Skip to content
 

Fake MIT journalists misrepresent real Buzzfeed journalist. (Maybe we shouldn’t be so surprised?)

It’s a funny thing.

MIT—the Massachusetts Institute of Technology—that sounds like a serious place. They do lots of excellent work at MIT. It’s the top engineering school in the world!

Buzzfeed—that’s some online newspaper, it’s not serious, right? We joke sometimes about Buzzfeed-style headlines.

But when it comes to journalism, Buzzfeed is the real thing, while MIT publishes misleading articles that violate journalistic ethics.

This is a tl;dr version of my earlier post from today.

Last month, I received the following query by email:

I have interviewed John Ioannidis and Eran Bendavid regarding the Santa Clara study. I am writing for Undark (MIT’s science magazine) and wonder if you would be willing to chat briefly today or tomorrow morning?

The other day, the article appeared. It misrepresented all sorts of thing about the famous Stanford coronavirus antibody study, in particular presenting its critics as politically motivated without even once noting that the critics were correct all the time that the data reported from the Stanford study were consistentwith near-zero prevalence rates (see also here from Will Fithian). Not once.

Also, Undark misrepresented the careful and thorough investigative reporting of Stephanie Lee at Buzzfeed (see here and here, here). Lee explains in this thread how the MIT article twisted and distorted what she wrote.

That’s odd. MIT (with the exception of its Media Lab) is known as a serious institution, not a purveyor of fake news?

What happened?

Here’s a clue. At the end of the article is offered this disclosure, which did not appear in the original article:

One of the authors of this op-ed . . . has previously co-authored published work with the subject of the piece.

I hate the scientist-as-hero narrative.

I really hate it when it’s used by “MIT’s science magazine” to misrepresent serious research and serious journalism.

This is just a little story, not such a big deal—except that we as a society absolutely rely on independent researchers and independent journalists to protect us from the caprices of our would-be scientific overlords. There are already lots of places for public relations spinners posing as journalists to spray their squid ink. No need for MIT to supply them with one more venue.

P.S. Here’s the kicker. It turns out that Undark, the publication where that article appeared, is not “MIT’s science magazine” at all! So the reporter was lying to me right off the bat. I wonder how MIT feels about this.

P.P.S. I had another post about the issue of whether Undark is really MIT’s science magazine, but I took it down because that particular issue is kind of irrelevant. My problem is not with the title of the magazine but with the content of the article. I’d be happy if Undark would fix the article, and then they and MIT can get together and decide whether they’re MIT’s science magazine or not.

57 Comments

  1. Psyoskeptic says:

    I know it’s not about Ioannidis but, just because someone turns out to be good at metascience doesn’t mean they’re good at science.

    • Joshua says:

      Psyoskeptic –

      > just because someone turns out to be good at metascience doesn’t mean they’re good at science.

      His meta-survey of the infection fatality rate of COVID-19 has come under a lot of criticism as well. Does that qualify as metascience or science?

      • Anything you can do I can do meta… 𝅘𝅥𝅘𝅥𝅘𝅥 𝅘𝅥 𝅘𝅥

        • Have never met a meta-scientist that understood science.

          Maybe I haven’t met enough or perhaps there are meta-meta-scientists that I haven’t met, that have pointed that out already.

          On a more serious note, there must a fair percentage of so called scientists that have just gamed their way into being seen as successful and a few super successful that are then funded to do meta-science. (Some who fund meta-science are aware of this issue – but it’s very tricky).

          • Hi Keith,

            A base rate neglect may be implicated in your observations about ‘meta-scientist’ b/c, to an extent, nearly all of us engage in meta-analytic that are well-conceived or less well-conceived.

            David Sackett was one of John Ioannidis’ mentors if I understand John’s speech at Lown Institute, where Shannon Brownlee, one of two authors of the UnDark articles is Senior VP. It was a moving speech and struck a chord among many different audiences specifically within the physician community, a remarkable response, given the prestige of allopathic medical education.

            https://www.youtube.com/watch?v=N63skNtYaJw

            John himself has a background in math too,

            I think. One of his recent arguments, which surprised me given his stance on stat-sig, was that mathematical abstract modeling results were being favored over concrete clinicians’ observations as, for example, of the data being collected on the Princess Cruise Ship passengers and crewmembers.

