Unit 2: Central tendency, variation & sampling distribution
Business Statistics notes · PTU syllabus (BBA 201-18)
On this page
Unit summary
A single figure often describes a whole data set better than a long list. This unit covers the measures of central tendency — arithmetic, geometric and harmonic means, median and mode with quartiles, deciles and percentiles — and the measures of variation: range, quartile deviation, mean deviation, standard deviation and variance, with the concept of a sampling distribution.
After this unit you can
- Compute the arithmetic, geometric and harmonic means, median and mode for ungrouped and grouped data
- Compute quartiles, deciles and percentiles
- Compute range, quartile deviation, mean deviation, standard deviation, variance and coefficient of variation
- Explain skewness and the relation between mean, median and mode
- Explain the sampling distribution of the mean and the standard error
PTU syllabus topics
- Arithmetic
- geometric and harmonic mean
- mode and median
- quartiles
- deciles and percentiles
- range
- quartile deviation
- mean deviation
- standard deviation and variance
- sampling distribution
Topic 1
Measures of central tendency
An average is a single value that represents the whole distribution. A good average is rigidly defined, easy to understand, based on all observations, not affected by extreme values, and capable of further algebraic treatment.
Mathematical
Arithmetic mean, geometric mean, harmonic mean
Positional
Median, mode, quartiles, deciles, percentiles
Arithmetic mean (AM)
Ungrouped data
x̄ = Σx ÷ n
Discrete or grouped (direct)
x̄ = Σfx ÷ Σf
Short-cut (assumed mean A)
x̄ = A + Σfd ÷ Σf, where d = x − A
Step deviation
x̄ = A + (Σfd′ ÷ Σf) × h, where d′ = (x − A) ÷ h
Weighted mean
x̄w = Σwx ÷ Σw
Combined mean
x̄ = (n₁x̄₁ + n₂x̄₂) ÷ (n₁ + n₂)
- Sum of deviations
- Σ(x − x̄) = 0
- Minimum square
- Σ(x − x̄)² is the least
- Effect of change
- If every value is changed by adding a, multiplying by b — the mean changes the same way
- Uses all values
- Therefore affected by extreme values
- Merits
- Easy, rigid, based on all data
- Demerits
- Affected by outliers; cannot be found for open-ended classes
Example
The marks distribution of Unit 1 has mid-values 5, 15, 25, 35, 45 and frequencies 5, 8, 15, 12, 10. Σfx = 25 + 120 + 375 + 420 + 450 = 1,390 and Σf = 50, so the mean = 1,390 ÷ 50 = 27.8.
Example
Weighted mean: a student scores 80 in Maths (weight 4), 70 in Statistics (weight 3) and 90 in English (weight 2): x̄w = (320 + 210 + 180) ÷ 9 = 78.9.
Geometric and harmonic means
Geometric mean (GM)
GM = (x₁ × x₂ × … × xₙ)^(1/n), or log GM = Σlog x ÷ n
Weighted GM
log GM = Σw log x ÷ Σw
Harmonic mean (HM)
HM = n ÷ Σ(1/x), or Σf ÷ Σ(f/x)
Relation
AM ≥ GM ≥ HM (equal only when all values are equal)
Two values
GM² = AM × HM
Uses of GM and HM
| Best used for | |
|---|---|
| Geometric mean | Average growth rates, index numbers, ratios and percentages — compound interest |
| Harmonic mean | Averaging rates and speeds over equal distances |
| Arithmetic mean | General purpose, additive quantities |
Example
For 2, 4 and 8: AM = 14 ÷ 3 = 4.67, GM = (64)^(1/3) = 4, HM = 3 ÷ (1/2 + 1/4 + 1/8) = 3.43 — confirming AM ≥ GM ≥ HM.
Example
A car goes to a city at 40 km/h and returns at 60 km/h: average speed = HM = 2 × 40 × 60 ÷ (40 + 60) = 48 km/h (not 50).
Example
Growth of sales 10%, 20%, 30%: the average growth factor is the GM of 1.10, 1.20 and 1.30 = (1.716)^(1/3) = 1.197 — about 19.7% per year.
Median
Median is the middle value of the data arranged in order; half the observations lie below and half above.
Ungrouped odd n
Value of the ((n + 1) ÷ 2)th item
Ungrouped even n
Average of the (n ÷ 2)th and (n ÷ 2 + 1)th items
Grouped data
Median = L + [(N ÷ 2 − cf) ÷ f] × h
Symbols
L = lower limit of the median class, cf = cumulative frequency before it, f = its frequency, h = class width
Median class
Class where cumulative frequency first reaches N ÷ 2
Example
Median of the marks data: N ÷ 2 = 25; cumulative frequencies 5, 13, 28 — median class 20–30. Median = 20 + [(25 − 13) ÷ 15] × 10 = 20 + 8 = 28.
