Correlation vs Causation

Don’t worry. This post will not be filled with terms like standard deviation, stanine, or scale scores. As educators, we spend a lot of our time hearing how important data is, yet more and more often, those who preach about the importance of data, have no idea what they are really talking about. Too many politicians, bureaucrats, and policymakers make decisions because they do not understand the most foundational idea of all statistics. I am here to help. Correlation does not mean causation.

A few years ago I had the opportunity to move to Florida to work as a building principal. As a man who has spent his entire adult life in Michigan, the idea of moving to the sunshine state with my family was something of a dream. Upon accepting my job offer, I immediately began looking for a home. As my family was preparing for our move and discussing the weather, my kids began to get excited and started asking about the possibility of getting a pool at our new Florida house. They were pretty convincing.

They talked about the idea of playing together outside all year long, making memories, staying healthy, and genuinely just being happy. I was just about convinced until I started doing some research. What I found was extremely alarming and as a dad and protector of my family, some of what I found was flat out scary.

It turns out that a few years ago a study was conducted and it was determined that in every state in America there is a correlation between the amount of ice cream kids eat and the number of childhood deaths resulting from drownings in pools. You read that correctly. The more that kids eat ice cream, the more likely they are to die from drowning in a pool. It may seem like a crazy connection to make, but it is totally accurate. What was super disappointing is that my kids, just like me, love eating ice cream, so as we were preparing to move across the country I was stuck with a difficult dilemma. My kids were going to have to choose between ice cream and getting a swimming pool.

Some of you reading this may be thinking this is absurd. I agree, but this is what happens when we begin to confuse correlations with causations. We see relationships and we jump to conclusions. Eating ice cream does not cause drownings, but there is a connection. The connection is actually pretty straightforward. More people tend to eat ice cream when it is hot outside. Likewise, more people tend to go swimming when it is hot outside, so as a result of more people eating ice cream and more people being in pools, there is a greater likelihood of accidental deaths by drowning. There is a correlation between the two, but there is not causation. This is similar to the old wives’ tale that children will get sick if they are outside in the cold. This was the belief for decades because we could see a correlation, a relationship, between childhood sickness and the winter months. As a result of this connection, some began to assume causation, but truth be told, we see increases in the common cold and the flu in the winter because children are back in school breathing and sneezing on each other. We spend more time indoors breathing recycled air. And we tend to be less active and more passive in our physical health. Sure there is a correlation to sickness and cold air, but that does not mean there is a causation.

Similarly, in the world of education, we hear statistics like:

Children who attend preschool are more likely to be reading at grade level in third grade. While there is a definite relationship between preschool attendance and reading proficiency in third grade, it is inappropriate to assume one causes the other. As we explore the students who do attend pre-school, we see other compounding factors such as parental marital status, exposure to life events, travel experiences, parental education levels, etc…that all impact student literacy. The scores are related, but this is not a guarantee of saying one causes the other.

At the secondary level, we see similar claims when looking at SAT and ACT scores and college readiness. There is a definite correlation between SAT scores and college freshman g.p.a.’s (that is actually as far as the correlation goes. Yup….only freshman year g.p.a). The truth is that the College Board designed the SAT to create a bell curve of results. On their website, we actually see an admission that they create an assessment that results in a bell curve because of the simplicity of analysis, not because it is predictive of future career or college success. The truth is, students who score well on the SAT may have an easier time participating in a lecture-style classroom that carries with it a requirement for independent study of low-level recall, but this does not mean these students are better prepared for long term success….as a matter of fact, there is not even a correlation of that, let alone a hint of causation.

This spring, there will be a number of people who begin making claims based on data that is reported. We will see student achievement scores, teacher evaluation ratings, school grades, etc…As the numbers roll in, pause before you make assumptions on what the data are telling you. Often the story is more complex than what we see at the surface. Question the policy writers and decision-makers. Do not let an editorial in the local newspaper, based on ill-informed reasoning shape the destination of you or the students in your care.

There is a correlation between the number of tests we give to students and the burn out of teachers in the field. I am not 100% sure I can claim causation yet, but I sure am willing to claim that the stress and negativity that comes from ill-informed judgment is a contributing factor. The more you know, the more you can grow. Want to be a part of the ride? Sign up for the Lasting Learners newsletter and be in the running for FREE giveaways every single month: http://eepurl.com/cQwHA1

Want to hear your stories and audio blogs every week? Listen to Lasting Learning: https://anchor.fm/david-schmittou