Class 11 Statistics Notes · GSEB

Meaning of Dispersion

Meaning of Dispersion — see why the same average can hide two completely different worlds, from the river-crossing story to same-mean companies. GSEB Class 11 Statistics notes with definitions and the six desirable characteristics.

Last updated: 5 Oct 2026

Notes

What Is Dispersion?

Words marked with a dotted underline carry the question they usually appear in — tap them to reveal it.

“Uncontrolled variation is the enemy of quality.”— Edward Deming

Dispersion — book definition

“A measure which shows how far the observations of the data are scattered from the measure of average is termed as dispersion.”

Spiegel — variation or dispersion

“The degree to which numerical data tend to spread about an average value is called variation or dispersion of the data.”

Central tendency alone is only part of what we need

An average tells you where the data sits. It never tells you how tightlythe observations cluster around that centre. Picture the chapter's two archers: both average 9 shots at 90 — one has a tight cluster at the bullseye, the other is scattered across the whole target. Same average, completely different shooters.

So when two groups are compared on their averages alone, the comparison may lead to erroneous and misleading conclusion.

The River-Crossing Story

The classic illustration of why an average on its own can be dangerous — told in four stages.

Stage 1 — Scenario

A river is reported to be 100 cm deep on average. The average height of the members of your family is 125 cm.

Two averages, both comfortably on the safe side of each other.

1 / 4

Think about it

The same logic runs through everyday numbers: an average delivery time of 30 minutes is meaningless if half your orders arrive in 15 and the rest in 70. Ask for the spread before you trust the average.

Same Average, Different Worlds

Three cloud-kitchen brands in Ahmedabad, each reporting an average annual profit of 30 lakh ₹ over five years. Their mean, median and mode are identical — their five yearly profits are not.

Average only

A founder pitches investors on “30 lakh ₹ average profit” and nothing else. Brand C actually ran from −5 lakh ₹ (a loss year) to 70 lakh ₹ — the average hid both the loss and the windfall.

Average + spread

The same pitch reports 30 lakh ₹ together with the spread (0 / 30 / 75 lakh ₹). Now the investor can see that brand A is a fixed-income machine and brand C is a bet.

The per-capita income trap

India's per-capita income is about ₹1.85 lakh a year — but that single number says nothing about how income is shared between the rich and the poor. Averages of national data hide inequality; only a measure of dispersion (and the distribution around the average) exposes it.

The Four Measures of Dispersion

This chapter studies four measures — each answers “how far?” in its own way, and each has a paired coefficient () for comparing across units.

Range

Highest observation − lowest observation

Quartile Deviation

Half of the inter-quartile range (Q3 − Q1)/2

Mean Deviation

Average distance of observations from the average

Standard Deviation

Root of the average of squared deviations

Range and Quartile Deviation are positional — they are read off the ordered data. Mean Deviation and Standard Deviation are built from the deviations of each observation from the average. The next topic sorts this family into a tree.

What Makes a Measure of Dispersion Good?

Six desirable characteristics of a good measure of dispersion

Two students working independently must get the same idea of what is being measured. A fuzzy definition produces fuzzy comparisons.

A measure nobody can compute by hand is a measure nobody uses — range is loved precisely because subtraction is all it needs.

Every observation should get a say. Range listens to only two of them, which is exactly why standard deviation is preferred.

The measure should survive being averaged, compared and combined across samples — a requirement that rules the quartile deviation out of advanced work.

Draw two equal-size samples from the same population and the measure should not jump around — sampling stability is what makes a statistic trustworthy.

One freak value should not rewrite the whole story. Quartile deviation survives this test; range fails it completely.

Best Measure, and the Constant-Data Rule

Standard Deviation is the best measure

Of the four, Standard Deviation is the best measure of dispersion: it is based on all observations, it is built from deviations taken about the mean, it is rigidly defined, and it is the measure that survives further algebraic and statistical treatment — which is why it heads every advanced topic in this chapter.

Constant data → every measure is 0

Take six identical observations: 20, 20, 20, 20, 20, 20. The mean is 20 and no observation is anywhere away from it — so every measure of dispersion of constant data is 0. A packaging machine that puts exactly 500 g in every pouch has a range of 0, a quartile deviation of 0 and a standard deviation of 0. Zero spread is not a boring result — it is the definition of consistency.

Worked Example Gallery

Meaning of Dispersion — solved practicals

1 problems

Same mean, different spread — book illustration

Full question

Suppose a financial analyst wants to study the performance of three companies A, B and C. He gets the following information of the profits of these three companies for last 5 years :

Year	1	2	3	4	5	Total
Profit of A (lakhs ₹)	30	30	30	30	30	150
Profit of B (lakhs ₹)	15	30	30	30	45	150
Profit of C (lakhs ₹)	-5	30	70	30	25	150

Identical centres — completely different internal variation
Company	Mean	Median	Mode	Spread (max − min)
A	30	30	30	0 lakh ₹
B	30	30	30	30 lakh ₹
C	30	30	30	75 lakh ₹

Spreads: 0, 30 and 75 lakh ₹ — same average, three different worlds.

Mean of A = (30 + 30 + 30 + 30 + 30)/5 = 30. The same computation gives 30 for B and 30 for C.
Order each set: A = 30, 30, 30, 30, 30 · B = 15, 30, 30, 30, 45 · C = −5, 25, 30, 30, 70.
Spread = highest − lowest: A = 30 − 30 = 0 · B = 45 − 15 = 30 · C = 70 − (−5) = 75.

Quick Reference

Idea	Statement
Dispersion	Scatter / spread of the observations around the measure of average
Spiegel	Degree to which numerical data tend to spread about an average value
Four measures	Range · Quartile Deviation · Mean Deviation · Standard Deviation
Best measure	Standard Deviation
Same average ≠ same data	A, B and C all average 30 lakh ₹ — spreads are 0, 30 and 75 lakh ₹
Constant data	For constant data, every measure of dispersion is 0
What it measures	Dispersion measures the internal variation of the observations

Key Takeaways

Dispersion is how far the observations scatter from the measure of average — central tendency is only part of what we need to know.
Two groups with identical averages can be completely different inside; judging on averages alone may lead to erroneous and misleading conclusions. So what: a "30 lakh ₹ average profit" means nothing until the swing around it is stated.
The four measures are Range, Quartile Deviation, Mean Deviation and Standard Deviation — each with a paired coefficient for unit-free comparison.
A good measure is clear, simple, based on all observations, algebraically usable, stable across equal-size samples, and not unduly moved by extreme values.
Standard Deviation is the best measure of dispersion, and every measure of constant data is 0. So what: zero spread is what a perfectly consistent machine, kitchen or study routine looks like.
So what: your own marks tell two stories — the average shows your centre across subjects, the spread shows how wildly your performance varies from subject to subject.