Skip to content

Correlation vs Causation

August 3, 2009

From xkcd:

Correlation does not imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing 'look over there'

In April, I took a micro course called “Scientist as Citizen,” which is basically a workshop for scientists who want to know about interacting with the public, specifically through journalism. It was taught by Cornelia Dean, Science Editor at the New York Times, so she has a lot of experience dealing with scientists and trying to communicate science to the public.

Unfortunately, many scientists are wary of science journalists. It’s tough to explain your research in a way that other people will understand, a process made more difficult because of the different ways that scientists think as compared to the general populace (and that includes science journalists who are by and large NOT scientists themselves). In this class, Ms. Dean brought up several examples of these differences, but there is one that I thought was especially insightful.

A lot of people tend to assume correlation = causation; cum hoc ergo propter hoc. My favorite illustration of this fallacy is in an open letter written to the Kansas City School Board regarding their fight to get Intelligent Design taught in schools

You may be interested to know that global warming, earthquakes, hurricanes, and other natural disasters are a direct effect of the shrinking numbers of Pirates since the 1800s. For your interest, I have included a graph of the approximate number of pirates versus the average global temperature over the last 200 years. As you can see, there is a statistically significant inverse relationship between pirates and global temperature.

Ms. Dean said something that I think bears repeating, “A correlation is not an answer, it’s an opportunity to ask a question.” Scientists generally know this when it applies to their research, but even scientists can be mislead by this fallacy in their day-to-day lives. It seems natural to assume that when two things follow a similar trend, they must be related in some way, especially when you have other reasons to believe it.

When two trends seem correlated, there are 3 possibilities:
1) One thing may cause the other: “Global warming is caused by the decline in pirates.” However, even in this case, it can be tough to determine the direction of causation – how can we be sure that the increasing global temperatures are not to blame for dwindling pirate activity?

2)  Both trends are caused by some third factor yet to be determined: “The use of fossil fuel powered ships causes global warming and makes it more difficult for pirates to catch them in their sail-powered frigates.”

3) The trends are caused by unrelated factors giving the illusion of correlation.

Ultimately, it can often require rigorous, controlled experimentation to accurately determine causal relationships, and that’s one of the places where science really shines.


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

%d bloggers like this: