← AS & A Level Mathematics

AS & A Level Mathematics · Probability & Statistics 1 · Chapter 2

Measures of central tendency

  • Average: a single representative value that summarises a dataset by identifying a central or typical location.
  • Three primary measures: the mode (most frequent value), the mean (arithmetic balance point), and the median (middle value of an ordered dataset).
  • Key A-Level skills: handling raw and grouped continuous data, combining datasets, using coded data transformations, linear interpolation for the median, and evaluating distribution skewness.

🎯 Learning objectives & Book exercises

  • Identify the mode for discrete data and determine when a dataset has one mode, multiple modes, or no mode.
  • Identify the modal class for grouped continuous data with equal or unequal class widths using frequency density.
  • Understand the practical use of the mode in business (e.g. stock sizing, popular demand).
  • Book exercises covered: Exercise 2A and matching review questions.

2.1 The mode and the modal class

  • The mode: the data value that occurs with the greatest frequency.
  • Unimodal vs Multimodal: a dataset with one peak is unimodal; with two peaks of equal highest frequency, it is bimodal; if all values have equal frequency, there is no mode.
  • Qualitative data: the mode is the only measure of central tendency that can be used for non-numerical categorical data (e.g. shoe style, eye color, favorite car).
1. Check DataUngrouped vs Grouped2. Count FrequenciesFind highest count3. Check Class WidthsEqual or Unequal?4. Frequency Densityfd = f / class width5. Modal ClassGreatest column height
Figure 2.1: Flowchart for identifying the mode of discrete data vs the modal class of continuous grouped data.

⭐ Key Point 2.1: The Modal Class

  • In histograms, the modal class is the class with the greatest column height (i.e. greatest frequency density).
  • Frequency density (fd) =

  • Crucial warning: do not simply pick the class with the largest raw frequency if the class widths are unequal!
  • When all class widths are equal, frequency density is directly proportional to frequency, so the modal class has both highest frequency and highest density.
Length x (cm) Frequency density (pencils/cm) 0 10 20 30 40 f = 100 fd = 25 f = 90 fd = 30 f = 80 fd = 40 3.5 7.5 10.5 12.5 Modal Class: 11–12 cm Tallest bar: fd = 40, despite smallest f (80)!
Figure 2.2: Histogram of pencil lengths. The modal class is 11–12 cm (10.5 ≤ x < 12.5 cm) because it has the greatest frequency density (40 pencils/cm).

Worked Example 2.1: Finding the modal class with unequal widths

Find the modal class of the 270 pencil lengths, given to the nearest centimetre in the following table:

Length (x cm)4–78–1011–12
Number of pencils (f)1009080
  1. Step 1: Determine true class boundaries and class widths
    • Because lengths are given to the nearest cm, class 4–7 has boundaries 3.5 ≤ x < 7.5, so width w = 7.5 − 3.5 = 4.
    • Class 8–10 has boundaries 7.5 ≤ x < 10.5, so width w = 10.5 − 7.5 = 3.
    • Class 11–12 has boundaries 10.5 ≤ x < 12.5, so width w = 12.5 − 10.5 = 2.
  2. Step 2: Calculate the frequency density for each class
    • Class 4–7: fd = 100 / 4 = 25 pencils per cm.
    • Class 8–10: fd = 90 / 3 = 30 pencils per cm.
    • Class 11–12: fd = 80 / 2 = 40 pencils per cm.
  3. Step 3: State the class with the highest frequency density
    • The highest frequency density is 40, corresponding to the interval 10.5 ≤ x < 12.5.
    • Answer: The modal class is 11–12 cm (or accurately 10.5 ≤ x < 12.5 cm).
    • Check: Even though 11–12 has the lowest raw count (80 vs 100), it packs 40 pencils per cm, making it the most concentrated interval.

Worked Example 2.2: Equal frequency densities (No modal class)

Two classes of data have interval widths in the ratio 3 : 2. Given that there is no modal class and that the frequency of the first class is 48, find the frequency of the second class.

  1. Step 1: Interpret the condition 'no modal class'
    • 'No modal class' means that all classes have the same frequency density.
    • Therefore, frequencies are in the exact same ratio as class widths: f₁ / w₁ = f₂ / w₂.
  2. Step 2: Set up and solve the ratio equation
    • Let the frequency of the second class be x.
    • Ratio equation: 48 / x = 3 / 2, which gives 3x = 48 × 2 = 96, so x = 32.
  3. Step 3: State and check the answer
    • Answer: The frequency of the second class is 32.
    • Check: Density 1: 48 / 3 = 16. Density 2: 32 / 2 = 16. Densities are equal ✔.

