Thursday, October 24, 2013

My thoughts for the panel on "open access and the future of scholarly publishing"

On Tuesday, as part of open access week I participated in a panel "Publishing, Authoring, and Teaching in the Evolving Open Access Environment: A Panel Discussion". While this is not a word for word write-up, this is more or less the gist of what I said.

When I was asked whether I would be willing to participate in a panel discussion here at Michigan State University on the role of open access journals and the future of academic publishing I said yes. While I am not convinced I am particularly knowledgeable about it, I thought that it  provided an opportunity to collect my thoughts, a manifesto of "how I communicate science, and why I do it that way".

While I do tweet and blog about aspects of open science, including open access publishing, I am not one of the most outspoken advocates, and only a moderate practitioner. I publish, review and edit in/for open access journals, but not exclusively. I continue to publish in many "subscription journals" that represent the journal of record for my field, or those with some inferred "prestige". I do happen to regularly discuss issues about open science, including open access publishing with many folks, but as you will see I am not sure where I fall down on it.

I work in the basic life sciences, at the interface of evolutionary biology, genetics and genomics. The norms of scholarly communication differ substantially from field to field, in terms of what is considered productive scholarship, books VS. articles, authorship and a host of other issues. Even within the natural sciences, scholarly communication differs between biology and say physics. So, my experience and understanding remains narrow and I claim no expertise.

I think that the future of scholarly publishing will be open access, in some shape or form. That is, the majority of published manuscripts will eventually be freely available to anyone with internet access. How do we get there? I have no idea. Will this be due to broad mandates from funding agencies and  Universities to deposit manuscripts into repositories? Will journals generally agree to make content freely available after a fixed amount of time (6-12 months) - so called green open access? Or will gold open access become the norm where authors pay to have work reviewed or published?  Likely a combination of these and other approaches, but I am not good at such guess work.

So why do we care about OA in the first place?

Several reasons.  

This has been discussed by many before, so my thoughts on this are brief. For more detailed thought, check out Peter Suber's book "open access".

If you happen to be on the MSU library site, and happen to click the faculty page you will see on the right hand side links to a number of things including "Crisis in Scholarly Communication". The discussion on these pages is about the increasingly difficult access to scholarly publications. The basic reason is that while academic library budgets tend to be relatively flat, the cost of subscriptions to academic journals continues to increase very rapidly. Often this is because many of the subscription (i.e. for profit) publishers are commonly practicing bundling of journals. So if you want journal A, you also have to subscribe to x (pick a country) journal of y (sub-field) of z (pick organism).

Why should scholarly work be behind a paywall, and thus inaccessible. 

In particular for scientific (and medically relevant work) it could benefit researchers, patients and doctors (who would otherwise not have access). Open access allows the whole public to look at the research if they so choose. This also removes one small barrier in the perception of the ivory tower, and rebuilding some trust with the public (more on this later).

Who is paying for the research? Generally not the publisher making money for the paper.

 At least in the sciences, research is usually paid from grants from federal agencies, and salaries are paid from those grants, or from the University (such as MSU) which in part comes from state allocations and tuition dollars. The manuscript is then reviewed by referees usually for free (as part of our scholarly role) including as scientific editor (which is also not usually a paid position, at least for associate editors). Under the current system most referees get nothing (neither money nor any other incentive) for this essential service, and their pay is from their institution (and does not depend upon them performing this service). The publisher may maintain the electronic system to shuttle the draft manuscript to the referees, and if accepted performs copyediting and typesetting.  There are exceptions to the rule (I have had absolutely excellent editing advice on both the writing and communicating the science for a recent paper in Trends in Genetics from the managing editor for instance), but this has not generally been the case for me.

Thus the publisher is making a great deal of profit, despite having only done a fraction of the work. They (not the authors) retain the copyright on the work. This is potentially a big problem.

How I got into this

I will tell you about how I got engaged with the ideas of open access publishing, as a small part of the larger endeavor to make science in academia more open, transparent, reproducible, and in such a way that scientific ideas and data are communicated more quickly and effectively. But I also will describe many of the remaining stumbling blocks that relate to views of open access journals specifically, and the nefarious concept of prestige in publishing and how that influences hiring, grants, and promotion.

The crisis in scholarship is much bigger than open access.

Why is there a crisis?

While open access of published work is certainly a factor in the crisis in scholarly communication, and the one I will speak the most about today, it is not the only factor (issues with peer review, reproducible research, sharing of data and code, etc..).  There have been a slew of articles including a few in the Economist and the Guardian in the last few weeks on aspects of this crisis. Essentially it is argued that scientific research is in a tailspin, there is no effort to do quality science, and everything is about quantity of papers and prestige (i.e. spin) with little effort to make sure the work is valid, well reviewed or replicated. That is the incentive system for scientists is completely out of whack with the process of good science

Is there a crisis? I am a skeptic and cynic, so I think there are some real concerns. However the little optimist voice in my head also points out that there are some great opportunities as well to help to not only resolve this crisis, but make science/scholarly research better, and far more dynamic. Indeed there is a vocal and active community trying to make this all much better.

