A population is the whole group we want to study, such as all Grade 12 students in Ethiopia. Studying everyone (a census) is often too costly or slow, so we study a sample, a smaller part, and use it to draw conclusions about the population. A good sample is representative: it reflects the population's characteristics. A numerical summary of the population is a parameter; the same summary for a sample is a statistic.
Key terms
- Population
- the whole group of interest
- Sample
- a subset studied
- Census
- collecting data from every member
- Simple random sampling
- equal chance for every member
- Stratified sampling
- random samples from each subgroup
- Cluster sampling
- randomly choose whole groups
- Sampling bias
- systematic over- or under-representation
Key formulas
Summary
- A sample should represent the population.
- Random methods: simple, systematic, stratified, cluster.
- Convenience samples are often biased.
- Bigger samples reduce sampling error but not bias.
In everyday life: The Ethiopian Statistics Service uses stratified and cluster sampling in household surveys so results represent both urban and rural regions.
Check yourself
- Distinguish a parameter from a statistic.
- Give one example of sampling bias in a phone survey.
- When is cluster sampling better than simple random sampling?
Practice questions
1. Dividing a population into regions and then taking a random sample from each region in proportion is called...
- A. cluster sampling
- B. convenience sampling
- C. systematic sampling
- D. stratified sampling
Show answer
Answer: stratified sampling
Stratified sampling splits the population into strata (here, regions) and samples randomly from every stratum, so each group is represented.
On WeLearn
The full lesson: reading notes, worked examples, class activities and the Unit 3 quiz
Study Sampling Techniques on WeLearnTeach Sampling Techniques with WeLearn
We will bring a demo to your school: a lesson on your interactive TV, a virtual lab and a classroom game, with your own teachers.