✏️ Try it: Mode & Modal Class

  1. Find the mode of the values: 19, 21, 23, 16, 35, 8, 21, 16, 13, 17, 12, 19, 14, 9.
  2. Continuous classes 0–4, 4–14, 14–20 have frequencies 5, 9, 8 respectively. Find the modal class.
  3. Four classes of continuous data are recorded as 1–7, 8–16, 17–20 and 21–25. Class 1–7 has frequency 84 and there is no modal class. Find the total frequency of the other three classes.
Answers
  • 1. Step 1: Count occurrences of each number: 16 appears twice, 19 appears twice, 21 appears twice, all others once. Step 2: Since three values share the highest frequency, modes are 16, 19 and 21 (multimodal).
  • 2. Step 1: Class widths: 0–4 (w = 4), 4–14 (w = 10), 14–20 (w = 6). Step 2: Frequency densities: 5/4 = 1.25; 9/10 = 0.90; 8/6 = 1.33. Step 3: Highest density is 1.33, so the modal class is 14–20.
  • 3. Step 1: Interval widths with continuous boundaries: 0.5–7.5 (w = 7), 7.5–16.5 (w = 9), 16.5–20.5 (w = 4), 20.5–25.5 (w = 5). Step 2: Density of class 1: 84 / 7 = 12. Step 3: Frequencies of other three classes: 9 × 12 = 108, 4 × 12 = 48, 5 × 12 = 60. Step 4: Total frequency = 108 + 48 + 60 = 216.

Next: The mean & combined datasets →

2.2 The mean: Ungrouped data & combined datasets

  • The mean (arithmetic mean x̄): the sum of all values divided by the total number of values.
  • Raw data formula: x̄ = where Σx is the sum of all x-values and n is the sample size.
  • Ungrouped frequency table formula: x̄ = =
  • Fulcrum property: the mean is the physical balance point of a distribution: Σ(x − x̄) = 0.
1. Choose FormulaRaw: Σx/n · Table: Σfx/Σf2. Calculate TotalsFind Σx or compute fx3. Divide by CountDivide by n or Σf4. Combined Data(n₁x̄₁ + n₂x̄₂) / (n₁+n₂)
Figure 2.3: Strategy for calculating the mean from raw data, frequency tables, and combined datasets.
2 −4 4 −2 Mean x̄ = 6 Fulcrum: Σ(x − x̄) = (−4) + (−2) + (+1) + (+5) = 0 7 +1 11 +5
Figure 2.4: The mean as the fulcrum (centre of gravity). Deviations from the mean balance perfectly to zero.

Combined sets of data

  • When two datasets are combined, their overall mean is not simply the simple average of their means unless the sample sizes are equal.
  • Formula for combining two datasets with sample sizes n₁, n₂ and means x̄₁, x̄₂:

    x̄combined = =

Group 1: 22 Boys Mean = 71% · Total = 1562 + Group 2: 28 Girls Mean = 76% · Total = 2128 Combined Mean = (1562 + 2128) / 50 = 73.8%
Figure 2.5: Weighted combination of two groups. The combined mean is weighted by the group sizes.

Worked Example 2.3: Algebraic equation using the mean

The mean mass of eight bags of cement is 24.8 kg. The masses of seven of these bags are 24.4, 24.6, 24.7, 24.8, 25.0, 25.2 and 25.2 kg. Find the mass of the eighth bag.

  1. Step 1: Express total mass using the formula Σx = n × x̄
    • Total mass of all eight bags: Σx = 8 × 24.8 = 198.4 kg.
  2. Step 2: Sum the known seven bags
    • Sum of 7 bags = 24.4 + 24.6 + 24.7 + 24.8 + 25.0 + 25.2 + 25.2 = 173.9 kg.
  3. Step 3: Subtract to find the unknown mass
    • Eighth bag mass = 198.4 − 173.9 = 24.5 kg.
    • Answer: The mass of the eighth bag is 24.5 kg.
    • Check: (173.9 + 24.5) / 8 = 198.4 / 8 = 24.8 kg ✔.

Worked Example 2.4: Combining two sets of data (Weighted average)

A confectionery shop sells bags of sweets. Group A contains 15 bags with a mean mass of 120 g. Group B contains 25 bags with a mean mass of 140 g. Find the mean mass of all 40 bags combined.

