Recently, the Committee for Economic Development released a report with the rather provocative title, “The Future of Taxpayer-Funded Research: Who Will Control Access to the Results?” Its author is Eliot Maxwell, a consultant with a strong background in technology and telecommunications, and many published reports advocating openness as a potential solution to sticky problems such as innovation, health care, economic growth, and teaching.
The title is provocative because it jumps quickly to equate “the results” to final, peer-reviewed, and edited manuscripts. This type of conceptual elision is a defining trait of the piece.
In other ways, the report is a bit unusual for something from the CED. It is written by one person and not the result of a subcommittee’s work. Maxwell has participated as a project manager or project director for other CED reports on openness, but not as the lead (and in this case, apparently sole) author. Also, in the press release accompanying the report, funding from the Kauffman Foundation is gratefully acknowledged, another unusual feature for a CED report, based on a couple of dives into their library. Maxwell doesn’t disclose within his bio his role with Science Commons (part of Creative Commons), which also receives some funding from the Kauffman Foundation. I’m not sure if this matters, but it’s worth noting in case it might somewhere down the road.
Whatever its slightly unusual pedigree, it’s immediately clear that the report represents a singular perspective on the arguments and evidence around access to reports coming from taxpayer-funded research, one that reflects an author with a long history of embracing open solutions to all sorts of social and economic challenges, yet one that often fails to contemplate the larger picture or provide cogent analogs for the STM market.
One example of this is his use of “open mice” as an analog for the benefits of openness in science communication. In this story, DuPont held patents on two strains of transgenic mice, patents they enforced onerously — restrictions on use, on the science derived from their use, and on any subsequent patents, royalties, or ventures were all part of DuPont’s approach. Yet another strain of mice offered by others had IP restrictions that were unenforced — the so-called “open mice.” (Forgive me for allowing dissection-table images to come to mind with this strange turn of phrase.) The open mice were used more often for obvious reasons, leading to more papers being published based on them, and of course more citations. Maxwell attempts to draw a parallel between open mice and open access, but I believe he fails. Subscription-based publishers do not restrict the use of papers except through pricing; do not attempt to control patents or enterprises associated with papers; and actively encourage further science by providing good places to publish interesting results.
In other situations, Maxwell makes sweeping statements that aren’t easily derived from the logic or evidence he expounds in his work. For instance:
The NIH public-access policy has substantially increased public access to research results with benefits as described below that far outweigh the costs.
In his argument, Maxwell states clearly that the economic calculations he uses to make these claims depend on “practitioners” having access to relevant scientific information. Practitioners are not the public. He also claims that any surprising science that might come from public access should be more valuable because it’s unexpected; therefore, public access makes sense. Yet he fails to cite a single compelling instance of this “dumb luck” or “amateur” science that goes beyond smart business people inventing things, and doesn’t touch on the instances in which public misunderstandings of scientific findings have created severe, even tragic, costs (excess fatalities from communicable diseases thanks to the “scientific” vaccine controversy, for example).
This could go on for a while. For a good critique of the report, I invite you to read Fred Dylla’s excellent response, in which he writes:
. . . the paper should not be viewed as an objective analysis of all of the available information on open access, but rather as a carefully crafted opinion piece, based on selected references, that is cleverly framed as a work of scholarship. . . . I simply cannot endorse a report that lacks an impartial examination of so many important factors germane to advancing an informed discussion on public access to scholarly communications.
Maxwell has also responded to Dylla’s critique. Through it all, Maxwell betrays a clear intellectual bias for “open” as the preferred solution. This shouldn’t be surprising — much of his career has been based on advocating for this position in other settings.
In the original report, Maxwell lets spill some interesting complaints, complaints that mirror those of many of the most ardent OA advocates. It made me wonder, What would the world look like if a few of Maxwell’s implicit and explicit dreams came true?
