Thursday, December 9, 2010

correlation vs. causation

Correlation does not imply causation. I believe that this statement means that just because events are related does not necessarily mean they are one causes the other. Correlation, by definition, is a statistic representing how closely two variables co-vary. For example, a woman who has not smoked her whole life develops lung cancer. Most people would say that smoking is the only way you can develop lung cancer, but people who do not smoke CAN develop lung cancer due to their genetic makeup. Lung cancer and genetic makeup must be correlated, while smoking throughout one's life will ultimately lead to the development of this disease (a cause and effect relationship).

Another example of correlation:
As rate of ice cream sales increase, the rate of drowning increases.
Ice cream sales increase during the summer months, as well as the popularity of the ocean, pool, or lake. Therefore, more people drown during the summer months, and more people eat ice cream during the summer. These two ideas are correlated, but in no way have a cause-and-effect relationship.
An example of causation:
If I don't study for a test, I will fail it. Not studying for a test will most likely cause me to fail a test.