[dado3212]

This is kind of an open problem, especially in the Internet age. Any retraction that’s published will have far less reach than the original incendiary claim. And it’s simply not viable to follow that hundred-chain length sourcing tree for everything used in an analysis. Not sure that there’s a solution that’s actually implementable.

[RetroTechie]

Life's too short to check or even verify the many factoids you encounter in daily life. With:

Check = check cited sources yourself. And

Verify = re-do a study yourself. Outside of hard science like physics (and relatively simple experiments), in most cases that's not even possible. Especially history: original artifacts disappear, witnesses die & the places where things happened are changed.

Written records are just that - what was written down. Accurate or not.

This is true for ordinary people as well as scientists. Though I'd expect scientists to make a bit more effort (+ using field-specific knowledge to sort things out).

But news reports tend to be notoriously unreliable imho. Especially popular media or (yuk) social media.

If some science study is reported on, it's often enlightening to read that study yourself. And find it says very different things than the news reports. Or does not even say what's reported elsewhere. Bias, sensationalism & clickbait are a thing.

[P_I_Staker]

Yeah, this has been a pet peeve of mine in online conversations where people call for sources. It's pointless, and maybe harmful. You can't just go looking to attach a citation to prove your point, without critical thought.

I've seen countless examples of "citations" that literally contradict their point. I can recall and exchange on reddit were someone smuggly claimed "they use science" when trying to make this crazy claim about how intermittent fasting is good for managing blood sugar and you should skip breakfast.

The citation said that people with blood sugar issues should always eat breakfast, and have small meals throughout the day. How did any of this help us get to truth?

[MichaelZuo]

> And it’s simply not viable to follow that hundred-chain length sourcing tree for everything used in an analysis.

Maybe not for the passing reader, but I would imagine it's totally viable for those especially interested. So why not make it available anyways?

[FreshStart]

Retractions should defacto be capable to create chain reactions?

[dado3212]

But then you have to link everything. Also, not all retractions will invalidate dependencies, so each cascading dependency (that will greatly increase in number) will need to check if the retraction affects their claim. Maybe they modify it. Now a dependency on that has to check if the modified claim still warrants invalidation. This is where the “implementable” problem comes in.