Sci-Hub and LibGen: what if… why not?

Abstract

Open Access is becoming the predominant form of getting access to peer reviewed articles. Many new non-traditional tools (institutional repositories, social media, and peer to peer sites) are available out there to retrieve the full-text of peer reviewed articles. Since the launch of Sci-Hub (2011) and Library Genesis (LibGen), several criticisms have been raised both from the library and publishing worlds. Some court decisions have also been rendered to block the access of such platforms in different countries. What can be said in terms of content, coverage, reliability, stability and currency of both Sci-Hub and LibGen?

In this era of severe budget constraints and economic recessions that libraries are facing, can we imagine of substituting some or most of our journal collection funds with the “open and free access” that Sci-Hub and LibGen is giving us? How much overlap between our collections and what is available through Sci-Hub and LibGen?

This article reports on preliminary results1 of a one year study with Sci-Hub, LibGen and of Google Scholar (GS) where 2,750 random samples (peer reviewed journal articles) coming from fifty-five different databases covering all disciplines (Arts & Humanities, Law, Music, Social Sciences and STM) were tested against those platforms. The samples have been searched on all three platforms at four different intervals during the year in order to evaluate the stability of content. Different data such as publication year, publishers, language of articles and OA are being looked at to see if content is affected by either or all of these parameters. To verify the currency of information in Sci-Hub, LibGen and Google Scholar, research articles from both Nature and Science (from current issues, Nature Advance Online Publication and First Release from Science) were queried on a daily basis.

Results are showing overall retrieval rates of 70% in Sci-Hub and 69% in LibGen across all disciplines. Most of Nature and Science new research articles were found within the same day or 24 hours from their first release on their respective web sites. This is also true within Google but strangely enough, a delay of between 4-5 days is seen in Google Scholar.

One conclusion that could be drawn from the present results is that the content is there in Sci-Hub and LibGen. Should we continue to ignore them totally?