Science Publishing Is Urgently In Need Of Major Reform

VOLUME 7, No:17 The Scientist September 6, 1993 Opinion Science Publishing Is Urgently In Need Of Major Reform AUTHOR: RUSTUM ROY, pp.11 The function of science publishing today is to get information about new findings in science to at least three different communities: Group A, the specialists working in the same field as that in which the findings were made (numbering anywhere from 10 to 1,000 scientists); Group B, the general community of scientists and engineers who, although not in that field, are nevertheless interested in major advances in scientific areas other than their own--advances that may, indirectly or in the long term, be significant to them (probably between 1,000 and 10,000); Group C, the general, attentive public and the policymakers who want or need to know of scientific developments that could have economic or social consequences (10,000-100,000). The classic media for science publishing--journals put out by societies and other discipline-oriented organizations--were designed only for Group A, while the media serving Group B are the short news articles in society house organs, such as Chemical Engineering News, Physics Today, and MRS Bulletin, and in such multidisciplinary publications as Science and Nature. Group C is served (very poorly, in my opinion) by the general media--including newspapers and magazines such as the New York Times, Omni, and Discover--which tend to overamplify and inappropriately dramatize the "breakthroughs" they consider newsworthy. Since maximum publicity--not accuracy--today serves both the scientist and the journalist, these and other such publications, in my opinion, tend increasingly to fall for exaggeration and hype. Meanwhile, such publications as New Scientist and The Scientist seek to bridge with both accuracy and social relevance the B and C groups. Confronted by such a menu of publishing alternatives, how does the responsible scientist announce effectively a discovery that she or he thinks would be of interest to one or all of the A, B, and C groups? The Establishment Position The route accepted by the establishment is that one should submit one's research report to a standard journal (the A group), have it go through the arcane ritual called "peer review," and--only after it has been accepted and published--send publicity releases or other notification of its findings to the wider press. The process is deficient, in my opinion, first of all because peer review--the first hurdle--is biased against innovation. In the path toward publication in the A-group journals, anything that extends beyond the current ruling paradigms of scientific thought will most likely experience inordinate delays in being considered; matters of routine science will get into print much faster. (There is an accompanying danger: that during the peer- review process, especially when an article is dealing with a dramatic discovery in a fast-breaking field and has potentially profitable application, two or three scientists--the so-called peers--and, hence, possibly their laboratories, companies, and colleagues may inappropriately be privy to advance information on a significant innovation that they can use to their own professional gain.) I strongly believe that when research results are really important, they deserve a wider academic audience than the specialized readers of the A-type journals and that, in line with this, the significant scientific innovations of today need new methods of publication. Beating The System A good example of how a creative scientist can circumvent the current publishing system in order to get his or her significant work out to the public is the case of Henry Heimlich, president of the Heimlich Institute in Cincinnati. Although Heimlich does not fit the traditional profile of the basic lab researcher working toward an esoteric advance, his approach is worthy of emulation: When he discovered his now-famous life-saving maneuver for discharging obstructive material from the windpipe, he wanted to get news of it without delay to the public who could use it. After agreeing with the editor of Emergency Medicine (a non-peer- reviewed journal) to simultaneously send out a story to the general press, he published it in the June 1974 edition of that publication (6:154-5, 1974). Newspapers all over the United States picked up and ran the news release sent to them. Although the Journal of the American Medical Association made note of Heimlich's technique in a news blurb a few months later, the technical article in JAMA did not appear until October of the following year (234:398-401, 1975). By this time, of course, the Heimlich maneuver was already saving lives all over the world. For the bench scientist, the effort to penetrate or circumvent the peer-review barrier in the way that Heimlich did can be perilous indeed to the prospect of ever seeing a research report in print. I know this from several personal, time-wasting experiences with leading science journals. For example, in October 1992, having made a scientific advance in diamond synthesis, I sent off a paper to Science, then held a press conference in Washington discussing my findings. At the press conference, I handed out copies of the paper, and I also distributed some copies by fax. After a month, Science returned the paper--unreviewed, yet rejected--on the