Open science: why is it so hard? • Hacker News Board

Re: Open science: why is it so hard?

#41

Post by polyfractal » Wed, Jan 11, 2012, 2:12 PM UTC

mechanical_fish wrote: post #36

polyfractal wrote: post #32

Earlier quoted context omitted.

Actually, and this is very surprising, most multi-billion dollar biopharmaceutical companies do not have access to scientific literature. I have several friends that work in big Pharma (Merck, Pfizer, etc). The common scientist working there has zero access to scientific literature. They routinely read abstracts from Pubmed and then guess about the methods/results based on the abstract. Scary, but true. Management at…

If true - and you'll forgive me if I haven't yet picked my jaw up off the floor - this is the biggest argument for open publishing that I've seen yet. The situation is way worse than I thought.

Yeah, I was aghast when I found out about this. Perhaps it is just the departments that my friends work in, but I've seen similar murmurings on the internet.

I actually found out because I wanted to build a startup that used biomedical literature and biopharmaceutical companies. I basically axed the idea after talking to my friends.

To be fair, the decision to cut R&D in big Pharma is unrelated to the decision to remove literature subscriptions. It is just one of the strings of mismanagement episodes that pharma has performed.

If anyone is interested, this thread at SA is pure gold: http://forums.somethingawful.com/showthread.php?threadid=339...

In particular, these are some quality posts to show the absolute inane nature of big Pharma:

http://forums.somethingawful.com/showthread.php?threadid=339...

Re: Open science: why is it so hard?

#42

Post by Karlsfriend » Wed, Jan 11, 2012, 2:15 PM UTC

roel_v wrote: post #5

It's not hard, it's that most researchers just don't care. Most researchers have access to all papers they need through their universities, and if they don't, most papers are only an email away. Publishing in those journals for which you need to pay (which is only a minority) is paid for by grands or the university. How many people do you know who couldn't get their papers published only because they couldn't afford…

In my experience, ALL the people who care about reading papers, also care about open access.

Because i am the only person i know, who want to read scientific papers. Which is probably because I have never seen an university from the inside.

So i am the guy wo would benefit from this.

Re: Open science: why is it so hard?

#43

Post by polyfractal » Wed, Jan 11, 2012, 2:21 PM UTC

zerostar07 wrote: post #39

polyfractal wrote: post #35

Earlier quoted context omitted.

But see, researchers just don't care . In biology, you want to publish in Nature, Science or Cell. Barring that, you grab one of the second tiers (Nature Neuroscience, etc). Barring that, you take a third tier like Molec Human Genetics or J Neurosci. If that paper is closed access, the average research does not give two shits . He just wants his paper in a quality journal that will gain him prestige and career advanc…

That's the point ; there's a whole system behind (grants, career advancement) that depends on "journal prestige"; that mechanism is based on "prestige factors" that were built decades or centuries ago. It's a self-reinforcing circle, authors-> journals -> author prestige that doesn't even have to do with the quality of review (i believe all well known journals are equally good at reviewing science, what differentiate…

I'm not disagreeing with you at all. I think all your points are accurate (with the exception of all peer-review being good - some journals really do have bad peer-review).

The life sciences would be a better place if everything was open-access.

My point is that there is zero incentive for those "in the power" to change. Scientists just want to keep running on the prestige hamster wheel, journals want to get money, universities pay the money because their scientists need the journals.

To break the cycle you need someone to get off the wheel first. Journals obviously won't. Universities can't (imagine a university cutting all of their scientists off from the data....). Government doesn't really play a role in this, they just hand out grant money. And they do require data to be open-access after a certain period of time if funded through the NIH.

So that leaves scientists. And very few are willing to stake their employment, livelihood and reputation on a "cause". The only people who realistically can do that are tenured faculty - and most of these are old PIs that like the system just the way it is, damnit.

So yeah, I agree with you. But I don't see it changing any time soon. Where is the incentive?

Edit: As an aside, you can lease the Medline abstract database from the NIH for free (http://www.nlm.nih.gov/databases/license/license.html). You can also dowload Biomed Central full text archive too (http://www.biomedcentral.com/about/datamining)

Re: Open science: why is it so hard?

