An author writes something, and then someone else reads it. There are some costs involved — who should pay them? The writer or the reader? Obviously, the one who gains from the “transaction”: if we are talking about useful or interesting information, then the reader; if we are talking about some form of advertising, then the writer.
The same principle should be true in academic publishing as well. It used to be that the readers paid to read academic journals. Publisher abuses of the system, together with the new possibilities opened by the Internet, caused academics to talk about open access journals that do not charge readers. The main associated costs are those of producing the research and these are anyway paid by government grants as well as professor salaries. Unfortunately, it seems that the tide is turning towards “open access” journals like PloS ONE in which the authors pay for their results to be published.
By definition, papers published in such journals are equivalent to advertising: you pay for others to notice you. The dynamics of such journals cannot maintain quality: they have strong incentives to increase quantity and have weak (or no) incentives to increase quality. Indeed, PLoS ONE does not hide it:
PLoS ONE will rigorously peer-review your submissions and publish all papers that are judged to be technically sound. Judgments about the importance of any particular paper are then made after publication by the readership (who are the most qualified to determine what is of interest to them).
From my point of view, this reduction in the role of referees makes the journal superflous: information that is unfiltered for quality, becomes useless since it is impossible to find the good stuff there — you might as well just put your paper on the arXiv (see also my post on the attention economy). I personally also don’t quite trust the referees for correctness, not in other journals and certainly not here — correctness is established after the “community” has chewed on the paper for a while (which will only happen if it is interesting). I find it hard to believe that any favorable reputation will be attached in the long run to publication in such places — they’ll mostly be a sink for readerless publications.
Let me just be clear: in no way am I defending the expensive “closed access” journals. It is not so difficult to just put your work openly on the web. If you also need a journal publication for your own promotion/grants, then do what you need to. But don’t forget: the “science” part is the making it available; the “journal” part is a bureaucratic hurdle.
John Ewing, former Executive Director of the American Mathematical Society, has written in detail of his fears about what the author-pay model holds for mathematics in particular and scholarship in general. I hope his predictions are wrong!
http://www.ams.org/notices/200803/tx080300381p.pdf
Click to access JournalsHeaded-v6.pdf
It seems that write-only journals could become even more common. Some people seem to think that we should make everything available on the arXiv, say, and use something like PageRank to determine quality. It seems unlikely to work, to me, but I would like to see some theoretical argument as to why it isn’t a good idea.
“Some people seem to think that we should make everything available on the arXiv, say, and use something like PageRank to determine quality.”
This is basically my point of view too, but we need a much more diverse and sophisticated method than PageRank alone to determine quality. In my view, journals, conferences, prizes, blogs, various ranking systems, various types of citation counts, expert opinion, etc. are all part of this more sophisticated system.
These three poles, closed access-reader pays, open access-author pays, internet-no refereeing,
are not the only options.
There is also the “charity” version. “Theory of Computing” as well as longer-standing journals such as the “Journal of AI Research” are examples of this latter form. As you mention there are some real costs involved in producing and maintaining such journals. In the case of ToC the server side is a voluntary contribution of some universities and the editorial process, including copy-editing* as needed, is entirely voluntary for the editorial board itself. JAIR is an official non-profit whose expenses are covered by corporate and institutional donations. (I believe that ToC should also follow the charity model more formally as JAIR does so that any needed copy-editing can be paid for from cash donations.)
I don’t see the same incentive problems with the charity model as with the other three models. In fact, the charity model seems to align well with other aspects of support for research. There is no incentive for a journal to pad its pages (to get larger user fees, either from reader or author). There is an incentive for a well-chosen board to try to maintain quality. The only issue is whether the charity can attract enough capital (human or otherwise) to run the journal.
* by copy-editing I include major corrections of serious formatting and grammatical problems, not the typical matter of making sure that the punctuation in some reference list is correct.
Stimulating post. Some caveats:
(1) Most open access journals are not author pays. Stuart Shieber estimated that 70% of open access journals use some other model to raise revenue. You’re talking about “Author pays open access”.
(2) Some author pays open access journals are not-for-profit. This is closely related to Paul Beame’s point: the incentives at a not-for-profit are not necessarily the same as you describe in your post.
