Logical Fallacy #1 – Correlation vs. Causation

This is a discussion that bears repeating, because the fallacy shows up in arguments I have seen on many and varied political discussions.

To state the premise simply:

Correlation does not necessarily mean causation.

Say it with me:

Correlation does not necessarily mean causation.

As a general example, let’s say as A increases, B increases. That does not necessarily mean that an increase in A causes an increase in B. There are many other possible explanations.

First one, which seems silly on the surface: it could be true that an increase in B causes an increase in A. Some studies have shown that an increase in violent video games correlates positively with an increase in real-life violence. Some people have concluded that violent video games cause an increase in real-life violence. How about the opposite – increases in real-life violence cause an increase in video game violence?

I’m not being a smart-ass when I say this. Video game makers obviously want to sell a product their key demographic will buy. Adolescent and teenage boys tend to like gun play, action movies, and rough sports. So, video game makers could be just filling a need. If adolescent and teenage boys were more preoccupied with sex than with violence, we would have more Leisure Suit Larry games (instead of the mere 8 that are already out there).

Another possibility is that an increase in C leads to an increase in both A and B. Think about this: An increase in ice cream sales correlates positively with an increase in drowning deaths. Therefore, ice cream sales causes drowning. Right?

Of course not. What is actually happening is that ice cream sales tend to rise in summer months. In those summer months, people tend to spend a lot more time outside, so drowning deaths will increase as more people play in the water. So, ice cream sales don’t cause drowning deaths; warmer weather correlates with an increase in ice cream sales and an increase in drowning deaths. C causes A and B.

Finally, perhaps A has nothing at all to do with B, and the fact that both increase is serendipity.

It’s important to look at conclusions being drawn from data, and see if there is in fact causation or merely correlation. This discussion was prompted by a Facebook post where the presenter implies (or at least I inferred, but I strongly think the implication was there) that increased handgun restrictions cause increased murder rates. I think I did a convincing job in Facebook comments showing that the one does not necessarily cause the other; in fact, New York has stiff gun control laws, and murders there have gone down.

Repeat after me:

Correlation does not necessarily mean causation.

One more example: Many proponents of the death penalty argue that it creates a general deterrence; people will be less likely to commit a capital offense if they know they might be put to death. Data suggests that this is not true: the murder rate is higher in states with the death penalty than without. This does not suggest that the death penalty leads to higher murder rates, but it does call into question the presumption that the death penalty causes lower murder rates.

One more time:

Correlation does not necessarily mean causation.

As always, comments are welcome.

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