← Back to course

Sampling Methods

You’ll be able to

Population, sample, parameter, statistic

A population is the entire group we want to learn about; a sample is the subset we actually collect data from. A parameter is a fixed number describing the population (like the true proportion of all voters who approve), while a statistic is computed from the sample and estimates it. We use statistics to infer parameters because measuring the whole population is usually impossible.

Four random sampling designs

In a simple random sample (SRS) every group of n individuals is equally likely to be chosen. In a stratified sample we split the population into similar groups (strata) and take an SRS within each — good when strata differ (e.g., sampling each grade level). In a cluster sample we split into groups (clusters) and randomly pick whole clusters to measure everyone in them — convenient when clusters are scattered (e.g., choosing whole classrooms). A systematic sample picks every kth individual from a random start.

Why randomize the selection

Random selection is what lets us generalize from the sample to the population. It avoids the bias a human chooser would introduce and makes the sample representative on average. Without random selection you may have a convenience or voluntary response sample, which cannot be trusted to reflect the population no matter how large it is.

Tip

Distinguish stratified from cluster by a simple test: strata are made homogeneous on purpose and you sample within each one; clusters should each resemble the whole population and you sample some entire clusters. Stratifying by grade, then sampling from every grade, is stratified; picking a few whole homerooms is cluster.

Worked example

A principal wants student opinions and worries that freshmen, sophomores, juniors, and seniors feel differently. She randomly selects 25 students from each grade. Name and justify the sampling method.

  1. 1.Identify the groups: the four grade levels, formed because they are expected to differ in opinion.
  2. 2.Note that a random sample is taken within every one of these groups (25 from each grade).
  3. 3.Sampling within groups that are formed to be internally similar is the definition of stratified sampling.
Answer: This is a stratified random sample, with grade level as the strata. Because opinions may differ by grade, stratifying guarantees every grade is represented and can reduce variability in the estimate.
Checkpoint

A quality inspector selects every 50th item coming off an assembly line, starting from a randomly chosen one of the first 50. This is an example of:

Watch out

The true proportion in the population is a parameter; the proportion you calculate from your sample is a statistic. Keep the labels straight: parameters are fixed and usually unknown, statistics vary from sample to sample and are what you actually compute.

Checkpoint

A news website posts an online poll and reports results from the 8,000 readers who chose to answer. Why might this large sample still be untrustworthy?

Answer the 2 checkpoints as you read.

Sign in to save your progress