Thoughtful Thursday: False cause and effect

I follow Uber Facts on Twitter, mostly because I know that quite a few of my students follow that, too. The facts are sometimes interesting, often a little raunchy, and, I have noticed, make some common mistakes in understanding relationships between cause and effect. Here are two things that are easy to misinterpret.

Correlation vs. causation

Correlation means that two things change at the same time. For example, you might eat more food and see your body weight increase. That’s a positive correlation. You might drive with underinflated tires and see your gas mileage decrease. That’s a negative correlation – as one thing goes up, the other goes down.

What correlation does not mean is that one thing has caused another. Necessarily. Sometimes things change at the same time by coincidence, for example. To really know if it is causation, and not correlation, you need two things to be true. The first is that one needs to precede the other – to happen first. First I eat more food, then I gain weight. The second is that the relationship is found even a carefully controlled environment. This means that you have ruled out competing explanations. If I tested my gas mileage after underinflating the tires, but also did the second test in stop-and-go traffic, I don’t know for sure if it was the inflation, the traffic or some combination of both. So I can’t really say what causes what.

Uberfact example: “Playing a musical instrument has been known to raise a person’s I.Q. by as much as 5 points.” It could be that smarter people are more likely to play musical instruments. Which came first, the music or the smarts?

Lurking variables

Sometimes it looks like one thing causes another, but really there is a third factor that you don’t know about that is the cause of the change. A classic example is ice cream and bicycle theft. You can look and see that both ice cream sales and bicycle thefts rise at the same time. Does this mean that eating ice cream makes you steal bicycles? Unlikely. The lurking variable in this case is hot weather, which encourages more people to eat ice cream and more people to ride bicycles, leaving them in the open and more susceptible to theft.

Uberfact example: “Statistically speaking, women are better drivers than men.” In other words, your gender causes you to be a better or worse drive. This may not be fair. Women do get in fewer accidents, but they also drive fewer miles. The driving less could be the lurking variable to explain the fewer accidents.

Learn more:
Did you know there is a whole scientific journal devoted to traffic research?

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s