utterly ridiculous spat: in tabloid headline style (“neuroscientist accused of going on date !!”), nature reports on a dispute over the interpretation of data between nikos logothetis at the MPI in tubingen and two of his former postdocs (amir shmuel and david leopold) now running independent labs at u.toronto and the NIMH. the article is written by alison abbott and entitled “Neuroscientist: my data published without authorization are ‘misleading’ “. just to provide context, alison previously authored a hagiographic article on logothetis a few years back.. my own take ? it is easy to argue that it is not logothetis’ data. while the exact contributions are not clear, it is likely that the data were collected in a lab in which he is the research director, by the two former postdocs who wrote the paper. and the situation is simple: one data set, two interpretations. both of which could have appeared in print. it seems that this spat has mostly blown up because Human Brain Mapping declined to publish logothetis’ alternative interpretation side-by-side with that of the authors’ own interpretation. whether that was because HBM considered logothetis’ objections unworthy of publication, is not clear… but the simplest option would have been for logothetis to write a review or opinion piece in some other journal discussing this issue. but commercial journals care for the bottom line and what is better than some tabloid drama to push copies ?

nature and plos: Declan Butler’s most recent salvo in his long-standing debate with open-access advocates has raised a real ruckus in the blogosphere; good places to start are on Bora’s blog and Open Access News. Also see Butler 2003 and 2006 for earlier articles, with resultant discussions here and here. The drama led to a clarificatory statement from Timo Hannay, that was much better argued, and Timo actually has some interesting things to say about open-access in this video over at open access news.

Many critics are complaining about either the appropriateness of Nature criticizing a competing journal (without explicitly discussing conflict of interest) or for criticizing open-access in general. I think it is entirely appropriate for Nature to write well-argued, well-reasoned articles on science publishing, even discussing competing models critically, but the Butler article under question does not pass these criteria IMO.

Declan makes the same points with new numbers: that PLOS has not yet broken even, they are surviving on philanthropy, PLOS’ operating costs are more than they ever expected or projected, i.e. they are now learning what it takes to publish a selective journal and we told you so in 2003 and in 2006. Declan and Timo are arguing against the PLOS claim that an open-access journal that only publishes a few articles per year can be financially viable on the basis of author-fees alone. Whether or not this is true is an open question, but one that is probably now purely academic because open-access is fast becoming mandatory for federally-funded research (e.g. NIH-funded research in the US and Wellcome-trust funded research in the UK). Besides, it is not easy to pin down what the exact PLOS business model is, since it is constantly evolving, and there  are widely-varying opinions even within the PLOS universe and certainly if one includes all open-access advocates (and supporters of PLOS). It is also not clear whether Nature Publishing Group’s own ever-expanding stable of ho-hum journals subsidize Nature’s more high-profile journals, just as PLOS One subsidizes the other PLOS journals; no discussion of NPG’s own financials seems likely to emerge anytime soon from NPG (but see Lars Juhljensen’s blog for a preliminary analysis of  NPG’s own “haute couture” publishing model (in Declan’s words, since I am not not entirely sure how that term works)).

But the authors from Nature then add one new twist, and Declan in particular uses some pretty strong language about PLOS One to make his case. Their claim is that PLOS has effectively dumped their model and are now using PLOS One’s revenue stream to stay afloat. And PLOS One, according to Declan is based on “bulk publishing of low-quality papers” that emerge from PLOS One’s “light-review” process. Whether intended or not, the aspersion being cast by the Nature writers (Declan much more so than Timo) is that PLOS One will publish pretty much anything, since publishing papers adds to its revenue stream and helps subsidize PLOS Biology and other selective journals in the PLOS stable.  These are serious accusations that are not backed up by any analysis or reporting on the quality of papers in PLOS One. Exactly what constitutes “light review” in Declan’s mind is hard to tell, but the main difference between the review process at PLOS One and other journals is that there is no subjective estimate of the article’s potential impact (novelty does matter, from the published guidelines). Perhaps there may have been some initial ambiguity about where the threshold for publication would be set in PLOS One, but last year’s record of papers appearing in PLOS One, as well as anecdotal evidence from those involved with PLOS One as reviewers or authors (example, here) suggests a perfectly rigorous and functioning peer-review system. And how exactly does “potential impact” assessed by peer-reviewers (or much more often, by Nature’s stable of fresh out of science professional editors, since the majority of submissions are not even sent out for peer-review) decide the quality of a paper ? Bjorn Bembs has made some excellent points on this topic.

