What's a Spurious Correlation?
Every sociology major learns about the concept of spurious correlation, but they don’t always fully understand it. This concept matters because when it occurs, two things look like they cause each other, but in reality they don’t.
Here’s how spurious correlation works. Suppose we have two things that are correlated. This means that when we see levels of one of them change, we usually also see levels of the other change. Because we’re academics, and not always very creative, we’ll call these things “A” and “B” (sounds like a Dr. Seuss book).
If we see “A” correlate with “B”
We might be tempted to assume that “A” causes “B” or that “B” causes “A.”
There may be causation, but there may not be. Before we go much further, we have to look for a third variable, boldly named “C”, that creates a spurious correlation between “A” and “B”. This happens when “C” causes both “A” and “B” and thus produces the observed correlation between them.
We might draw it as follows:
C
Okay, this is where presentations of spurious correlation usually stop. Some students get it, some don’t. I’m teaching research methods this semester, and I wanted to make this concept more understandable, so I asked the students to come up with examples of spurious correlation based on things that they observed in their everyday life. After reading these examples, you should have a better understanding of how spurious correlation works.
- One student had gone out partying the weekend before, and while sitting in the bar watching his friends during the evening, he noticed that people who had the most fun dancing were also those who were most likely to throw up by the end off the evening. It’s not that dancing made them sick (“A” causes “B”), or that being sick make them have fun dancing (yuck—“B” causes “A”), rather there is a third variable, alcohol consumption (“C”) that leads to both fun dancing and sickness.
- A student works as a nurse at a local hospital. He noticed that the patients who received radiation therapy were also those most likely to die (A and B). Why? Cancer (C) leads to both radiation and death.
- A student has noticed that the mornings when she has the toughest time getting out of bed are also the mornings with the most car accidents on the roads around campus. It’s not that her staying in bed causes car accidents (A-> B), or that she stays in bed out of fear of car accidents (B->A); rather bad weather causes both car accidents and her wanting to stay in bed.
- A student lives in the dorms next to some guys who are varsity athletes. The guys were explaining to her that women are attracted to them (A) because they are athletes (B). She, however, pointed out that it could be because of their muscular bodies (C). Maybe being strong increases their athleticism and it also attracts women.
- A student is from the Caribbean island of St. Thomas. He has noticed that the more people there are on the island, the more crime there is. It’s not that simply having a greater population leads to crime, or that crime attracts more people. Instead, the population swells when lots of tourists arrive, and the tourists commit crimes.
- A student explained that when she burps (A), she also gets the chills (B). She doesn’t think that she burps in response to the chills, or that she gets chilled because she burps. Rather, she just thinks that something weird is happening in her body (C) that causes both.
- Finally, a student and his friends formed a rock band some years ago, and they are trying to make it big. It doesn’t look promising though. As he explains it, the more they practice, the better they sound. Unfortunately, he thinks the relationship between practice time (A) and quality of sound (B) is spurious. The band practices in a friend’s unheated garage and during winter its pretty cold, so they don’t like to practice much. When they do practice, their fingers are so cold, that they have trouble playing their instruments. As such, the cold temperature (C) both decreases practice time (A) and performance quality (B), leading to correlation between them. The student doubts that they are actually getting better, and he may need to plan a career as something other than a rock star.
I posted a link to this article on my class blog. I'm teaching two introductory sociology courses and we discussed research methods for a short while. Some students commented on what they thought about this post. I put a link to their comments, just click on my name above and you can see what some students said about the helpfulness of this article.
Posted by: Steve | May 03, 2008 at 07:29 PM
Blogs are good for every one where we get lots of information for any topics nice job keep it up !!!
Posted by: proquest digital dissertations | December 30, 2008 at 02:30 AM
I am currently in an online sociology class and I think this post is a great way to explain spurious correlation. We're reading about the surveying portion of sociology and it does not talk about spurious correlation. The examples made the idea a lot more clear. They were very straight forward and were very helpful. Great post!
Posted by: Jaclyn McNallly | February 06, 2009 at 08:08 AM
Wow! I read about spurious correlation in my Sociology textbook and couldn't understand how to apply the concept to their example of juvenile delinquency and church attendance. The student examples you shared were more practical and I could easily understand the concept with your simple diagrams. I'll be following your other blogs now...I'm hooked!
Posted by: Mary Graca | February 15, 2009 at 09:06 PM
Could this be an example of a spurious correlation ?
An individual is attacked/mugged and beaten badly, days later in the hospital after numerous stitches, medical treatment, cat scans and x rays the doctor diagnoses that the patent will survive the beating, but signs of cancer were found on one of the x rays which the patent never knew he had. If it had not been for the beating/ mugging he would never have known he had cancer or the chance to now begin fighting it.
Spurious Correlation, am I in the neighborhood ?
Posted by: Martin Kelleher | September 12, 2009 at 03:05 PM
This really did help me better understand spurious correlation. I am currently in an online Sociology class and the textbook does not explain it clear enough for me to understand. Your many examples really did clear up everything that I was not understanding.They were easy to believe too because they were every day examples. This was an awesome post!
Posted by: Heather Chapman | September 27, 2009 at 09:14 AM
This article made so much sense. I didn't really understand spurious correlation from the definition in the book, but the visual examples using the arrows, and the applie examples using real life made it really clear and understandable. Thanks!
Posted by: Aleksis Landers | October 19, 2009 at 08:07 PM
Hi,
This is a new issue for me, but I must say that you have tried it well. Good work!
Posted by: Dissertation Help | October 23, 2009 at 06:35 AM
I'm in an online Sociology class and I actually just read a section in our textbook that talked about this subject. I didn't get it at first, feeling like I was missing something. When reading, I felt a sudden realization. This inturn caused me to get a better understanding. The Way you approached this topic made it more understandable and the examples were real. I think thats what made it so easy to understand. GOOD WORK!
Posted by: Ashlyn Mancini | September 20, 2010 at 02:37 PM
i am 83 years old, and am amazed at how little i know. I am retired, but busy all my life earning a living, raising a family, and so on. whenever I would see a subject such as a Spurious Correlation, I would stop reading the article, and go elsewhere, pretending I was not interested. In truth, I WAS interested, but thought I might not be smart enough to understand what I was reading. A is related to B. But now that I am retired, I find that I am a C, and able to read and understand A to B with myself as the 3rd variable. Or, am I mistaken, after all these years, tell me I am spot on, please. Even if I am not!
Posted by: scott denton | October 10, 2011 at 08:16 PM
The examples given for spurious correlation helped a lot with the explanation of the concept. It was very interesting to see the chain of reasoning and how A does not always cause B or B cause A in the examples you gave. The concept was much easier to understand and more interesting with obvious real-life examples.
Posted by: John Kim | April 09, 2012 at 05:18 PM