I want to point out that there are costs associated with high-quality publishing that I don't think have been mentioned thus far. Peer reviewers do their vital work for free, but someone has to administer the process -- track the submissions to a journal, get the articles out to reviewers, remind reviewers to submit reviews, correspond with authors through revision/acceptance/rejection. The scientist (usually "editor-in-chief" of "chief editor" of the journal) who does this on behalf of a for-profit publisher is normally paid; in other cases, the scientist-editor must hire an assistant to help, and someone must pay the assistant. If the journal is pub'ed by, say, a non-profit academic society, the society staff may do this work, and that staff need to be paid.

Also: copy editing. It has value, when done right. Styling (making all headings look alike, making elements like equations look right and alike, etc.): Reasonable people may disagree on the value of this; I think it helps. "Production": If you have copy editing and styling, you must have checks and back-and-forth between author and proofer. Printing: Until all journals go to all-electronic, someone's gotta pay to produce the print copies. Mailing: As long as there is print. . . Copyright enforcement. Etc.

Now, these services need not be done by for-profit publishing companies. Academic societies do some of it and tend to charge much less for their journals -- which are often of equally high quality or even much better quality than those produced by the for-profits -- than the for-profits. I just hope that the employees who do this work are paid decently. (Non-profits are known for their long hours and low pay.)

Any useful discussion of the future of the academic journal must account for these realities. If/when publishers are out of the mix, some other sorts of entities will need to step up.

It is very important to distinguish between society and commercial publishers. Society publishers, by and large, publish both the less expensive and the more cited journals.

Commercial publishers are at the root of the problem, having largely destroyed the author-page-charge/low-subscription publishing model, prevelant until recently, by introducing their no-page-charge/high-subscription publishing model.

NIH funding doubled over a short period of time (1999-2003) and I looked for the corresponding increase in scientific productivity. The simplest metric is the number of scientific publications. A natural control is the number of publications in other scientific fields where the budget did not double. The total number of publications cited in PUBMED shows a 7% annual increase between 1999 and 2005 (Table 1). The ISI database for in biology yielded an increase of 13%. In contrast, chemistry, without a budgeti doubling, increased 7%. Key words for different areas such as “DNA” and “Neuron” produced similar results. The annual increases reflect primarily increased coverage of journals by the databases.

PUBMED "Biology", ISI "DNA", ISI "Neuron"

7%, 13%, 0%,

ISI "Physics", ISI "Chemistry", ISI "Arts and Humanities"

3%, 7%, -2%,

Table 1. Increase in the number of publications/year for different databases and search terms before (1996-1999) and after the budget doubling (1999-2005).

As another control, I compared productivity in US biology labs to that of other counties(Table 2). Doubling the NIH budget had little effect. Only China showed a significant increase in scientific productivity over that period.

ISI: kw=neuron, USA, Japan, Germany, China, UK

1996-1999 2.9%, 7.6%, 12.3%, 42.7%, 1.2%,

1999-2005 3.6% 1.2% -1.4% 38.9% 1.4%,

ISI: kw=DNA

1996-1999, 0.00%, -0.01%, 0.01%, 0.22%, 0.01%,

1999-2005, 1.55%, 0.26%, 5.43%, 39.49%, 0.22%

Table 2. The annual increase in publications from labs in five different countries for the indicated key words before (1996-1999) and after the budget doubling (1999-2005).

Perhaps the articles are now twice as good as they were in 1999. but an anecdotal sampling of scientists suggests otherwise.

The lack of correlation between NIH funding and scientific productivity is surprising. To lobby Congress to provide more money for the NIH, we need evidence that more money means more productivity. I encourage the NIH to supply this data.

I also reported this issue on my blog in Chinese.

But I care more about when will have Chinese OA platform.

I learned a lot

Thank you all

The issue is not peer review and the issue is not profit. One can be confident that the publishers will succeed or fail like any other company when confronted by the (relatively) free market. If the scientific, medical and technical communities perceive a benefit to a certain journal, then it will prosper. If they do not, no Orwellian PR firm or corporate welfare bonanza from vest pocket Senators will save them.

Peer review is just one example of building and conveying confidence in the fruits of research. There are other trust models, and more are evolving every day on the web. Amazon has user ratings. Ebay places a direct dollar value on the most obscure items. The most fundamental trust model pertaining to journals themselves is not peer review, but rather the citations from other works in a community's literature. Trust ultimately derives from the entire community, not from individual reviewers - these are the true peers. Peer review per se serves simply as the gatekeeper to ensure a minimal level of compliance with the norms of a particular field. Truly revolutionary research often overturns those norms, of course, and may have a hard time getting published.

The journals have been wrestling with the realities of electronic publication for ten or fifteen years now. Ultimately, the existence of ubiquitous computing, of the web and of cheap digital media will dominate future public policy questions regarding publishing. The final disposition of this issue will lie somewhere between completely open access and completely closed access. Few publishers will find that they have the clout to use the denial of access to their product as leverage to demand payment. Rather, the contravening demand inherent in most types of technical and scientific communities for open access will cause traditional publishers to find ways to provide at least semi-open access to their own publications (and not just to abstracts).

One suspects that the lobbyists and PR firms will be long gone before the journals reach a new accommodation with the communities they serve (not to mention with the public interest). It is the journals who are squandering their budgets rather than wrestling with the inherent facts of the situation who are suffering here.