Change the variable you're asked about, and nothing else
- The idea
- Predict
- See it
- Why
- Check
The idea
“Any two groups that differ in only one thing make a fair test of the question.”
Spotting two groups that differ in just one thing is a good habit. That one thing still has to be the variable in the question.
The short version
The mistake: Any two groups that differ in only one thing make a fair test of the question.
What's actually true: A comparison only tests the variable its groups differ in. That has to be the variable the question names.
- A pair that changes one thing tests that one thing. A and B change only light, so they say nothing about nitrate.
- A pair that changes two things tests neither. You can't tell which change caused the result. The two variables are confounded.
On the exam: Find the variable the question names. Pick groups where it changes and nothing else does.