Outliers
Not affected by extreme values
Ignores the actual values at the extremes
Graph
Can be found graphically from an ogive
Needs arranging the data
Open classes
Can be found
Not suited to further algebraic treatment
Mode
Mode is the value that occurs most frequently.
Ungrouped
Value with the highest frequency
Grouped
Mode = L + [(f₁ − f₀) ÷ (2f₁ − f₀ − f₂)] × h
Symbols
L = lower limit of the modal class, f₁ = its frequency, f₀ and f₂ = frequencies before and after
Empirical relation
Mode = 3 Median − 2 Mean
Types
Unimodal, bimodal, multimodal
Example
Marks data: modal class 20–30 (f₁ = 15, f₀ = 8, f₂ = 12). Mode = 20 + [(15 − 8) ÷ (30 − 8 − 12)] × 10 = 20 + 7 = 27.
In a symmetric bell-shaped curve all three averages coincide at the centre
The long tail on the right pulls the mean towards the tail, so Mean > Median > Mode
The long tail on the left pulls the mean down, so Mean < Median < Mode
Choosing the average
| Use when | |
|---|---|
| Mean | Data is symmetric and has no outliers |
| Median | Data is skewed, has outliers, or classes are open-ended (incomes) |
| Mode | Most common size or category is needed (shoe size, popular brand) |
| GM | Rates of growth |
| HM | Rates such as speed or price per unit |
Quartiles, deciles and percentiles
Partition values divide an ordered distribution into equal parts.
Quartiles
Q₁, Q₂ (= median), Q₃ divide the data into 4 equal parts
Deciles
D₁ to D₉ divide into 10 parts
Percentiles
P₁ to P₉₉ divide into 100 parts
Formula for grouped data
Q = L + [(k × N ÷ 4 − cf) ÷ f] × h, with k = 1, 2, 3
Deciles
D = L + [(k × N ÷ 10 − cf) ÷ f] × h
Percentiles
P = L + [(k × N ÷ 100 − cf) ÷ f] × h
Example
Marks data: Q₁ position N ÷ 4 = 12.5 lies in 10–20: Q₁ = 10 + [(12.5 − 5) ÷ 8] × 10 = 19.38. Q₃ position 37.5 lies in 30–40: Q₃ = 30 + [(37.5 − 28) ÷ 12] × 10 = 37.92.
Topic 2
Measures of dispersion
Dispersion measures how widely the values are scattered around the average. Two data sets can have the same mean but different spread.
Both curves are centred on the same average; the narrow tall curve has small variation, the wide flat curve large variation
Absolute
Range, quartile deviation, mean deviation, standard deviation
Relative
Coefficient of range, of quartile deviation, of mean deviation, of variation
Range
R = L − S ; coefficient of range = (L − S) ÷ (L + S)
Quartile deviation
QD = (Q₃ − Q₁) ÷ 2 ; coefficient = (Q₃ − Q₁) ÷ (Q₃ + Q₁)
Mean deviation
MD = Σf abs(x − A) ÷ Σf about mean or median ; coefficient = MD ÷ average
Standard deviation
σ = square root of [Σf(x − x̄)² ÷ Σf]
Short-cut SD
σ = square root of [Σfx² ÷ Σf − (Σfx ÷ Σf)²]
Variance
σ²
Coefficient of variation
CV = (σ ÷ x̄) × 100
- Based on all observations
- Uses every value
- Change of origin
- SD does not change if a constant is added or subtracted
- Change of scale
- SD is multiplied by the same constant
- Combined SD
- σ² = [n₁(σ₁² + d₁²) + n₂(σ₂² + d₂²)] ÷ (n₁ + n₂)
- Least
- Mean square deviation is minimum about the mean
- Unit
- Same as the data
Example
Marks data: Σf(x − 27.8)² = 7,608, so σ² = 7,608 ÷ 50 = 152.16 and σ = 12.34. Coefficient of variation = (12.34 ÷ 27.8) × 100 = 44.4%.
Range
Simplest; quick
Depends only on extremes; ignores the middle
Quartile deviation
Not affected by extremes
Ignores the first and last 25%
Mean deviation
Uses all values; easy
Ignores signs — not algebraic
Standard deviation
Best; algebraic treatment; used in sampling
Affected by outliers; harder to compute
Example
Comparing consistency: Team A scores mean 50 with SD 5 (CV = 10%); Team B scores mean 40 with SD 6 (CV = 15%). Team A is more consistent because its coefficient of variation is smaller.
