How not to write a review of a journal paper

My excellent PhD student submitted a paper to a special issue of a journal dedicated to network science. One of the reviewers wrote a report that included the following choice observations. We have since been told that the writer is a very senior figure in the field, and the editor apologized for the tone. This is a fairly outrageous abuse of power and not likely to nurture future researchers. There were some useful observations that made sense but not as many as I would have expected from someone with such alleged competence.

I regret to claim that, despite the enthusiasm with which I accepted the invitation to review this manuscript, I am appalled by it. My sense is that you do not comprehend the nature of signed networks, their subtle dynamics and have produced a manuscript that lacks any understanding of the substantive concerns surrounding the study of signed networks. You appear to be clueless – sorry, but I must be blunt – about important substantive and technical issues in this area. Substance matters when studying social networks not you give an airborne copulation about substance. Nor do you appear to care about the interplay between substance and technique in studying social networks.

….

These figures are meaningless. While moving onto the realm of biological networks has potential interest value, the exposition adds nothing. I am sure there are substantive issues in this literature that dwarf a narrow-minded concern with frustration indices as if this is the entry point to the kingdom of eternal life.

…

Please pay attention to substance! But I suspect, given this manuscript, you do not care about substance. As I read your manuscript, substance just gets in the way of a narrow focus on methods.

…

You have got to be kidding! … your claim, given the equivalence of two measures, strikes me as false and designed to promote your own work.

AlCo vs JACo – a stark comparison

Journal of Algebraic Combinatorics has been published by Springer since 1992. It was founded by Chris Godsil, Ian Goulden, and David Jackson. It has been a well regarded specialised journal.

In June 2017 the four editors-in-chief gave notice to Springer that they would resign at the end of the year. The entire editorial board (except for two members who decided to retire on the grounds of age) followed the EiCs to a new home. The new title is Algebraic Combinatorics, currently published by Centre Mersenne. Note that this journal is run according to the Fair Open Access Principles, so that any subsequent change of publisher will not require a change in the title of the journal.

Springer is attempting to continue the old title J. Algebraic Combinatorics with new editors. I know of many people from the algebraic combinatorics research community who were approached and refused. Eventually Ilias Kotsireas has accepted the offer to be EiC, despite explicitly being asked not to by the former EiCs.

The entire editorial board of Algebraic Combinatorics, including the 4 current EiCs, consists of 43 people. JACo, on the other hand, has 15 including 4 Advisory Editors.. The new editors of JACo have very little to do with the subject of algebraic combinatorics. Using the American Mathematical Society’s invaluable (and paywalled) resource MathSciNet, we can look for at papers written by various editors, having either primary or secondary classification 05E (Algebraic Combinatorics). We find the following data for JACo.

EiC – Kotsireas 0
Advisory board (4 people) 1
Editors (10 people) 9

However for AlCo:

EiCs (4 people) 69
Editorial board (4 of 39, almost randomly selected) 110

AlCo has published 12 papers since January 2018 and received 140 submissions since July 2017. According to one of the EiCs, the quality of submissions has risen since the move from Springer (although some subfields have reduced in quantity, which he attributed to authors waiting until AlCo is fully indexed).

It is completely clear that Algebraic Combinatorics is the continuation of the original journal founded in 1992, and the journal currently called JACo is a “zombie”.

Kotsireas’ recent editorial states: “…the research area of algebraic combinatorics is vibrant enough to sustain two high- quality journals”.

The obvious response is: “Why not set up a new journal elsewhere if there is so much room in the market?” If there were room, why didn’t Springer, or the new editors, do this? It seems clear that they only seek to use the reputation created by the old editors (and authors and reviewers) to improve their own personal standing, at the expense of the research field they purport to represent. Whatever that research field is, it seems that it is not really algebraic combinatorics.

Kotsireas also says: “I would like to thank the previous Editors- in Chief of JACO, with whom we had a very professional, productive, cordial and effective collaboration, during the transition period.”

This ignores the fact that the following message was received from Springer by the EiCs, who had agreed to work until 31 Dec 2017 and to see through all papers in their editorial queue.

