David ThielDec 19, 2022

I wrote a bit about an effect that I'd seen for a while but had difficulty explaining: we call it "Content Moderation Survivor Bias", and it's an effect that can muck up social media analyses and lead to dubious conclusions.

I define it thusly: in a retrospective sample of moderated social media platform, ToS-violating or inauthentic content tends to appear most prevalent in the immediate past. This appearance is misleading, however.

https://cyber.fsi.stanford.edu/io/news/content-moderation-survivor-bias

Content Moderation Survivor Bias

Show threadGabi PurcaruDec 19, 2022

@det under the hood, content moderation is actually a whole array of parallel mechanisms, from simple list-based filtering, machine learning, and delayed actioning. Generally, the longer you wait, the more confidence you have in your actions (lower false positives).

To get accurate numbers, you need to take a daily sample of recent content and compare those daily-generated figures.

But this is very expensive.

Show threadDavid Thiel
@gabipurcaru Yup — to get accurate numbers you ideally define your queries ahead of time and use a real-time API, but that requires foresight we don’t always have.
Dec 19, 2022 at 9:03pmWeb