#44

Post by polyfractal » Wed, Jan 11, 2012, 2:30 PM UTC

kitsune_ wrote: post #9

I find it really aggravating that schools such as the ETHZ get billions in taxes a year yet refuse to publish the majority of their papers online.

To be fair to the university, many journals have legal contracts that forbid you from publishing your manuscript online within X days of publishing in their journal. Basically, they get first publication rights.

After a certain period of time most journals let you publish on your own site (although few authors actually do, which is a different issue)

Re: Open science: why is it so hard?

#45

Post by brianprogrammer » Wed, Jan 11, 2012, 3:02 PM UTC

polyfractal wrote: post #43

zerostar07 wrote: post #39

Earlier quoted context omitted.

That's the point ; there's a whole system behind (grants, career advancement) that depends on "journal prestige"; that mechanism is based on "prestige factors" that were built decades or centuries ago. It's a self-reinforcing circle, authors-> journals -> author prestige that doesn't even have to do with the quality of review (i believe all well known journals are equally good at reviewing science, what differentiate…

I'm not disagreeing with you at all. I think all your points are accurate (with the exception of all peer-review being good - some journals really do have bad peer-review). The life sciences would be a better place if everything was open-access. My point is that there is zero incentive for those "in the power" to change. Scientists just want to keep running on the prestige hamster wheel, journals want to get money, u…

This reminds me of the debate in the media about high-frequency trading after the flash crash. The media and general public were pounding the table demanding data and analysis and solutions, and on the inside of the industry, no one cared. Even the SEC did not have the tools or expertise to do a proper analysis.

The demands for change have to come from outside, and they have to be forceful. I would go further than this article and demand that not only should the papers be available to the public, but so should all of the data from which the conclusions were drawn. And furthermore there should be some open repository of all the data, along with the analysis, with api's etc... If taxpayers are paying for research then why not also pay for a way to actually consume that research?

Re: Open science: why is it so hard?

#46

Post by andreadallera » Wed, Jan 11, 2012, 3:07 PM UTC

zerostar07 wrote: post #39

polyfractal wrote: post #35

Earlier quoted context omitted.

But see, researchers just don't care . In biology, you want to publish in Nature, Science or Cell. Barring that, you grab one of the second tiers (Nature Neuroscience, etc). Barring that, you take a third tier like Molec Human Genetics or J Neurosci. If that paper is closed access, the average research does not give two shits . He just wants his paper in a quality journal that will gain him prestige and career advanc…

That's the point ; there's a whole system behind (grants, career advancement) that depends on "journal prestige"; that mechanism is based on "prestige factors" that were built decades or centuries ago. It's a self-reinforcing circle, authors-> journals -> author prestige that doesn't even have to do with the quality of review (i believe all well known journals are equally good at reviewing science, what differentiate…

> Taking a step back you have to ask: Why are these publishing companies given such power

Publishing companies have not that much power. What they publish is decided based on how many references the paper gets and how prominent in the academic society the authors (or the professors backing them) are. They just enjoy their role as "gatekeepers" - basically, they get paid big money to put a barrier at the entrance of the whole system.

Re: Open science: why is it so hard?

#47

Post by mechanical_fish » Wed, Jan 11, 2012, 3:16 PM UTC

roel_v wrote: post #24

route66 wrote: post #17

Earlier quoted context omitted.

> My point is: while ... theoretically nice idea, the practical realities are ... not a real problem for the vast majority of people. ... philosophical objection ... best. Which is not enough to instill enough urgency in those whose support is needed to change things. What you basically say is : "It is the way it is because that is how things are. period" Do you agree with the complaint? Or do you think it's useless…

"What you basically say is : "It is the way it is because that is how things are. period"" No, what I'm saying is "it's irrelevant how we got to where we are, but there just isn't enough demand to drive a change, and therefore I suspect things will remain the way they are for at least a while longer, until the boundary conditions change (more)." My overall point vis-a-vis the article is that it's wrong in its premise…

It's simply the market at work.

This is something of a mantra in US culture, but it's a poor model, because in fact the "market" for research papers may be the least "market"-like thing I know. Papers are funded by grants. There's a handful of agencies that give out significantly-sized grants. The most important of them in the USA - NIH, NSF, DoD - are agencies of the federal government, politically driven and centralized. In the research business as it stands today, the public's "demand" for an esoteric manuscript on, say, the mental health of Drosophila is barely relevant. All that matters is the opinions of your peers, and the opinions of your grant administrator. And these groups don't constitute anything like a "market"; they're more like a family, if perhaps a slightly dysfunctional family. Find a researcher with funding, and you'll find that he or she is almost inevitably well-connected to the research community by a web of patronage and mutual favors.

