and whiskers are approximately equal in length.
    • Positively skewed (skewed to the right): Q₃ − Q₂ > Q₂ − Q₁, and the right whisker is noticeably longer than the left whisker. High values are stretched out.
    • Negatively skewed (skewed to the left): Q₂ − Q₁ > Q₃ − Q₂, and the left whisker is noticeably longer than the right whisker. Low values are stretched out.
    Symmetrical Positive Skew Negative Skew
    Comparing distribution shapes: Symmetrical, Positively skewed (right tail), and Negatively skewed (left tail).

    ✏️ Try it · Quartiles & box plots

    Question 1: A dataset has Q₁ = 34, median Q₂ = 42, and Q₃ = 58. Find the interquartile range (IQR) and calculate the lower and upper outlier fences.

    View solution

    Step 1: IQR = Q₃ − Q₁ = 58 − 34 = 24.

    Step 2: 1.5 × IQR = 1.5 × 24 = 36.

    Step 3: Lower fence = Q₁ − 36 = 34 − 36 = −2.

    Step 4: Upper fence = Q₃ + 36 = 58 + 36 = 94.

    Question 2: Would a score of 98 in the dataset above be classified as an outlier? Explain your reasoning.

    View solution

    Step 1: The upper outlier boundary is 94.

    Step 2: Since 98 > 94, the score lies beyond the upper fence.

    Step 3: Therefore, 98 is classified as an outlier.

    Question 3: In the dataset with Q₁ = 34, Q₂ = 42, Q₃ = 58, state with a reason whether the distribution is positively skewed, negatively skewed, or symmetrical.

    View solution

    Step 1: Calculate lower spread: Q₂ − Q₁ = 42 − 34 = 8.

    Step 2: Calculate upper spread: Q₃ − Q₂ = 58 − 42 = 16.

    Step 3: Since Q₃ − Q₂ > Q₂ − Q₁ (16 > 8), the upper half is more spread out. The distribution is positively skewed.

    Next: Comparing representations →

    1.5 Choosing the best representation

    • The choice of representation depends on four factors:
      • Data type: qualitative vs quantitative, discrete vs continuous.
      • Dataset size: small sets preserve raw values (stem-and-leaf); large sets require grouped summaries (histograms, cumulative frequency).
      • Purpose: showing individual values vs showing overall distribution shape vs estimating medians and percentiles.
      • Audience: clarity and honesty; avoiding misleading features such as truncated scales or visual distortions.
    Select Representation Qualitative Bar chart / Pie chart Discrete Quantitative Small: Stem-and-leaf Continuous Quantitative Histogram / Ogive Histogram Shape & density cf Graph Quartiles & %
    Decision flowchart for selecting the most appropriate graphical representation.

    Recognising misleading representations

    • Truncated vertical axis (Displaced zero): starting the vertical axis at a value far above zero exaggerates minor differences (as in Tom's renovation work hours in Exercise 1D).
    • Unequal class widths plotted as raw frequency: when bar widths vary, plotting raw frequency on the vertical axis severely misleads because wider bars gain visual prominence unfairly.
    • Overlapping class intervals: data with overlapping classes cannot be represented in a histogram because items would be double-counted.
    ⚠️ Misleading (Truncated Axis) 49 50 51 52 Exaggerates 49 → 52 as a 400% surge ✅ Honest (Zero Origin) 0 25 50 Shows virtually constant work hours
    Comparison of a truncated vertical axis vs an honest zero-origin scale for identical data.

    ✏️ Try it · Comparing data representations

    Question 1: An advertiser claims that Brand A outsells Brand B by 5 to 1, showing a bar chart where Brand A's bar is 5 times taller than Brand B's. However, the vertical axis begins at 90 units, with Brand A at 100 and Brand B at 92. What are the true sales figures and why is the chart misleading?

    View solution

    Step 1: Read the actual numbers: Brand A sold 100 units, Brand B sold 92 units.

    Step 2: True ratio = 100 : 92 ≈ 1.09 : 1, a modest 8.7% advantage.

