Confusing correlation with causation is a common mistake for people to make. Often, this mistake is the result of sloppy thinking, such as the case with this article here.

Today, President Bush gave an eight-minute speech on the economy, which was meant to reassure the markets. “This is an anxious time,” said Bush. “But the American people can be confident in our economic future.” However, ABC News notes that during those eight minutes, the Dow Jones dropped another 107 points. The Dow was down 78.70 at the beginning of Bush’s address, but at 185.66 by the end.

Correlation occurs between two variables when there is a strong relationship between A and B, such that when A changes, B also changes. The tendency for most people is to assume that since there is a correlation between A and B, A must cause B. However, as anybody with high school statistics could tell you, it is a mistake to assume a causal relationship because of a correlation. It might be the case that A causes B, but it could also be the case that B causes A, or even the third possibility that there is some variable, C, that is causing the change in both A and B. Because of the number of possibilities, we cannot draw a conclusion about probable causation with the observation of a single correlative study.

In this case, we have good reason to assume that the causal relationship is incorrect. By the time Bush had made his address, the stock markets had already been losing value at a steady rate for weeks. Furthermore, he made his address specifically to address the financial crisis. In this case, it would seem that the most reasonable assumption to make would be that there is a third variable (or group of variables) that caused both the stock market to fall, and Bush’s speech.

This post’s author may or may not have been serious about the argument, but I have seen similar arguments made regularly enough to know that many people do not understand the difference between correlation and causation. I hope that the next time you hear somebody make an argument like this, you will recognize that it is incorrect, and demand more proof before accepting a causal relationship.