Back when I was a new graduate student, more than a dozen years ago, nearly all scientific journals in my field had a website, but that didn’t mean you could always get the papers you needed online. Often, I had to go to the library with a handful of quarters for the photocopier in order to get the print version of an article that was neither online nor pay-walled. Because this was time consuming, I would only do this for articles I really needed to read. If an article wasn’t accessible online and didn’t seem particularly important, I wouldn’t bother to track it down—and I wouldn’t cite it in my own work.
Publishers of academic journals, whose primary reason for existence is to facilitate communication between scientists, were clearly failing to take full advantage of the Internet to disseminate the papers they published. The situation has improved since then—I haven’t used a library photocopier in years—but, in many ways, academic publishers still aren’t taking full advantage of the possibilities offered by the Internet, especially the more recent developments in social media. Most academic journals, despite having Facebook pages and Twitter accounts, still look and operate much like a traditional print journal that has merely been transferred online.
They also compared papers uploaded to Academia.edu against those freely available online elsewhere, such as a journal site or researchers’ homepage, and found that papers on Academia.edu had 75 percent more citations.
Part of the problem is that nobody knows what uses of social media will be effective. How should we use the Internet to improve on a model of scientific publishing and communication that has been in place for centuries? Many traditional publishers are reluctant to tinker with a system that is working for them, but some researchers and publishers are thinking about how to better use the Internet for scientific communication—and doing what researchers are supposed to do: conduct experiments. They are trying out radical new models for academic publishing that take advantage of the powerful communication tools of social media. If these experiments work, they could dramatically alter how scientists publish, share, and discuss their work.
To know whether an experiment works, you need a way to measure success. In academic publishing the metric is citation, because researchers want others to read, recognize, and build on the work they publish. An early study published in 2001 suggested a simple way that the Internet could have an impact on citations: Make papers freely available online, and they will get cited more often. Examining the citation rates of papers in computer science, the study reported that the advantage of being freely available online was big: Accessible papers were cited nearly twice as often as those that were not. This study coincided with the rapid growth in new, open-access journals that abandoned the traditional subscription-based business model and made all published articles freely accessible. Open access has since become the most widely adopted publishing innovation. Most major scientific publishers now run their own open-access journals or at least give authors the option of going open access in a pay-walled journal for an additional fee. The National Institutes of Health now requires grant recipients to deposit their papers in the National Library of Medicine’s open archive within six months of publication, and the National Science Foundation just announced plans to follow suit.
While the idea of making papers freely available online is catching on, the notion that this leads to more citations hasn’t held up very well. Though several studies reported the citation advantage of freely accessible papers, a 2011 review of this literature reported that these studies didn’t properly account for confounding variables—such as the possibility that more highly cited researchers are more likely to publish in open-access journals—and concluded that “there is little evidence that open access status has an independent effect on citation counts.” This conclusion has been challenged by other studies, but the idea of an open-access citation advantage remains controversial.
Perhaps this shouldn’t be surprising. The biggest innovation of open access has more to do with rethinking the economics of academic publishing than it does with disseminating research. Simply removing the paywall is not an especially innovative use of the Internet. But several radical experiments in scientific publishing and communication are underway, and they explore ways of using social media to help researchers review, share, and discover research papers. Some of these experiments, including Academia.edu and ResearchGate, are social networking platforms designed to make it easier to share work published elsewhere. Other ventures, such as PeerJ and Faculty of 1000, are hybrids that offer a social media platform and also publish new work using an open and innovative peer-review process. And then there’s radical upstart The Winnower, a DIY scientific publishing platform that lets people upload papers and then recruit their own peer-reviewers who openly discuss the paper online.
One of these ventures, Academia.edu, recently released a provocative study showing that there is a citation advantage for research papers shared on its site. Academia.edu provides researchers a place to create a personal profile and upload their published papers, and also offers social networking tools somewhat along the lines of what you find on Facebook and LinkedIn. Researchers there discover and share papers using the site’s equivalent of newsfeeds and retweets—in other words, as the site’s founder Richard Price put it to me, Academia.edu offers a “push network with viral properties” that lets researchers re-share favorite articles.
In their study, the Academia.edu team looked at citation rates for 34,000 newly published papers uploaded to their site between 2009 and 2012 and compared them with a random set of papers published in the same journals during the same years but not uploaded to Academia.edu. The results were dramatic: Three years after publication, uploaded papers had, on average, 37 percent more citations, and, after five years, 85 percent more. They also compared papers uploaded to Academia.edu against those freely available online elsewhere, such as a journal site or researchers’ homepage, and found that papers on Academia.edu had 75 percent more citations. They used a statistical model to estimate the impact of some of the confounding variables that have hampered other studies—such as researchers uploading only their best papers—and found that the citation advantage still held up.
The Academia.edu team was duly cautious about overgeneralizing from these initial results. Since they did not directly compare their site with any other single site that offers similar features, it’s difficult to say what social networking tools produce a citation advantage, and whether that advantage is unique to Academia.edu. But they hope to provoke a discussion that moves beyond the impact of simply making research papers available online and begins to explore “what features of open access repositories or databases make articles easier to discover, and to what extent that leads to increased citations.”
It’s a timely question. Most scientific papers are still published and shared through relatively traditional journals, and the substantial communication advantages of social media are largely untapped. The current publishing experiments are fascinating, but to truly succeed they need to address two big questions: First, do they make academic publishing work better—do they help researchers distribute, discover, and evaluate research more effectively? And second, can they get researchers to actually use them? If there really is a robust citation advantage to platforms that make better use of the power of the Internet, then researchers may flock to them. If that happens, the effect could be truly transformative, as important work by less well-known researchers becomes recognized sooner while flawed or fraudulent work is exposed more quickly. This process of sorting sound claims from specious ones is central to science, and it will improve as academic publishing finally embraces the modern communication tools offered by the Internet.
Inside the Lab explores the promise and hype of genetics research and advancements in medicine.
