Census and sampling

The population (or universe) is the whole group being studied. If we want the average monthly wage of all 600 workers in a factory, the 600 workers are the population.

There are two ways to collect data from a population:

• A census (complete enumeration) covers every element of the population. India's Census, held every ten years, goes from house to house to count every person and record details about them. A census is thorough, but it takes huge cost, time and staff.
• A sample survey studies only a part of the population, chosen carefully. A representative sample (one that reflects the population well) is much smaller, yet gives reasonably accurate results, at lower cost and in less time.

So most surveys are sample surveys. Besides being cheaper and quicker, they allow more detailed questions to be asked, and the smaller team of investigators is easier to train well.

How should a sample be chosen?

• Random sampling: every unit of the population has an equal chance of being chosen. This can be done by the lottery method (names on slips, drawn from a box), by random number tables, or by computer programmes. Because nobody's preferences decide who is picked, the sample is fair.
• Non-random sampling: units are chosen by the investigator's judgement, purpose, convenience or a quota. The units do not all have an equal chance, so the sample may not represent the population well.