“Due to the fact that a competing journal in 2017 (Algebraic Combinatorics) has been formulated which comprises the board of JACO, Ilias Kotsireas will be installed as an EIC as of Sunday October 15. It is also our understanding that current EICs of JACO (or at least some of you) will also become EICs of board members of Algebraic Combinatorics in 2018. Due to these extraordinary circumstances, we want Ilias Kotsireas to have final input on the acceptance and rejection of all articles that are in process until the end of your terms on December 31. If you feel you cannot comply with this measure and cooperate fully with Illias on the disposition of all papers, then it is best to part ways at this point and terminate your contracts early.”

Personally, I find this to be inconsistent with the description by Kotsireas. The fact that the original editors of JACo did not cause problems when presented with a fait accompli is not the same as having a “very professional, productive, cordial and effective collaboration, during the transition period.”

At first sight, it seems that it would have been so much easier for Springer to just agree to Fair Open Access, and publish the journal at a reasonable price. But the business model of such big commercial publishers involves running down the reputational capital generated by decades of work by authors, editors and reviewers, and investing as little as possible, and maintaining huge profit margins. Why work for a living if you can be a rentier capitalist? The company founded in 1842 by Julius Springer had a long history of service to the mathematical community, but, like journals such as JACo, the current entity usually called “Springer” bears little relation to the original entity. Since its takeover by Bertelsmann in 1998 it has gone through several incarnations, currently being called SpringerNature after a merger in 2015.

My guess is that the current JACo EiC is being paid about $8000-10000 per year, getting another CV item, and perhaps getting local recognition at his institution, so motivation there is clear. Exactly what the advisory board members and some associate editors, distinguished mathematicians from other fields, think they are doing is very unclear to me. I call on them to let the title die a natural death. Clear public statements by senior members of the algebraic combinatorics community are desirable. And the founding editors should insist on their names being removed from the JACo site. They founded a journal, and it continues under a new name, with them as editorial board members. The zombie journal must die.

Cybermath column NZ Math Society Newsletter Apr 2018

We focus (yet again) on a few developments in scholarly publishing, with a strong mathematical flavour. One is the establishment in late January 2018 (by me, with help from Jonathan Klawitter and Dmitri Zaitsev) of the Free Journal Network. We are all familiar with “diamond” open access journals, with no author fees, typically run by volunteer academic labour. Examples include Electronic J. Combinatorics, Australasian J. Combinatorics, New York J. Math., NZ J. Math.. Such journals are often of a very good standard as far as editorial processes go, but sometimes lack a few desirable features (e.g. DOIs, mobile-readable websites), can be seen as wasteful of researcher time, and are run on very low budgets (typically zero, with subsidy from a university providing the website). The FJN has been established to help promote such journals, attract small amounts of funding to fund luxuries such as those described above ($\varepsilon > > 0$ in this case), and allow sharing of best practices. We intend it to act as a whitelist for people searching for well run, ethically acceptable journals with reasonably high standards. So far there are 22 member journals of which 14 are in mathematics. Many of the latter were mentioned in the August 2017 column. One of these is Algebraic Combinatorics, the thriving new incarnation of J. Algebraic Combinatorics (see more on this below).

Of course, there are hundreds of mathematics journals that fall under the “diamond” label (many but certainly not all can be found in the Directory of Open Access Journals). However FJN has some formal requirements, namely that members satisfy the Fair Open Access Principles. These recently formulated principles are intended to formalise the intuitive idea of journals that run like those mentioned above, with no financial barriers to authors or readers, and community control of the journal. Some diamond journals (for example Annales de l’Institut Fourier and Acta Mathematica) have copyright transfer agreements that are inconsistent with FOAP, but are otherwise fine. Many diamond journals are very small, regional or otherwise not high priority for FJN to invest effort into. The ultimate aim of FJN is to build a portfolio of open access journals that is strong enough that libraries will pay to support them by redirecting subscription funding, so we can compete head-on with Elsevier, Springer, et al.

Readers having suggestions for journals to consider for membership should contact info@freejournals.org.

Another recent project is the establishment (with Dmitri Zaitsev) of an online forum Publishing Reform (there is also a private strategy forum). The idea is to centralize discussions and collaborate on useful documents and concrete actions to improve journal publishing. Mathematicians are again well represented here, including G\”{u}nter Ziegler, negotiator with Elsevier for DEAL in Germany; Martin Gr\”{o}tschel, President of Berlin-Brandenburg Academy of Sciences and Humanities; Timothy Gowers, Fields Medallist and Elsevier boycott initiator.

