Brown: Correlation is not causation

Phil Brown

Some things always seem to happen together. When the weather gets colder, everyone starts wearing long pants and jackets. When students study for tests, they get better scores. When ice cream sales go up, more people drown.

Now for those first two, one could make a fairly strong argument that one causes the other. People are going to wear clothes that keep them warmer because it is getting colder outside. Students are going to know more relevant information if they study, and that will generally cause them to get a higher test score.

For the third example, however, it probably is not the case that ice cream sales cause people to drown, nor vice versa. Even though these two events occur together, one does not cause the other. In this case, the correlation can probably be chalked up to the warm weather of summer, which sees both an increase in ice cream demand, and an increase in swimmers.

To be sure, that is quite an obvious example, but many people can be fooled into believing that one event is caused by another just because they occur together. Examples abound, but some of the more dubious correlations are presidents with the economy and race with intelligence.

As far as presidents and the economy go, there is unquestionably some form of relationship. The president, while not having total power over our government, does hold quite a lot of political clout. The policies of the nation, in turn, do affect the country’s economy. As any economist will tell you, however, there are more than a few other factors.

George W. Bush was famously linked with spiking gas prices, and some of his administration’s policies may have had a hand in such a change, but there were certainly outside forces that caused prices to rise. Some of the factors that jump out are instability in oil-producing regions and increased global demand, both of which can hardly be attributed solely to Bush. I would bet my bottom gallon that Al Gore and John Kerry would have also had presidencies that saw rising gas prices.

Race and intelligence are personal qualities that also appear to be linked. For example, members of certain not-so-upstanding organizations have long claimed that the historically higher average intelligence of white Americans relative to black Americans shows that one race is scientifically smarter.

That is not science, it is a convenient explanation.

Another explanation is that one group has been historically privileged to have better education and better economic opportunities, and has been taught by society that they have more intellectual potential. It is also worthy to note that for simplicity, intelligence is quantified by IQ tests, which only measure a narrow range of social and cognitive aptitudes. They are by no means a full and complete test of “intelligence.”

The logical fallacy of assuming that correlation implies causation can create some serious problems, but it is still extremely prevalent. One reason for this may be that people intentionally misuse information to suit their needs. It would certainly behoove the political adversaries of Bush to tie him to unpopular gas prices, just as it makes sense for political adversaries of President Barack Obama to falsely blame the economic downturn of late completely on the president’s administration.

Another, more subtle reason could simply be a lack of thought. It can be very easy to look at two events that always seem to occur together and take for granted the idea that one causes the other. It may take too much time to think about other possibilities, or they may just elude detection.

Such was the case with the correlation between “good” (HDL or high-density lipoprotein) cholesterol and lower risk for heart disease. Many medical professionals thought that having good cholesterol would lower a patient’s chance of heart disease. Dr. Peter Toth outlines this hypothesis well in “The Good Cholesterol,” found in the American Heart Association’s Circulation journal.

There has since been pressure to find drugs that would increase HDL cholesterol levels. It has been proposed, however, that both are actually caused largely by genetic factors, so artificially raising HDL cholesterol alone will not substantially aid heart health. Such was the decision of a National Health Institute clinical trial that ended HDL-inducing treatments 18 months early.

Correlation and causation are two inextricably linked concepts, but they are not the same. Only by separating the two can individuals get a more accurate picture of the world, and all of the complex relationships in it.