Tuesday, July 15, 2008

Creating an organization for open access publishers - but should we let big publishers dominate?

Dave Solomon has published a draft of possible By-laws for the proposed Open Access Scholarly Publishers Association (OASPA) (Archived at WebCite on 15/07/2008 at http://www.webcitation.org/5ZLR1YeVL ).

My personal interest in being possibly one of the founding members of this organization (which also may include BioMed Central, Copernicus, Co‐Action Publishing, Hindawi, Medical Education Online, Journal of Medical Internet Research, PLoS, and presumably others) is the recognition that some sort of organization is needed to set, promote and enforce professional and ethical standards and OA publishing practices (see also the post and responses to my post "Black sheep among Open Access Journals and Publishers").

But it is important that such an organization is set up and run not to primarily defend the commercial interests of large commercial publishers, but to act as much as possible in the interest of scholarly communication (these are not mutually exclusive goals - but they are not always aligned - if there is a conflict, then responsible publishing practices should always take precedence over commercial goals. For example, pricing or marketing policies of OA publishers should not primarily be designed to maximize revenue, but also keep the long-term interests of scholarly communication in mind).

In this context I am surprised and concerned about some of the language in the bylaws, which seem to have resulted from lobbying activities of (at least one) larger commercial open access publisher who apparently wishes to get more "votes" than other publishers.

As a result of these lobbying efforts, the bylaws contain the following odd regulation ("odd" because I do not know any other professional organization or trade association where certain members get more votes based on their size, revenue, or other criteria):

Organizations who publish large numbers of OA articles will be allowed to designate additional representatives per table 2.01 below. Each representative will be allowed to vote in OASPA elections and be allowed to hold office.

More specifically, Dave Solomon and Caroline Sutton are contemplating the following rules:

Each publisher gets one voting member and an additional voting member for every additional 250 OA manuscripts published in the previous calendar year up to a maximum of 10.

In other words, this will ensure that large publishers have 10 times more votes than small publishers. One possible scenario is that the 3 largest publishers together can comfortably vote themselves into key positions and vote for policy decisions against up to 30 smaller publishers.

In my view, setting up a Open Access Publisher Association which provides voting rights based on the number of open access articles published is a bad idea. I am not aware of any other publishing association / organization where voting rights are weighted by the number of articles published. In my view, quality is more important than quantity.

I am surprised to hear that one of the larger OA publishers is pushing for this - apparently one large OA publisher is so concerned about being "overwhelmed by single journal publishers" (whatever that means in the context of creating an association of OA publishers) that they have indicated that they may not join the organization unless they have more votes than other members.

But there are opposing views to this. In fact, everybody who I have spoken to agrees on the simple fact that we must (at least initially) have a straightforward, simple, transparent, unbiased membership scheme (1 publisher = 1 vote), as is the case in
ALL other publisher organization I am aware off (can anybody give me an example of a publisher organization where this is not the case?. I am not aware of a single publisher organization out there where larger publishers get multiple votes based on their size, manuscripts, revenue etc).

The reason why I am (and others are) so opposed to this are twofold:

1) definitional problems: which articles do you count? When setting the threshold of 1 vote per 250 articles, would one count news articles, tutorials, viewpoints, reviews, original papers? Who is determining what counts and what doesn't? Or do we mean only peer-reviewed papers? If so, how do we verify if an article has indeed been peer-reviewed (not all publishers are transparent on that)? And does the level/rigor of peer-review matter? For example, PLoS One published over 2000 articles which are (to my knowledge) not "peer-reviewed" according to the peer-review definition in the "professional conduct" section of the proposed bylaws. In addition, innovative models of peer-reviews are emerging (esp post-publication peer-review with no peer-review
upfront), so how do you treat these articles? JMIR will shortly make submitted papers available for open peer-review. Given that "publication" is a continuum (Richard Smith), I would argue that it is naive and fraught with problems to introduce "published papers" as a metric which determines the number of votes a publisher will have. One only has to look at the problems and discussions Thomoson/Reuters has each year in determining the denominator for their journal impact factor... - what should be counted as a peer-reviewed original article is not such a clear-cut decision as one might think.