(3) I disagree with your criticism of PLoS ONE. Unlike most other journals they’re putting a lot of effort into making post-publication filtering work. This is partially through efforts like their article-level metrics, which should make it easier to find high impact work (and penalizes low impact work). But it’s also simply by being open access. I see PLoS ONE papers much more often in Google searches than I do articles from most other journals. If Google is doing their job well, the articles I’m seeing are presumably the better articles from PLoS ONE. These effects create an incentive for authors to submit their best work to PLoS ONE. It’s possible that this incentive is weaker than the countervailing incentives you describe, but I don’t think it’s yet clear how this will play out.
With all that said, I agree with your main point: this is potentially a failure mode for author pays open access, and something that needs to be watched carefully.
That Shieber post is here:
http://blogs.law.harvard.edu/pamphlet/2009/05/29/what-percentage-of-open-access-journals-charge-publication-fees/
Although he’s done an impressive job, I haven’t thought carefully about how accurate his 70% figure is. He uses a screenscraper, which is not always a reliable method of getting data. Still, that work and related work makes it seem highly likely that many and probably most open access journals are not author pays.
I’m afraid that my title was a bit exaggerated: my concern is with “author-pay” OA, not with OA in general. I support OA journals that don’t require author payments, especially overlay ones which have very low associated costs.
I feel that these costs are so low now (given that the authors, the referees, and the editors, all work for free), that FREE access is the logical choice (with various possible ways of generating the little revenue needed.)
I think freemium model would be useful to consider too.
[…] “Open Access” Journals are Advertising « Algorithmic Game Theory […]
You suggest PLoS ONE provides information ‘that is unfiltered for quality’. That’s not the case. PLoS ONE does filter for quality— through the reviewing process. However we do not make the ‘importance’ judgment call (and hope that post-publication comments and metrics will increasingly provide that). I am spending a bunch of time this weekend reviewing manuscripts for PLoS ONE and considering other reviewers’ comments regarding their quality so it frustrates me that people think PLoS ONE does not screen for quality!
The distinction I made is the same that PLoS did: correctness vs. significance. It is the latter that I call quality.
I went through and tried to understand PLoS. It is not as bad as this blog makes it sound. The primary down side, which is actually an upside in my own opinion, I see is that it decreases the importance of leading authors in the community. So Noam is a leading algorithmic game theorists does not have much importance. This would be the case in any system which chooses to publish a lot of papers (i.e. supply of publishing slots increasing), since as the supply increases, the lower quality papers would start getting accepcted. Therefore the importance of leaders goes down.
In fact, if we go by the game theory Noam is trying to say, our conferences are worst. They also charge, so called publishing charges. I would love to see if the next Stoc/focs says, in case a speaker does not have money to travel, the conference will cover it and the fee will be charged by the audience, who should be getting value out of a talk. We know this is senseless.
But PLoS is clearly saying that they will waive the fee for those who can’t afford it. They will also give discount to the participating institution.
The rest of the blog post is simply misleading. Because the quality of an articles published is defined by the market equilibrium. Since Noam works for Google. Suppose Google changes its advertising business models as follows. Each advertisement is fixed cost. But if you search camera, only one camera ad will be shown? In another case, 100 camera ads will be shown?
Is not the quality of the ad in this case would be the function of the supply of ad-slots and the demand of those ad-slots? Sure the authors being asked to pay decreasing the demand, so for the same number of publishing slots, the quality will be lower vs if the publishing is free (and everything else remains the same).
But everything is not remain the same. The publication is free for the readers. Therefore, if everything else remain the same, except the cost of reading the material, then PLoS is increasing the value of each publication.
Of course, PLoS is new and seek to broaden the horizon of sciences, and that could be the reason of the lower quality.
But I would be willing to argue against the rest of the post. In the following two models, it is not clear to me which one is better. Noam seem to be saying the first one, and I think he is mostly wrong.
1. N articles per month are published. Authors do not pay. Readers pay.
2. N articles per month are published. Authors pay, and readers do not pay.
Everything else is constant. The editorial board is the same in both the cases. N is the same. The cost of publishing is the same etc. I think, which one is better whether 1 and 2 depends upon the other variables, i.e., how the value for the readers and the authors are distributed.
