Much has been written about citation indexing since Eugene Garfield's seminal article in 1955, "Citation Indexes for Science: A New Dimension in Documentation through Association of Ideas."1 In the 1960s, the discussion was about Garfield's Science Citation Index and later its successor, Web of Science. For almost half a century, these were the only available tools for tracing scholarly discourse forward in time. Now there are two new tools. SCOPUS is a commercial product from Elsevier and was launched in November 2004. Google Scholar also arrived on the scene in November 2004. Because Google Scholar is freely accessible from the Google site, students and faculty are finding and using it. They are beginning to ask librarians for their professional opinions of its efficacy. Practicing reference and instruction librarians need to understand the strengths of both Google Scholar and Web of Science so that they can appropriately recommend them for use by their patrons—whether they are undergraduates, graduate students, or faculty.
Since April 2005, 10 studies have been published that directly compare the citation features of Google Scholar to those of Web of Science. In an effort to better understand the advantages and disadvantages of these tools, the author undertook an investigation of this research. Although the parameters of the studies varied greatly with respect to the disciplines, dates, and sample sizes, analysis of these preliminary studies can help us gain an initial—albeit sketchy—impression of the relative strength and weakness of each service. The results of this effort are reported here as an informal meta-analysis, followed by some recommendations for utilizing these tools across a range of needs.
The coverage of each database is both overlapping and complementary. Web of Science is comprised of a known list of highly prestigious journals. The extent to which disciplines are covered is well known, but it is limited to these journal holdings. Although Google Scholar overlaps with Web of Science in some of its coverage, it also includes conference proceedings, books, preprints, and a variety of versions of articles available in open access databases and institutional repositories. Search results do, however, include many "false hits"—non-scholarly sources or titles that are similar to, but not [End Page 243] exact matches for, the target citation. These spurious links cause the researcher to spend much more time analyzing and evaluating the sources.
Date coverage is also complementary, with Web of Science consistently providing older articles. Google Scholar often returns more current results due to its ability to access early versions of works in progress and open access articles available on the Internet. Although the controlled vocabulary in Web of Science is less than perfect, it wins hands-down over Google Scholar, which is totally lacking in any of the finer points of indexing. [End Page 244]
The findings from these initial studies suggest a variety of uses for both of the databases' citation features in academic reference and instruction. For instructing undergraduate students in the use of the "cited by" features, Google Scholar's lack of advanced search functions may actually be a boon. Most students today are familiar with the look and feel of the "one box searching" of Google. Having a similar uncluttered look, Google Scholar appears less intimidating to novice users, allowing them to focus on the concepts involved in citation searching and analysis. Google Scholar acts as a bridge from the known quantity of Google to more advanced instruction required for Web of Science. Discussing the strengths and weaknesses of both databases is a good way to begin a dialog about when Google Scholar might be an appropriate tool [End Page 245] for research and also allow librarians to make a stronger case for why other databases such as Web of Science need to be used as well.
For many of these same reasons, Google Scholar...