A result means little without its control
- The idea
- Predict
- See it
- Why
- Check
The idea
“The treated group changed a lot, so the treatment caused the change.”
A big change in the treated group looks like proof on its own. Only the control shows whether it would have happened anyway.
The short version
The mistake: The treated group changed a lot, so the treatment caused the change.
What's actually true: A treatment's effect is the difference between the treated group and a control that matched it in every way but the treatment.
- A control group gets everything the treatment group gets except the one variable being tested. So it shows what happens without that variable.
- The effect is the difference between the two. In the light, G's 93% against water's 91% is too small a gap to count on.
- In the dark, water gave only 12%. Without that dish, you couldn't tell whether G caused the 88% or the seeds would have sprouted anyway.
On the exam: Find the control first. The effect is the gap between the treatment and its control.