Maybe I’m old-fashioned, but I feel that if you publish fact checks, you should be able to interpret data correctly.

Reuters released this “fact check” about a CDC study concerning masks. Let’s analyse what is actually being said here.

The TL;DR seems straightforward enough:

Posts on social media say a study of the U.S. Centers for Disease Control and Prevention (CDC) showed that a high percentage of people who wore masks became infected with the new coronavirus and that masks are ineffective. The results of this study, whose aim was to assess situations of community and close contact exposure, rather than the use of masks, have been misrepresented.

Let’s dive in further. At first, when publishing anything these days, it’s important to blame Donald Trump. Always blame Trump. ORANGE MAN BAD!

The results of the September CDC study have also been misrepresented by media outlets and U.S. President Donald Trump. Fox News’s Tucker Carlson said the study showed that “clearly (wearing a mask) doesn’t work the way they tell us it works.” Trump has erroneously referred to the study at least three times, including during his town hall on Oct. 15.

I guess that’s when Reuters decided to take a look at the study.

While the numbers mentioned in these claims are authentic (see page 3), the data is missing context. It is true that 85% of people in the study who tested positive for the coronavirus were reported to have worn a mask always or often. It is also true that mask-wearing levels were reportedly high (88.7%) among people who did not catch the virus.

OK. So how is it wrong to say that, apparently, a lot of people who end up SARS-CoV-2-positive were wearing a mask?

Jason McDonald, a spokesman for the CDC, told Reuters via email that “the aim of the study was to assess possible situations for community exposure, not mask use.” The study, McDonald noted, found that “going to places where mask use and social distancing cannot be maintained”, like restaurants, coffee shops and bars, “might be an important risk factor for COVID-19.”

That is nice. But what does the aim of the study have to do with the data gathered and the conclusions being drawn from that? Wouldn’t it be bad scientific practice to have your study goal influence your conclusion? Shouldn’t your conclusion be solely determined by the data you collected and analysed?

Authors of the study, which analyzed the activities of two groups of people in the 14 days prior to showing possible COVID-19 symptoms, also acknowledge that the results might not be representative of the U.S. population, as it only considered adults that had been tested in 11 health care facilities.

In total, 360 people tested for the coronavirus in July were part of the analytic sample: 154 that tested positive (case patients) and 160 that tested negative (control patients).

And what are you saying here, Reuters? It seems that you are now making the argument that the study wasn’t representative. So what is it? Has the study been misinterpreted – as you claimed on the outset of your “fact check” – or is the study bullshit? I am confused.

It seems the authors are deliberately trying to confuse readers. After this bit, the article goes into several paragraphs paraphrasing results from the study – I thought it wasn’t representative and therefore not trustworthy – that have nothing to do whatsoever with the question that this “fact check” set out to answer.

The article then quotes the CDC guy again:

McDonald added that the CDC has clearly stated that “that wearing a mask is intended to protect other people in case the mask wearer is infected, rather than the wearer”.

Yes, OK. Fair point. But what does that have to do with the result of that study? At all?

Reuters now tries to hedge their bet a bit:

Health authorities have never said masks alone will slow the spread of the coronavirus, but have an impact when combined with other prevention methods, such as hand washing and social distancing.

Again: What does that have to do with the CDC study that was apparently so egregiously misquoted? For most of your article, you have not even tried to show why the conclusion you’re trying to prove wrong is actually wrong. But hey, let’s reiterate this bullcrap in the verdict, shall we?


Missing context. Both groups of people with and without the novel coronavirus, demonstrated high levels of mask use (85% and 88.7% respectively).

So, seeing as both people who got infected and those who didn’t wore masks a lot, how exactly is it wrong to conclude that wearing masks obviously didn’t help the people who got infected? And to further conclude that depending on a mask for protection is ineffective?

So your actual criticism of Trump and Carlson should have been that they misunderstood what masks are for. That they are worn to protect others, not yourself1. But they didn’t misinterpret the study in question. Actually, they interpreted it better than your writers.

The study, which aimed to analyze community and close contact exposure and not mask use, found that more infected people had been to locations where masks cannot be worn effectively, like restaurants and bars.

Again, it doesn’t matter what a study aimed to do. You are showing that your writers don’t know how to read scientific publications efficiently. You don’t read them from top to bottom. You start at the bottom. The conclusion is the most important part. The introduction and the aims of the study don’t matter if you are trying to analyse what the results mean.

How is this shit acceptable? How can you in good conscience, as a journalist, for a huge news organisation like Reuters no less, write a “fact check” and then not check the fact in question? Instead you waffle around aimlessly, quote people without understanding what they are saying and show that you don’t even have the most basic understanding of how to work scientifically. And you’re claiming the authority to educate the public on these things? You should be ashamed of yourselves.

Cover photograph: Franki Chamaki

  1. If that actually works is something that clearly hasn’t been comprehensively studied, either. Especially considering that a significant portion of the public is wearing masks that do not protect other people (including, but not limited to, N95 masks with exhalation valves). The CDC itself, which apparently is so adamant about masks being there to primarily protect others, is obfuscating the issue in their FAQ on PPE: “An N95 respirator with an exhalation valve does provide the same level of protection to the wearer as one that does not have a valve. The presence of an exhalation valve reduces exhalation resistance, which makes it easier to breathe or exhale. Some users feel that a respirator with an exhalation valve keeps the face cooler and reduces moisture build up inside the facepiece. However, respirators with exhalation valves should not be used in situations where a sterile field must be maintained (e.g., during an invasive procedure in an operating or procedure room) because the exhalation valve may allow unfiltered exhaled air to escape into the sterile field. If you only have a respirator with an exhalation valve available, cover the exhalation valve with a facemask (surgical or procedure mask) that does not interfere with the respirator fit.” — With other words: Wearing such a mask doesn’t protect others. But we are burying that information in a lot of superfluous nonsense so that people don’t notice that none of our mask guidance is consistent. ↩︎