Thursday, March 6, 2008

2. Correlation does not imply causation.

2. Correlation does not imply causation. If two variables move together, it may not be the case that one variable caused the other variable to change. This is a common mistake in all areas of research, not just economics.

A Latin phrase expresses a similar concept. Post hoc, ergo procter hoc translates as "it happened after, so it was caused by".

An extreme example to illustrate this point involves the sunrise. If a person wakes up before sunrise every day for many years, it does not imply that if the person oversleeps, then the sun will not come up. The person arising does not cause the sun to rise.

Similarly, just because a person is a country leader when the economy does especially well or poorly, it does not imply the leader is responsible for the economic performance. The leader indeed may influence the economy. The correlation does not imply it, however. One needs to look closely at the actions of leaders to determine how much credit and blame they deserve.

Politicians, such as the President of the United States, probably receive more credit and blame than is justified for economic events that occur during their years of service.

Political publicists, sometimes called spin doctors, use favorable correlations to trumpet the great value of their candidates while ignoring correlations that put their candidates in an unfavorable light.

No comments:

Post a Comment