Grades 6–8 · Societies
How can who is included change a survey result?
A survey sample is a subset used to learn about a larger group, called the target population. Its result depends on which members are included as well as how many respond. A large response count cannot by itself correct a rule that excludes part of the group of interest.
In the diagramSame sample size, different selectionInvented target: 12 members • 6 yesMorning groupEvening groupYesNoOutlined sample: 5/6 yes • population: 6/12
The invented target population has six morning members and six evening members. Five morning members answer yes, while only one evening member does. Across all twelve, six answer yes, giving a population proportion of one half in this complete fictional record.
A morning-only selection includes six responses and gives five yes, or about 83.33%. An evening-only selection also includes six responses but gives one yes, or about 16.67%. The sample sizes match, yet each rule leaves the other group out. When selection favors members with different answer patterns, it can create selection bias.
The both-groups illustration takes three members from each group and happens to contain three yes answers. That match to the full record is constructed, not guaranteed by group coverage. Coverage and selection both need attention; these fixed records supply no probability model, confidence interval or forecast of real opinions.
About this explanation
Starting knowledge: Read counts and fractions, and distinguish a whole group of interest from a subset.
Sources
- OpenStax, Rice University — 1.2 Data, Sampling, and Variation in Data and Sampling
- OpenStax, Rice University — Chapter 1 Review