  1. Step 1: Calculate total mass for each group
    • Group A total: Σx_A = 15 × 120 = 1800 g.
    • Group B total: Σx_B = 25 × 140 = 3500 g.
  2. Step 2: Combine totals and total number of items
    • Combined total mass: Σx = 1800 + 3500 = 5300 g.
    • Combined total number of bags: n = 15 + 25 = 40.
  3. Step 3: Calculate the combined mean
    • x̄ = 5300 / 40 = 132.5 g.
    • Answer: The mean mass of all 40 bags is 132.5 g.
    • Check: Notice that 132.5 is closer to 140 than 120 because Group B has more bags (25 > 15). A simple unweighted average (120 + 140) / 2 = 130 would be incorrect.

Worked Example 2.5: Finding an unknown frequency from a combined mean

The mean duration of 18 drama films is 110 minutes, and the mean duration of n action films is 95 minutes. The mean duration of all (18 + n) films together is 101 minutes. Find the value of n.

  1. Step 1: Write the combined mean formula
    • x̄ = (18 × 110 + n × 95) / (18 + n) = 101.
  2. Step 2: Clear fractions and expand
    • 1980 + 95n = 101(18 + n)
    • 1980 + 95n = 1818 + 101n
  3. Step 3: Collect terms and solve for n
    • 1980 − 1818 = 101n − 95n, giving 162 = 6n, so n = 27.
    • Answer: The number of action films is 27.
    • Check: (18 × 110 + 27 × 95) / (18 + 27) = (1980 + 2565) / 45 = 4545 / 45 = 101 minutes ✔.

✏️ Try it: Mean & Combined Data

  1. The mean of 15, 31, 47, 83, 97, 119 and p² is 63. Find the possible values of p.
  2. An examination was taken by 50 students. The 22 boys scored a mean of 71% and the 28 girls scored a mean of 76%. Find the mean score of all 50 students.
  3. A company employs 12 drivers with mean monthly salary $1950. A new driver is employed and the mean salary falls by $8. Find the monthly salary of the new driver.
Answers
  • 1. Step 1: Total sum for 7 numbers = 7 × 63 = 441. Step 2: Sum of the 6 known numbers = 15 + 31 + 47 + 83 + 97 + 119 = 392. Step 3: p² = 441 − 392 = 49. Step 4: p = ±7 (7 or −7).
  • 2. Step 1: Boys total = 22 × 71 = 1562. Step 2: Girls total = 28 × 76 = 2128. Step 3: Combined total = 1562 + 2128 = 3690. Step 4: Combined mean = 3690 / 50 = 73.8%.
  • 3. Step 1: Initial total salary = 12 × 1950 = $23 400. Step 2: New mean = 1950 − 8 = $1942 for 13 drivers. Step 3: New total salary = 13 × 1942 = $25 246. Step 4: New driver salary = 25 246 − 23 400 = $1846.

Next: Grouped data & mid-interval values →

2.2 The mean: Grouped frequency tables & class midpoints

  • Why an estimate? When data is grouped into classes, the exact raw values are lost. We assume every value in an interval is represented by its mid-interval value (midpoint x).
  • Midpoint formula:

    x =

  • Grouped mean formula:

    x̄ = =

  • Watch for boundaries:
    • Continuous intervals with strict inequalities (e.g. 10 ≤ x < 20) have boundaries 10 and 20, midpoint (10 + 20) / 2 = 15.
    • Discrete or rounded intervals (e.g. 20–29 to the nearest integer) have boundaries 19.5 and 29.5, midpoint (19.5 + 29.5) / 2 = 24.5.
Step 1: Check Interval BoundariesIdentify continuous limits (e.g. 20–29 with integers has limits 19.5 to 29.5)Step 2: Calculate Class Midpoint (x)x = (lower boundary + upper boundary) / 2Step 3: Multiply by Frequency (fx)Multiply each class midpoint by its frequency to form the fx columnStep 4: Sum TotalsCalculate total frequency Σf and total product sum ΣfxStep 5: Estimate the Meanx̄ = Σfx / Σf (estimate because individual raw values are unknown)
Figure 2.6: Five-step procedure for estimating the mean from a grouped frequency distribution.
Lower boundary: a Upper boundary: b Unknown raw data values inside class Class Midpoint x = (a + b) / 2 All frequencies in this interval are represented by this single value x
Figure 2.7: The midpoint approximation. The calculation assumes data values average out to the class center.

Worked Example 2.6: Calculating the estimated mean for continuous data

The masses, x grams, of 60 coconuts are recorded in the following frequency table. Calculate an estimate of the mean mass.

