A recent article entitled “Electronic Publication and the Narrowing of Science and Scholarship,” in Science (v.321; 395-399, July 18, 2008) by James A. Evans of the University of Chicago’s Sociology Department shouts out to the library acquisitions, journal publishing, and database/aggregator communities something they are not likely to want to hear aloud, even though they may all secretly whisper it among themselves.
The message is this: Adding electronic back files of journal runs in the sciences hardly increases the chances that the articles within them will be read. In fact, the opposite appears to be true. The more you provide by way of deeper back files of journals in a wider variety of fields, the less likely it is that they will actually be cited.
This is not because the older content is incorrect, but because of the changed way at which scientists arrive at their personal research agendas, and the changed way in which they calculate that they have a big enough safety net of journal articles, so that they feel they know all they need to know in order to proceed.
A lot of romantic folklore, some still re-told in library skills instruction classes about how scientific discoveries are based on a combination of serendipity and a deep, sound traditional literature search, are going to lose even more of their dwindling band of believers.
The Decline of Journal Browsing as a Means of Shared Awareness of a Wider Research World
Browsing a fairly wide variety of journals in print, and even now electronically, encouraged scientists to be aware of a wide range of ongoing research activities in not only their own fields but in some reasonably affiliated disciplines, a trait that was thought to be one of the hallmarks of a competent scientist.
But browsing in the electronic journal world is dying habit. This is despite the fact that publishers are trying to make their journal’s home pages more colorful and animated. They incorporate RSS feeds of headlines in the discipline, feature interviews with leading figures, offer streaming videos, and most common of all, they highlight what the editors regard as this issue’s “hot articles. ”
But study after study shows the most commonly used feature on the page is invariably the Search button. Despite the fact that current issues of journals in electronic formats can be navigated very easily, they are now being used to search narrowly rather than to explore panoramically. Never before has it been easier not to notice what else is going on, and still be considered a very good research scientist in terms of being current with the literature that actually matters to getting a job done.
Ironically the fewer journals which do get browsed, assume an ever-increasing importance despite the fact that they are less used for perusal, because they now also function as de-facto first choice portals to electronic links in other journals.
There May be Deans of Interdisciplinary Research on Campuses Everywhere,
But They’re Going to Have Less & Less to Do With the Articles That Actually Get Read by Anyone
The often heralded hope for inter-disciplinary academic Nirvana furthered through cross-browsing or even direct online searching of the home pages of journals covering the fields of others, is therefore also becoming a dying art, and was probably actually still-born in the first place. Library users typically ignore the platitudes and do what is practical and less complicated. This is demonstrated by the fact that while more scientists have access to more journals in more fields than ever before, today’s journal citations actually have markedly less scatter to journal’s outside their primary target fields, and therefore more concentration on the fewer titles that are believed to be most prestigious in their particular field. The more that’s available, the more important are the key arbiter journals.
After having scanned and searched their own small number of favorite journals, scientists self-search databases for additional articles for research projects they are considering. And today, an absolute abundance of articles can be found, and most importantly, found rather easily, even if the initial result is not always as precise an answer set as might be found in the old complicated Boolean search, using carefully controlled vocabularies. Even after being told this, most self-searchers just don’t care, because, as will be seen, online search culture has evolved its own definition of what constitutes a good and sufficient search, and this new paradigm is accepted by their peers, even if not always by their librarians.
The very abundance of results that comes with many searches has driven browser makers and even some traditional indexing abstracting firms to do instantaneous relevancy ranking before the searcher even asks for it, so as to save the researcher time, and reduce their despondency at having to plow through so many articles.
Browser makers have clearly seen that searchers simply start with those results at the top, and go on down only until they feel they’ve found enough. Browser makers do not, like many bibliographic instruction librarians, hector the customer to be as thorough as possible “for their own good.” Browser makers understand that they are rated for ease of use, not for the depth of character formation they encourage in their users.
While multi-specialty databases like Google Scholar and the Web of Science would seem to have the effect of enabling searchers to look beyond their disciplinary borders, the fact is that both proprietary and open-source searching/indexing search engine weighting algorithms zero in on the common language and buzz words of subspecialty fields and niches, and post them first. Unless all the possible fields use the same terminology and juxtaposition within the text or titles or abstracts in pretty much the same way, articles from other disciplines are going to fall farther and farther down the results list and probably never get looked at. And the customers simply do not care, because finding everything no longer seems important.
Dissertation Style In-Depth Searching Has Been Killed Off by “More Results Like This”, “See Related” & CrossRef ‘s Capabilities
Traditional online searching led the scientist back as far as the database went, and then the scholarly tradition of the field was such that one then searched paper indexes and abstracts until virtually all the relevant, and particularly historically classic papers had been found. Literature searching was about time travel to original sources from which you could document the progress made by successive researchers up to the point of providing a rationale for one’s own current proposed research project.
