Generalising from a sample that is too small may lead to conclusions about a larger population that lack credibility. There is, however, no need to sample every element in a population to make credible, reliable conclusions. Recognising the variation between samples and establishing when samples are representative of a population ensures sampling methods are efficient and effective.
Students may have strong beliefs about what makes a fair sample. They often have difficulty understanding the characteristics of an unbiased sample, the naturally occurring variation within samples and the need for samples to be representative.
As a starting point, classroom discussions should focus on possible bias. This might relate to:
- survey questions
- sample size
- sampling methods.
Biased samples favour one way of thinking over another. Activities should challenge students’ notions of the nature of sampling and the purpose of a sample in relation to the questions asked. They should also consider the context of the data in order to appreciate how sampling can influence the claims made.