Mass (x g)200 ≤ x < 300300 ≤ x < 400400 ≤ x < 500500 ≤ x < 600
Number of coconuts (f)8182410
  1. Step 1: Determine the mid-interval value (x) for each class
    • Class 200 ≤ x < 300: x = (200 + 300) / 2 = 250 g.
    • Class 300 ≤ x < 400: x = (300 + 400) / 2 = 350 g.
    • Class 400 ≤ x < 500: x = (400 + 500) / 2 = 450 g.
    • Class 500 ≤ x < 600: x = (500 + 600) / 2 = 550 g.
  2. Step 2: Multiply each midpoint by its frequency (fx)
    • 8 × 250 = 2000
    • 18 × 350 = 6300
    • 24 × 450 = 10800
    • 10 × 550 = 5500
  3. Step 3: Sum the products and total frequency
    • Σf = 8 + 18 + 24 + 10 = 60
    • Σfx = 2000 + 6300 + 10800 + 5500 = 24600
  4. Step 4: Divide to estimate the mean
    • x̄ = 24600 / 60 = 410 g.
    • Answer: The estimated mean mass is 410 g.
    • Check: 410 g lies comfortably in the largest frequency modal class (400–500 g).

Worked Example 2.7: Unequal intervals with rounded age boundaries

The ages of 40 volunteers are grouped into classes: 20–29 (f = 12), 30–39 (f = 16), 40–59 (f = 8), and 60–79 (f = 4). Calculate an estimate of the mean age.

  1. Step 1: Establish true continuous class boundaries
    • Ages in whole years have boundaries at .5:
    • 20–29 gives [19.5, 29.5), midpoint x = (19.5 + 29.5) / 2 = 24.5
    • 30–39 gives [29.5, 39.5), midpoint x = (29.5 + 39.5) / 2 = 34.5
    • 40–59 gives [39.5, 59.5), midpoint x = (39.5 + 59.5) / 2 = 49.5
    • 60–79 gives [59.5, 79.5), midpoint x = (59.5 + 79.5) / 2 = 69.5
  2. Step 2: Compute fx for each class
    • 12 × 24.5 = 294.0
    • 16 × 34.5 = 552.0
    • 8 × 49.5 = 396.0
    • 4 × 69.5 = 278.0
  3. Step 3: Calculate totals and divide
    • Σf = 12 + 16 + 8 + 4 = 40
    • Σfx = 294 + 552 + 396 + 278 = 1520
    • x̄ = 1520 / 40 = 38.0 years.
    • Answer: The estimated mean age is 38.0 years.

✏️ Try it: Grouped Data Means

  1. Classes 0 ≤ x < 2 (f = 8), 2 ≤ x < 4 (f = 9), 4 ≤ x < 8 (f = 11), 8 ≤ x < 14 (f = 2). Calculate an estimate of the mean.
  2. For the grouped data: 13 ≤ y < 16 (f = 7), 16 ≤ y < 21 (f = 17), 21 ≤ y < 28 (f = 29), 28 ≤ y < 33 (f = 16), 33 ≤ y < 36 (f = 11). Estimate the mean of y.
  3. Explain why the true mean of a grouped dataset might differ from the estimated mean calculated using class midpoints.
Answers
  • 1. Step 1: Midpoints: 1, 3, 6, 11. Step 2: fx values: 8 × 1 = 8, 9 × 3 = 27, 11 × 6 = 66, 2 × 11 = 22. Step 3: Σf = 30, Σfx = 123. Step 4: x̄ = 123 / 30 = 4.1.
  • 2. Step 1: Midpoints: 14.5, 18.5, 24.5, 30.5, 34.5. Step 2: Σf = 7 + 17 + 29 + 16 + 11 = 80. Step 3: Σfy = 7(14.5) + 17(18.5) + 29(24.5) + 16(30.5) + 11(34.5) = 101.5 + 314.5 + 710.5 + 488 + 379.5 = 1994. Step 4: ȳ = 1994 / 80 = 24.925 (or 24.9 to 3 s.f.).
  • 3. Step 1: The calculation assumes all data values in each class are equal to the mid-interval value. Step 2: In reality, values may be clustered towards one side of the interval (e.g. skewed within the class), causing the true mean to be higher or lower than the midpoint estimate.

