The turn to online research is narrowing the range of modern scholarship, a new study suggests
FOR SCHOLARS - ESPECIALLY scholars who like to wear pajamas - the Internet has been a godsend. It allows instant communication with colleagues around the globe, and makes tracking down published research a matter of seconds.
But perhaps the greatest boon is the sheer quantity of readily accessible knowledge. Millions of journal articles are available online, enabling scholars to find material they never would have encountered at their university libraries. From classic psychology studies to the most esoteric literary theory, it's all just a few clicks away.
A recent study, however, suggests that despite this cornucopia, the boom in online research may actually have a "narrowing" effect on scholarship. James Evans, a sociologist at the University of Chicago, analyzed a database of 34 million articles in the sciences, social sciences, and humanities, and determined that as more journal issues came online, new papers referenced a relatively smaller pool of articles, which tended to be more recent, at the expense of older and more obscure work. Overall, Evans says, published research has expanded, due to a proliferation of journals, authors, and conferences. But the paper, which appeared in July in the journal Science, concludes that the Internet's influence is to tighten consensus, posing the risk that good ideas may be ignored and lost - the opposite of the Internet's promise.
"Winners are inadvertently picked," says Evans. "It drives out diversity."
This study adds weight to concerns, shared by other Internet analysts, that the rise of online research has costs as well as benefits. Internet search tools are not neutral: they tend to privilege the new and the popular. And for all the frustrations of older research methods, their very inefficiency may have yielded rewards. Leafing through print journals or browsing the stacks can expose researchers to a context that is missing in the highly targeted searches of PubMed or PsychInfo. The old-fashioned style of browsing, some say, can provide academics with more background knowledge, and lead to serendipitous insights when they stumble upon articles or books they weren't necessarily looking for.
Yet there is vigorous debate over the Internet's effects, and the Evans research has proved controversial. A University of Quebec researcher, Vincent Lariviere, has coauthored a forthcoming paper that challenges some of its conclusions. (Evans plans to publish a rebuttal.) Another researcher, Carol Tenopir at the University of Tennessee at Knoxville, says that she has not studied citations, but that her surveys of reading patterns show the reverse of a narrowing effect.
"Electronic journals, I can say with confidence, have broadened reading," says Tenopir.
This debate has important implications for the academic world, but it also has wider significance. More and more, the Internet dominates everyday life. Our daily experience - what we watch, listen to, and read; the people we date and the friendships we maintain - is increasingly shaped by the vast information landscape of the Internet, and how it is filtered for personal use.
Different interpretations notwithstanding, many experts agree that we have only begun to understand the repercussions of our Internet consumption. This dim awareness, despite the interactive ethos of Web 2.0, leaves us more passive than we may feel, in the grip of seismic change.
"We have an opportunity to maximize the good effects, and minimize the bad effects," says Katrina Kuh, a law professor at Hofstra University. "We're missing that opportunity."
. . .
Inevitably, the discussion of these questions turns to the theory of the "long tail," articulated by Chris Anderson, editor in chief of Wired magazine. Anderson's argument focuses on consumer choices, positing that the Internet reduces the "blockbuster effect" - whereby consumers settle on a few big hits - and disperses attention over a wider range. Web users gain access to obscure books, movies, and other products that might have escaped their notice without the Internet, and that conventional "brick-and-mortar" stores wouldn't find it worthwhile to stock. As a result, the theory goes, consumers can satisfy their idiosyncratic preferences rather than following the herd. This is a key part of how we think of the democratic potential of the Internet: We each find our own niche, hierarchies are abolished, and diversity thrives.
Yet in a recent article in the Harvard Business Review, Harvard Business School professor Anita Elberse challenged that premise. Examining recent music and DVD sales, she found greater concentration, not less. For example, from 2000 to 2005, the number of titles in the top 10 percent of weekly home video sales fell significantly, by more than 50 percent. Her paper concludes: "The importance of individual best sellers is not diminishing over time. It is growing."
