Correlation Versus Causation
We live in a consumer driven economy and this means enduring endless advertising everywhere we go. It’s actually kind of jarring to think about how every commercial we watch on television or hear on the radio, every billboard we see on a street corner, and every ad we are subjected to on the internet is designed to make us buy something. Many of us are helpless pawns against the sophisticated psychology embedded in the advertising of the 21st century.
In this article we are going to talk about just one trick in the vast universe of tricks that companies use to coerce you into buying their products. Having a solid grasp of the difference between correlation and causation will make you more immune. We will use some examples to help you understand the distinction clearly. Usually, by taking a step back and considering all the moving parts you will be able to quickly recognize when you are mistaking correlation for causation.
Let’s take peanut butter and jelly as a simple example. When you see peanut butter on a sandwich, you might expect jelly. And when you see jelly, you might expect peanut butter. But peanut butter and jelly are actually correlates to one another, not predictive of one another. People eat peanut butter all the time without jelly, and they eat jelly all the time without peanut butter. Peanut butter and jelly do have a very high correlation to each other though. When you see one you often see the other. The true causation occurs at the higher level where the type of food being made is decided upon. If the intention is to make a peanut butter and jelly sandwich then we will be able to predict with almost 100% certainty that peanut butter and jelly will be in the mix.
Causation means that one variable predicts another variable to some degree. Advertisements always want you to believe there is very high predictive value to their products. ‘If I buy x product, y will happen for me.’ In the case of a fancy car maybe the commercial intimates that you will be happy, confident, savvy, and a good lover. But at most these are correlates of owning a fancy car. Some people who own fancy cars are actually quite miserable, uncertain, and terrible lovers. Many others would be happy, confident, savvy, and good lovers even if they didn’t own a fancy car. In the cases where people who own fancy cars exhibit these attributes we can see that they are correlative, and it’s questionable how strong the correlation actually is.
Companies will resort to all forms of trickery to blur the line between correlation and causation. One reason it’s so easy to do is that discerning which variable is actually predictive can be difficult. Some information is always going to slip through the net, however fine the meshing. Let’s take a billboard I saw on Pinterest, which claims that adding a google +1 button to your website increases web traffic by 350%. It’s possible that the single act of adding the button gets these results. However, what is much more likely is that the person or company who has added the +1 button has become highly motivated to generate more traffic. The button will be only one of several other actions, such as writing better content more frequently, creating more connections on a wider variety of social networks, and paying for advertising to get more visibility on the site. The +1 button is just one correlate amongst many. There probably is some causation there but not nearly as much as the ad wants you to believe.
When you start thinking this way more often the shroud will be lifted in many situations in your life and you will be able to see your relationships and yourself more clearly. Humans often want to see causation even when there is none. This is one reason why we are so susceptible to many types of advertisements. Uncertainty creates anxiety, and we all like to feel like we understand the forces operating in our lives.