Correlation VS. Causation (Back to Basics: Tools We Need to Understand Health Information)

I recently saw a really distasteful and misinformed post on Instagram – ok I realize that could refer to like millions of IG posts really, but there was one in particular which I do not want to link (so it doesn’t get more reach) that used the word correlation to reference a relationship between two different variables (variables aka two different things, concepts, occurrences – like smoking and lung cancer to give a very basic example). The author clearly was harnessing the fact that many people don’t understand the true difference between correlation and causation – or, the author themselves was simply unaware of the difference. I’m not sure. But it felt both malformed and of harmful intent.

Whatever their purpose was, it got me thinking about how correlation and causation are frequently misunderstood and misinterpreted. One is often mistaken for the other (correlation for causation) yet they are significantly different when it comes to reading studies, examining data and being informed. These terms are important (in my opinion) for anyone who is doing their own “health research” to know.

So I thought why not write a quick blog post and take us back to the basics of informed healthcare. Understanding these two terms and the differences between them is truly crucial to being able to be informed – because, as I often encourage people to use reliable resources to look into topics of interest/importance for themselves (in addition to asking their healthcare provider always), I should also make sure that people understand these terms so they can fully understand the information they’re reviewing.

How to Identify Reliable Resources Online – Read It Here Now!

Disclaimer: this is not medical advice. Please seek medical advice from your healthcare professional.

What are “correlation” and “causation”?

Both correlation and causation are terms that speak to the relationship between two variables or things.

Correlation implies that there is a relationship between two things, but how the two things are related we can’t be sure without a doubt. We may hypothesize or think that one thing causes the other, but we just can’t be sure (I’ll get to the way we could be sure in a few).

Causation implies that two things are related because one causes the other to occur. Therefore, it highlights a “cause and effect” relationship. This is a lot harder than it seems to prove.

It can also be hard to see, when two things are correlated, which direction causality flows in (does A cause B, or B cause A?).

Both correlation and causation can be examined in scientific studies. Remember that one single scientific study is not usually enough to render any type of sure result – repetitive and transferrable (across multiple studies, populations, etc.) results are key. For example, many studies have been done to show a very strong correlation between certain types of lung cancer and smoking cigarettes. For some time, it has been widely accepted that smoking cigarettes is the #1 risk factor for lung cancer development as a result of the extensive amount of studies done around this relationship.

Controlled studies are able to show that smoking cigarettes creates an environment that is conducive to the development of lung cancer (and other cancers for that matter).

But what kind of studies can be used to examine correlation and causation?

Correlation can be seen more easily than causation. A study that simply observes two things can usually conclude that there is a correlation between the two (again however, multiple studies would be needed to support this claim)- however, the nature of the correlation or the nature of the relationship is what is harder to determine. This is why proving causality (if A caused B or B caused A, or if neither was the cause of the other but an unknown variable not yet considered!) is particularly challenging.

A more controlled study is needed to prove this – usually done in a lab where we can control for confounding variables, which are outside factors/things that may impact the things we are examining, studying or testing. But with human studies, this is hard to do – we humans are simply affected by too many things in a day! What we eat, where we sleep and the place we live for example, can all affect outcomes for different things. This makes human studies in a controlled environment for many things difficult – which is why we unfortunately don’t have a lot of “sure” answers to things, but many “probable” ones. There is also of course the issue of ethics in studying human beings – but that is a topic for another day.

Notably, an observed relationship may also be a coincidence or the way we think a cause and effect relationship flows may actually be reversed. See how many things can get in the way of coming to a conclusion?

The takeaway point here is – it is wrong to assume that because two things/events are related or seem to be, that one causes the other. As you can see, it is much more complex than that.

You should assume that most relationships – unless specifically noted – are just that, relationships. Some are strongly related – like with risk factors for certain diseases; but actually being able to say one thing solely causes another is very hard to do when we’re talking about human health and well being. Humans are complex so it makes sense that the science related to us is too.

Why are these terms important to know?

There is a lot of information out there. Some of it is very useful and valuable, but it has to be interpreted correctly. Some people make a business out of misinterpreting data from studies on a public platform to fit their own needs! Seriously, they do. Be weary of that.

My goal here is not to make you fearful – it is to make you aware. I started this blog because of the immense amount of misinformation that exists in today’s world – and it spans much further than maternal, reproductive and family health topics. If you’ve been keeping your eyes open during this pandemic, you’ll see that misinformation is absolutely rampant and it has the potential to be very dangerous. I want people to be able to recognize information and interpret it as friend or foe; valuable or misinformed.

Knowing the difference between these two terms – and understanding the difficulty of truly being able to say one individual thing causes another – is very important when looking at data and information. As I said – we are complex beings, and the world is complex. One thing is not usually the sole cause of another – even with smoking and lung cancer, though it certainly increases the risk, we see that some types of lung cancer develop without a person ever touching a cigarette to their lips. It’s because the relationship is more complex than it appears at face value. And that’s what I want you to take away from this.

Statements that make two things seem easily connected often need to be scrutinized unless they’re backed up by a significant amount of data. Think of the statement “vaccines cause autism” which we know to be extremely false – beyond all the other issues with this statement, it’s making the relationship sound too easy. That is a red flag for misinformation, because relationships are not often simple in reality.

I also want to highlight that information on risk factors is incredibly important – I’m not devaluing the relevance of a strong relationship. I am shining a light on people who take a relationship and infer that it is causal – to fear monger and scare people. Two unrelated things can be related – either by coincidence, or sometimes because there is a confounding (third) variable/thing that we cannot see causing one or the other. A relationship does not have to be a cause-and-effect relationship.

The post I saw mentioned a correlation between two things – but they made it seem sinister. They didn’t say – by the way, this was from an observational, single study where one of the noted limitations was that confounding variables could not be entirely controlled for (because it was not done in a lab, etc.). They were using the concluding phrase to their advantage, to push their agenda, and nothing more.

I hope that the posts on my blog help you to do your own research, know when to ask for expert advice (and your healthcare professionals should be there to answer your questions), and when to turn a blind eye to social media posts that are trying to sell you on one person’s view.

Just because someone shares a study link along with a claim does not make it true or relevant.

And more importantly, not everyone knows how to read or interpret studies – we can learn. I am still learning and I’ve spent about 8 years of higher education learning to do research and read studies!! Also, some people can read studies in their own field very well, but may not know the terminology or correctly interpret conclusions outside of this field. There are so many things to consider.

Be on the lookout for statements that claim causal relationships (look into them!) or too easily connect two concepts. Studies should back up large claims, but not just one study.Also be mindful of where you are seeing the post – is it a reliable resource? Just because someone puts “Dr.” (or any other grand title) in their name does not make them reliable.

Moral of the story – be informed about how to be informed, ask for help from experts (real experts!) and beware of misinformation that is out to make you feel bad, guilty or worse. Happy scrolling!

Additional Resources

When Correlation Does Not Imply Causation – Harvard University

UCSF Evaluating Health Information


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: