The Headline “According to Science” is almost always a lie

Brigitte Dreger
5 min readJul 2, 2021


Photo by camilo jimenez on Unsplash

If there’s one headline that sets my alarm bells ringing, it’s this one. Why, you ask?

Let’s start with a quick look at the top headlines I see when I google, “According to science”.

#1. Real Simple

This is an article claiming that it can share a science-backed way to take a bath. Sorry to disappoint, friends, but there’s…no science in this article.

Oh, you’re surprised there’s no science in this article?

Are you surprised the authors didn’t run a double-blind, randomized control trial, randomly assigning subjects to baths with different temperatures? Baths with bubbles vs without?

Honestly, I was surprised too. (kidding).

Basically, this article quotes a dermatologist for credibility (who basically says, “don’t make it too hot”).

#2. Lifehacker

In the first paragraph, they refer to cortisol as a toxin. I literally need to read no further.

Let’s first break down cortisol:

  • Cortisol is a hormone that is naturally produced by your body.
  • It’s often referred to as the “stress hormone”, but cortisol is actually involved in a number of processes throughout the body, including immune response, stress response, and managing levels of glucose.

While it’s true that too much or too little cortisol can cause disorders in the body, it’s vital in many of your body’s natural processes.

A toxin, on the other hand, is a poison or venom that causes disease when present at low concentration in the body. If this was true of cortisol, none of us would be here.

#3. Eat This, Not That

So you don’t waste your time scrolling endlessly for the answer, let me spoil it for you.


Yes. The number one cause of arthritis. is. AGE.

So…stop aging when you hit 40. And if you’ve already hit that milestone…I’m sorry.

#4. Body and Soul

Again, let me spoil it for you. The best time of day to exercise?

Whenever the f*ck you have time.

We should be dubious of scientific information we read from the media

Media is a business.

The media makes money on you reading their articles and watching their videos. That’s why clickbait exists. Everyone wants your attention — and if they can get you to click, you just made them money.

And we’re all victims of it. Most of us are more likely to read an article titled, “Cry this often, according to science” than one that says, “The neurobiology of human crying.”

I don’t like it, but it’s true.

To me, “according to science” is a pithy, vague attempt to add credibility to under-researched nonsense by calling upon the highest possible power: science.

Saying “according to science,” would be like saying, “according to history,” or “according to English”.

But we don’t say those things.

Science is not some niche subject. It’s not two men in lab coats with pipettes doing all of the world’s research.

Science is massive. There has been an explosion of scientific research in the last century. Global scientific output doubles every nine years. We know more today than at any point in history. There are potentially 100s of branches of science, so many that scientists in different fields couldn’t possibly keep up with all the data, let alone non-scientists.

…And science is facing its own crisis of repeatability

The replication crisis has been haunting science for over a decade. On a large scale, scientists aren’t able to reproduce the findings of other scientists in published papers.

Bad papers are being published. And other scientists are citing them. And some science journalist is scouring the web for new science and does a write-up on it for some major publication.

And then you or I read it. And we don’t fact check it.

And there’s a decent enough chance that it may be false, misleading, too small a sample to be accurate, or done in mice (which doesn’t necessarily translate to people).

(for a fun Twitter feed that explores that last one, check out Justsaysinmice.)

So what is the solution? Is there one?

Nassim Nicholas Taleb says to watch the news less, for the more involved you are with current events, the less you know about the world.

While a laudable goal, that’s just not really realistic for a lot of us.

I say, critically assess what you’re reading.

In university biology, one of the first things you’re taught is how to determine whether a study is credible.

Let’s go back to the first example — the one about having a bath. How can we determine whether this is scientifically factual?

We’re going to ask these questions while we read it:

  1. What was the sample size? While there’s no hard and fast rule for this, generally, the bigger, the better. If they tested 2 people having a bath, we wouldn’t trust those results. But if they tested 2000 people, that’s a more reliable data point. Even better if it’s an aggregation of multiple studies (called a “review”).
  2. Where is it published? If the Harvard Business Review publishes an article about how baths may help you concentrate, that’s a more reliable source than Lifehacker or EatThisNotThat.
  3. Who funded the study? If this was an independent research project taken on by an academic team who were curious about baths, I’d be more likely to trust it. Conversely, if it was funded by Dove, I’d be skeptical: they have a vested interest in me taking baths (especially bubble baths with their products).
  4. Was the study done in humans? If we’re curious about when we should bathe, it’s not super relevant to learn about how baths impact mice or rats.

For a non-science article, I’d probably stop there. If you want to dive deeper though, you can:

  1. What were their methods? Ideally, you want something “double-blind and randomized” — basically meaning that participants are randomly assigned to a group (bath or no bath); and the researchers don’t know who’s getting what (bath or no bath). At the end, the scientists review the outcomes of all participants, and draw conclusions based on the data alone — without being biased toward one outcome or another.
  2. Were there confounding variables? Basically, were there things that could have influenced the study? Like, did the people assigned to the “bath” group all go for a run right before, which may be responsible for any perceived benefit?

While I could go on, I don’t really think I need to. For the most part, these will be enough to let you know whether you can trust a study or not.



Brigitte Dreger

Talking about the things people are afraid to talk about. LGBT+, Startup Culture, Diversity.