Correlation is Not Causation!

We hear it or read it often: “there’s a correlation between (enter two things here)” – poverty and crime, tobacco use and lung cancer, family income and academic success, and (according to a study reported by PLOS Genetics) even between height and IQ (you can read about that study here: The way these things are often communicated, it’s either implied or stated that correlation is causation – poverty causes crime and smoking causes lung cancer, dontcha know! (I promise I’m not from Minnesota and have never been to Minnesota).

Frankly, I’m getting tired of hearing it. Let’s get something straight here: CORRELATION IS NOT CAUSATION! But before I go further, maybe some definitions (from Merriam-Webster) are in order.

CORRELATION: a relation existing between phenomena or things or between mathematical or statistical variables which tend to vary, be associated, or occur together in a way not expected on the basis of chance alone.

CAUSATION: the act or process of causing; the act or agency which produces an effect.

So, basically, a correlation is a relationship (or at least presumed relationship) between two things or among two or more things – crime tends to occur more often in lower-income neighborhoods, cigarette smokers seem to get lung cancer more often than non-smokers. We can just about come up with any two or more things and find a relationship between or among them.

Causation is something that causes something else (or at least something that is likely to cause something else) – exposing your skin to excessive heat causes the skin to burn.

Yet, so many people – and many of them should actually know better – insist on saying, in effect, that correlation is causation. Probably the most common example is the idiocy that poverty causes crime. NO! Poverty doesn’t cause crime. Most poor people (“poor” in America is any income level below a seemingly arbitrary level) aren’t out there committing crimes and there are plenty of rich people who are committing crimes. People making conscious (or not-so-conscious) choices causes crime – and “crime” is nothing more than the act of violating a law, statute or ordinance.

There’s a statement that is often attributed to Mark Twain: “There are lies, damned lies and statistics.” The field of statistics very often deals in correlations and statisticians appear to believe they actually mean something (like the study in PLOS Genetics I linked earlier). Thankfully, I’m not a statistician because I would very likely anger the statistics community with my insistence that correlations are pretty much meaningless. It doesn’t really matter that there’s a supposed relationship between crime and poverty or cigarettes and lung cancer, what matters is what causes what. (For the record, polls are a form of statistics and, I must confess, they’re what turn me off to the whole field of statistics because polls can be manipulated to achieve pretty much any result you want).

So, next time you read a newspaper article or hear a news report about some study saying something causes something else, based on some correlation between the two somethings, don’t believe it! Correlation is not causation.