Example
Marks data: MD about the median 28 = Σf abs(x − 28) ÷ Σf = 518 ÷ 50 = 10.36. Range = 50 − 0 = 50. QD = (37.92 − 19.38) ÷ 2 = 9.27.
Meaning
Lack of symmetry
Peakedness of the curve
Measure
Karl Pearson: Sk = (Mean − Mode) ÷ SD, or 3(Mean − Median) ÷ SD
Mesokurtic (normal), leptokurtic (peaked), platykurtic (flat)
Sign
Positive — tail to the right; negative — tail to the left
—
Topic 3
Sampling distribution
A sampling distribution is the probability distribution of a statistic (such as the sample mean) over all possible samples of the same size drawn from a population.
- 1Take a population with mean μ and SD σ
- 2Draw many random samples of size n
- 3Compute the mean x̄ of each sample
- 4List the sample means — they form a distribution
- 5Study its mean, spread and shape
Mean of sample means
E(x̄) = μ
Standard error
SE(x̄) = σ ÷ square root of n
With finite population
SE = (σ ÷ √n) × square root of [(N − n) ÷ (N − 1)]
Standard error of a proportion
SE(p) = square root of [pq ÷ n]
Central limit theorem
For large n (30 or more), x̄ is approximately normal whatever the population shape
Both are centred on μ, but the sampling distribution is narrower — larger samples give a smaller standard error
- Estimation
- Confidence interval x̄ ± Z × SE
- Hypothesis testing
- Compare x̄ with a claimed μ
- Sample size
- SE falls as n rises — quadruple n to halve SE
- Accuracy
- SE measures the reliability of x̄
Example
A population has μ = 60 and σ = 12. For samples of size 36, SE = 12 ÷ 6 = 2, so about 95% of sample means lie between 60 ± 1.96 × 2 = 56.08 and 63.92.
Key terms
- Average
- A single value representing a distribution
- Arithmetic mean
- Sum of values divided by their number
- Weighted mean
- Mean in which values carry different importance
- Geometric mean
- nth root of the product of n values
- Harmonic mean
- Reciprocal of the mean of reciprocals
- Median
- Middle value of ordered data
- Mode
- Most frequent value
- Quartile
- Value dividing the data into four equal parts
- Standard deviation
- Square root of the mean squared deviation from the mean
- Coefficient of variation
- (SD ÷ mean) × 100 for comparing consistency
- Standard error
- Standard deviation of a sampling distribution
Quick revision
- AM ≥ GM ≥ HM; Σ(x − x̄) = 0.
- Median = L + [(N/2 − cf)/f] × h; Mode = L + [(f₁ − f₀)/(2f₁ − f₀ − f₂)] × h; Mode = 3 Median − 2 Mean.
- Positive skew: Mean > Median > Mode; negative skew reverses.
- QD = (Q₃ − Q₁)/2; SD = √[Σf(x − x̄)²/Σf]; CV = SD/mean × 100 — lower CV means more consistent.
- Standard error of the mean = σ/√n; central limit theorem gives a normal sampling distribution.
Important exam questions
Practice questions written to the PTU exam pattern for this unit's syllabus: short answers (Section A style) and long answers (Sections B and C style).
Short-answer questions
- Q1.State four requisites of a good average.
- Q2.Find the mean of 4, 8, 12, 16, 20.
- Q3.Differentiate between AM, GM and HM; state their relation.
- Q4.When is the geometric mean used?
- Q5.Write the formula of the median for grouped data.
- Q6.What is the empirical relation between mean, median and mode?
- Q7.Define a quartile.
- Q8.Define standard deviation and variance.
- Q9.What is the coefficient of variation used for?
- Q10.What is a sampling distribution? What is the standard error?
- Q11.State the central limit theorem.
Long-answer questions
- Q1.Calculate the mean, median and mode for the given frequency distribution.
- Q2.Calculate Q₁, Q₃, D₆ and P₉₀ for grouped data.
- Q3.Calculate the range, quartile deviation, mean deviation and standard deviation of the given data.
- Q4.Compare the variability of two series using the coefficient of variation.
- Q5.Explain skewness and the relationship between mean, median and mode with diagrams.
- Q6.Explain the sampling distribution of the mean and the standard error with an example.
Stuck on this unit?
Message SBS on WhatsApp for help with Business Statistics, or to ask about studying BBA at Synetic.
