Web citation persistence over time: a retrospective study

Purpose - By studying a large number of citations in the US field, this paper seeks to examine carefully the persistence status of web resources specified by their domains and type of files. Design/methodology/approach - All 2005-2008 volumes of six LIS journals ranked by ISI Thomson Reuters were selected. From 1,181 papers, 37,791 citations were recorded. Only original articles, which had a list of references, were included in the study. The persistence of web citations was checked by directly following the cited URLs. Findings - Of the 37,791 citations, 4,840 (12.8 percent) were web citations. The means per articles of web and print citations were 4.09, and 27.9, respectively. Of all web citations, 4,617 (95 percent) were readily persistent, and 5 percent returned errors and thus were not originally accessible. The relationship between the print and web citation over time (year) was significant. The most prevalent domain of citations was html and the most favorable and persistent file format was pdf. Practical implications - The web resources are used for their easy accessibility and the support they provide for a scientific content. While direct accessibility to a web citation is not provided, many strategies are adopted to recover the dead citation. The issue is to what extent the authors rely on web resources and are they finished with citing paper-based materials? Are web resources becoming replaced with their print counterparts? The study showed that scholars still rely more on print resources than on the web materials. Originality/value - Tracking current trends in scholars' communication behavior shows a shift from print to web resources. The paper examines web citations persistence in some prestigious journals to show whether the web citations are reliable enough and always accessible in the digital world.