Data Science · Beginner level

Beginner level · Module 3: Statistics for Decisions · Lesson 2 of 3

Distributions and Sampling

Understand that a sample gives an estimate of a population and that different samples give different answers.

In this lesson you will

  • Distinguish samples from populations
  • Explain sampling variation and why sample size matters

Key concepts

population and samplerandom samplesampling variationsample size

The lesson

A population is everyone you want to know about — all households in a county. A sample is the households actually surveyed. Statistics from a sample are estimates of the population's true values.

Different random samples give somewhat different estimates; this is sampling variation. Larger samples vary less, but a large sample that is not representative — for example only households near the road — can still be badly wrong.

How the sample was chosen matters more than how big it is. Random selection gives every household a known chance of being included.

Step by step

  1. Simulate drawing many samples from a population data set.
  2. Plot the spread of sample means for sizes 30 and 300.
  3. Draw a biased sample and compare it with the population.
  4. Discuss the survey's sampling method.

Practice activity

Explain, with the simulation results, how reliable estimates from 30 and 300 households would be.

Expected result: A clear explanation showing smaller variation for larger samples and the risk of biased samples.

Check your understanding

  1. What is sampling variation?

    Show answer

    Different samples giving different estimates by chance.

  2. Can a large sample still be wrong?

    Show answer

    Yes, if it is not representative.

Taught live by an instructor — online or in person. Free course lessons are for self-study and do not lead to a certificate — see registration options.