- Authors would be paid for their research reports. Currently, authors are generally not paid by publishers when they submit research reports. This is a complaint Maxwell makes (” . . . manuscripts . . . are submitted by their authors without any compensation from the publishers”); others have noted the same complaint here repeatedly in comments in order to make publishers look unfair. Let’s play this out. First, authors get plenty of indirect benefits from publication. Second, the same people who make this complaint usually make another complaint about scholarly journals — that they are too expensive, and that prices are rising faster than inflation. Another companion complaint has to do with journal “bundling” so that smaller journals can survive in a market that might otherwise kill smaller or more specialized titles. If authors were to be paid, the cost basis of nearly every journal would increase, driving prices higher throughout the journals market. More importantly, prices for the most prestigious titles would increase the fastest, as these journals would be willing and able to pay more for the best papers, with many bidding wars occurring as authors of the best papers fielded multiple offers. (I can just see a major cardiologist on ESPN in a one-hour interview called, “The Submission,” talking about why he signed a five-year publishing exclusive to the New England Journal of Medicine instead of the Cleveland Clinic.) To offset the expenses involved with paying authors (including the payments themselves, the processing overhead, and the overhead involved with negotiations), the highest impact and most prestigious titles would charge much more, causing their prices to increase exponentially. Smaller journals in less competitive areas would suffer from even greater pricing inequalities than they do now, bundles would become even more divergent intramurally, and subscription models would shred under the strain. Libraries would be forced to devote their shrinking budgets to fewer journals, maybe only a few. And don’t think this would only affect subscription journals. Author-pays OA journals (the most common and successful kind) would cease to exist, as their current financial model primarily hinges on authors paying them. Even funder-sponsored OA journals would face new economics. So, if suddenly the author market expectation were to shift in this manner, OA publishing would come to a screeching halt. The net effect? Less availability of the literature, a few high-end journals pricing everything else out of existence, the end of OA publishing, less access, and less information availability. Yet Maxwell casually tosses off this complaint, without contemplating its consequences, a clear sign of thoughtless bias.
- Publication should include not only reporting, but data, tools, and protocols. Maxwell states “. . . researchers should be able to access the manuscript and its sub-parts—underlying data, protocols, tools utilized for analysis etc.” This is a dream of machine-readable data that users can access with tools of their choice, text that can be translated into other languages whenever desired, and so forth. Let’s take just those two as examples, and ignore the disincentives a research lab might have for sharing underlying protocols or analytical tools. Providing machine-readable data has implications on storage, format, standards, presentation, preservation, provenance, and review that makes such a dream non-trivial in reality. Many research papers now arrive at the publisher with tables that aren’t right, and the Excel files some authors send in are disorganized and erratic. Yet the proposition is to pile on vast sets of raw data? In the same paper, Maxwell talks about the importance of “high-quality” research reports. High-quality research reports require a lot of review, revision, and editing. The authors often have rushed their writing, have experienced internal role-definition problems, and stepped on each others toes. Editors and reviewers help them sort these things out through an iterative and more reflective process. If authors were forced to throw on top of their troubled paper scads of data from different labs with the inevitable compromises and customizations creeping in, you’re looking at a mess. As for translating text into any other language? Because translations are expensive to do well, translations typically only occur if there is a commercial reason to pay for a translation. Because commercial entities would be the most likely translators of scientific studies, more scrutiny of these translations would be wise. Who will provide the scrutiny? Who will pay for an independent reverse-translation to make sure there’s no funny business going on, and/or that quality is sufficient? Who will monitor the copyright? Who will pay for the archiving, monitoring, and updating?
- Copyright remains with the authors. The US Constitution provides copyright protections to authors and other creative types. More importantly, it is a constitutional right specifically linked to the goal of promoting the progress of science and “the useful arts.” The specific laws around its execution and limits are debatable, but its existence is not, nor is its goal as stated in the Constitution. Maxwell seems to lean toward the notion that authors should keep their copyright, but proves himself a bit naïve when he states, “Copyright is protective of authors’ rights to control the use of their own work in someone else’s creation.” This is not entirely true on a practical level. Copyright is not a simple thing to defend in the wild. You need deep pockets. You also really want to have registered your copyright if you want to defend it, because this allows you to claim statutory damages. Otherwise, you have to prove actual damages. How can a researcher prove actual damages from a copyright violation? Did they miss out on tenure? Lose a grant? Miss out on citations? Even if a researcher does register her or his copyright, what then? They have to monitor the wider world for infringements, something that is clearly undesirable and untenable for a productive researcher (and something that definitely hurts the public’s interests if more research supports their interests). If Maxwell’s dream of a free-wheeling translation machine takes off, authors defending a useful copyright goes from untenable to impossible. Copyright itself is not protective except in the abstract — having the wherewithal to fully establish, monitor, and defend it matters a great deal in the real world.