2) making the quantity of published manuscripts a factor in determining the number of votes a publisher has might encourage "quick & dirty" publishing with high acceptance rates and sloppy/superficial peer-review. I am not saying that all publishers publishing a large number of OA articles are low-quality publishers violating the professional codes of conduct. I am just saying that it makes little sense to penalize a publisher who perhaps handles far more than 250 submissions per year, but decides to be selective and only "publish" are fraction of it (e.g. 40), compared to another publisher who might be less selective and publishes nearly all submissions they get. Why should the latter publisher be "rewarded" by having more votes and influence in a publishers' association for being less selective?

Quality is more important than quantity, so if anything, than a measure determining voting power should be based on "quality" and impact, but as there are no easy and transparent measures for quality (the acceptance rate would be a good measure but can not be independently verified by a third party) or impact (the number of citations is discipline specific and also not a good metric to determine voting power), I would strongly suggest to keep the original bylaws simple: 1 publisher = 1 member = 1 vote. Once the organization is created, the members in the association can debate the issue further and decide whether adjustments should be made in the future.

Frankly, I also cannot see any rationale for why larger publishers should have up to 10 times more votes than smaller publishers. Larger publishers have (by their
sheer media presence and/or marketing budget) already more channels to influence policy than smaller publishers. The association should, in my opinion, strive to develop and defend best practices on a level playing field. The voices of large publishers will always be louder than other members' voices, and this phenomenon is independent from any formal weighting of their votes.

In addition, I am also concerned about the optics of these bylaws - in fact, it looks VERY bad to create an organization which primarily serves the interests of the large publishers.

Please leave a comment to let me and Dave know what you think about the idea of giving larger publishers more voting power.

Another (unrelated) second critique I have about the current draft bylaws is the code of conduct:

# Direct marketing to individuals shall be relevant;
# Recipients of direct marketing materials shall be provided with an opt-out option;

This is a much too weak statement which (rather than combating spam) actually seems to endorse spamming activities ("Direct marketing to individuals shall be relevant").

What the bylaws must say here is very simple and straightforward:
"Spamming (mass emailing) individual researchers is not an acceptable way to solicit submissions."

There is a clear definition on what "spamming" means (in most jurisdictions), and if there is any uncertainty, one may adopt the definition of Spamhaus.

"Opt-out options" are not a solution. In all three high-profile spam
cases (Bentham, Dove Medical Press, Libertas Academica) the spam emails contained a unsubscribe option ("email us to opt out"), but these requests were never honored. It is simply too easy to use "human error" as an excuse and to refer to the fact that an unsubscribe option was provided.

Spamming (mass emailing) individual researchers is not an acceptable way to solicit submissions - I think this is a rule we could (and must) all agree upon. If publishers are endorsing spamming as a marketing strategy then researchers will all find daily messages from open access publishers in the inbox asking for their manuscripts, which will completely damage the reputation of open access journals and put them on par with internet pharmacies selling viagra.

I do hope that publishers can agree on fixing these major flaws in the draft bylaws, which otherwise would steer the ship into the wrong direction.

Update added 02/08/2008. The bylaws are currently being revised and apparently none of the founding publishers has raised objections to get rid of the controversial clause on giving large publishers more votes.
Update added 20/08/2008.
Update: One of the founding publishers has raised objections and insists on having more votes.
Update added 19/09/2008.
Update: We finally seem ready to go ahead with a one-publisher-one-vote system. Hurray! OASPA is scheduled to launch on OA day.


dsolomon said...

I would like to respond to Gunther’s comments. First, it is not true and I think unfair to characterize the professional publishers who are involved in forming the Open Access Scholarly Publishers Association (OASPA) as attempting to dominate the organization and imposing their agenda.