I agree with the spirit of what Bjorn has written (as well as many others). PLOS One or another archive of material peer-reviewed solely for methodological soundness is the future. For those who wish to get an external assessment of a paper’s impact in a field, appended commentary, ranking by peers, no. citations, H-index can all serve as useful proxies in case one is unable to get expert opinion on the paper. Importantly, the current method of judging the impact of a paper by the impact factor of the journal it appears in, is a particularly bad one, as Chris Surridge at PLOS has pointed out. In addition, journals like Nature can then exist as news magazines commenting on the scientific output deposited there; the excellent science communication skills of Nature’s editorial staff will come in very handy when they discuss the peer-reviewed research published in PLOS One and enable non-experts to further understand and evaluate papers published in PLOS One.

This brings up another issue: Why have so many PLOS journals, and not just PLOS One, as Biocurious has asked ? It is unclear from the official statements (example here) how PLOS intends these different journals to evolve; perhaps it is too early and the future is too uncertain for PLOS to make any definite statements about how this will all look like in 2015. The only rationale I can understand is that PLOS Biology and others are there for pragmatic reasons and eventually PLOS One will be the future, if all works well. There is a hint that this is indeed the case from reading between the lines of the FAQ on PLOS’ web page. I hope it is true !

ps1. On an unrelated note, I am sure NPG believes that it is the ahem, high quality of its publication process and editorial staff that enabled its new journals (and they are multiplying like rabbits) titled “Nature xxx” to rise to the top of the impact factor pile, and not the unfortunate susceptibility of scientists to branding. Pedro Beltrao makes the same point here. This, i hope and pray, will change, as a result of such discussions and PLOS One should certainly accelerate this process.

ps2. Mike Dunford has a nice piece on this entire fracas.

ps3. I seem to remember that Nature has taken an editorial position somewhere that treating publication in nature (by extension, the decisions of Nature editors whether or not to publish) as the holy grail for scientific career decisions represents a failure of decision-making by the scientific community, and that it is not Nature’s intent to be used in this manner. I agree with them entirely on this issue.

more on nature’s own publishing model: So Nature has been claiming all along that it is not possible to successfully running an open-access, yet selective journal that only publishes a few articles. Well, now that the era of funder-mandated archiving has begun, what is Nature going to do ? Is it going to change its business model, or shut shop ? Or are we suddenly going to see how it was indeed possible, all along, to run a “selective” journal that allows free access to its scientific content ?

In my mind, one of the major flaws in declan butler’s serial analyses of PLOS’ finances is that he does not compare them to those of other groups, including NPG.  If declan is reporting on this dispassionately and wants to be taken seriously by people interested in well-substantiated and comprehensive arguments, then perhaps he should skip the gonzo journalism and attempt a more serious analysis. There is precious little data available (at least, to amateurs like me) on NPG’s business model, but Greg Laden has just provided a link to some interesting numbers from a 2004 presentation from NPG to the UK House of Commons (friendfeed discussion here).

The written submission to the House, while very sparse, is sufficiently interesting to merit some consideration. I will stick to the numbers alone; we each have our own views on Nature as a journal, so I will not quibble with the detailed descriptions of their “superior product”. Also, since I am an amateur at reading business statements, if my analysis has errors, I am eager to be corrected ! In what follows, indented text in italics is from Nature’s written submission to the House of Commons.

1. You stated that, under a pay-to-publish system, Nature would have to charge authors between £10K and £30k per article. Can you supply a breakdown of the costs that would necessitate this charge? (Q16)

The £30,000 figure was arrived at simply by dividing the annual income of Nature (£30 million) by the number of research papers published (1,000). This assumes that author fees would be the only source of revenue (ie. a totally open access model). The £10,000 figure is derived from estimates of the costs of selecting, reviewing, editing, designing and producing the research article element of Nature and the amount NPG would need to charge in order to cover these expenses adequately. For this calculation, it has been assumed that the very significant and extensive costs of producing the `front half (news, comments, reviews etc) would continue to be predominantly covered by a combination of subscription and advertising revenues. <…> The profit margin before interest and tax in 1984 was 15% and will be 19% this year. However, NPG’s profit margin can vary considerably by year and has fallen to 7% in some years between 1984 and 2002, as a result of market conditions and investment decisions