The peers don't care about open publishing, but apathy works both ways: They won't push for it very hard, but they won't push against it very hard either. What does a researcher really care if the paper comes out in Nature or in PLoS? Sure, there are considerations of culture and fashion - impact factor is not irrelevant to your life as a researcher - but fashion is fickle; old scientists are more adaptable than the average old person and in any case they don't live forever.

Meanwhile, if we gathered the bosses of all the grant administrators in the USA into a room, that room would be the size of an elevator. Convince just a critical mass of agencies - say, the NSF and the NIH - to mandate open publishing and only an act of Congress will prevent the rest of the system from falling in line.

Maybe this isn't right. But, tellingly, the opponents of open publishing have just proposed that act of Congress. A threat to deploy such a blunt instrument doesn't seem like a sign of strength. It seems like a last-ditch attempt to hold back the tide.

(Alas, I agree that the act of Congress could well pass, or even just threaten to pass, and that if it does my whole argument could go into reverse and the status quo will tend to dominate.)

Re: Open science: why is it so hard?

#48

Post by yummyfajitas » Wed, Jan 11, 2012, 3:33 PM UTC

polyfractal wrote: post #41

mechanical_fish wrote: post #36

Earlier quoted context omitted.

If true - and you'll forgive me if I haven't yet picked my jaw up off the floor - this is the biggest argument for open publishing that I've seen yet. The situation is way worse than I thought.

Yeah, I was aghast when I found out about this. Perhaps it is just the departments that my friends work in, but I've seen similar murmurings on the internet. I actually found out because I wanted to build a startup that used biomedical literature and biopharmaceutical companies. I basically axed the idea after talking to my friends. To be fair, the decision to cut R&D in big Pharma is unrelated to the decision to rem…

Is it mismanagement?

The medium term outlook of the pharma industry is not that good. This is true for various reasons - a lot of low hanging fruit has been plucked, regulatory requirements are becoming more burdensome, political risk is high, and older (read: out of patent) drugs are increasingly competing with newer drugs.

Under these circumstances, it just seems like a good idea to cut back on R&D and focus on exploiting existing revenue streams. I've advocated that a lot of other companies in other industries should do this as well (e.g., MS should scrap Bing, pay dividends, and shareholders can spend the dividends on GOOG if they want a long position in search).

Re: Open science: why is it so hard?

#49

Post by roel_v » Wed, Jan 11, 2012, 4:07 PM UTC

Karlsfriend wrote: post #42

roel_v wrote: post #5

It's not hard, it's that most researchers just don't care. Most researchers have access to all papers they need through their universities, and if they don't, most papers are only an email away. Publishing in those journals for which you need to pay (which is only a minority) is paid for by grands or the university. How many people do you know who couldn't get their papers published only because they couldn't afford…

In my experience, ALL the people who care about reading papers, also care about open access. Because i am the only person i know, who want to read scientific papers. Which is probably because I have never seen an university from the inside. So i am the guy wo would benefit from this.

How many have you read over, say, the last year?

Re: Open science: why is it so hard?

#50

Post by woodson » Wed, Jan 11, 2012, 4:15 PM UTC

jamii wrote: post #38

andrewfelix wrote: post #12

Earlier quoted context omitted.

Specifically which subscriptions would be prohibitively expensive to interested parties?

Prices vary wildly. IEEE starts at $5695 per 350 downloads per year [1]. ACM membership + library access is a very reasonable $200 [2]. I can't find an all-access subscription for Springer but individual journals range from 100 EUR to 20000 EUR per year [3]. JSTOR only sells subscriptions to academic institutes and libraries [4] but you can buy 14 days access to a single article for $10+ [5]. [1] http://www.ieee.org/…

Just to clarify, individual IEEE membership is much less expensive. Full membership in the U.S. is $181/year and includes online access to all IEEE articles, dues for other parts of the world are even lower (for whatever reasons) [1]. For students it's $32/year.

[1] https://origin.www.ieee.org/membership_services/membership/j...