Home >> Science & Nature >> Science

What Are the Hallmarks of a Good Experiment?

When asked what makes a good experiment, physicist Richard Feynman noted that an experiment need not be shocking, ground-breaking or even interesting. For an experiment to be good, it needs only to be well-controlled, reproducible and honest. No matter the scale, from the atomic collisions in the largest particle accelerator to a student's first science fair project, a good experiment that conforms to these criteria has the potential to impart knowledge and experience in the use of the scientific method.

The Scientific Method
- Experiments conducted at any scale begin with careful, detailed observation.
  
  Performing an experiment is the fourth and final step in the scientific method, which begins with observation. After observation comes the formation of a hypothesis, followed by making predictions. Once a phenomenon is observed, you can form hypotheses to explain it. A hypothesis is essentially an educated guess, given what you know about the world and what you see, that attempts to explain a phenomenon. Hypotheses are used to make predictions about the effects a certain variable will have on the phenomenon, and following up predictions with experiments will either prove or disprove the hypothesis from which they sprang. An experiment that consistently supports a given hypothesis produces a theory, which may then be accepted as a law of nature. If the experiment disproves its respective hypothesis, it's time to return to the observation phase and begin again by forming a new hypothesis.
Control
- A single variable is introduced to the experimental subject or group.
  
  An experiment that lacks control will yield results of questionable value. Only one variable should be tested in a single experiment, in which subjects are placed either in a control group or an experimental group. The control group is left unaltered while the experimental group will be exposed to a single changed factor, called an independent variable. Maintaining control of the factors that influence an experiment's results ensures that no unintentional variables are introduced. The independent variable is then introduced to the experimental group and the outcome is observed, then compared to the control group. If the result is a difference between the two groups, the difference is noted as a dependent variable, or effect.
Reproduction
- If an experiment cannot be reproduced with the same result, its results cannot be trusted. Maintaining control during the experiment helps to ensure that the experiment is reproducible. Successful reproduction of a good experiment will consistently yield the same result, and consistent results support its hypothesis.
Scientific Integrity
- At the conclusion of an experiment, you summarize patterns in your data to determine whether or not your hypothesis is supported. Honesty is vital in maintaining the integrity of the experiment's results. A scientist should be willing to let others test his results by reproducing his methods. If he isn't, he stands to raise questions about the integrity of his work, his findings and his honesty in reporting. Even if the results are sound, a scientist's failure to demonstrate integrity can cause other members of the scientific community to doubt them.

Science