12,000 Retweets on an Unverified Anecdote; 2 Retweets for the Retraction
Misinformation alert: On Twitter Friday morning, a well-meaning doctor from the Pacific Northwest has posted a long “heard from a front-line Seattle doc” thread that contains really specific treatment advice and *sounds* of real.
But original post can’t verify the source, other Seattle docs are doubtful, and the information doesn’t track with the data. (The post claims that there are critically ill patients in their 20s; there aren’t any at this point). The post has 12K retweets.
The OP posted this apology.
It has 2 retweets. Please, please, please - even the most specific, jargon-laden threads should be suspect until their source can be verified. Do not rely on anecdotes from anonymous threads that you find in closed
Turns out that the same anecdote has been circulating about a hospital in China, and in Italy. So the info may be from a real person, taken wildly out of context. Please - repeat and amplify - do not spread anecdotes about what’s happening in hospitals unless you witnessed them. More on that below.
—
In the space of 96 hours, the residue of American civil society thickened, fortified: (a) state/ local governments stepping up, largely in a vacuum, (b) by journalism providing its core function, (c) by technology, (d) by individuals doing amazing things (and having horrible things happen to them), (e) by seeing others do right (f) , and healthy fear. Animating all of this has been rapid, resilient, often ad hoc, communication. Communication between and among; from top to bottom. A flow of enough solid, concrete, accurate information to topple my modified Gresham’s law – the wash of bad info driving out the good – and a platform – mostly Twitter, I think, that has facilitated it just efficiently enough.
Sitting from a position of relative comfort and privilege, I am guardedly optimistic that people are beginning to do what will required of them to flatten the curve of the coronavirus.
My focus here is on communications lessons learned. These are all provisional.
First, Twitter. I remember writing something very dumb about Iranians using Twitter in 2009, as it seemed like, just for a moment, hidden energies of the Arab Spring might prevail. My biggest failing was that I knew nothing about how Iranians used Twitter, what they used it for, and so I dramatically overestimated how effective individual protests were. Add in the Dunning-Kruger effect, a bias for optimism, a lack of personal stakes, aside from genuinely wanting Iranians to be free, and a few other cognitive heuristics, and you’ve the recipe of a hot take that spoils instantly.
This time, I feel more comfortable in saying that Twitter has been a catalyst for the type of creatively destructive interventions that local officials, public health mandarins, leaders of corporations, scientists, reporters, and countless others are making. We are building a new set of ethics for a society where a large number of sociable mammals are actively choosing to separate themselves, rather than to come together. There is a new ethical code, and Twitter seems to significantly magnify the social pressure that works as a compliance and even coercive force. It also compresses the time it takes for a decision to turn into an intervention. Judd Legum reported that Darden restaurants had no plan to give its 170,000 employees paid sick leave. Within moments, Darden reversed course. Legum + Twitter + social pressure = enough dissonance to change the minds of corporate executives. There are numerous examples of this; sports team after team deciding to offer paid leave to employees, the state of Tennessee regaining a spine and sharing testing statistics; Georgia’s state university system deciding hours after news spread on Twitter about its refusal to close classes for two weeks that it might be a better idea to follow others. The University of Washington’s Virology Department tweeted that it was running out of the thin tubes that transfer and parcel out fluids. An emergency. Offers of help came instantly.
That said, this system of social trust can be easily gamed, and this thin bond of trust can be broken when well-meaning people spread anecdotes without context that turn out not to be true. (See above). Twitter doesn’t have a way of verifying seemingly accurate anecdotes. Only crowd-sourced caution can. Social pressure can be brought to bear against people who willfully spread misinformation, but it can also help people take a second look before they share.
Many, many problems can’t be solved by Twitter, of course, and the social safety net is quite patchy. I worry about those in immigration detention, and those in jail – people who can’t distance themselves and can’t tweet for help. But in this pandemic, Twitter has become a public utility for the efficient distribution of information and the application of social pressure.
And I’m very curious about the externalities from all of these major decisions - things we don’t anticipate or hadn’t foreseen - rapidly made manifest. These will be positive and aversive. We will probably find out about them more quickly.
Second: the rise of #epiTwitter, the sudden privileging of a new stan for virologists, epidemiologists, microbiologists, ER docs and public health specialists. Nicholas Christakis (@nchristakis), Jennifer Nuzzo (@jennifernuzzo). Laurie Garrett (@laurie_garrett). Trevor Bedford (@trvbd), who is working in the middle of the Washington State outbreak, and many more. Researchers are sharing their provisional findings, their notes from conference calls, their guesses, their admonitions and cautions. There is plenty of bad information out there; bot-driven disinformation, non-experts looking at raw numbers and making predictions… but hours of looking through the spread of this convinces me that the good is ratioing out the bad. A corollary: Twitter helps scientists with different specialties collaborate. There are other collaboration platforms but there might not have been reasons for epidemiologists to work so closely with, say, front line mental health providers, and vice-versa.