Stratified sampling is all about using a smaller sample to collect data
.. and then using the information to make conclusions about the whole population. It’s usually cheaper and quicker – so, if you want to find out how many students, in a high school, like football – you can just ask a few and decide the outcome.
That’s the theory… but there is a bit of a problem. Suppose you
- just ask girls whether they like football, would that be a fair sample?
- Or, just boys in year 7 – would that be fair?
The whole idea of stratified sampling is that
- the population is split into groups (age, gender etc)
- a random sample is taken from each group
- the random sample is proportional to the size of the group.
Watch the video and try the quick test.
Click here to try the Quick Test stratified sampling
This video is based on an A grade GCSE question:
“Andrew is going to carry out a survey of these students. He uses a sample of 50 students, stratified by year group and gender. Work out the number of Year 13 girls that should be in the sample.”
Another example that appeared as a GCSE question, although could be a real survey was:
“The government wants to survey students, studying science, about their views on becoming teachers. The University of Surrey is chosen as there are 2371 students.
The cumulative percentages of students studying each science subject is:
- The government decides to use a 10% stratified sample. Write down the numbers from each category they should sample. (3 marks)*
- Give one other factor they should take into account when selecting, to ensure an unbiased sample. (1 mark).
- geographical surveys to look at number of animals, types of plants, rocks or soil
- factory products – working out the likely number of defective items
.. and so on.
Please add a comment below if you can think of any others!
View the video on YouTube:
* Calculate 18% of 2371 and so on… then work out 10% of each number.