Two things that change together are correlated. That tells you a pattern exists, but not that one thing causes the other. Exam questions often ask for this limit directly, sometimes in the form “suggest why the data does not prove…”.
This lesson follows explaining how a substance alters signalling and keeps the same careful reading habits from interpreting stimulant evidence, inside drugs and biological effects.
What does correlation show, and what does it not show?
A correlation shows that two variables move together. It does not show three other things:
- that the first variable causes the change in the second
- that the direction is what you assumed
- that the pattern would still appear with a larger sample
A controlled experiment changes one factor and keeps the others the same. That design gives much stronger evidence of cause than a table of things people already do.
A reliable way to write the limit
- Describe the pattern with numbers: “as X rises, Y falls”.
- Say the pattern is an association, not proof of cause.
- Give one reason: a third factor, the reverse direction or a small sample.
- Say what would help: a controlled trial or more data.
Worked example
Invented data, for teaching only. Six students recorded how many caffeinated drinks they had each day and how many hours they slept.
| Drinks per day | 0 | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|---|
| Sleep (hours) | 8.5 | 8.0 | 7.5 | 7.0 | 6.5 | 6.0 |
Step 1, pattern. As drinks rise from 0 to 5, sleep falls from 8.5 to 6.0 hours. Each extra drink goes with 0.5 hours less sleep. Mean drinks = 15 ÷ 6 = 2.5. Mean sleep = 43.5 ÷ 6 = 7.25 hours.
Step 2, association. The data show a negative correlation: more drinks, less sleep.
Step 3, limit. Three explanations fit the same table.
- Caffeine may reduce sleep.
- Students who sleep less may drink more to stay awake, so the direction is reversed.
- A third factor, such as exam week, may cause both.
Step 4, what would help. A larger sample, and a controlled trial in which volunteers are given caffeine or a plain drink, would give stronger evidence.
Conclusion: the data is consistent with caffeine shortening sleep, but it does not prove it.
The mistake to watch for
Mistaken answer: “The table proves that caffeine makes students sleep less.”
“Proves” is too strong. The answer also ignores the reverse direction and any third factor, and there are only 6 students.
The correction is to write “is associated with” or “is consistent with”, then name one specific alternative explanation. Naming a real alternative scores better than writing “there could be other factors” with no detail.
Check yourself
1. Use the worked table. By how many hours does sleep fall between 1 drink and 4 drinks per day?
Show answer
Sleep falls from 8.0 to 6.5 hours, so it falls by 8.0 − 6.5 = 1.5 hours. That matches 3 extra drinks × 0.5 hours.
2. In an invented survey, towns with more ice cream sales also had more sunburn cases. Give one reason why ice cream does not cause sunburn.
Show answer
A third factor, such as hot, sunny weather, can increase both ice cream sales and time spent in the sun. The two variables are correlated but not causally linked.
3. Rewrite this claim more carefully: “Energy drinks cause poor marks.”
Show answer
“Energy drink use is associated with lower marks in this data, but the data does not show cause, because students with less study time may drink more, or a third factor may affect both.”
Where this leads next
The next lesson, avoid turning a biology answer into personal medical guidance, applies the same care to how an answer is worded. To see how small samples can differ from a model, try the inheritance model board and the probability tree and counting board. Then test all five skills with the practice set.
If “suggest one limit” questions often feel vague, our teachers can help you build specific limits in online one-to-one Biology tuition.