Correlations: How The Media Ruin Science

We’ve all seen those headlines and articles claiming something along the lines of “Scientists prove that [food or beverage] causes [adverse medical condition].” For example, meat and cancer (which I debunked here).

The problem with a lot of these claims are that most of these things are not “proven” and there is usually not enough evidence to establish causality (one thing or action causes a certain effect to occur every time you have that thing or do that action). Many of these headlines pull their “facts” from a type of research known as observational study.

These studies essentially follow people and record a certain behavior or pattern and look to find any common trends between the people being observed to develop an association or possible link between “behavior X” and “outcome Y”. These are your meat and cancer studies or red wine and heart health studies (you’ve probably seen basic moms post about that one to justify their alcoholism, I know I have!).

Why does this matter you might ask? Well science gets reported often in the news and state and local representatives, in an ideal world, want to protect their constituents, so they will do what they can to service their community. If a study gets misreported by media claiming that meat causes cancer or dihydrogen monoxide is harmful to our health, then misinformation may be spread by the representative who is simply trying to help their community by banning water or meat because the public who watches the news demands they do something about this travesty!

Here’s another example of why we can’t rely on solely observational studies to make our decisions. Ice cream and drowning. Maybe you’ve heard this before. Ice cream sales are highly correlated with drowning deaths. This means that as ice cream sales increase, so do deaths from drowning. Does ice cream lead to drowning? Or do drownings lead to buying ice cream? Probably not for either. There’s something else at work here known as a “confounding variable” that is clouding the conclusion. Ice cream sales increase likely during the Summer and more people are also likely in the pool during the Summer, also leading to increased potential for drownings.

We have to be careful with news stories that reference research studies because observational research is popular among journalists because they can make bold claims like the ones mentioned before; however, as we’ve seen, those claims don’t actually hold up/are not what the study is actually saying.

On a related note: Be very careful of the word “proven”. Rarely is something proven in science by research and when it is, it has been studied over and over and over and over and over and over and…you get the point. It takes a long time, potentially decades, to establish causality or proof that one thing causes another. We can make associations about things, but causality is a completely different concept because it has to happen 100% of the time. If meat really did “cause” cancer, then there would be an actually be an outright ban on meat to protect the public; however, it’s only an association and it’s weak at that. Not everyone who eats meat will get cancer.

2gqdkh.jpg

To demonstrate how hard it is to prove something, consider this: Gravity is still a theory. Our explanation of why we observe the effects of gravity is only a theory, meaning we’re pretty sure but not 100% sure. So when someone says that science has proven that processed foods are cancerous or that this causes that, put on your skeptic hat because that person is probably speaking in hyperbole to argue their point rather than present fact.

You may have also heard that insulin causes obesity or “this one thing” causes obesity. Put your skeptic hat on for a minute: There are so many aspects to our lives that could contribute to weight and fat gain, how is it possible that we can pin all the weight gain for the billions of people that gain weight and fat every year? I’m willing to say that it’s impossible. While this may start to sound like a rant, I think it’s appropriate to say because many people will blame one issue when there’s a plethora of other potential things going on.

If you approach someone’s weight problem thinking that there’s only one underlying issue and you try to treat just that one thing, you’re probably going to fail because they could have other factors or behaviors in their life that could be adding onto the complexity of obesity. It’s also belittling to the person to say they just need to stop doing “x” and when they do it, they’re still obese. Don’t be a dick. Be an empathetic non-dick and realize obesity is a complex issue that has many potential causes and it’s up to the trainer or care provider to find those causes and help the person get through them instead of blaming insulin, processed foods, or their lack of motivation. That’s being a dick.

All of this leads to the famous phrase said by many people in the sciences: Correlation does not equal causation. I hope I’ve burned into your mind why this is. I also hope that, if you’ve stuck around for this long to read this (thank you by the way), then you understand to be more skeptical when you see a sweeping headline claiming causality or that science proved something.

Because if we can’t even be 100% sure about gravity, how the hell are we going to be sure about a food causing cancer? 

Thanks for reading! What questionable shit have you seen online or on the news?

Contribute to the conversation

This site uses Akismet to reduce spam. Learn how your comment data is processed.