The findings of medical research are disseminated too slowly
economist.com
economist.com
Don't give them more money! Journals already have far too much power. They already publish research in a way that serves their own financial needs. They are too slow. They make you do all the writing, editing, formatting and figure design for free, then they charge people to read your publicly funded research. They make you put your data in an excel spreadsheet or a PDF. They fail miserably at peer review. At my institution, editors of nature journals are going around and giving seminars on how to publish in their journal - journal editors giving SEMINARS to ACTUAL SCIENTISTS. Absurdity.
All of this will sink journals. They just don't know it yet. Their strategy seems to be to become more like news outlets, publishing opinion pieces on how to fix scientific publishing, yet never actually changing what they do. The only reason it survives is because more than anybody, scientists learn to put up with all sorts of shit.
Shouldn't research scientists be the ones who write? I mean, who else knows the intimate details about your experiments that need to be communicated to posterity? And, do they really make you do the formatting? My understanding is you basically just give them an easily readable pdf version of your paper... in standard double spaced 12 font arial, and when it gets accepted they are the ones who choose the pretty fonts (probably with Indesign) and revise your figures so they don't look crappy.
That said, I'll say one nice thing about the system (partly in the spirit of contrarianism, as hardly anyone has good things to say about top journals:)). If you really want to be published in a top journal, there are two routes: 1) make something really ground-breaking... and then you can even submit a crappily-formatted paper with ugly figures and you'll still get accepted -- and they'll make good figures for you! 2) have something decently good... make good figures, make sure you follow formatting guidelines to a tee and make everything as easy as possible for the editors, and then hope for the best. I think this situation is not too bad.
I've realised that scientists will put up with almost anything. You have to be that way, otherwise you won't last. But when Elsevier makes billions of dollars in profit, and the person that reviews the papers for them lives in a box subsisting on instant noodles with no job security, you have to ask some questions. We don't actually need Elsevier. If they disappeared tomorrow science would march on largely unperturbed. Elsevier needs the scientists, and bless them, they have found a way to put us to work for free and sell us the product at the same time.
Given that the subject deals with trying to "fix" complex, organic systems, it's worth taking time to get things right before designing interventions. It is important to realize that a flood of rapid communication often makes it harder to settle on the underlying truth (as clearly evidenced by the recent atmosphere) by rendering the search process unstable.
As some other articles (on the HN front page) have recently argued, the rush to faster publication is one of the biggest causes of research debt. And I think that is a much more important problem to solve. So, by all means let's have open-access and also more distilled dissemination. But optimizing for speed (eg: race to publish) often gets in the way of correctness.
It's extremely hard to move the needle on that particular issue.
Alternative is not just a faster publishing, but more feedback, and better discussion around the subject. You could call it a "Wikipedia" approach.
In the "Wikipedia" approach, you don't rely on the publisher's prestige, nor citations, as a meter of trustworthiness. Before accepting the results you review it's discussions and critique.
A similar switch happened with traditional news publishing. Years ago you had to rely on the fact that an article was published in Economist, and vetted by their good editors. Nowadays, there is a lot of independent content, and there are communities (like HN, or Reddit's AskScience), and non-profits (fact checkers, or Snopes) that point out the potential issues.
For refs, Google Paul Glasziou's research in this area.
Having survived three and a half centuries, scientific journals will no doubt be around for a long time yet. With luck, though, they will return to being science’s servants, rather than its ringmasters.
I have myself published a few papers with PhD students that should have been OA but we used the PhD students' allocated funds to send that student to a relevant conference, that's a better prospect for their career.
Would love to see rate of adoption of new standards of evidence based medical care as a function of institutional access to top journals or simply Uptodate (the tool people actually use nowadays). Access to curated guidelines may be more clinically important than primary lit.
One reason prestigious journals are great is you get the imprimatur of success for further funding/promotion without needing to wait for citations and scientific consequence. Perhaps an opensource analytics team will over time convert the community to other metrics like reads, shares, and other h-factoresque metrics (granted all are contrived and gamable to some degree), perhaps tied to some funding consequences. Cough, ResearchGate, cough.
Regarding thread title: The small number of good medical findings are disseminated far too slowly; the large number of irreproducible, contradictory, poorly supported, and statistically misinterpreted findings are disseminated far too quickly. Couldn't resist.