            Shannon Brownlee has written an excellent book, Overtreated. She doesn’t form opinions prematurely. I admire that quality in her. And aspires to a collaborate model for Right Care Alliance organization.

            It could be that some of us have a better handle on the behind the scenes sociology of expertise.

            I myself want to present it in the most efficacious and helpful way.

            As for Andrew and his blog, I have been relieved to find such interesting and intelligent posters. It’s been a lot of fun for me too.

        • Martha (Smith) says:

          Thank you, Daniel and Keith, for your meta-play.

  2. Abe says:

    So now the Undark article has a correction disclosing a conflict of interest that was initially left out:

    “…one of the authors of this op-ed, Jeanne Lenzer, has previously co-authored published work with the subject of the piece, John Ioannidis. This should have been disclosed at the time of initial publication.”

    Even better, that co-authored article itself has a correction disclosing a conflict of interest that was initially left out:

    “The authors of this Feature (BMJ 2013;347:f5535, doi:10.1136/bmj.f5535) would like to declare the following competing interest, which should have been declared at the time of publication: Jerome R Hoffman has consulted on lawsuits involving allegations of negligence related to non-use of thrombolysis in stroke.”

    Icing on the cake: the co-authored piece is on “Ensuring the integrity of …”. Irony is dead.

    Links:
    Undark article: https://undark.org/2020/06/11/john-ioannidis-politicization/
    The correction to the co-authored article: https://www.bmj.com/content/348/bmj.g1335

  3. Roy says:

    There can be all the analyses in the world, but there are two things I have pointed out before:

    1. “It’s just the flu”: https://www.nytimes.com/interactive/2020/06/12/opinion/coronavirus-doctors.html

    2. 117,000 deaths in the US in 4 months and counting (compare with any seasonal flu year of your choice)

    Somehow you have to square your analysis with those facts. The series of preprints from Stanford do not.

    • Anoneuoid says:

      There was quite the spikes in deaths but cumulative deaths so far in the US are only about 1.1-1.2x higher than last year. Tell everyone to stay home where they sit around watching doom porn on the news all day, then they have panic attacks with the similar “shortness of breath” symptom, get presumed covid and put on a ventilator in a negative pressure room filled with concentrated virus in the air.

      I mentioned here before I now know two people with serious health issues. One can no longer walk after throwing out their back when just standing up straight and the other started having seizures from anti-anxiety/depression medications. Both are young but terrified of covid and havent left the house for months. I can only imagine what its doing to the elderly in terms of blood pressure, heart attacks, etc.

      As of week 20 (5/16/2020, lets ignore the latest few data points because those are typically missing a lot of deaths that get updated later) there were 1,134,468 deaths in the US in 2020 and 1,230,118 in 2020. So thats 95,650 extra deaths or about an 8.5% increase (btw, 83,575 covid deaths in the US are reported by that date). Lets just round up to 10% because surely more data will come in for 2020 for week 20 and earlier. So that’s ~115k “extra” deaths:
      https://xayadata.com/covidstates.pdf
      https://xayadata.com/mortstates.pdf

      Is it plausible that anxiety, sitting at home, wearing masks, etc increase the chance of dying in these circumstances by 1.1-1.2x? Maybe 1.05x?

      I also note the timing of the peak in deaths was april 11th. This is about two weeks after doctors in NYC started raising the alarm that “early intubation” was a very bad idea: https://www.youtube.com/watch?v=k9GYTc53r2o

      Of course lots of other stuff was changing too (lockdowns, weather, etc).

      Also out of those 95k extra deaths so far reported up to week 20 its about 15k in NJ, 10k in NY (excluding NYC), and 25k in NYC. So 50k, or about half the excess deaths were all in once concentrated area thats only about 8% of the population.

      • Ian Fellows says:

        If you asked people if they would be willing to make major behavioral changes for one year in order to increase their life expectancy by 10-20%, I think that quite a lot of people would say yes. 10-20% is pretty huge when you are talking about death.

        • Anoneuoid says:

          Yes, people were asked to make behavioural changes and theres reason to think these changes put them at some higher risk of death (but of course deaths from stuff like car accidents and occupational hazards would be expected to drop).