- Faster commercialization of research. This is one of the more novel fantasies in Maxwell’s long editorial piece — that access speeds the commercial application of research results. Odd, I thought funding, manufacturing systems, and entrepreneurs did that. Perhaps the $5 million per year it costs to run PubMed Central would certainly fund a few great studies and maybe even a nice start-up or two. And if we’re connecting Maxwell’s dots and spending all this money paying authors, where’s the funding coming from to drive applied research? Maybe the authors made enough money from their publications that they’re off on vacation and can’t be bothered. In any event, the larger question is, Is the rapid commercialization of research an unvarnished good? The story of the past few decades or so has been about regulators increasing the barriers to commercialization because of increasingly complex science, after learning harsh lessons caused by someone cutting corners, or both. And this hints at the real drivers in the application of scientific findings. It’s not access, but regulation and safety concerns, that throttle commercialization, and perhaps rightly so. After many mishaps with drugs and devices (Vioxx, ICDs, metal-on-metal hips, just to name a few), regulators are being more careful. You can pour all the literature out on the floor, but only a few cups of it will ever be seen by and approved by those groups regulating commercial research markets. If information access were to drive commercial application solely, we would live in a very unsafe and unpredictable world indeed. Maxwell misunderstands the drivers of commercialization — regulation, funding, and commercial appeal. To him, openness seems to be the solution to nearly everything. This doesn’t strike me as pragmatic or practical.
There are other problems with Maxwell’s editorial piece. In one instance, he goes on at length about the virtues of open science, citing a study of an open source approach to a drug for schistosomiasis. The paper he cites is equivocal and quite preliminary in its findings, yet overtly cheerleading in its tone. A related education Web site the paper trumpets apparently lasted two months before being abandoned. Maxwell is uncritical or unaware of these things.
But the fact remains that despite the scholarly shortcomings of this CED report — shortcomings Dylla exposes in a well-reasoned, well-written, and well-cited critique — the “solutions” inherent to the diagnosis are likely worse than the symptoms. In fact, paying authors would drive up the prices and pricing discrepancies around scholarly materials while pushing OA publishers out of business; expecting data and vast translation options to be part of scholarly communications would drive up costs, lower quality, introduce erroneous datasets, or distract authors — likely all four; leaving copyright with authors would create vulnerabilities and distractions, and likely create punishing financial burdens, in cases where copyright needed to be defended; and faster commercialization of research, if it stemmed from access to information alone, would create an unsafe and unpredictable world of snake oils and magic elixirs — just look to the dietary supplements market for a hint of how unregulated science works. (Taken your Airborne lately? It was invented by a teacher!)
Perhaps the problems with this report aren’t all that surprising. After all, Maxwell notes in his response to Dylla that, “I am not and have never been employed by, nor have I represented or have any financial interest in, any STM publisher, for-profit or not-for-profit, subscription-based, open-access or hybrid.” You’re left to wonder what made the CED believe he was qualified to be the sole author of this report. For a contrast, see the recent report in the Chronicle of Higher Education about predatory OA publishers. That’s a report by someone with a healthy dose of skepticism, something science really benefits from. Cheerleading isn’t science’s natural tone.
Not only does the CED report lack a balanced evaluation of the current environment, it proposes solutions that would make scholarship even more expensive, publication more distracting to researchers, and the world we live in less safe. Other than that . . .
1 Thought on "Fleshing Out the World Dreamed Of — What If the Idealistic Recommendations Passed?"
I recently learned a useful German word that keeps coming to mind when reading various schemes to replace current publishing methods:
Verschlimmbesserung: an attempted improvement that makes things worse than they already are.
The current state of publishing is the result of a long evolutionary process. It’s very difficult to replicate that from scratch. Proposed systems quickly get complicated and confusing, and as you note, lead to many unintended consequences.
A few thoughts here:
2. Publication should include not only reporting, but data, tools, and protocols.
When you speak with researchers, they all agree on this point. But when you ask them to spend time (and time is a researcher’s most valuable commodity) writing up detailed protocols, annotating and organizing data for others to use, and writing up instruction manuals for their tools, most balk. That’s a huge time investment that could instead be spent on doing the next rounds of experiments. On every journal where I’ve worked, you also see a minimal use of supplementary material–very few readers will care about the extra details for a given paper.
After the writing process is complete, authors continue to press publishers for increased speed of publication. Speed has become a strongly defining part of the journal selection process. If I have a funding deadline or a tenure/thesis committee meeting coming up, I want a journal that can get my paper out in time to use it for those events. I’d wager that for a large percentage of PLoS ONE authors, the 70% acceptance rate and the assurance that the paper is likely to be published after only one round of journal submission is much more important than the journal’s open access status.
If you require complete protocols/data/tools, you massively slow the writing process and you massively slow the review process.
I would suggest that there is great value in making protocols/data/tools publicly available. The question is whether the published research paper is the right place for this material.
3. Copyright remains with the authors.
Many journals leave copyright with the authors, but license the material for publication. Usually these licenses give the journal the right/responsibility to enforce copyright on behalf of the author. So there is likely a reasonable way forward on this point.