I think we can ignore the 30,000 pound number; the scientific community does not have to subsidize all of NPG’s profits. Exactly how much profit is necessary in order to keep improving the services provided by the journal is a complicated question (see discussion of margin below), but that requires more data and analysis. The 10,000 pound number is interesting because  if we take Nature’s word for it, then this is their estimate of how much it costs to produce a zero-profit high-end journal that publishes about 1000 articles per year; of course, this means that PLOS’ is substantially under-estimating its expenses by only charging authors 2500 dollars per submission, assuming a similar no. publications per year. And this is the crux of the long-standing debate between NPG and PLOS; is PLOS naive, or is Nature inefficient ? If Nature is indeed working at a high level of economic efficiency, then how are they going to function now with federally-mandated open-access in the UK and the US ?

The margin figures are critical, but hard to interpret. Perhaps Nature’s analysts should take a Stats 101 class on how to report a distribution ? All we get is the minimum, and the most recent value (rather than say, the median or mean)… Is that the profit margin for Nature alone, or for the entire NPG publication stable that definitely appears to be following the “haute couture” model of publishing that Declan accuses PLOS of ? If it is the profit margin for Nature alone, then it appears that since the income is 30 million pounds per year, the costs for Nature amount to about 26.3 million pounds per year (assuming a 14 % profit margin). Of this amount, 10 million is the cost of publishing the research articles, so that means it costs 16.3 million pounds per year to produce the front-half of the journal. Is there anyone with experience in publishing who knows if these numbers make sense ?

Further, there is an important assumption in the numbers Nature gives above: Nature seems to be assuming that the advertising and subscription revenues (which, after going open-access, may be higher, lower or the same as they are now) will just about cover the cost of producing the front-half of their journal. Since the current income (from advertising and subscription) is 30,000 pounds, this seems to suggest that Nature assumes that the revenue from advertising and subscription (to the front-half) will drop substantially to a level that just about covers the production costs for the front-half. And if the production costs for the front-half are about 16.3 million pounds, then that is a 50 % drop in subscription and advertising revenues. How do they justify this ? Also, if this assumption is true, then it seems that Nature’s front-half material will just about allow it to break even, and it is the research articles that allow it to make a nice profit. Therefore, I think one could argue that much of the “value added by Nature” to the research it publishes (something they like to brag about) does not generate much profit, and is therefore not of much value (since in their capitalist model, distinctions between inherent value and economic value are tenuous). It is the free material supplied by the scientists and reviewed by scientists that the readership considers valuable. Nature could argue that it is their selection process, which allows readers to focus on a filtered subset of articles, that gives the research material Nature publishes such value. If this is the case, then Nature should be able to derive the same benefit from simply listing and discussing interesting papers that are published elsewhere; this, I contend, is a great role for Nature in an ideal world for scientists. In other words, Nature could function as a high-end ScienceDaily.

I agree with Bjorn Brembs: submitting articles in a serial fashion to journals until they are accepted is highly inefficient, wasteful and an anachronism. It leads to repeated reviews, delays the publication of science, and wastes valuable resources. (Btw, David Zetland, an economist formerly at Davis, has yet another interesting alternative to this publication merry-go-round: let journals bid for the right to publish data placed in an E-bay like repository).

Site licences for Nature-branded journals are generally negotiated directly between NPG and the institution, without the involvement of subscription agents. The price charged per title depends, firstly, on the number of full-time employees and, secondly, on discounts. Increasingly large discounts are applied for each additional title taken.

So yes, the bulk published NPG titles do subsidize Nature to at least some extent.

The mix of subscription and advertising revenues has provided stability by spreading the costs of publication across advertisers and Nature’s hundreds of thousands of readers, rather than the much smaller number of published authors.

From a cursory glance, it appears that this might be true if there was keen competition for papers and Nature was operating with a very small profit margin, with prices dropping further as the number of subscribers increased. But in the absence of meaningful  competition, having a lot of readers merely improves Nature’s bottom line, since there is no pressure to reduce prices. More serious analysis would be extremely useful here …

ps. There is much interesting material on this whole issue at Greg Laden’s blog (e.g. here and here). I don’t share greg’s fervent anti-capitalism and am less emotional about these issues, but that is merely incidental.