Next: Coded data & linear transformations →

2.2 The mean: Coded data & linear transformations

  • Why code data? Coding transforms large, awkward numbers into small, friendly numbers (or gives them a mean of zero), simplifying manual calculation and algebraic manipulation.
  • Translation (Addition/Subtraction coding):
    • If y = x − b, then every data value is shifted by −b.
    • The mean shifts by the exact same amount: ȳ = x̄ − b or x̄ = ȳ + b.
    • Sum of coded values: Σ(x − b) = Σx − nb.
  • General Linear Coding (Scaling and Shift):
    • If y = (x − b) / a or y = ax − b, both addition and multiplication affect the mean:
    • ȳ = ax̄ − b or x̄ = (ȳ + b) / a.
    • Sum expansion: Σ(ax − b) = aΣx − nb.
Step 1: Choose or Identify CodingTranslation: y = x − b · Linear: y = (x − b) / aStep 2: Work with Coded DataCalculate coded mean: ȳ = Σy / n or use given coded sum Σ(x − b)Step 3: Decode to Original MeanFor translation: x̄ = ȳ + b · For linear: x̄ = aȳ + bStep 4: Algebraic ExpansionsRemember: Σ(x − b) = Σx − nb and Σ(ax − b) = aΣx − nb
Figure 2.8: The coding and decoding pipeline for linear transformations.
x 101 103 104 109 113 Mean x̄ = 106 Shift by −100 y = x − 100 1 3 4 9 13 Coded Mean ȳ = 6 ⇒ x̄ = 6 + 100 = 106
Figure 2.9: Subtracting 100 translates every value and their mean by −100. Adding 100 back decodes the mean.

Worked Example 2.8: Combining coded datasets

Exactly 5 years ago, the sum of the ages of 10 boys was 127.0 years, so Σ(b − 5) = 127.0. In exactly 5 years' time, the sum of the ages of 15 girls will be 351.0 years, so Σ(g + 5) = 351.0. Find the current mean age of:

  • (a) the 10 boys
  • (b) the 15 girls
  • (c) the 10 boys and 15 girls combined.
  1. Step 1: Expand or decode the boys' total
    • Method A: Σ(b − 5) = Σb − 10(5) = 127.0, so Σb = 127 + 50 = 177.
    • Current mean age of boys: b̄ = 177 / 10 = 17.7 years.
    • Method B (Mean shift): b̄ − 5 = 127.0 / 10 = 12.7, so b̄ = 12.7 + 5 = 17.7 years.
    • Answer (a): The 10 boys have a mean age of 17.7 years.
  2. Step 2: Expand or decode the girls' total
    • Σ(g + 5) = Σg + 15(5) = 351.0, so Σg = 351 − 75 = 276.
    • Current mean age of girls: ḡ = 276 / 15 = 18.4 years.
    • Answer (b): The 15 girls have a mean age of 18.4 years.
  3. Step 3: Combine the current totals for boys and girls
    • Total current sum of all 25 people: Σb + Σg = 177 + 276 = 453.
    • Combined mean: x̄ = 453 / (10 + 15) = 453 / 25 = 18.12 years.
    • Answer (c): The combined mean age is 18.12 years.
    • Check: (10 × 17.7 + 15 × 18.4) / 25 = (177 + 276) / 25 = 18.12 years ✔.

Worked Example 2.9: Grouped coded data

Forty values of x are coded in the table below where y = x − 3. Calculate an estimate of the mean value of x.

Coded class (x − 3)0–1818–2424–32
Frequency (f)91318
  1. Step 1: Calculate midpoints of the coded classes
    • 0–18 gives y₁ = (0 + 18) / 2 = 9
    • 18–24 gives y₂ = (18 + 24) / 2 = 21
    • 24–32 gives y₃ = (24 + 32) / 2 = 28
  2. Step 2: Calculate the coded mean ȳ
    • Σfy = 9(9) + 13(21) + 18(28) = 81 + 273 + 504 = 858
    • ȳ = 858 / 40 = 21.45
  3. Step 3: Decode to find x̄
    • Since y = x − 3, we have x̄ = ȳ + 3 = 21.45 + 3 = 24.45.
    • Answer: The estimated mean of x is 24.45.
    • Check: Decoding individual midpoints first (12, 24, 31): Σfx = 9(12) + 13(24) + 18(31) = 108 + 312 + 558 = 978. 978 / 40 = 24.45 ✔.

Worked Example 2.10: Solving a linear coded equation

For 20 values of x summarised by Σ(2x − 3) = 104, find x̄.

  1. Step 1: Method A — Using the coded mean directly
    • Let y = 2x − 3. The mean of y is ȳ = 104 / 20 = 5.2.
    • Since ȳ = 2x̄ − 3, solve for x̄:
    • 2x̄ = ȳ + 3 = 5.2 + 3 = 8.2, so x̄ = 4.1.
  2. Step 2: Method B — Expanding the sigma notation
    • Σ(2x − 3) = 2Σx − 20(3) = 104
    • 2Σx − 60 = 104, so 2Σx = 164, which gives Σx = 82
    • x̄ = 82 / 20 = 4.1.
    • Answer: The mean of x is 4.1.

🧮 Interactive Data Coding Transformer (y = (x − b) / a)

Enter a comma-separated list of raw data values x, a shift constant b, and a scale divisor a. Observe how the mean transforms linearly.

