Ever seen an ad for #GroundNews and wondered where they get their data? Well, they too get it from Dave. Depending on the metric, they average his scores with either 1 or 2 other so-called "trusted ratings systems."

Yet upon Dave rest the conclusions of hundreds if not thousands of studies (>500 on arxiv alone), representing unknowable amounts of computation and thousands of hours of labor. This is the spectacle of computation:

theluddite.org/post/spectac...

He uses a comically naive rubric to generate the data. It is the exact political theory that just-some-guy-named Dave would have. Had the site been DavesOpinions.com, no one would use it. Transmorgify them into data, pick a neutral name, etc., then researchers use it without even knowing who Dave is

Digging further, we find that MediaBiasFactCheck.org isn't run by, say, a lab, or a panel, but some guy named Dave. Dave is a healthcare worker in North Carolina with a passion for bias in the media working with a couple volunteers. According to the site, he makes all final editorial decisions.

Our paper's sources leads us, through several schema transformations and an intermediary, to MediaBiasFactCheck.com, a common source for those studying misinformation. It rates NYT as further left than FAIR.org, which regularly criticizes NYT for being a right-wing rag (my words, not theirs).

Merci ! Je suis toujours plus inquiet pour l'état de la science des données, ça ne serait pas le seul article sur le sujet. Maintenant, tout le monde parle de l'IA, mais les problèmes comme le positivisme non critique, les biaises, etc. viennent d'avant et ont des racines plus anciennes.

How can you label millions of tweets as true/false? If we pick a paper and follow the provenance of its data, I think that we can learn about not just this field, but also about the power of tech companies and the role of data and computation in society.

theluddite.org/post/spectac...