I recommend that readers check out the discussion site above and contribute as they see fit. More details about the ecosystem of community-controlled journals can be found in an article I recently wrote, to appear in August 2018 Notices of the American Mathematical Society.

Speaking of Elsevier, negotiations in Germany are dragging on, with no end in sight. The slightly bizarre spectacle of institutions agreeing to cancel access to journals and refusing to pay, but Elsevier providing access for free (presumably for fear of researchers working out just how easily they can cope without a subscription) shows just how dysfunctional the journal publishing market is. I have heard rumours that the Australia/NZ negotiations with Elsevier have also stalled, over the issue of nondisclosure agreements.

In Europe in particular there is substantial government support for the idea of open access, but almost without exception the wrong choices are being made over and over, and legacy publishers (presumably because they can pay lobbyists) are being given unfair advantages.The latest missteps are a Call for Tenders for the European Open Research Publishing Platform, which excludes organisations not already having a turnover of at least 1 million euros, and the EU Open Science Monitor giving Elsevier a contract to monitor the progress of open science.

Finally, the flipped journal Algebraic Combinatorics, published by Centre Mersenne, is thriving, having published 12 papers since January and having had 140 submissions at time of writing (anecdotal evidence from one editor-in-chief is that the quality has risen since the breakaway from Springer). An analysis of the board of the zombie Journal of Algebraic Combinatorics shows exactly what is going on. Springer apparently intends to capitalise on the reputation of the journal, built up over 25 years by the editorial board that has departed, by using new editors and ignoring what they do. A systematic look at MathSciNet shows that the editor-in-chief, advisory editors and editorial board of JACo (14 people in total) altogether have fewer papers published in the field of algebraic combinatorics (AMS classification 05E primary/secondary) than does almost any one individual editor of AlCo, and about one-third as many as just one of those editors.

As Marcus Tullius Cicero apparently apparently ended his speeches in the Roman Senate by calling for he destruction of Carthage, I call here for zombie JACo to die, and the big publishers to be abandoned until they actually provide service at reasonable price, and start to care about quality. These issues are too important to ignore – do something about it (ask me how if you don’t know)!

We have met the enemy, part 3: out of touch patriarchs

It is easy to argue that the problem of commercial control of scholarly journals is largely the fault of previous generations of academics. If they had not been so naive as to cede control of journals, and to fall for the wiles of Robert Maxwell, giving away valuable content and labour for free, we may have avoided the mess we are in now. This is probably unfair. After all, the publication and marketing of journals was difficult, researchers wanted to focus on research, funding was increasing, and academics were not used to dealing with unscrupulous businesspeople.

However, it has been abundantly clear since at least the late 1990s that the current system in which Elsevier and other large companies sell back the fruits of our labour at exorbitant and ever-increasing cost, while the overall value added decreases, is not sustainable. It is also clear that a major reason it has not collapsed and been replaced by a simpler, fairer, more efficient and higher quality system is the lack of leadership by senior figures in the research community. The examples of Donald Knuth and Randy Schekman are impressive precisely because of their rarity.

Here is an anecdote. I recently approached the editor in chief (call him H) of a strong specialized mathematics journal (call it J) with a carefully worked out proposal for leaving its publisher E. After 3 attempts over many months by email and a promise by another editorial board member to raise the matter with H, nothing happened. I then approached B, the founder of the journal and indeed of much of the research field. There followed a few rounds of civilized discussion by email during which I believe that I dealt with some of B’s rather cliched concerns about the effect on junior researchers. It all seemed to be going well, until B took offence at my claim that H was behaving unacceptably, and claimed that I was arrogantly judging H for his refusing my proposal. The explanation that I was in fact judging H for his refusal to engage with the proposal (a refusal would have been an improvement on being ignored repeatedlly) was apparently not understood, and discussion ceased.

I simply cannot understand the attitude of B, who exited in the 1990s the role of EiC at the journal he created in the 1980s. The sensitivity to criticism is unworthy of someone of such stature. The excessive loyalty to H who whatever his other merits may be, has clearly not acted professionally in this situation, is strange. All that would be required is for B to ask H to consider this seriously and give an answer, and surely this would get some results. If even this is too much, then perhaps E is right that academics do not deserve to run their own journals.

