|
The decision not to include citations to books and to other non-indexed
items in the analysis for ISIHighlyCited.com centered around the need
for complete and correct bibliographic information to ensure accurate
citation attribution. The analysis is limited to items indexed by ISI
within the specified time-interval, and citations to those items, also
limited to the specified time interval. (For more information on our
analysis see: How do we identify Highly Cited Researchers [link to
essay]?) Thus, for each citation considered, we have a fully indexed,
linked source item.
For books, the bibliographic information presented in an
article's reference list often contains first author only, or a partial
list of authors (eg: Smith, JA et al.). Linking to the source item
allows us to expand the reference to include the complete list of
authors. Lacking this expanded information would decrease the accuracy
of our analysis by under-representing the contribution of additional
authors. For articles and books with a small number of authors, this
would result in a grievous inaccuracy. The ISI Citation Databases
represent a wide range of subjects, and these differ in publication,
citation and author attribution practices. In some fields, the
principle investigator is listed as the final author; in others the
first author listing is considered to be the position of most
prominence. For articles with a small number of authors, attributing
citation credit only to the first author would be a grievous
inaccuracy. Book references occasionally compound this problem by
referencing the volume editors rather than the author of a specific
chapter. Citations to a book chapter could be attributed
erroneously to a the editor of the volume rather than any of the authors
of a chapter.
When we can link a reference to a specific source item,
we also gain a significant body of additional information on the item
(full article title, category, correct year of publication, list of
author addresses) that facilitate our efforts to identify the authors of
the work.
The volume of references to books varies greatly by
field. Surveying recent articles in a journal in Cell Biology
showed that less than 3% of the references were to books, at least half
were to laboratory manuals. A journal with a similar number of
articles and citations but in the subject category Political Science had
nearly 60% book references. Books and journals fulfill different
roles in the archival history of scholarly progress, and their use in
different fields of research often reflects differences in the time line
and process of research in these fields. The concentration of our
analysis on the scholarly journal literature will not be sensitive to
the contributions that outstanding, even ground-breaking books, but will
detect the commentary and expansions on a book that an author will
subsequently contribute to scholarly journals in his/her field.
Books are an important avenue of scholarly communication
in many fields, and Web of Science indexes the citations to books and
other items outside our source database. Although ISIHighlyCited.com
did not use book citations in its analysis, we do not imply that the
book literature is of lesser scientific influence, any more than our
limiting the data to 1981-1999 implies that no important research was
published before 1981.
When a Highly Cited Researcher is identified based on
citations to their journal articles, we invite the researcher to send a
complete publication list, including books and book chapters. All of
these items are then included in their publication record in the product
to provide a more complete look at their scholarly contribution.
|
