Understanding the difference between correlation and causation is essential for anyone who’s interested in becoming science literate. In order to appreciate this, let me present you with a shocking fact: When ice cream sales increase, so do homicide rates (1).
Before we start banning the sale of ice cream to prevent the loss of innocent lives, let’s define both correlation and causation. Correlation is a mutual relationship or connection between two or more things. Causation, on the other hand, is the action of causing something. In short, the difference is that two things can be correlated, but that doesn’t mean that one thing caused the other.
Going back to the ice cream example, it’s true that ice cream sales and murder rates are positively correlated. However, ice cream doesn’t cause murder. It turns out, both ice cream sales and increased homicide rates have a third variable in common: heat. When temperatures rise, ice cream sales increase. Unrelated to the ice cream, summer heat also causes murder rates to increase. *
The main takeaway of this post boils down to one sentence: correlation is not causation. While it might be easy to remember that when someone tells you that ice cream and homicide have been statistically linked, it becomes trickier when the claim is less outrageous. Maybe a friend tells you that they read an article saying people who eat kale live statistically longer. You want to live longer, so you start eating lots of kale. After all, eating kale will give you a few extra years, right? While those kale salads probably aren’t bad for you, you’ve forgotten that correlation isn’t causation. I think it’s significantly more likely that people who eat kale are more concerned with health and wellness, and therefore live a bit longer, than kale alone being responsible for a longer life.
Discard that kale salad and go buy yourself some ice cream, because correlation is not causation. **
* Heat is thought to cause an increase of murder rates for a variety of reasons. One reason is that heat makes people more irritable and quick to anger, and therefore more likely to commit an act of violence (2). If you’re interested in reading more, this article provides a well-cited discussion of the heat-murder phenomenon (3).
** Health disclaimer: Kale is healthier than ice cream. All I’m saying is that eating ice cream won’t get you murdered, and eating kale won’t make you live to one hundred and eighty-two.
- K. Scheidegger, Rebutting the myths about race and the death penalty (2012) (available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2178223).
- C. A. Anderson, Heat and Violence. Curr. Dir. Psychol. Sci. 10, 33–38 (2001).
- W. Moser, Heat and Crime: It’s Not Just You Feeling It. Chic. Mag., (available at http://www.chicagomag.com/Chicago-Magazine/The-312/March-2012/Heat-and-Crime-Its-Not-Just-You-Feeling-It/).