This view tallies with Evans's observations about scholarship, and those of several other analysts about the Internet more generally. The explosion of online materials has two, somewhat contradictory effects. The scope of available information expands, remarkably so; but as a consequence, the information needs to be filtered somehow.
To make sense of this overwhelming sea of data, search tools must present results in some kind of order. Scholars, like other Internet users, rely on tools that rank results primarily in two ways: in reverse chronological order, and by popularity. (Google's algorithms, for example, take into account the number of times a website is linked from other websites.)
Social Science Research Network (SSRN), the widely used Internet resource, offers lists of "top papers," "top authors," and "top institutions." A paper titled, " 'I've Got Nothing to Hide,' and Other Misunderstandings of Privacy," by George Washington University law professor Daniel J. Solove, has hovered at the top of the rankings for months, with a total of 65,846 (and counting ) downloads. If you click on a given paper, you can even see a graph depicting the paper's "raw score" (total new downloads) over time. On many other academic websites, it is standard to present the newest articles first.
These search tools clearly have the potential to open up research. Sean Franzel, a professor of German studies at the University of Missouri at Columbia, studies the effect of different media on scholarship. In his own experience, online searches often bring up results from minor journals he never would have thought to consult. In this way, Internet use "takes you further afield than you otherwise might have gone." And many Internet users protest that online serendipity is certainly possible, indeed common. Just as researchers may come across unexpected articles in a table of contents, they might see intriguing but not directly relevant articles in lists of top papers and sidebars of related papers.
Even the narrowing effect Evans diagnoses can have advantages. There are benefits to sharing common knowledge and reference points. In this sense, online winnowing could restore the "water cooler" culture of which the Internet and other technologies have supposedly robbed us. In scholarship, convergence facilitates communication and progress. As Evans says, it's "not so different from the effect of shared language."
But several observers perceive losses as well. According to Alex Bentley, an anthropologist at Durham University in England, this tendency "makes academic research a popularity contest. My hypothesis is that the way that we latch onto ideas is going to become more fashion-based."
Naturally, papers that are ranked high in "top papers" lists are more likely to get downloaded again, in part because it's easiest, but also because their position enhances their legitimacy. "When people become more aware of each other's choices, they factor those choices into their own activities," says Evans. One threat is that these decisions can accumulate to amplify an initial choice that might have been arbitrary.
Some scholars lament other lost aspects of print resources. Indexes and tables of contents provide a context, giving a broader snapshot of the field at a given time. And after consulting them, the researcher must make a considered decision to take the next step. By contrast, in online searches, the researcher tends to follow hyperlink to hyperlink, in a journey that resembles "a plunge down a rabbit hole," in the words of Robert Berring, a law professor at the University of California at Berkeley who has studied the impact of electronic media. "If you get to an index, a table of contents, you see the environment that surrounds it. In the culture of paper, a lot of these signals are important."
If the narrowing effect is real, what is to be done about it? One possibility is that online tools will emerge to counteract it. Evans himself is working on developing a system that would use sophisticated statistical analysis to find papers including statements that agree or disagree with other statements, "rather than treating papers as bags of words," Evans says.
Some experts are skeptical that innovative search tools will enter widespread use soon. Instead, they say, it's up to individuals to be more conscious of the limitations of the current tools.
This caution has far-reaching importance. Harvard law professor Cass Sunstein has written about the ways that, contrary to mythology, the Internet can have detrimental influences on democracy, as people retreat to their virtual bubbles.
Many Internet users customize their consumption of news sources and other information in a way that fosters polarization. This, it could be argued, has elements both of the narrowing effect and the long tail. Americans seek out sources that reflect their personal beliefs, consistent with Anderson's vision. But, akin to the narrowing Evans observes, large groups - liberals and conservatives - converge on different reference points, resulting in mutually unrecognizable versions of reality. The common lesson of all of these phenomena is to be cognizant that the tools we use affect us in ways we may not fully appreciate. We should always be searching, the findings suggest, for new ways to search.
Rebecca Tuhus-Dubrow is a contributing writer for Ideas. She can be reached at firstname.lastname@example.org.