Read on Twitter
For all the hate on @FiveThirtyEight for incorrect predictions their much larger journalistic failure was not even making an effort to report on, measure, and model the effects of disenfranchisement on the outcome. This was a story they were uniquely qualified to tell.
I started following @NateSilver538 back in the Blogspot days because he did such a great job explaining the statistical intricacies of polling and specifically how much uncertainty was involved. It was a masterclass in epistemic modesty.
But the 2016 and 2020 version of the site is just another form of horse race reporting. And it’s a particularly toxic one because it produces constant and authoritative seeming information on a subject where people have deep fears and anxieties.
Instead of channeling those fears into increased knowledge about candidates or issues or activism — or even increased understanding of polling methodology and its inherent uncertainty — it just feeds off of them for clicks.
Like a lot of the internet @FiveThirtyEight has become an opportunity for people to LARP as experts without gaining any of the modesty or caution that comes with real expertise. It teaches its readers jargon and shows them graphs that give them a false sense of control.
I’m sure the thing I want from them — actual reporting on voter suppression that they could integrate into their predictions — would be more expensive than what they do now. And I know the news business has been decimated by the internet over the years culminating with Facebook.
Maybe @propublica or someone else in the non-profit news space should start an election polling analytics site that specifically focuses on issues of disenfranchisement and how they affect likely outcomes.