          The question is whether that can account for a non-negligible portion of the increased deaths we are seeing.

      • Phil says:

        Anon,
        You point out correctly that about half of excess deaths were in an area that had only 8% of the US population and had a really bad outbreak. You seem to be suggesting that if there were no shutdown, that wouldn’t change anything or at least wouldn’t change much: if the rest of the country had not shut down, we would still have had ‘only’ those 100K deaths (so far). That doesn’t seem like a defensible position.

        Also: “Is it plausible that anxiety, sitting at home, wearing masks, etc increase the chance of dying in these circumstances by 1.1-1.2x?”. No, this is not remotely plausible. We have about 100K deaths among people who were diagnosed with coronavirus either before or shortly after death. Most of these people died of coronavirus-related complications. For one thing, if there were _another_ 100K people who died of anxiety etc., the number of excess deaths would be 200K instead of 100K. There are several other reasons this is not plausible but I think this is the strongest one so I’m not going to dilute it with the others.

        I agree with you that there are public health costs to the shutdown, not simply a benefit from reducing coronavirus spread, and that there are good arguments to be made for relaxing the shutdowns in various ways. I don’t agree with some of those arguments but at least I agree with the point. But the idea that there has already been a big increase in premature deaths due to anxiety is not supported by the data.

        • Mendel says:

          Health costs and benefits are confounded in various ways.
          a) People are afraid to go to the hospital. But with no lockdown, wouldn’t they be even more afraid, as the hospitals are full of Covid-19 patients? How much anxiety is caused by conspiracy theories going around?
          b) There are health benefits from social distancing:
          — less contacts & pubs closed & no sports games means less brawls,
          — less traffic means better air quality and associated illnesses (e.g. stroke)
          — but due to higher speeds on the highways, potentially more severe accidents?
          — less non-Covid respiratory illnesses (data shows this!) including pneumonias of any kind
          These benefits need to be weighed against the potential costs. You need to find data and look at them comprehensively.

          Euromomo data shows a higher all-causes death rate in Sweden (no hard lockdown) than in Germany (restricted public meetings to 2 persons, closed most shops and venues, but still allowed people outside by themselves). German restrictions clearly had an overall benefit compared to Swedens less restrictive regime. You need to know which situations exactly you are evaluating.

          As a public decision maker, you have a choice of instituting restrictions early or late. For Covid-19., the cost of being days late is enormous. Compare the UK (who weren’t “surpised” like Italy was) with Germany, I think the UK was about 5 days later, and they have many more cases and bigger problems balancing case load with economic reopening. When you have a doubling rate of 3 days in the growth phase but a halving rate of 2 weeks in the mitigation phase, being a week late with the restrictions may mean a month more of them.
          For the suggested (unproven) health costs (anxiety, fitness etc.), do a few days earlier or later matter? Does the cost of locking down longer if you miss the “window of opportunity” matter? It seems it does, so being cautious and restricting early seems to have an overall net positive value over missing that window and restricting late.

        • Anoneuoid says:

          But the idea that there has already been a big increase in premature deaths due to anxiety is not supported by the data.

          Well if someone had a panic attack and complained they had trouble breathing I do wonder if back in March/April in the NY region they were likely to be presumed covid, get put with other covid patients, and then on a ventilator.

          I havent seen this data, only anecdotal accounts saying that was going on.

      • It would be helpful to track the individual’s symptoms over the course of the infection with greater granularity because the constellation of symptoms for each individual is unique, depending on the age and location of the individual.

        We’ll see how the weather plays out since it has been found that sunlight can kill the virus on hard surfaces within two minutes or so. Not sure how sunlight can affect the aerosols and droplets landing on masks or on clothing.

  4. MK says:

    Just because an author from Undark calls themselves “MIT’s Science Magazine” doesn’t mean they are. MIT is not mentioned anywhere on their website, and Undark doesn’t seem mentioned anywhere at MIT (at a quick search). What is the basis for their claim? Could you please update your title to write “Not only is Undark misrepresenting my comments, they are also misrepresenting their affiliation with MIT”. To put it converseley, an author who misrepresented themselves as writing for “MIT’s Science Magazine” also ended up misrepresenting your comments. And you’re falling for the same trap, and misrepresenting Undark as “MIT’s Science Magazine” to draw attention to your criticism of Undark. I agree with your criticism of Undark. But i disagree with their mis-representation, and your mis-representation of Undark as “MIT’s Science Magazine”. Please don’t try to right a wrong with another wrong. You were cheated, and I feel for you, but that doesn’t make them “MIT’s Science Magazine”, and it doesn’t make MIT responsible for their slight. Thank you!