The point is really in the comparison that you make:
1. N articles per month are published. Authors do not pay. Readers pay.
2. N articles per month are published. Authors pay, and readers do not pay.
The point is that you will NOT get the same N papers in both cases: in the first case you will tend to get those that readers want; while in the second case, those that authors want. I would say that the first is much better.
You have to say. I am assuming in both cases, N papers are chosen by the referees.
In fact, one could argue that, based on the information theory, that if the authors pay chosen N would be much better.
In fact, some of the N that readers want to read but authors do not even want to write, readers won’t ever see those.
First thing is that you change your position depending upon the context. I would love you to prove your point that what you write here you mean it. Start a workshop culture, not directly sponsored by the public, but paid by the listeners which is free for the speakers (including the travel). So the listeners pay for the travel of themselves and the speakers (indirect subsidy of the workshop by the government is fine, in the sense that listeners could use their grants). In some sense you could make the same argument.
Writing a paper is just one part of research, doing the research is another and in fact more expensive part. You could ask the same thing, researchers have to pursue N projects. Better these should be free, and the readers pay for benefeting for the same N projects. In that case researchers would choose the better N projects. I have yet to see such a world. That definitely won’t have a tenure system, because the readers would want to have a control.
Which one is better between 1 and 2, also depends upon the information theory structure, that is who has the most information to make the match between a reader and an article.
Imagine if you 1 Turing award to bestow, then whose opinion would be matter the most – the top leaders in the community. That also means that the most deserving persons gets it.
Imagine if you have 100 Turing awards to bestow, then the opinion of everybody counts, and therefore the quality of the winners also go down too.
But that is fine, if there is a new platform which lets the community eventually figure out who is the real one among the 100. I think that’s what PLoS is thinking. Papers as long as they looks reasonable, and let the readership consider which are the most deserving ones.
My ideal system be that, both publishing and reading should be free. A paper should be published as long as it is reasonable. Let the community decides the ranking after that. Web Search works reasonably well demonstrate that community’s signals could be used to figure out the gems from the stones.
“My ideal system be that, both publishing and reading should be free. A paper should be published as long as it is reasonable. Let the community decides the ranking after that.”
This is my ideal system too, but it exists, and we just have to use it: arXiv/ECCC + Google Scholar/other-measures. Why do you need journals in this scenario then if not to be part of the way that “the community decides the ranking”? Also, note that if journals publish according to this scenario then publication in them provides no useful signal what so ever for academic promotion/grants — just like arXiv (or TRs) today.
Frankly speaking, I do not think, we need journals in this world. They exist for historical dynamics. A system like you have, is much better. That is you recommend the papers on your blog you liked on arXiv. But this syste, should gain volume, acceptability, and accountability. The accountability is the part which is tricky. With accountability, you would get fairness etc.
I sit on number of PCs, in case of subjective issues where my perspective is different than that of authors, I can’t justify myself, that I have a better judgment than the authors on the topic. So I usually face this dilemma, whether to recommend rejection on my own perspective, or recommend acceptance on authors perspective. I usually give a lot of weight to the authors and oppose other reviewers who give full wait to their own perspective, and in the real sense of the word “judge”. This real sense of the word “judge” is something of a concern.
[…] 5, 2009 · Leave a Comment Noam Nisan has a new post equating open access journals to advertising. While Nisan says he does not support […]
I second Noam: Open Access” Journals are indeed for Advertising. When I hear “open” I imagine something like ToC or El. J. of Combinatorics. They are open (=free accessible) and they are journals. I think, we should not take all this “open journals” (=busines as usual) stuff very seriously. Journals are for archiving the findings for next generations. Referees try to make this “heritage” a bit more trustworthy. (Spending a lot of their time.) Now, there comes “open access guy” and says: you pay, we put your paper in internet, and people will vote … In this case I also prefer ArXiv, ECCC or my homepage, as Noam said.
I am for “charity” version, Paul mentioned.
I hope very much that ToC will make an example.
The biggest problem is not journals, but rather their need for promotion/grants. Would a dean ask several experts in the field “is a candidate strong in your field”, instead of counting publications, the world would be in order. But I’d bet, this will never happen … In a former USSR countries the main criterion is now “publishing in ISI journals”, not what you have really done. Will some “open access” journal come in ISI list, it will be automatically become richer, just because of these “ISI pints counting” reason …