  1. Enter parameters above and click Transform Data.

✏️ Try it: Coded Data

  1. Twenty-five values of z are such that Σ(z − 7) = 275. Find z̄.
  2. For 15 values of x, Σ(3x − 2) = 528. Find x̄.
  3. Data x has x̄ = 45. If the data is coded as y = (x − 15) / 6, find ȳ.
Answers
  • 1. Step 1: Σ(z − 7) = Σz − 25(7) = 275. Step 2: Σz = 275 + 175 = 450. Step 3: z̄ = 450 / 25 = 18. (Or: z̄ = (275/25) + 7 = 11 + 7 = 18).
  • 2. Step 1: ȳ = 528 / 15 = 35.2. Step 2: 3x̄ − 2 = 35.2 gives 3x̄ = 37.2. Step 3: x̄ = 37.2 / 3 = 12.4.
  • 3. Step 1: ȳ = (x̄ − 15) / 6. Step 2: ȳ = (45 − 15) / 6 = 30 / 6 = 5.

Next: The median & linear interpolation →

2.3 The median: Raw data & linear interpolation

  • The median: the central value of an ordered dataset, splitting the data into a lower 50% and upper 50%.
  • Positions:
    • Discrete / Raw data: Median position is . If n is odd, it is an exact middle data value; if n is even, it is the mean of the two middle values.
    • Continuous grouped data: On a cumulative frequency distribution, the median position is taken at: . Using n / 2 ensures exact symmetry whether counting from the bottom up or from the top down.
Step 1: Locate Median PositionContinuous data: use rank n / 2 on cumulative frequencyStep 2: Identify Median ClassFind interval where cumulative frequency reaches or passes n / 2Step 3: Set Up Linear Interpolation(m − L) / (U − L) = (n/2 − F) / f where L = lower boundary, F = prev cf, f = class freqStep 4: Solve for Medianm = L + ((n/2 − F) / f) × class width
Figure 2.10: Procedure for estimating the median of continuous grouped data using linear interpolation.

⭐ Linear Interpolation Formula

When the median falls in the class [L, U) with frequency f, class width w = U − L, and cumulative frequency prior to this class F:

=  ⇒  Median = L + () × w

Variable x Cumulative freq (cf) f = class frequency w = class width (U − L) n/2 F F + f L Median m U
Figure 2.11: Similar triangles in linear interpolation: (m − L) / w = (n/2 − F) / f.

Worked Example 2.11: Discrete data from a stem-and-leaf diagram

The numbers of customers served during 15 hours on Saturday are: 22, 23, 24, 26, 35, 35, 36, 38, 39, 39, 40, 41, 43, 47, 49. Find the median number of customers.

  1. Step 1: Determine the median position for discrete ordered data
    • Sample size n = 15 (odd).
    • Median position = (n + 1) / 2 = (15 + 1) / 2 = 8th value.
  2. Step 2: Locate the 8th value in the ordered list
    • Values 1 to 4: 22, 23, 24, 26
    • Values 5 to 7: 35, 35, 36
    • 8th value is 38.
  3. Step 3: State the median
    • Answer: The median number of customers is 38.
    • Check: Exactly 7 values lie below 38 and 7 values lie above 38 ✔.

Worked Example 2.12: Linear interpolation on grouped continuous data

The masses of 148 objects are recorded in the table below. Calculate an estimate of the median mass using linear interpolation.

Mass (m kg)0 ≤ m < 0.20.2 ≤ m < 0.30.3 ≤ m < 0.50.5 ≤ m < 0.70.7 ≤ m < 0.8
Frequency (f)161292244
Cumulative frequency (cf)1628120144148
  1. Step 1: Find the target cumulative frequency rank
    • Total n = 148. Median rank = n / 2 = 148 / 2 = 74.
  2. Step 2: Identify the median class
    • Class 0.2 ≤ m < 0.3 ends at cf = 28.
    • Class 0.3 ≤ m < 0.5 ends at cf = 120.
    • Since 28 < 74 ≤ 120, the median lies in the class 0.3 ≤ m < 0.5.
  3. Step 3: Extract parameters for interpolation
    • Lower class boundary L = 0.3
    • Class width w = 0.5 − 0.3 = 0.2
    • Class frequency f = 92
    • Cumulative frequency before this class F = 28
    • Values needed inside class = 74 − 28 = 46
  4. Step 4: Substitute into the interpolation formula
    • Median = L + ((n/2 − F) / f) × w = 0.3 + (46 / 92) × 0.2
    • Notice 46 / 92 = 0.5 exactly.
    • Median = 0.3 + 0.5 × 0.2 = 0.3 + 0.10 = 0.40 kg.
    • Answer: The estimated median mass is 0.40 kg.
    • Check: 0.40 is exactly halfway between 0.3 and 0.5, which matches rank 74 being exactly halfway between 28 and 120 ✔.

