Lesson 1 of 7 · 12 min

Populations, samples and random samples

Tell a population from a sample and a parameter from a statistic, and explain what makes a sample random and why randomness matters.

What you will be able to do

  • Identify the population, the sample, the parameter and the statistic in a context.
  • Explain what a random sample (and a simple random sample) is.
  • Explain why a random sample gives estimates you can trust, and a non-random one may not.

What proportion of Queensland Year 12 students have a part-time job? There are more than 50 000 of them. You can't ask them all, but you could ask 500. The question is whether the answer from 500 students tells you anything about all of them.

It can — if the 500 are chosen well. This topic is about what "chosen well" means, and about how much an answer from a sample can be expected to wobble.

You observe the sample and calculate a statistic such as p^\hat{p}. You use it to estimate a parameter of the population, such as pp.

Definition

The vocabulary of sampling

  • The population is the whole group you want to know about.
  • A sample is the part of the population you actually observe.
  • A parameter is a number describing the population, such as the population proportion pp. It is fixed, but usually unknown.
  • A statistic is a number calculated from the sample, such as the sample proportion p^\hat{p} (say "p hat"). It changes from sample to sample.

In a random sample of 500 Queensland Year 12 students, 0.46 have a part-time job. Which statement is correct?

Choose one answer

A sample is only useful if it is representative: if it looks like the population in the ways that matter. You can't check that directly (you don't know the population — that's why you're sampling), so instead you make sure the method of choosing the sample is fair. The fairest method is to let chance choose.

Key idea

Random samples

A sample is random when it is chosen by a chance process that gives every member of the population the same chance of being included.

In a simple random sample of size nn, every possible group of nn members is equally likely to be the one chosen. Drawing names from a well-mixed hat, or using a random number generator on a numbered list, gives a simple random sample.

Randomness does two jobs. It removes the chooser's preferences, conscious or not, so no part of the population is systematically favoured. And it makes the sample's results behave like a random variable whose behaviour you can predict with probability — which is exactly what the rest of this topic does.

A random sample

The school office numbers all 1200 students and uses a random number generator to choose 80 of them.

Every student has the same chance, 801200\frac{80}{1200}, of being chosen. The sample's mix of year levels, interests and backgrounds will be close to the school's, give or take chance variation.

Not a random sample

A student asks the first 80 people who walk into the library at lunch.

Students who are never in the library have no chance of being chosen. People who study at lunch may differ from everyone else, so the sample may systematically misrepresent the school.

Worked example

Population, sample, parameter and statistic

  • Foundation
  • Technology free

Queensland Health wants to know the proportion of Queensland adults who are fully vaccinated against influenza this year. It phones a random sample of 2000 Queensland adults, and 840 of them say they are vaccinated.

(a) Identify the population and the sample.

(b) Calculate the sample proportion and say whether it is a parameter or a statistic.

(c) Explain why another random sample of 2000 adults would probably give a different answer.

5 steps

Try it yourself first, then check the working one step at a time.

Method

  1. Population: everyone the question is about. Sample: those actually observed.

  2. A number calculated from the sample is a statistic; the matching population value is a parameter.

Watch out for

  • Calling the 2000 adults the population: they are only the part that was observed.

  • Calling 0.42 "the proportion of Queensland adults who are vaccinated". It is an estimate of that proportion.

A gym with 2400 members wants a random sample of 100 members. Which method gives a random sample?

Choose one answer

Common mistake

"Random" doesn't mean "haphazard"

Picking "whoever I happen to see" feels random but isn't: you are more likely to see some people than others. A random sample needs a genuine chance process, such as random numbers, applied to a list of the whole population.

Note

Why not ask everyone?

Surveying the whole population is a census. The ABS runs one every five years, at enormous cost. Between censuses, samples are cheaper, faster and — if they are random and reasonably large — accurate enough. Sometimes a census is impossible anyway: you can't crash-test every car or taste every jar of Vegemite.

Summary

  • The population is the whole group of interest; the sample is the part you observe.

  • A parameter (such as pp) describes the population and is fixed; a statistic (such as p^\hat{p}) comes from a sample and varies.

  • A random sample is chosen by a chance process that gives every member the same chance of selection; in a simple random sample every group of size nn is equally likely.

  • Randomness protects against the chooser's preferences and lets probability describe how the sample results vary.

Finished the lesson?

Try the 2 quick checks above and the lesson completes itself, or mark it complete now.

Lesson 1 of 70% read

Save your progress with a free account