What is an experiment?

Photo source: Wikimedia Commons

What is an experiment? Here is my definition:

An experiment is an event that potentially tests a hypothesis

When thinking about experiments, I divide experiments into two types: artificial and natural. An artificial experiment is a man-made procedure, while a natural experiment is a real-world event. Natural experiments are controversial because there will always be an abundance of confounding factors.

For example, the impact of the American Recovery and Reinvestment Act of 2009 is not a clear experiment because there were many other variables affecting the economy at the time, making it difficult to isolate the impact of the stimulus package. That's not to say that all natural experiments are bad, but they certainly require a higher level of analysis to assess the results.

Conversely, with artificial experiments it's much easier to isolate the impact of a treatment. With an artificial experiment you can have two identical groups and implement a different treatment to each group. Since both groups are theoretically identical, any difference in the results can be assumed to be caused by the treatment.

License: CC BY-SA 4.0