Skip to content
VJ Classes

One place to learn science, commerce, arts, computers and life skills

A survey of a few thousand people is used to say what a whole country thinks. A batting average helps pick a cricket team. A medicine is approved after testing on only some patients. How can a part tell us about the whole? Statistics is the science of learning reliably from data that is incomplete, varying and sometimes misleading.

Why we need statistics

Data are facts and measurements: marks, prices, rainfall, heart rates. A list of a thousand numbers is unreadable, and a single number rarely tells the whole story. Statistics gives us tools to collect data properly, summarise it, and draw careful conclusions while being honest about uncertainty.

The reason we need such tools is variation. No two tube lights burn for exactly the same number of hours, and no two monsoons bring the same rain. Statistics is the science of variation.

You meet it everywhere:

  • Government: the Census counts every person, usually once in ten years, while national sample surveys by the National Statistics Office study employment, spending and health by asking a sample of households.
  • Business: forecasting demand and checking the quality of products.
  • Medicine: comparing a new treatment with an old one.
  • Sport and media: averages, strike rates, opinion polls and exit polls.

Mahalanobis and Statistics Day

Prasanta Chandra Mahalanobis founded the Indian Statistical Institute in Kolkata in 1931 and pioneered large sample surveys in India. India marks his birthday, 29 June, as National Statistics Day.

Two branches and some key words

Descriptive statistics summarises the data you already have, using tables, graphs and numbers such as the average. Inferential statistics uses a sample to draw conclusions about a larger group, with probability measuring how sure we can be.

WordMeaningExample
PopulationThe whole group we care aboutAll 4,000 students of a college
SampleThe part of it actually studied200 students chosen at random
VariableA characteristic that differs between membersMonthly spending
ObservationOne recorded value₹3,200
ParameterA number describing a populationThe true average spending of all 4,000
StatisticA number describing a sampleThe average spending of the 200

The statistical process

Good statistical work follows a path:

  1. Ask a clear question.
  2. Collect data with a sound plan (Lesson 2).
  3. Organise and display it in tables and graphs (Lesson 3).
  4. Summarise and analyse it with averages, spread and probability (Lessons 4 to 7).
  5. Interpret and communicate the result, including its limits (Lesson 8).

Why an average is not enough

Two batters have scored in five innings each:

BatterScoresMean
A40, 42, 38, 40, 40200 ÷ 5 = 40
B0, 10, 100, 5, 85200 ÷ 5 = 40

The means are identical, yet A is steady and B is wildly unpredictable. To describe data fully we need a measure of the centre and a measure of the spread. In Lesson 5 you will find that the standard deviation of A is about 1.3 runs and that of B about 43 runs.

Population or sample?

A college wants to know the average monthly spending of all its 4,000 students. It asks 200 students chosen at random. (a) What is the population? (b) What is the sample? (c) Is the average of the 200 a parameter or a statistic?

Answer: (a) all 4,000 students; (b) the 200 chosen; (c) a statistic, because it describes a sample. It is used to estimate the parameter, the true average of all 4,000.

Key points

  • Statistics is the science of learning from data in the presence of variation.
  • Descriptive statistics summarises data in hand; inferential statistics draws conclusions about a population from a sample.
  • A parameter describes a population; a statistic describes a sample.
  • Two data sets can have the same mean but very different spread, so we need both.
  • The process runs from asking a question, through collecting, organising and analysing, to honest interpretation.