Sampling means collecting data from some places or people instead of everywhere. Random sampling gives every location an equal chance, systematic sampling uses a fixed interval, and stratified sampling divides the area into groups and samples each group in proportion to its size.
You choose the method to suit the scenario. This lesson follows stating an investigable question and comes before designing the recording sheet.
How do the three methods work?
- Random: number all possible sites and pick some using random numbers. This removes personal choice but may leave gaps.
- Systematic: choose a start point, then sample at equal intervals. It covers the area evenly, but can miss a pattern that repeats at the same interval.
- Stratified: split the area into zones that differ, then sample each zone in proportion to its size. It is fair when the zones are very different.
How do you decide which to use?
- Look at the area: is it one even area or made of different zones?
- Look at your question: do you need even coverage along a line, or fair coverage of different zones?
- Look at the time: how many readings can you take?
- Choose the method that matches, then say why in one sentence.
Worked example
Scenario A: A class has 10 sites to sample along an invented 500 m beach, Pantai Biru, to measure pebble length.
Step 1, choose: a straight line suits systematic sampling.
Step 2, interval: 500 ÷ 10 = 50 m.
Step 3, random start: pick a start within the first 50 m, say 20 m.
Step 4, sites: 20, 70, 120, 170, 220, 270, 320, 370, 420, 470 m. The last is 20 + 9 × 50 = 470 m, which is inside the beach.
Scenario B: A 1,000 m² park has three zones: a path edge (600 m²), a lawn (300 m²) and a shaded corner (100 m²). The class has 20 quadrats and wants to measure trampling.
Step 5, choose: the zones differ, so stratified sampling suits.
Step 6, proportions: path edge 600 ÷ 1,000 × 20 = 12, lawn 300 ÷ 1,000 × 20 = 6, shaded corner 100 ÷ 1,000 × 20 = 2. Check: 12 + 6 + 2 = 20.
The mistake to watch for
A common slip is to give every zone an equal number of samples when the zones are very different in size.
Mistaken plan: 7 quadrats in each of the three zones.
The shaded corner is only 10% of the park but would provide a third of the data, which would distort any overall result.
The correction is to take samples in proportion to the area (12, 6 and 2), or to explain clearly why equal numbers are wanted, for example when comparing zones with each other.
Check yourself
1. A footpath is 800 m long and you want 8 sites using systematic sampling with a random start of 35 m. List the sites.
Show answer
Interval = 800 ÷ 8 = 100 m. Sites: 35, 135, 235, 335, 435, 535, 635, 735 m. The last is 35 + 7 × 100 = 735 m, within the 800 m path.
2. A field has zones of 50%, 30% and 20% of the area. How many of 10 quadrats go in each zone in a proportional stratified sample?
Show answer
50% × 10 = 5, 30% × 10 = 3, 20% × 10 = 2. Total = 10.
3. Why might random sampling be a weak choice for a very narrow 100 m stream with 5 sites?
Show answer
Random numbers may cluster, for example all within the first 30 m, leaving most of the stream unsampled. Systematic sampling would spread the sites evenly.
Where this leads next
Once the sites are chosen you need a sheet to record data. Move on to designing an observation record, and use the fieldwork sample and graph planner to see how sites are laid out. Try the statistics and distribution explorer to see how sample size affects an average.
A teacher on online one-to-one Geography tuition can set new scenarios and ask you to justify each choice aloud.