Science Puzzle
Can You Spot the Bad Experiment?
A student wants to prove her new fertiliser makes plants grow. She puts fertiliser in Pot A, stands it on a sunny windowsill and waters it daily. Pot B gets no fertiliser, sits in a dark cupboard, and is watered once a week.
After a month, Pot A is thriving and Pot B is a wilted stick. She concludes the fertiliser works. What is wrong with her experiment?
The Answer
Her experiment cannot tell her anything about the fertiliser, because she changed three things at once: fertiliser, light and water. Pot A grew, but it may have grown because of the sunlight, or the extra watering, or the fertiliser, or any combination of the three. There is no way to separate them.
A fair test changes exactly one thing, the independent variable, and holds every other condition identical. Both pots should have the same soil, the same amount of sunlight, the same watering schedule on the same days, the same size pot, the same species of plant grown from the same batch of seed. The only difference permitted is the fertiliser. Then, and only then, does a difference in growth point at the fertiliser.
Adding more pots would not save her, because every pot would still have three differences. Repeating a flawed experiment simply gives you a flawed result more precisely. Design first, then repeat.
The principle: Controlled variables. In a fair test, only the independent variable changes, so any observed effect can be attributed to it and nothing else.