This was a tradition fostered by the extensive process of writing up a dissertation proposal and reworking the justification into an extensive literature review that made up a substantial portion of the resulting dissertation. But the dissertation per se is increasingly a dead letter. Today, in some schools, the graduate student simply prints off about five articles that have appeared in journals that really matter, and on which he or she is an author or coauthor, and has them bound as a kind of ersatz dissertation without all the 19th century procedural crap involved. Who cares what the committee thinks if the papers are already published in Cell or Science? Certainly it’s not the committee at those schools because they likely served as coauthors on the papers. Perhaps the Dean of the Graduate School is dubious, but what brings in the grant money and makes the reputation of the school, rules or results? (This is not an entirely new development. Linus Pauling put together 5 of his reprints in 1925 and told Caltech to take it or leave it. They took it and eventually got credit for having given a Ph.D. to a Nobel Prizewinner.)
Today, literature searching is about finding whether or not there is a consensus of other contemporary scientists working in an area about what articles seem to matter. This tends to be driven not only by the search engine using a shared vocabulary algorithm in their weighing of results, as mentioned above, but by the search engines’ concurrent provision of weighing search results by shared patterns of citations. In other words, articles that cite the same or similar articles are quickly pointed out to the searcher, by being put at the top of the results list.
The searcher then typically stops drilling downwards through the strata of results and starts building a platform of contemporary articles that seem to matter. This can be done in two convenient ways.
First, many databases now flat-out tell the searcher that with a single click they can get “More Results Like This”, or “See Related.” These are overwhelmingly recent articles that cite other recent articles.
Second, with so many search engines enabling the searcher to directly link to the full text of a promising article in a search, the searcher will often simply starts reading a single promising article and through hypertext CrossRef links within the article, navigates to more and more articles that share citations. Once today’s searcher has found enough papers that seem to be centering around a common core of key published papers over and over again, the searcher tends to assume that he or she has found all that really matters, and that the rest is tangential.
The More Journals You Invent The Less Likely It Is That They Will Matter
(Unless they’re from the Nature, Cell & Maybe The PLoS families)
Despite the lack of faithful browsing of a wide variety of journals, and particularly in the face of all sorts of new Open Access journals that are supposed to bring down what some critics of for-profit publishing call the Evil Empire(s), the fact of the matter is the final arbiter of what gets counted by working scientists today to be a really worthwhile paper depends to a greater degree than ever on its being published in a handful of journals. By this I mean journals like Science, Nature, PNAS, the leading journals of the disciplinary societies, some of the for-profit journals that compete with those society journals, and in particular the members of the extended Cell and Nature families.
The talk that having libraries support online institutional archives that preserve and e-publish in advance a school’s vast output is feckless in light of the way searching and consensus building about what papers matter is done today. The larger the number of choices that a scholar has show up in a search, the more dependent he or she is on relevancy ranking and on the newer working style of “enough is enough,” rather than on “everything that can conceivably found in history is essential,” and the more likely it is that reliance on a close circle of sources universally recognized today as excellent is going to be that final arbiter.
So stop and consider before you expand your holdings of electronic back files, particularly in any quest for accommodating increased multidisciplinarity. It may well be good for some slower moving fields that preserve a kind of in-depth searching (Paleontology and Natural Products Chemistry are two that come to mind), but in today’s molecular and cellular biology, and in a lot of clinical medicine as well, you’re paying for something that is unlikely to be used, even if using it is easier than ever. Because what matters in literature searching now has passed them by.
Tony Stankus, [email protected] Life Sciences Librarian & Professor
University of Arkansas Libraries MULN 223 E
365 North McIlroy Avenue
Fayetteville, AR 72701-4002
Voice: 479-409-0021
Fax: 479-575-4592
Are Online and Free Online Access Broadening or Narrowing Research?
http://openaccess.eprints.org/index.php?/archives/443-guid.html
Excerpts: Before OA, researchers cited what they could afford to access, and that was not necessarily all the best work, so they could not be optimally selective for quality, importance and relevance...
When everything becomes accessible, researchers can be more selective and can cite only what is most relevant, important and of high quality. (It has been true all along that about 80-90% of citations go to the top 10-20% of articles. Now that the top 10-20% (along with everything else in astrophysics), is accessible to everyone, everyone can cite it, and cull out the less relevant or important 80-90%...
Are online and free online access broadening or narrowing research? They are broadening it by making all of it accessible to all researchers, focusing it on the best rather than merely the accessible, and accelerating it.
Posted by: Stevan Harnad | August 04, 2008 at 06:40 PM