Thread by @cphoffmann

This type of reporting is so irresponsible and a big part of the problem. A political group hires some consultancy to run the type of analysis that has been refuted again and again - just to claim strong disinformation effects without any evidence. 1/
theguardian.com/world/article/…

This analysis, which is hardly documented and (of course) not peer-reviewed, basically categorizes a bunch of accounts that follow each other as troll accounts. They may well be. But that tells us absolutely nothing about the reach, let alone impact of these accounts. 2/

These types of analyses are all about big numbers that sound scary („thousands of accounts with hundreds of thousands of views“, „flooded“). Yet these numbers don’t tell us if real citizens actually saw this content or if this content had any impact at all. 3/

The analyses never provides scales: that’s because the seemingly scary big numbers turn out to be vanishingly small when compared to the total number of active users, number of tweets, shares, views etc. amassed on a single day on X. 4/

The analyses describes the troll network as „densely connected“. Sounds scary, right? Well, that basically means that these accounts follow each other and don’t reach any real users. But the authors claim real users are „surrounded“. Sounds scary, doesn’t mean anything. 5/

The reality is, that this content was likely seen by a tiny minority of X-users and constituted a miniscule percentage of the content even these exposed users saw on X. It is exceedingly unlikely that this content had any effect at all on anyone. 6/

But that’s not the message these sources want to put out there. They want to fuel fear. It’s a toxic business model. We need to be more outspoken in criticizing these misleading contributions to public discourse on misinformation. 6/

This type of shoddy, misleading, fear-mongering PR & journalism is itself misinformation, if not disinformation and harmful to our public discourse. It undermines trust, sows discord, and weakens our institutions (here: acceptance of election results). /end