Communicating About Correlation and Causation


CommunicateHealth
Alt: A newscaster doodle on TV says, “Studies show that putting a cat on your head causes cat head!” A doodle watching TV at home puts a cat on its head and looks confused.

Here at We ❤ Health Literacy Headquarters, we love to back up our health advice with solid research. And that means we’re always on guard against that most menacing of research sins — mistaking correlation for causation.

Of course, as health communicators, we like to think we can spot the difference a mile away. But let’s be honest: it’s easy — and even tempting — to conflate them. Spend enough time mining a broad data set, and you’ll see all sorts of tantalizing correlations that are just too good to be true. Red wine linked to heart health! Chocolate associated with winning Nobel prizes! And before you know it, you’re trying to guzzle and munch your way to a tip-top ticker and a major academic award.

So this week, we’re offering tips to keep ’em straight. First, a quick refresher:

  • Causation means we know that X causes Y to happen. In health research, proving causation is a pretty big deal. Think cigarettes-and-lung-cancer-level certainty.
  • Correlation means there’s a relationship between 2 things, but we don’t know what kind. Is X causing Y? Is Y causing X? Is there a sneaky Z factor waiting in the wings to muck up the whole equation?! We just don’t have the evidence to say for sure. Think the murkier waters around breastfeeding and breast cancer risk.

With that out of the way, there are 2 main things that will help you keep your causes far from your correlates in health materials. First, check the original source. Online news is notoriously click-baity when it comes to reporting the latest health-related studies. So don’t take someone else’s word for it! Instead, look at the original study and consider:

  • Were the researchers aiming to prove causation?
  • Was it a randomized controlled trial in a peer-reviewed journal (the gold standard for proving cause and effect)? Or was it an observational study that noticed an interesting relationship, but couldn’t prove the “why”?

This may well send you down an internet rabbit hole here and there, but it’s well worth it to make sure you’re not accidentally leading your dear readers astray, dear readers. (And this type of dutiful sleuthing can also help you avoid falling prey to pseudoscience.)

And once you’re sure you know what you’re working with, choose your words carefully.

  • If you’re writing about causation, knock yourself out with strong verbs like “Some types of HPV cause cervical cancer.”
  • If you’re writing about correlation, tone down the causality with a nice hedge-y approach like “breastfeeding may lower your risk for breast cancer” or “X is linked to Y.” (Don’t be alarmed — we know it’s passive voice. But every once in a while, the dreaded passive can serve a purpose.)

And while you’re at it, be upfront about any uncertainty. It’s better to tell your readers that experts just aren’t sure than to have to pull a U-turn in your health advice later on.

The bottom line: When writing about cause and effect in health materials, keep a tidy line between correlation and causation.

Tweet about it: It can be tough to draw a tidy line between causation and correlation in health materials. @CommunicateHlth has #HealthLit tips to help you write about cause and effect: https://bit.ly/3jnJO99 #HealthLiteracyMonth





First Published at medium.com on 2020-10-02 01:36:37

Leave a Reply

Your email address will not be published. Required fields are marked *

Scroll to Top