✏️ Try it: Median & Linear Interpolation

  1. For the discrete frequency table of t: 7 (f = 4), 8 (f = 7), 9 (f = 9), 10 (f = 14), 11 (f = 16), 12 (f = 41), 13 (f = 9). Find the median.
  2. Continuous data is grouped into classes 10–20 (f = 8), 20–30 (f = 14), 30–50 (f = 18). Calculate an estimate of the median by linear interpolation.
  3. A cumulative frequency polygon has coordinates (20, 24) and (30, 62) for total n = 100. Find the median by linear interpolation.
Answers
  • 1. Step 1: Σf = 4 + 7 + 9 + 14 + 16 + 41 + 9 = 100. Step 2: Rank = (100 + 1) / 2 = 50.5th value. Step 3: Cumulative frequencies: 4, 11, 20, 34, 50, 91, 100. Step 4: 50th value is 11, 51st value is 12, so median is (11 + 12) / 2 = 11.5.
  • 2. Step 1: n = 8 + 14 + 18 = 40. Target = 40 / 2 = 20. Step 2: cf values: 8, 22, 40. Median is in class 20–30. Step 3: L = 20, w = 10, F = 8, f = 14. Step 4: Median = 20 + ((20 − 8) / 14) × 10 = 20 + (12 / 14) × 10 = 20 + 8.571 = 28.6 (to 3 s.f.).
  • 3. Step 1: Target rank = 100 / 2 = 50. Step 2: In class 20–30, L = 20, w = 10, F = 24, f = 62 − 24 = 38. Step 3: Median = 20 + ((50 − 24) / 38) × 10 = 20 + (26 / 38) × 10 = 20 + 6.842 = 26.8.

Next: Comparing averages & skewness →

2.3 Comparing averages & skewness

  • Skewness: describes the asymmetry of a distribution.
  • Positively skewed (skewed to the right):
    • Long tail stretches towards the larger values (to the right).
    • The mean is dragged towards the long tail: Mode < Median < Mean.
    • Examples: household incomes, house prices, number of goals in football matches, marathon running times.
  • Negatively skewed (skewed to the left):
    • Long tail stretches towards the smaller values (to the left).
    • The mean is dragged towards the low tail: Mean < Median < Mode.
    • Examples: ages of retirement home residents, marks on an easy exam.
  • Symmetrical (no skew):
    • Bell-shaped or uniform: Mode = Median = Mean.
    • Examples: adult heights, weights of packaged manufactured goods.
SymmetricMode = Median = MeanPositive SkewLong right tail: Mode < Median < MeanNegative SkewLong left tail: Mean < Median < ModeChoose AverageMode: popular/stock · Median: skewed · Mean: math calc
Figure 2.12: The relative ordering of Mode, Median, and Mean across symmetrical and skewed distributions.
Mo Med x̄ Positive Skew Mode < Median < Mean Symmetrical Mo = Med = x̄ All three coincide x̄ Med Mo Negative Skew Mean < Median < Mode
Figure 2.13: The Skewness Trio. Notice how extreme tail values pull the mean away from the peak mode.

When to use each measure: Advantages & Disadvantages

MeasureAdvantagesDisadvantagesBest Used When...
Mode
  • Unaffected by extreme outliers.
  • Easy to identify.
  • Works for qualitative data.
  • Ignores most data values.
  • May not exist, or multiple modes may exist.
  • Unsuitable for further algebra.
Finding most popular shoe size, clothing stock, or categorical data.
Median
  • Robust against extreme outliers.
  • Can be found even when open-ended classes exist (e.g. >100).
  • Easy to interpret (50% above, 50% below).
  • Only uses positional rank order.
  • Does not use the numerical value of most data points.
  • Awkward to combine for two datasets.
Skewed distributions (incomes, house prices, wait times).
Mean
  • Uses the exact numerical value of every data point.
  • Easily manipulated algebraically.
  • Readily combined across groups.
  • Heavily distorted by extreme outliers.
  • Cannot be calculated if boundaries are open-ended.
  • May not correspond to an actual possible score.
Symmetrical data without outliers, or when further calculations (e.g. variance, standard deviation) are required.

🕹️ Interactive Simulation: Skewness & Averages Explorer

Move the skewness slider below to see how the curve shifts between negative skew, symmetry, and positive skew, and watch the relative movement of the Mode, Median, and Mean.

🧮 Grouped Data Mean & Median Interpolation Calculator

Select a preset or edit interval limits and frequencies below. The tool calculates midpoints, products, modal class, estimated mean, and linear interpolation for the median with explicit step-by-step working.

