- Practice
- About 18 min
- 5 practice questions
Standard deviation of grouped data and the coefficient of variation
Find the standard deviation of discrete and continuous frequency distributions by the mean, assumed mean and step deviation methods, then use the coefficient of variation to compare the consistency of two data sets.
If something looks wrong, please tell us.
By the end of this lesson you can
- Find σ for a frequency distribution using σ = √(Σfd²/N) when the mean is a whole number.
- Use the step deviation method σ = c √[Σfd²/N − (Σfd/N)²] for class intervals.
- Find the coefficient of variation C.V. = (σ/x̄) × 100%.
- Decide which of two data sets is more consistent.
Large data sets are usually given as frequency tables: how many students scored each mark, or how many families fall in each income class. The ideas of lesson 1 carry over, with each value counted as many times as its frequency f. Then we meet a measure that lets us compare spreads fairly, even when the data are in different units.
Grouped data: the formulae
For values (or class mid-points) xᵢ with frequencies fᵢ and N = Σf, every sum of lesson 1 becomes a sum weighted by f. For class intervals, x is the mid-point of each class.
| Method | Deviation d | Standard deviation σ |
|---|---|---|
| Mean method | d = x − x̄ | √(Σfd²/N) |
| Assumed mean method | d = x − A | √[Σfd²/N − (Σfd/N)²] |
| Step deviation method | d = (x − A)/c | c × √[Σfd²/N − (Σfd/N)²] |
Worked example 1: a discrete distribution (made-up data)
Hours of homework per week reported by 20 students: x = 4, 6, 8, 10, 12 with f = 2, 5, 6, 5, 2. Σfx = 8 + 30 + 48 + 50 + 24 = 160, so x̄ = 160/20 = 8, a whole number, so use the mean method. d = x − 8 = −4, −2, 0, 2, 4; d² = 16, 4, 0, 4, 16; fd² = 32, 20, 0, 20, 32, so Σfd² = 104. σ = √(104/20) = √5.2 ≈ 2.28 hours.
Step deviation method for class intervals
- Mid-points Find x, the mid-point of each class.
- Choose A and c A is a middle mid-point; c is the class width.
- Find d d = (x − A)/c gives small whole numbers.
- Columns fd and fd² Add them to get Σfd and Σfd².
- Formula σ = c √[Σfd²/N − (Σfd/N)²].
Worked example 2: class intervals (made-up data)
Marks of 40 students: 0-10: 5, 10-20: 8, 20-30: 15, 30-40: 8, 40-50: 4. Mid-points 5, 15, 25, 35, 45; take A = 25 and c = 10, so d = −2, −1, 0, 1, 2. fd = −10, −8, 0, 8, 8, so Σfd = −2. fd² = 20, 8, 0, 8, 16, so Σfd² = 52. σ = 10 × √[52/40 − (−2/40)²] = 10 × √(1.3 − 0.0025) = 10 × √1.2975 ≈ 10 × 1.139 = 11.39 marks. The mean is A + c × Σfd/N = 25 + 10 × (−0.05) = 24.5.
Coefficient of variation
Standard deviations in different units (kilograms and rupees) cannot be compared, and even in the same units a larger mean tends to bring a larger σ. The coefficient of variation, suggested by Karl Pearson, divides σ by the mean and expresses it as a percentage, giving a relative measure of spread.
Coefficient of variation C.V. = (σ/x̄) × 100%. The data with the smaller C.V. is more consistent (more stable).
Worked example 3: C.V. of the marks
For worked example 2, C.V. = (11.39/24.5) × 100% ≈ 46.49%.
Worked example 4: who is more consistent? (made-up data)
Over a season, batsman X averaged 50 runs with σ = 10, and batsman Y averaged 40 runs with σ = 6. C.V.(X) = 10/50 × 100% = 20%; C.V.(Y) = 6/40 × 100% = 15%. Y has the smaller C.V., so Y is more consistent, even though X scores more on average.
Worked example 5: working backwards
A data set has mean 30 and coefficient of variation 15%. Then σ = C.V. × x̄ ÷ 100 = 15 × 30 ÷ 100 = 4.5, and the variance is 4.5² = 20.25.
Key takeaways: for frequency data, weight each deviation by its frequency and divide by N = Σf; for classes use mid-points, and the step deviation method keeps numbers small. The coefficient of variation (σ/x̄) × 100% compares spreads fairly: smaller C.V. means more consistent.
Check your understanding
-
1. A data set has mean 40 and standard deviation 8. What is its coefficient of variation?
- 5%
- 32%
- 20%
- 8%
Show answer
Answer: C. C.V. = (σ/x̄) × 100% = (8/40) × 100% = 20%.
-
2. Data set A has a coefficient of variation of 12% and data set B has 18%. Which statement is correct?
- A is more consistent than B
- B is more consistent than A
- A has a larger mean than B
- They are equally consistent
Show answer
Answer: A. The data set with the smaller coefficient of variation is more consistent. 12% < 18%, so A is more consistent. The C.V. alone tells us nothing about which mean is larger.
-
3. Find the standard deviation of the distribution x = 1, 2, 3 with frequencies 2, 6, 2.
- 0.4
- 0.63
- 0.2
- 0.8
Show answer
Answer: B. N = 10 and Σfx = 2 + 12 + 6 = 20, so the mean is 2. d = −1, 0, 1, so fd² = 2, 0, 2 and Σfd² = 4. σ = √(4/10) = √0.4 ≈ 0.63. (0.4 is the variance.)
-
4. The coefficient of variation of a data set is 25% and its standard deviation is 5. What is the mean?
- 1.25
- 125
- 4
- 20
Show answer
Answer: D. C.V. = (σ/x̄) × 100, so 25 = (5/x̄) × 100. Then x̄ = 500/25 = 20.
-
5. In the step deviation method, c = 5, Σfd/N = 0 and Σfd²/N = 2.25. What is σ?
- 7.5
- 11.25
- 1.5
- 2.25
Show answer
Answer: A. σ = c × √[Σfd²/N − (Σfd/N)²] = 5 × √(2.25 − 0) = 5 × 1.5 = 7.5.
Sign in to keep track of the lessons you finish.