Two new initiatives in journal publishing reform

I have started (with help) two new initiatives intended to improve scholarly publishing.

The Free Journal Network is in its initial recruitment phase. It aims to foster independent open access journals to the point where direct competition with the big commercial publishers is possible, and will also help with starting new journals.

An online discussion forum and file repository is intended to become *the* place on the internet to discuss journal reform with emphasis on the Fair Open Access model.

Get involved now!

Cybermath column NZ Math Society Newsletter Dec 2017

As the deadline for each column approaches, I hope to write about something other than journal publishing, but lately there has been so much news from that direction that it is hard to ignore.

In 2014 Timothy Gowers and others used Freedom of Information laws to discover the amounts paid by UK universities for journal subscriptions to Elsevier. The reason they did this was that Elsevier (and SpringerNature, and maybe other publishers) insist very strongly on confidentiality agreements when they sign contracts with universities. The presumed reason for such insistence is that this makes their job of profit maximization much easier by lowering the bargaining ability of the universities.

The UK data showed that not only was each university spending a large amount, these numbers varied substantially even between universities with very similar size and research profile.

Earlier work in the USA and later work in Finland and Netherland have confirmed this overall picture. In 2014 I wrote toall NZ universities except Lincoln (for no really good reason, and I should rectify this, although it is only 1/4 the size of the next smallest university), requesting information of subscriptions paid to Elsevier, Springer, Taylor \& Francis, and Wiley (the first three are actually divisions of larger companies RELX, SpringerNature, Informa). These are the top 4 publishers in terms of expenditure by most libraries, although they account for considerably less than half of total journal expenditure. The universities concerned have around 8400 EFT academic/research staff and 130000 EFT students.

As expected, all the universities refused, and it was clear from the similarity of their answers that they had help from the publishers. Unlike the situation in UK there was no right of review of these refusals at a university level, so I complained next to the Ombudsman, citing the Official Information Act 1982.

After over 3 years of delays of all types, the Ombudsman’s final report unambiguously ruled in my favour, and the universities eventually supplied the information. So now we know how much they have spent, and the results are illuminating. Because of the fact that payments were made in various currencies, I have had to make some assumptions on exchange rates based on historical data. The raw data is available on Figshare.

For just these 4 publishers, the 7 universities paid NZ$19.4 million in 2016 in order to rent access to journal articles.
This amounts to $2300 per academic/research staff member.
For comparison, the Marsden Fund awarded $84.6 million this year, a big increase on previous years.
In the period 2013-2016, the amount paid rose by 17%, whereas CPI inflation in NZ and most other developed countries was around 3% over that period.

Longtime readers of this column will have no doubt about my opinion on these data. A huge waste of public money is occurring – independent estimates of the real cost of production of journal articles by modern publishers put it around US$500 per article, at most, while the current setup yields income 10 times that for the large publishers. These publishers make profits of around 40%, unmatched in any other legal industry.

The big publishers realise that the current subscription model is not sustainable. Although the way they market journal bundles — “Big Deals” — helps to insulate them from cancellations, such cancellations by academic libraries are slowly increasing, because the cost increases year on year re simply too much for budgets to bear. The publishers have seized upon the author-pay open access model as a way to protect their revenue. This model has serious resistance from researchers in fields such as mathematics.

Readers interested in learning about how we got to such an unpleasant situation should read this article. Readers interested in helping to get us out of the situation could do worse than to contact me at info@mathoa.org.

Big Deal journal bundles: price information from New Zealand

“Big deal” journal contracts by libraries with commercial publishers have been controversial for many years. Such contracts consume a large fraction of university serials budgets, and annual price increases are unsustainable in the long term. One main cause of such market dysfunction is price secrecy, whereby some publishers (including Elsevier and Springer, certainly) insist on confidentiality clauses in contracts (other causes include bundling of journals and the apparent inability of the research community to stop using historical journal reputation to evaluate researchers). While these companies are never short of justification for their actions, I believe that the main reason for these clauses is to facilitate differential pricing and weaken the negotiating situation of buyers.