ClassLower Boundary (L)Upper Boundary (U)Frequency (f)
  1. Click Calculate Averages to see complete step-by-step calculations.

✏️ Try it: Comparing Averages

  1. Homes in a neighborhood sold for $220 000, $242 000, $236 000 and $3 500 000. Which average should a buyer consider?
  2. A teacher recorded quiz marks of eight students as: 11, 13, 15, 15, 17, 18, 19, 20. If mark 11 was a typo and changed to 1, what effect does this have on mode, mean, and median?
  3. In an examination, Chemistry marks have mode > median > mean and Physics marks have mean > median > mode. Describe the skewness of both exams.
Answers
  • 1. Step 1: $3 500 000 is an extreme outlier (mansion) compared to the other three around $230 000. Step 2: The mean = (220 + 242 + 236 + 3500) / 4 = $1 049 500, which is completely unrepresentative of typical homes. Step 3: The median ($239 000) is far more helpful because it resists extreme outliers.
  • 2. Step 1: Mode: 15 remains the most frequent value (unchanged). Step 2: Median: previously between 15 and 17 = 16; now the 8 values in order are 1, 13, 15, 15, 17, 18, 19, 20; middle two values are still 15 and 17, so median remains 16 (unchanged). Step 3: Mean: decreases by (11 − 1) / 8 = 10 / 8 = 1.25.
  • 3. Step 1: Chemistry: mean < median < mode, so Chemistry marks are negatively skewed (long tail of low marks). Step 2: Physics: mean > median > mode, so Physics marks are positively skewed (long tail of high marks).

Next: Summary & mini-quiz →

Chapter summary

  • Measures of central tendency: the mode, the mean, and the median represent the center of a distribution.
  • Mode: the most frequently occurring value. For grouped data, the modal class is the class with the highest frequency density (fd = f / w).
  • Mean: x̄ = Σx / n for raw data, and x̄ = Σfx / Σf for frequency tables. For grouped continuous data, mid-interval values x = (L + U) / 2 are used to estimate the mean.
  • Combined datasets: x̄ = (n₁x̄₁ + n₂x̄₂) / (n₁ + n₂).
  • Coded data:
    • Translation: x̄ = mean(x − b) + b = (Σ(x − b) / n) + b. Also Σ(x − b) = Σx − nb.
    • Linear coding: if y = (x − b) / a, then x̄ = aȳ + b. Also Σ(ax − b) = aΣx − nb.
  • Median:
    • Discrete data: position (n + 1) / 2.
    • Continuous grouped data: rank n / 2 on cumulative frequency, evaluated using linear interpolation:

      Median = L + () × w

  • Skewness relationships:
    • Symmetrical: Mode = Median = Mean.
    • Positively skewed: Mode < Median < Mean (tail pulled right).
    • Negatively skewed: Mean < Median < Mode (tail pulled left).

📄 Practice questions

  • All exercises from the coursebook (Exercises 2A, 2B, 2C, 2D, 2E and End-of-chapter review 2) are available in the printable chapter assignment.
  • It includes original "Check yourself" exam-style questions and the printed Chapter Quiz.

Mini-quiz

1. What is the mode of the dataset: 3, 7, 8, 5, 12, 7, 9, 3, 7, 4?
2. For grouped continuous data with unequal class widths, how is the modal class identified?
3. A set of 12 numbers has mean 15, and a set of 8 numbers has mean 25. What is the combined mean?
4. Why is the mean calculated from a grouped frequency table called an estimate?
5. For 20 values of x, Σ(x − 15) = 140. What is the mean x̄?
6. Data x is coded using y = (x − 50) / 4. If ȳ = 6.5, what is the mean of x?
7. For 10 values, Σ(2x − 3) = 64. What is the value of Σx?
8. In an ordered set of 25 discrete items, at which position is the median located?
9. When estimating the median of n values from a cumulative frequency curve, which rank is used?
10. A distribution has a long tail to the right (positive skew). How are the averages usually ordered?
11. Which measure of central tendency is least affected by an extreme outlier in a dataset?
12. Dataset A has 10 values with Σ(x − 5) = 30. Dataset B has 15 values with Σ(y − 5) = 45. Combined mean?

Source: Cambridge International AS & A Level Mathematics: Probability & Statistics 1 Coursebook (Cambridge University Press, Dean Chalmers), Chapter 2 (Measures of central tendency). Explanations, extra examples, diagrams, the interactive simulation, the tools, the Try-it answers and the quiz are Guru Jazzy's own. This page is not made or endorsed by Cambridge University Press or Cambridge Assessment.