Magazine article Online

ISI Web of Science, Scopus, and SPORTDiscus

Magazine article Online

ISI Web of Science, Scopus, and SPORTDiscus

Article excerpt

My two picks, the newest version of ISI Web of Science (and ISI Web of Knowledge) and the new Scopus database, represent the state of the art in indexing/abstracting databases. Not only are the databases huge and interdisciplinary, but both are also endowed with cited references, make superb use of citation indexing, and facilitate citation searching, which reduces the need to rely on controlled vocabulary. The pan is SPORTDiscus, which is huge (for a single discipline database), but in an unhealthy way, due to the duplicate and erroneous records that are not removed even after correction. It contains a befuddling thesaurus, which illustrates how frustrating controlled vocabulary searching can be with an inappropriate tool.


This summer, Thomson Scientific's ISI Web of Knowledge (WoK) platform and ISI Web of Science (WoS) service got an "interfacelift" that would make Botox aficionados green with envy. The databases also received an overall rejuvenation treatment, with a couple of substantial brain-booster features. One improvement, the Analyze feature, allows users to rank a results set of up to 2,000 records by author, institution name, source title, publication year, subject category, language, or document type, and to see the results in a tabular format along with a histogram. The distribution figures instantly and clearly show the most productive authors, institutions, journals, and subject categories for the topic of your search in a way that would have been appreciated by Bradford and Lotka (who formulated theories about the scatter-distribution of information in scholarly journals and the concentration of authors writing about a specific subject). I can only wish for one more enhancement--an option for analyzing the set by cited journals, cited authors, and cited year.

The other very important cerebral option is the Related Items feature. This now shows not only the list of papers that have common citations with any of the records selected by the user, but also gives the number of common (shared) citations, which can also be displayed. These features will make even the most technophobe users sit up and pay attention. ISI Web of Science demonstrates the ideal combination of power of use and ease of use. The acquisition of BIOSIS by Thomson and the licensing of scholarly databases, such as INSPEC, FSTA, CAB Abstracts, and PsycINFO, present especially promising potential if integrated with WoS on the much-enhanced ISI Web of Knowledge platform.


I tested only the beta version of Elsevier's new mega database, Scopus (the official launch of the product is scheduled for November 2004). It has about the same number of records as WoS has for the 1975-2004 time period (approximately 27 million), fused from four sources: the abstracting/indexing (A/I) databases of Elsevier and MEDLINE, the ScienceDirect archive of the Elsevier journals, some open access Web components of Scirus, and its partner publishers' archives. Enhancing the Elsevier A/I databases with cited references must have been a massive project. The result shows splendidly in the novel presentation format that I have been awaiting for many years.

The grid format of the short results list is conducive to quick scanning and, most importantly, the results matrix prominently shows the number of times the paper was cited by journal articles whose records appear in one or more of the component resources of Scopus. The grid can be re-sorted almost instantly by any of several bibliographic data elements, including the citedness count of the items. Sorting by citedness count (and some other elements) is also possible in WoS. However, the citedness count is not displayed in the WoS short results list, only in the detailed record format. This layout in Scopus will motivate users to choose the most-cited documents. Obviously, an article published 4 years ago will likely have been cited much more often than an article published a year ago, but it could be offset by selecting another column to show the adjusted (relative) citedness count. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.