Before I delve into the specifics about the need for open access (and what might be stopping some of us from diving into it completely), I want to speak about the larger crisis of the scholarly enterprise in general, and how open science initiatives can fix this. In my mind this comes down to an issue of trust. Trust between collaborators. Trust between scientists working in the same field. Trust between researchers at different stages of their careers (graduate students, post-docs and PIs). Finally, we need trust between scientists and the general population. Not just how the public perceives scientists (and scholars in general) and the work they do, but that we do our science in an open way that leads to the appropriate self-correcting mechanisms. However, even beyond the large number of anti-science and anti-intellectual movements out there, there has been a substantial loss of trust in how scholars operate, and what motivates us.

Between the large number of research articles that get picked up by the popular press about "cures" or of genes for "this that and the other", only to have such research shown to be largely (or entirely) incorrect a few months later.  Combine this with the many news articles that point out the lack of repeatability of scientific studies, or examples of scientific fraud and misconduct, our lay audience (and those that ultimate help pay for our research and salaries) are perhaps becoming quite skeptical. The lack of access to the scientific literature for the public due to paywalls (from subscription based journals)  is simply another large nail in the coffin for the trust that scientists and science communicators have been trying to build with the public.

Frankly our motives are questioned, and not just by the public at large. They are also questioned by our graduate students too. As undergraduates (or from watching nature shows) they see this amazing wonderful universe to study, but then come to do research as graduate students and realize that a business model has taken place with a culture of "scientific stardom" being the goal for many. Worse than that, they see a perverse incentive system where quantity of publications and prestige over where articles are published has taken hold, and the overall quality of scientific research is the perceived victim (has this been evaluated?). There are many folks to blame, including university administrators, the "high prestige" journals, etc, but we first need to look at ourselves (practicing scientists) for accepting and adopting this system as it has developed into the status quo.

So how do we fix it?

Let's think about the aspects of open science. Not only do we need to communicate what we do more effectively, but we need to make everything far more transparent.

Open Science

Much has been written about open science, and the open science movement. It has many goals, but I would say in general the two most important ones are to increase transparency of the scientific process and to speed (and open) up science communication. There are many aspects going all of the way from open "lab notebooks" (here and here). Submitting pre-prints (papers prior to peer review and formal acceptance at a journal) to repositories to speed up science communication. Increasing openness in peer review, so that all can see the comments of the peer referees and editors. This can include pre and post publication review (pubpeer, pubmed commons). Sharing of raw data associated with research papers as well as all of the details of how the analysis was performed (the computer code associated with it). It all comes down to the fact the published paper is not the science itself, but a precis (or an advert) for the actual scientific work. All of the data, the work and even the peer review process itself is part of the scholarship of science. Making some or all of this available will not only help with transparency, but will speed up the scientific process. Having access to the raw data may also help to answer all sorts of new and interesting questions. I have always loved this quote by Sir William Bragg:

"The important thing in science is not so much to obtain new facts as to discover new ways of thinking about them"

Open Access and the entanglement of scholarly publishing with prestige and other incentives

Crossroads for publishing

At least in the life sciences, it is clear that publishing articles in "high profile" journals like Nature, Science, PNAS (and a few others) can make or break a career. The prestige associated with such articles can trump many things. Having such publications can open many doors in terms of jobs, grants, tenure, invited talks and more. Indeed last week in Nature, articles were written about just this phenomenon, the so called "golden club". Not surprisingly most of the traditional journals with such cachet are subscription based, although there are at least two open access journals that are certainly up there (PLoS Biology and ELife).

Since most of the prestigious journals are subscriptions based, and there are such strong incentives to publish in them, it makes it very difficult for many researcher to move to publishing in open access journals (although they can still submit papers to institutional or disciplinary repositories). If I had the opportunity to publish in Nature or Science, would I? Yes, precisely because I know that having such a publication will open doors, aid in getting grants, promotions and raises. While I feel strong support for open access, I am frankly not above such concerns. Some of this may simply reflect my petty needs for external validation of my science (which I can get over), but grants and raises potentially influences the quality of my life and my work. It is hard to pass that up.

Perceptions of open access journals

There remains a common misperception that many open access journals are nothing more than predatory or "vanity" publications with little or no rigorous peer review. A recent "sting" by the science journalist John Bohannon in Science has done little to help this perception. Too much has already been written about this article, mostly highly critical of his methods, biased sampling approach and lack of a control group. While It was presented as a news piece (not a scholarly article), Bohannon  has stated on several occasions that his original plan was to submit this to PLoS One (an open access article with peer review) so this remains an issue.

