Hacker News Reader

Comment by gjsman-1000 | Hacker News Reader

Does it matter? Many scientists in the know say that peer review is as good as a trashy spam filter these days. This isn't to say that it is absolutely useless, more that it is very overrated.

View on HN

I like to think of "peer review" in the same way as "peer code review". Just because someone looked at it and approved it doesn't mean that it is free of shortcomings and defects.

Not just "peer code review" from a good friend or someone working on the same project as you - a review from somebody who has a ton of stuff to review at once, in many different areas, and very limited time to do it.

Still better than something someone didn’t even bother trying to peer review, usually.

Not always. As academia becomes increasingly autocratic and contaminated by power, peer reviews become a marker for group-think, and placing relevance into peer review is a mechanism to empower the oligarchy and suppress technical dissidence, therefore anti-scientific.

I do peer review, and know colleagues that do as well. I'd say the median number of articles reviewed per person in a given year is 10. I don't claim to represent most reviewers, but I also don't think a presumption that reviewers primarily only have time to half-ass it is fair either.

> I'd say the median number of articles reviewed per person in a given year is 10.

That's a pretty light annual workload. How do you get to be one of those people?

Or do those people have other work to do?

They are other researchers in the field. Literally, your peers.

Yes, they have other work to do.

Yes, I'm aware of those things.

I'm pointing out that the number of papers someone reviews in a year is not even distantly related to the amount of attention they have available to devote to the job.

That hasn't been my experience in academia. In peer review you can get very high quality feedback and critics. The system has been great for the overall quality.

Someone is wrong on the internet!

With the volume of research being churned out of "publish or perish" institutions, I expect a peer review to inform me on the "Is this complete garbage" metric rather than give some assured level of quality / correctness. Ie: if its peer reviewed, it /probably/ isn't /complete/ garbage.

I think the other direction is more useful, if it isn't peer reviewed (and is written in this format) it almost certainly is complete garbage. Otherwise it would be peer reviewed.

Yes, it matters very much *if* the paper is in a reputable publication.

Reply on news.ycombinator.com