Free online availability substantially increases a paper's impact

Sir — The volume of scientific literature far exceeds the ability of scientists to identify and use all relevant information. The ability to locate relevant research quickly will dramatically improve communication and scientific progress. Although availability varies greatly by discipline, more than a million research articles are now freely available on the web. Here we investigate the impact of free online availability by analysing citation rates. Online availability of an article may not greatly improve access and impact without efficient and comprehensive search services; a substantial percentage of the literature needs to be indexed by these search services before scientists consider them useful. In computer science, a substantial percentage of the literature is online and available through search engines such as Google (www.google. com), or specialized services such as ResearchIndex (www.researchindex.org) — although the greatest impact of online availability is yet to come, because comprehensive search services and more powerful search methods have become available only recently. We analysed 119,924 conference articles in computer science and related disciplines, obtained from DBLP (dblp.uni-trier.de). In these fields, conference articles are typically formal publications and are often more prestigious than journal articles, with acceptance rates at some conferences below 10%. We estimated citation counts and online availability using ResearchIndex, excluding self-citations. The figure shows the probability that an article is freely available online as a function of the number of citations to the article, and the year of publication of the article. The results are dramatic, showing a clear correlation between the number of times an article is cited and the probability that the article is online. More highly cited articles, and more recent articles, are significantly more likely to be online, in computer science. The mean number of citations to offline articles is 2.74, and the mean number of citations to online articles is 7.03, an increase of 157%. We analysed differences within publication venues (the proceedings of a conference for a particular year, for example), looking at the percentage increase in citation rates for online articles. When offline articles were more highly cited, we used the negative of the percentage increase for offline articles: hence if the average number of citations for offline articles is two, and the average for online articles is four, the percentage increase would be 100%. For the opposite situation, the percentage increase would be ǁ100%. Averaging the percentage increase across 1,494 venues containing at least five …