As Peter Suber (among many others) describes in his book there are currently two models of open access. The first (so called green OA) means that while papers may be published in subscription based journals, a free version of the accepted manuscript (usually without final copyediting and typesetting provided by the journal) is placed in a repository such as pubmed central or an institutional archive. Often this "free" version has a 6-12 month delay before being released. Since this model of green OA is a required stipulation for projects funded by organizations such as the NIH, some journals (where the majority of authors are funded by such agencies) are now just making all of their content open access after a 12 month embargo).

The other major model for open access publishing is gold open access. In this case, once  a manuscript has gone through peer review by expert referees and academic editors, and it is accepted, then the authors of the manuscript are charged a fee for typesetting and (usually online) publication.  Thus the model is that the authors, not the readers are charged. In the life sciences the funds for this usually (more on this in a second) come from granting agencies, although many journals have fee waivers (or no fees at all).

The concern with this of course is this may create a perverse incentive system, such that journals would increase their acceptance rate to increase profit (~1-2K/paper accepted). Thus the rigor of peer review could be negatively impacted, resulting in so called "vanity publications" that have the veneer of scientific rigor and peer review, but in fact do not. Couple this with so called predatory (scam) journals (that are much like other scam spam).  Before open access journals existed, such vanity journals already existed among the subscription models. And as the Bohannon sting has shown us, journals that are published by well known publishers like Elsevier, are not above being "stung" and by accepting faux articles with obviously flawed methodologies as well. Beyond that, somewhere on the order of 70% of all open access journals have no author side fees.


Despite this issue, and the existence of predatory and possible vanity journals (such as many of those found on Beall's list), the Open access scholarly publishers association has a code of conduct for journals aiming to maintain reputable scholarly journals from predatory ones. From my perspective, the journal PLoS One which in many ways represents the flagship of open access journals (peer review entirely based upon technical soundness of the experiments and interpretation, not upon subjective assessment of novelty) was noted for how thorough the review process (and rejection was). The other worth while point is to take a look at the re-analysis done by Brian Wood. Seems like the one thing that journal impact factor might be useful at doing is predicting whether a reasonable amount of peer review might take place.

It is also worth pointing out that problems with sufficient peer review occurs with subscription based journals as well. In addition there has been a history of these so called vanity journals even among subscription publishers. In addition, many journals with subscription based models also have page and figure charges that the authors must pay, so some of the same incentives also apply to these journals. In my own personal experience, these page charges end up being about the same cost as publication in open access journals. So many of the same charges against OA journals can equally be leveled against such journals.

Why have I not embraced open access completely?

It is probably clear from my perspective on all of this, that I am firmly in favour of open access models of publishing, and like I stated from the outset, I do think that this is where everything is going to, although by what model I am not sure.

 So given all of this, why don't I publish exclusively in open access journals? Well there are two reasons, or possibly one reason arising from two different parts of my mind.

The first relates to "establishment" journals. In my field, there are several well regarded journals that have persisted for a very long time, some for over a century such as Genetics. In my field, publishing my work in journals such as Genetics, or in Evolution means that A) It has a natural readership. B) These are the same journals that shaped my understanding of the field during my intellectual development and so I have a fondness for them and C) While they may not have the cachet of Nature, Science and the like.. there is no doubt that in my field they are considered well regarded journals. D) These also represent the journals for my professional societies, which I actively support and promote above and beyond their role in scholarly communication.

As for the second set of reasons.. I am not sure I am willing to be a martyr. In other words, I may be acting with a great deal of cowardice. Despite having been an editor at PLoS One for many years, and I stand by the rigor of reviews by my referees and myself, there is no doubt that there are many in the community who still believe that it (as a journal) accept anything. If I choose to publish all of my work (and that of my students and postdocs), I risk losing readership. If such views are held by university administrators I risk loss of salary raises promotions and grants (depending on the panel).

Thus until the incentive system has changed, and this can only change by concerted effort between university administrations, grant program officers and well established scholars in each of our fields embracing such changes, many researchers like myself will continue using this screwed up system, because of the incentives, risking further erosion of public trust.  Is my half way attitude a cop-out. Yes. Some horrible mix of rationalization, cowardice and avarice I suppose. Am I likely to change my behaviour? Probably not until my mortgage is paid off and my kids have finished university.


2 comments:

  1. You are too hard on yourself.
    Looking out for your students is noble.
    As you point out, there's lots of factors in play here. Even if everyone collectively solves the prisoner's dilemma, it could just create other issues as the system's tenuous grasp and appraisal of research struggles to catch up to a system that would require being informed.

    ReplyDelete
    Replies
    1. I don't view it as being hard on myself, as much as recognizing and trying to balance the competing forces. Perhaps I am overly cynical, but I do think that those competing forces are "what is good for careers (mine and my students)" and "What is going for science". Some realignment of incentives might help, but I try to discuss these issues with my students and at least let them make the choices about where their papers go.

      Delete