Several months ago I put forth a proposal with Gunther’s help and support to form an OA professional publishing organization. Shortly after that, Caroline Sutton from Co-Action Publishing contacted me. Co-Action and a number of other professional OA publishers had been discussing forming a similar organization. She suggested that we talk about either working together or at least keep in touch and not work at cross purposes in forming organizations. After talking, we decided to try to work together and form a single organization. The other professional publishers noted in the post above have been very supportive of this effort. They did not have to be. They were well on the way to forming an organization of their own. They have adequate resources plus financial support from SPARC Europe and could easily have told us thanks but no thanks and gone on their own way.

Over the last several months I have worked very closely with Caroline in developing the proposed bylaws. It’s become clear that though there is a great deal of overlap in our interests and goals for the organization, there are some differences between professional OA publishers and people like myself (I’ll call scholar-publishers) that publish single journals on a non-professional basis largely on volunteer effort. There are also society publishers, research libraries, university presses and other types of OA publishers with somewhat different interests and perspectives. We have tried to design the bylaws to ensure that no one group dominates the organization and to the extent possible the organization represents everyone’s interests. We have proposed that there be several classes of members including professional publishers and scholar-publishers as separate classes of members. The board of directors that will largely govern the organization can have no more the 49% of the board members from any one class of membership. No one publisher of any type can have more than one member on the board. That in itself should go a long way to ensuring no one group dominates the organization.

While I have worked mainly with Caroline on this project, I have had some contact via e-mail with the other professional publishers. Nothing in any of my interactions with Caroline and the other publishers has suggested they are trying to dominate the organization. They have some concerns about the organization mainly reflecting the interests of scholar-publishers due to our overwhelming numbers but to my knowledge none of the professional publishers had threatened not to participate without additional votes.

Caroline and I had proposed to allow larger publishing organizations to have additional voting members. This may be unique but every organization is unique in some ways. The group of publishers we are seeking to represent have a different set of demographics than an organizations such as ALPSP.

We proposed to base additional voting members on the number of peer-reviewed OA articles published in the preceding year. While clearly not perfect it seemed the best option and at least easily measurable. The last version of the bylaws that has been circulated publicly stated this as a general concept. This part of the bylaws was highlighted in red with the hope that it would generate some feedback. We received several comments and concerns on a variety of issues, but none as yet have addressed this issue other than the concerns above brought by Gunther.

The latest proposal which has not with the exception of this blog been made public is mine and mine alone. It was written as one possible implementation of the general concept which in my view would provide some balance between a few large publishers that distribute a lot of OA scholarship and a large number of scholar-publishers many of whom publisher only a few articles a year. It is not written in stone but as a starting point for discussion as is the whole issue of a formula for basing votes on size. The details need to be worked out if this concept itself were to be adopted. The issue of what is “peer-review” needs to be addressed anyway since it is a requirement for voting membership.

The second issue Gunter brings up on the language on direct marketing is going to be addressed in the next round of revisions.

If you have thoughts on this issue, please let me know.

David Solomon

Gunther Eysenbach said...

Dear Dave: Thanks for clarifying that “the professional publishers who are involved in forming the Open Access Scholarly Publishers Association (OASPA) [are not] attempting to dominate the organization and imposing their agenda.”. I was not implying that they wish to impose their agenda, but there is no question that if larger publishers get up to 10 times more votes, their points of view will be “artificially” amplified and it gives at least the impression that they strive to dominate the association. While you are now implying that this is not what they want, you wrote in an earlier email to me: “We have a provision for the larger publishers to get additional voting members. Without this, it would probably have been a deal breaker for the larger publishers. (…) It was I believe [name of a larger publisher] that had the biggest concern about being overwhelmed by single journal publishers. (…) Caroline and I added this to the bylaws and updated it with a specific proposal to address what we felt to be a general concern among the large publishers, not just [name of a larger publisher]. Their concerns are not a mandate but these large publishers are important to make this organization work. (…) ”). I am still confused on whether the idea to give up to 10 times more votes to large publishers was “preemptively” added by you/Caroline, or whether this was the result of pressure from larger publishers or one specific publisher - but whatever the process was, my critique in the end-result remains the same – it is simply a bad idea, and must be taken out of the draft bylaws.
In terms of the process, it would have been better to draft the bylaws in a collective manner (e.g. on a wiki) – allowing for transparent exchange of arguments and counter-arguments and where it stays “on record” who makes which suggestions – as opposed to a process where you and Caroline act as an information filter and where it is unclear whose opinions should be attributed to whom.

