Hacker News Reader

Comment by dfxm12 | Hacker News Reader

According to the article, bad papers are getting criticized and retracted. It appears "science" is robust enough to work as intended, even with some bad actors involved.

View on HN

There is evidence to suggest that corrections and retractions do not even effect citations.

What practical solution would you offer to this problem? Without changing the entire publishing software ecosystem, it’s not like articles are living in github repositories where pull requests are a thing.

In many cases, citing articles still go to print media.

All of them? That worked well for Alzheimer research, didn’t it

Obviously not all of them. And obviously there is corruption and mistakes with anything involving humans. What's so funny is that when people make criticisms like this, they always leave out the alternative. What's the alternative? Trusting mostly uneducated influencers and quacks who do even less research and don't even attempt peer review?

The alternative is not trusting the science or the quacks.

Most people cannot tell those apart...

And those who are not supposed to be quacks will commit fraud anyway. Which is why you shouldn’t trust anyone.

Good idea, trust no one, get nothing.

Why not move back into a cave and stab your own meat with a sharpened stick next.

At least we wouldn't have Alzheimer

I'm sure if, god forbid, you were diagnosed with a curable cancer you wouldn't go "trust the science" and go get treatment. I'm really sure you would do that.

Except it did? Fraud was identified. Science moved forward. Literally working as intended.

You can say that of all cases of fraud, but that’s a survivorship bias fallacy, since we can only talk about cases we know to be fradulent, i.e. where fraud has been identified.

Reply on news.ycombinator.com