    Step 3: The chart is misleading because the truncated vertical axis displaces zero, exaggerating the visual difference by a factor of 5.

    Question 2: Why would a stem-and-leaf diagram be unsuitable for displaying the weights of 5000 airline luggage bags?

    View solution

    Step 1: A stem-and-leaf diagram records every individual observation as a digit.

    Step 2: With 5000 items, the rows of leaves would be hundreds of digits long, creating clutter.

    Step 3: Grouped representations like a histogram or cumulative frequency curve are far more appropriate for large datasets.

    Question 3: State one key statistical feature visible in a box plot that cannot be seen in a histogram.

    View solution

    Step 1: A histogram shows class frequencies and distribution shape.

    Step 2: A box plot displays exact five-number summary positions (Median, Quartiles) and explicitly highlights identified outliers beyond the 1.5 × IQR fences.

    Next: Summary & chapter quiz →

    Chapter 1 Summary

    • Data taxonomy:
      • Qualitative: verbal descriptions/categories (eye colour, vehicle type).
      • Discrete quantitative: counted numbers or discrete steps with gaps (family size, shoe size).
      • Continuous quantitative: measurable along an unbroken scale (mass, time, length).
    • Stem-and-leaf diagrams: preserve every raw value; rows must have equal class widths; leaves must be ordered and equally spaced; key is essential.
    • Histograms:
      • Used for continuous grouped data with no artificial gaps between bars.
      • Column area ∝ Frequency.
      • Frequency density = (Class frequency) ÷ (Class width).
      • When data are rounded to whole units, class a–b has true boundaries (a − 0.5) to (b + 0.5).
    • Cumulative frequency graphs:
      • Points plotted at upper class boundaries with running total frequency.
      • Starts at (L₁, 0) where L₁ is the lower boundary of the lowest class.
      • Used to estimate medians (Q₂), lower quartiles (Q₁), upper quartiles (Q₃), percentiles, and counts above/below given thresholds.
    • Box-and-whisker plots:
      • Five-number summary: Min, Q₁, Q₂, Q₃, Max.
      • Interquartile range (IQR) = Q₃ − Q₁.
      • Outlier boundaries: Lower = Q₁ − 1.5 × IQR, Upper = Q₃ + 1.5 × IQR.
      • Skewness: Symmetrical (Q₂ − Q₁ ≈ Q₃ − Q₂), Positive skew (Q₃ − Q₂ > Q₂ − Q₁), Negative skew (Q₂ − Q₁ > Q₃ − Q₂).

    📄 Printable Chapter Assignment

    Download the complete printable assignment PDF containing every exercise question (1A to 1D), the End-of-chapter review 1, and Check yourself questions:

    📥 Download S1 Chapter 1 Assignment PDF (35 pages)

    Chapter Quiz

    1. Which of the following is an example of continuous quantitative data?
    2. Exam marks are rounded to the nearest integer. What are the true boundaries for the class 60–69?
    3. What is the frequency density of a class with interval 20–50 and frequency 75?
    4. In a histogram, what does the area of each column represent?
    5. A histogram bar has width 4 cm and height 6 cm. If 1 cm² represents 5 items, find the class frequency.
    6. In a stem-and-leaf diagram with key 3 | 5 = 35, why must the leaves in each row be equally spaced?
    7. When constructing a cumulative frequency graph from grouped data, at which points are frequencies plotted?
    8. For 120 data values, at which cumulative frequency is the lower quartile Q₁ located on a curve?
    9. A dataset has Q₁ = 30 and Q₃ = 42. Using the 1.5 × IQR rule, which of these values is an outlier?
    10. A box plot has a much longer whisker on the right than on the left. What does this indicate about the data?
    11. Why cannot data with overlapping class intervals (such as 20–40 and 30–50) be displayed in a histogram?
    12. What is a major advantage of a stem-and-leaf diagram over a histogram?

    Based on Cambridge International AS & A Level Mathematics: Probability & Statistics 1 Coursebook by Dean Chalmers (Cambridge University Press). Check yourself and quiz questions are Guru Jazzy's own.