How to Design a Controlled Experiment to Test a Hypothesis

Brainstorm all of the variables that may have an effect on the system or process your hypothesis addresses. For example, if you were studying the effects of watering frequency on rate of plant growth, some of the relevant variables would include ambient temperature, type of plant, soil composition, exposure to light and water quality.

Consult others or research similar experiments to see if there are any factors you are missing.

Devise a way to keep each variable constant aside from the variables that your hypothesis addresses. In the plant example, the hypothesis addresses plant height and volume of water per unit time. You could control other variables by growing the plants in a room with a controlled temperature, placing the plants by the same window and using the same type of soil, seeds and water for each plant.

Design a way to vary the independent variable in a meaningful and systematic fashion, such as watering five plants with zero, 25, 50, 75 and 100 grams of water per day, respectively. If you are testing the effects of an unusual addition to a system, such as the effects of a new fertilizer on plants or a new drug on mice, design a control group that does not receive the addition but otherwise undergoes the same experimental design as the other groups.

Design a way to measure the dependent variable that will not affect the experimental results. Decide ahead of time on the frequency, precision and method by which you will measure the results.

Write up your experimental method, including necessary apparatus, setup, reasons for the design and steps in the experiment. Perform the experiment, watching for ways to improve it the next time someone performs it.