Cited Reference Variations

Each year, Thomson ISI processes more than 1 million articles from 8,500 journals for the three citation databases. More than 25 million cited references are captured from the reference lists or bibliographies of these articles.

When a cited reference is captured, it is condensed, and, to some extent, its component parts are unified. No attempt is made to include all of the bibliographic information that may be supplied in a reference. Rather, the goal is to represent the cited reference by a string of characters that, taken together, uniquely identify the cited work.

Different authors may cite the same work in different ways. For example, the author of one article might cite a particular page number and edition of a book, while the author of another article might cite only the book's author and title. Other common variations include differences in author initials, volume numbers, page numbers, and cited publication years. In the citation database, each version of the reference will appear as a separate reference.

It would be impossible for Thomson ISI to consult citing authors and journal editors about the accuracy and completeness of every one of the 25 million references captured from source articles every year. Even changing transposed digits in a volume number without such verification would lead to error.

Consider these references:

You might assume that these are two variations of a citation to the same article, when in fact they are references to two different articles published in the same issue of Journal of the American Chemical Society.