My overarching concern is if larger publishers (BTW - I object to the term “professional publishers”, because it gives the impression that smaller publishers - in your definition those who publish less than 250 OA manuscripts per year - are “unprofessional”) already at this stage express concerns about “being overwhelmed by smaller publishers”, how will this later translate to policy decisions within the association, starting with simple decisions such as membership fees (will they try to keep smaller publishers out by voting for higher membership fees?), and possibly continuing with other important decisions such as membership criteria or accreditation criteria? If the association would have to decide on membership fees or develop some sort of an accreditation program, then larger publishers could easily use their voting power to implement conditions that would make it impossible for smaller publishers to implement such conditions, essentially pushing smaller publishers out of the association or even out of the market.
Yes, presumably the smaller publishers together collectively have more votes than larger publishers (“long tail”) unless you give larger publishers more votes, but how exactly would this be a threat to larger publishers or impede the mission of the association itself? Yes, joining such an association means having to deal with a "mass of scholar-publishers" - this process is called democracy! The issue is that the "mass of scholar-publishers" does not necessarily speak with one voice unless there is some consensus building process (which is sometimes painful, but necessary in a democracy, and for which the association should be a platform), but if you accumulate 10 votes to certain classes of publishers, they will obviously speak with one voice and will dominate any decisions made.
Now, if you say the arguments could be turned around, then can you give me an example for a scenario where smaller publishers could use their collective voting power to push larger publishers out of the association or out of the market? I guess such scenarios would be much harder to come up with, as anybody knows they are needed for the association to work.

Although I do not believe that “pushing smaller publishers out of the market” is on the minds of the larger publishers (or the reason why they want to create such an association) – or vice versa -, I do believe that such a provision and the resulting “rich gets richer” mechanisms are simply threats to the integrity of the association. If this argument doesn’t convince you, then perhaps you can at least look at the impact of such a decision on the perceived integrity of the organization. In terms of the “perceived integrity” of the association, you should also consider that any policy decisions or lobbying efforts by the association would carry a lot more credibility if the public wouldn’t have the impression that these are the voices of a few large commercial publishers, but represent a broad consensus of all open access publishers.

In summary, giving larger publishers more votes is a bad idea, for at least four reasons I gave above: 1) measurement/definition problems (the unit of "published, peer-reviewed article" is not as straightforward as one might think), 2) providing the wrong incentives (quantity instead of quality), 3) giving the impression that large publishers want to keep small publishers out of the organization or override their votes, and resulting trust issues and artificial divisions between “first class” and “second class” members of the association, 4) the optics and credibility of the association. While I gave you four reasons for not doing it, I have yet to hear any concrete arguments that speak for more votes for larger publishers other than saying they are necessary (with a similar logic you could also argue that citizens who pay more taxes should have more votes). Nobody denies that larger publishers should be part of the association, but to preemptively “bribe” them with more votes (which you are now saying they didn’t ask for) so that they agree to work with smaller publishers in their association is not the right way to go forward and in my mind sets off several warning bells. Their voices will always carry more weight regardless of how many votes you give them, and if larger publishers don’t agree to form an association unless they get more votes than smaller publishers, then it is even better to create separate associations. But you say this is not the case, so it still remains a mystery to me why this is in the bylaws in the first place and why we don't use a simple 1 member - 1 vote structure as a basis for discussion, until a publisher stands up and provides some concrete arguments for why he thinks he needs more votes than others (which I think is an arrogant notion, but I would be interested in the arguments brought forward).