In Fair Tests you changed one thing at a time and kept the rest the same. Biologists do exactly that, with two more ideas on top: a clear hypothesis to test, and enough repeated trials to trust the result.
Key idea
A hypothesis is a testable explanation that makes a prediction. In a controlled experiment you change one independent variable, measure a dependent variable, hold the controlled variables constant, compare with a control group, and repeat the trials.
- Independent variable: the one factor you change on purpose (the cause you are testing).
- Dependent variable: what you measure (the effect).
- Controlled variables: everything else that could matter, kept the same for every group.
- Control group: the group that does not get the treatment, so you have something to compare with.
Why it works
If two things change at once, you cannot tell which one caused the result. Holding everything else constant leaves only one possible cause. Repeating trials matters because living things vary: one bean plant may grow tall just by chance, but the mean of ten plants is far less likely to be a fluke.
Worked example: Name the variables
A student thinks fertilizer helps duckweed grow. She sets up twelve tanks with 20 duckweed plants each. Three tanks get 0 drops of fertilizer, three get 1 drop, three get 2 and three get 4. All tanks sit under the same lamp at the same temperature, and she counts the plants after two weeks. Name the variables.
- What did she change on purpose? The drops of fertilizer: the independent variable.
- What did she measure? The number of duckweed plants after two weeks: the dependent variable.
- What did she keep the same? The starting number of plants, the light, the temperature and the time: the controlled variables.
- The tanks with 0 drops are the control group, and using three tanks per group repeats each trial.
Answer: Independent: fertilizer drops. Dependent: plant count. Control: the 0-drop tanks.
Worked example: Write a testable hypothesis
Turn the idea "light color matters to plants" into a testable hypothesis.
- A hypothesis names what changes and what you will measure.
- It makes a prediction you could check: "If bean seedlings grow under red light, then they will be taller after three weeks than seedlings under green light, because leaves absorb more red light than green."
- Data could support it or show it is wrong, so it is testable.
Answer: An if-then prediction with a reason, about a quantity you can measure.
Common mistake
Saying one experiment that agrees with a hypothesis "proves" it. Data can support a hypothesis or refute it. Agreement builds confidence, especially when other people repeat the test and get the same result, but a later experiment could still show a problem. One careful result that disagrees counts heavily, because a true hypothesis should never fail a fair test. Check your wording: "the data support the hypothesis," not "the hypothesis is proved."
More plants per group and more repeated trials make a result more reliable. A test with one plant per group tells you very little, however careful it is.