Sample design limits inference because a conclusion can only go as far as the sample behind it. The number of readings, where and when they were taken, and how sites were chosen all shape what you can claim.
This is the final skill in fieldwork design and sampling. It uses your question from lesson 1, your sampling choice from lesson 2 and your recording from lesson 3.
How do you judge the limits of a sample?
Ask four questions:
- Size: how many readings or people? Few readings make the pattern less reliable.
- Location: were sites spread across the area, or only in one convenient place?
- Timing: one day, one season, one hour? Conditions change.
- Method: was the choice random, systematic or stratified, or based on convenience?
For each limit, state its effect and suggest an improvement.
Worked example
A class stands at the car park entrance of an invented viewpoint, Bukit Cahaya, on a Saturday morning. They ask 40 visitors how they arrived. 30 say by car.
Step 1, calculate: 30 ÷ 40 = 0.75, so 75% of the sample arrived by car.
Step 2, state what the data shows: “In this sample of 40 visitors, 75% travelled by car.”
Step 3, test a larger claim: “75% of all visitors to the region use cars.” The sample was taken at the car park, so people who walked or came by bus in other ways were less likely to be asked. It covers one morning, so weekday visitors are missing.
Step 4, write the limits with effects: “The survey was at the car park, so car users are probably over-represented. It was one Saturday morning, so the result may not hold on weekdays.”
Step 5, suggest improvements: “Survey at the car park, the bus stop and the footpath, on weekdays and weekends, and ask more visitors.”
The mistake to watch for
A common slip is to claim more than the sample supports.
Mistaken conclusion: “75% of tourists in Malaysia travel by car.”
The sample was 40 people at one viewpoint on one morning, so it cannot describe a whole country.
The correction is to keep the claim to the sample and the place, then say what extra data would be needed to widen it.
Check yourself
1. In a survey of 25 shoppers, 12 said they walk to the shops. What percentage is that?
Show answer
12 ÷ 25 = 0.48, so 48%.
2. A group measures river width at 3 sites on one dry day and concludes that the river is widest at site 3 all year. Give one limit.
Show answer
Only one day was sampled, and river width changes with rainfall and season. Repeating the measurements in different conditions would test the conclusion.
3. A student interviews only people outside a sports shop about leisure habits. Why is it biased?
Show answer
People outside a sports shop are more likely to be interested in sport, so the sample may overstate sporting activity. Sampling in several places and at several times would reduce the bias.
Where this leads next
You have now covered the full enquiry route: question, sample, record, safety and limits. Test it all in the mixed practice set, and use the statistics and distribution explorer to see how an average changes with sample size.
A teacher on online one-to-one Geography tuition can mark your evaluation answers and show how to tie each limit to an improvement.