  5. Mark Palko says:

    Lots a established publications and institutions have adapted badly to the 21st century while newer and/or less respected players have upped their game. NYT is my go-to example, particularly with the hype economy, tech messiahs, conservative messaging strategies, Trump statements and pretty much any story west of Pittsburgh.

    Andrew raised eyebrows suggesting Gawker>NYT and I still kinda surprise myself when I cite a CNN story to critique something from the Gray Lady.

  6. Undark is not an MIT’s Science Magazine. Not only did they misrepresent you, they misrepresented themselves.

    • It seems a little murky… they claim to be funded by the Knight Science Journalism Fellowships which is an organization that is run out of an MIT dept. or something like that. So they seem to have some kind of affiliation at least, but it seems clear that they don’t speak for MIT in any official way.

      • Andrew says:

        Daniel:

        I looked at the Undark webpage and I don’t see any mention of MIT at all. So, while they might have some tenuous connection to the institution, it can’t be right to call it “MIT’s science magazine.” That was just a ploy to get me on the phone.

        The ironic thing is that the story didn’t use a single thing from my phone interview. The reporter kept wanting to talk about Ioannidis and I kept saying it wasn’t about Ioannidis, it was just a research paper with bad statistics, etc. But that didn’t fit the story. The data didn’t fit the model, so the reporters threw out the data. We’ve seen that before!

  7. Jordan Anaya says:

    This isn’t the first time Undark has been critical of Stephanie’s work. Here they referred to her piece on Wansink’s emails as “piling on”.
    https://undark.org/2018/03/13/brian-wansink-data-masseur-science/

  8. Ernest Davis says:

    Andrew —
    Undark is a publication of the “Knight Science Journalism Program at MIT”
    https://ksj.mit.edu
    which “is part of the university’s acclaimed Program in Science, Technology, and Society (STS) located in the MIT School of Humanities, Arts and Social Sciences.”
    https://ksj.mit.edu/history/
    So it seems perfectly reasonable to me for Undark to claim that it is (one of) MIT’s science magazines, and only a slight stretch for Lenzer and to claim that they are thus affiliated with MIT. You ask how MIT feels about this. As far as I can tell my dear alma mater MIT doesn’t object to anything up to and including Jeffrey Epstein, as long as the soft money is good.
    — Ernie

    • Andrew says:

      Ernie:

      Yes, see P.P.S. My problem is not with the title of the magazine but with the content of the article. I’d be happy if Undark would fix the article, and then they and MIT can get together and decide whether they’re MIT’s science magazine or not.

  9. Shiva Kaul says:

    The Stanford paper was heavily scrutinized for good reasons, completely unrelated to personal agendas.

    1. Attention. It was the first result of its kind, on a contentious topic, with wide media coverage.
    2. Content. It published a statistical appendix with details on methodology, whereas other papers just say “we calculated some stuff”.
    3. Posterity. Due to (1) and (2), the study had the potential to serve as a template for future work. Oh well.

  10. Rahul says:

    “MIT publishes misleading articles that violate journalistic ethics.”

    So which University doesn’t? Is this clickbait again?

  11. Dan F. says:

    The culture where people are paid 200K annually for generating citations is the problem.

  12. Mendel says:

    “But when it comes to journalism, Buzzfeed is the real thing, while MIT publishes misleading articles that violate journalistic ethics.”
    Is that based on statistical or anecdotal evidence? ;-)
    “Columbia publishes misleading blog posts.” :-P

  13. Zad says:

    > Here’s a clue. At the end of the article is offered this disclosure, which did not appear in the original article:

    “One of the authors of this op-ed . . . has previously co-authored published work with the subject of the piece.”

    I’m only recently seeing all this but I always did say that disclosures like that would be super helpful/important! Unfortunate about that Undark article.

Leave a Reply