In 2014 Timothy Gowers and others used Freedom of Information laws to extract the relevant price information from UK universities. See here for more detailed information. Earlier (2009), less extensive, work in the USA had also been done by Ted Bergstrom and others. Inspired by this, I tried the same thing in New Zealand (for 7 of the 8 universities – representing around 8400 academic/research staff and 130000 students, so far (Lincoln University, very much smaller than the others, was omitted owing to an oversight). Whereas Gowers was able to obtain the requested information within a few weeks, it has taken me 3.5 years. In both countries universities originally refused to release the information. However, in the UK there is an automatic right of review of such decisions, undertaken by an academic. In NZ, no such right exists, and my next step was to complain directly to the Ombudsman, the government official charged with determining whether information from the state sector should be publicly disclosed (all NZ universities are public).

The process was long and required persistence. I count at least 36 emails and several phone conversations. I commented on the preliminary report earlier this year, and the large publishers certainly had considerable input. The final report was released on 8 October 2017, more than 3 years after my first complaint. Gratifyingly, it ruled unambiguously that the commercial interests of the publishers and universities were outweighed by the public’s right to know. The universities have all complied, supplying me with the amounts spent on journals from Elsevier, Springer, Wiley and Taylor & Francis for years 2013-2016 inclusive. There are several other problematic publishers, notably the American Chemical Society, but I had to stop somewhere. I hope that others can continue this work in NZ and other jurisdictions.

There are some subtleties (such as exactly what products from the listed publishers the money is spent on, different currencies and exchange rates) that need more clarification than space permits here. I present basic derived data here, with the almost raw data also available. Prices have been converted at an exchange rate of $NZ = EUR0.63 = US$0.77 = $A0.90 for the entire period, which is obviously not completely accurate but is my best estimate based on looking at exchange rate graphs over the period 2012-2017 from x-rates.com.

The results are qualitatively similar to those found in USA and UK. At first glance, there are some major points:

The total amount of money spent on just 4 publishers is substantial, around US$14.9M in 2016.
The mean expenditure per academic/research staff member in 2016 is around US$1800.
University of Canterbury is getting a much worse deal than the others, 35% above the mean.
The rate of increase of subscription costs (17%) over the period clearly exceeds the Consumer Price Index inflation rate over the period (2-3% in NZ, USA and Europe).
The publisher with highest percentage increase over the period was Taylor & Francis (33%).

Universal open access to (largely publicly funded) research will remove barriers to readers, but still has costs that must be paid, presumably by reallocating money currently spent on subscriptions. The GoldOA model with author-paid APCs has been popular with traditional publishers, who often set the APC level in the $2000-3000 range. The analysis above implies that wholesale conversion to such APCs will not save substantial money for NZ universities. This is of course the aim of the publishers who try to exert their market power to prevent real competition. In order to provide market price controls of APCs, it is necessary to decouple ownership of a journal title from provision of publication services. This reclaiming of community control is the most fundamental of the recently formalized Fair Open Access Principles. New organizations such as MathOA, PsyOA, LingOA and the Fair Open Access Alliance have been set up precisely in order to facilitate large-scale conversion of subscription journals to an open access model with community control of journals and no direct author payments. We expect that savings of at least 75% can be made by using modern publishing providers. What is the research community waiting for?

Vienna

I spent almost 2 weeks in Vienna, Austria in November, visiting the Schroedinger Institute. The work environment was excellent (maybe the blackboards on the toilet walls were overkill) and the city is really impressive – no wonder it ranks so highly in the standard international quality of life surveys. In addition to the ease of getting around by foot, tram or U-bahn, and the high quality music (I went to 3 operas, a highlight being a 3 euro Stehplatz L’Elisir d’Amore experience at the Staatsoper), both of which I expected, the well priced and varied restaurants and cafes, and the cosmopolitan feeling were a nice surprise and different from what I remember from 25 years ago. I highly recommend a visit to Vienna!

Banff

I attended a workshop at Banff International Research Station last week. The setup there is very conducive to productive work, although there is too much good food available (and because of the signs about bears, I didn’t do any any serious walking). The workshop itself was very well put together with a lot of interesting talks. Really good video of talks is available. Thanks to the organizers Mireille Bousquet-Mélou, Stephen Melczer, Michael Singer and Marni Mishna and to the participants.