Sampling
S3 Topic 2, spec point by spec point (Sampling). Each spec point has the method, worked examples written the way the mark scheme wants them, how the marks go and the mistakes that cost the most marks.
Spec2.1Simple random sampling
Methods for collecting data. Simple random sampling, and the use of random numbers for sampling.
The idea
In a simple random sample of size \(n\), every possible sample of size \(n\) has the same chance of being chosen (so every member has an equal chance too).
How to take one with random numbers
- Get a sampling frame (a list of the population) and number every member, using the same number of digits (e.g. 00 to 69 for 70 people).
- Read random numbers of that many digits from a table or calculator.
- Ignore numbers outside the range and repeats.
- Continue until you have \(n\) members.
| Advantages | Disadvantages |
|---|---|
| Free from bias; easy to use; each member has a known, equal chance | Needs a sampling frame; awkward for large populations |
using a random number list
Choose 5 of 70 students (numbered 00 to 69) using the random numbers 23 87 05 23 61 71 40 69 12.
Take 23, ignore 87 (too big), take 05, ignore the repeat 23, take 61, ignore 71, take 40 and 69.
Sample: 23, 05, 61, 40, 69.
- M1 for seven numbers with at least four correct, A1 for all correct with no repeats (WST03 Jan 2021 Q1a).
- For 70 students you need two-digit random numbers (00 to 69); a table of one-digit numbers isn't suitable as it stands: B1 B1 (WST03 Oct 2020 Q4d).
- Reasons for random selection: to reduce bias or favouritism (WST03 Jun 2025 Q1).
Often asked as: short questions: continue a selection from a list, explain a method, or say why a method isn't random.
Numbering 1 to 70 and then using two-digit numbers 00 to 99 without saying what to do with 00 and 71 to 99.
Spec2.2Stratified, systematic and quota sampling
Other methods of sampling: stratified, systematic and quota. When each might be used, with advantages and disadvantages.
The methods
Stratified: split the population into groups (strata), then take a simple random sample from each stratum in proportion to its size.
\[\text{number from a stratum} = \frac{\text{stratum size}}{\text{population size}} \times \text{sample size}\]
Systematic: take every \(k\)-th member of an ordered list, with \(k = \frac{\text{population}}{\text{sample size}}\), starting from a randomly chosen member among the first \(k\).
Quota: an interviewer chooses people to fill a set number (quota) from each group, e.g. 20 men under 30. Selection is not random.
| Method | Advantages | Disadvantages |
|---|---|---|
| Stratified | reflects the population structure; gives results for each group | needs a sampling frame and clear strata |
| Systematic | simple and quick; spreads the sample over the list | needs a list; biased if the list has a pattern every \(k\) |
| Quota | no sampling frame needed; quick and cheap; easy to compare groups | not random, so it can be biased; interviewer chooses who to ask; non-responses aren't recorded |
Which to use: distinct groups that matter → stratified. Long ordered list → systematic. No list available (e.g. stopping people in a street) → quota.
a stratified sample
A company has 6000 full-time and 4000 part-time staff. Describe how to take a stratified sample of 200.
Full-time: \(\frac{6000}{10000} \times 200 = 120\). Part-time: 80.
Number the full-time staff 0001 to 6000 and use random numbers to choose 120 of them (ignoring repeats). Number the part-time staff 0001 to 4000 and choose 80 the same way.
a systematic sample
Take a systematic sample of 10 from 40 records. \(k = 4\). Choose a random start from 1 to 4 (say 3), then take records 3, 7, 11, …, 39.
- M1 for numbering within each stratum (numbering all 10000 together is M0), M1 for using random numbers, A1 for 120 and 80 (WST03A Jan 2026 Q6a).
- Systematic: B1 for every 4th record, dB1 for a random start (using the wrong population size scores B0 B0) (WST03 Jun 2018 Q4b).
- Advantages must suit the method: for stratified, "reflects the population structure" or "results for each group" (WST03 Jan 2024 Q2c).
Often asked as: describe a method in context with the numbers; give an advantage or disadvantage; name the method someone used.
Describing a stratified sample without random selection inside each stratum.
Calling quota sampling random.