procedural ground that preprints had been distributed, which, in Science's terms, constituted publication. An Obsolete Process For the A group, the classic model of writing a paper, submitting it to an editor, and having it go through the arcane peer-review system described above may once have been suitable for science done by an elitist minority (known to each other) with relatively high ethical standards. But the A community itself is no longer well served by this process because of several factors--and the process is clearly inadequate to address the needs of groups B and C. Among the factors are: The volume of the current literature is so large that peers and editors have greater difficulty in determining whether work reported is really significant or has been done before. (Indeed, I believe that significant numbers of working academic scientists essentially ignore the literature as they go about their investigations.) The peer-review process as now practiced in most journals is biased against real innovation, in part because of problems associated with selection of knowledgeable and objective reviewers. Even the key word in the term "peer review" is difficult to define: Who, for example, is a "peer" of Linus Pauling? Any assistant professor of chemistry at a university? His arch-enemies in the anti-vitamin C camp? Yet, although no clear definition of "peer" has as yet been created, the term continues to be used by the science publishing industry as its needs require. Most devotees of the process forget that the peer- review system cannot be any guarantor of correct or good science; all its practitioners really can do is check the plausibility of the case and determine whether the author got the sums right. The current process in a peer-reviewed journal also gives an unfair advantage of from three to six months in a given field to two or three groups or individuals who are chosen to review an article. This may be no problem for pedestrian science, but it is unacceptable in many cases, especially those of interest to groups B and C. Devotees of peer review commonly assert that there are no alternatives to the system. For research proposals, however, there are dozens of alternatives in use all over the world. At the U.S. Department of Defense, for example, a knowledgeable R&D manager seeks out the very best researchers and takes the responsibility for judging the best proposals with no authority at all being given to peers. Every mid-level or higher manager in industry does the same every day with no help from peers. For publishing papers, the science community, in theory devoted to "experimentation," often appears unwilling to try new approaches for the new situations, although there have been exceptions. In 1965, I participated in the founding of the Journal of Materials Research Bulletin. This journal, which has a good impact factor in citations, runs open review either by associated editors or by anonymous reviewers. Until recently, the Proceedings of the National Academy of Sciences did not mandate peer review. Quality presumably was maintained by examining an author's previous track record. However, this is no longer true for PNAS. Recommendations I believe that new journals should be started (or new sections started in established journals) for non-reviewed papers by authors who establish their credentials by their track record. In this process, a scientist who submits a list of published works with a minimum of, say, 50 or 100 papers in the regular peer- reviewed journals could publish a paper with only a brief delay in the process for copy editing. Or any scientist with the same or an even better track record could officially communicate a colleague's paper. This scheme would achieve several goals at once. First, it would open up the system to genuinely new ideas. Second, it would speed up the process by several months. Third, it would bring some lively debate into science by having authors use the media Group B to discuss or critique the work. Fourth, it would protect the new work from some of the dangers inherent in the current system--such as the prospect of the paper's never being published. In the case of a paper reporting on what the author considers an especially impor- tant research advance that would be of interest to audiences B and C, as well as A, standard practice should entail the author's first sending the manuscript to a journal, then distributing preprints by mail and fax, or calling a press conference or doing whatever is necessary to get the word out. Meanwhile, reporters covering such an announcement should examine carefully the claims and explicitly report the author's previous track record in the field. This is by far the best predictor of reliability. And the reporters should seek comment on the record--in radical contrast to the anonymity of peer review--from other knowledgeable colleagues on the available preprint version. The reporters should be skeptical of any claims the paper makes concerning applications of technological or economic impact, preferably ignoring them. Rustum Roy is Evan Pugh Professor of the Solid State at Pennsylvania State University's Intercollege Materials Research Laboratory in University Park.

Science Publishing Is Urgently In Need Of Major Reform

Opinion

Science Publishing Is Urgently In Need Of Major Reform

Interested in reading more?

Become a Member of