Page 123 - Becoming Metric Wise

114 Becoming Metric-Wise

not strictly) a special subset of ORCID. Virtual International
Authority File is another international authority file, mainly used by
National Libraries and by OCLC. Moreover, MathSciNet, the online
version of Mathematical Reviews, maintains a name authority file.
• Synonyms (different names referring to the same person). Citations
will be scattered unless a unique name is used for an author. Examples
of such “synonyms” are author names with a variable number of initi-
als, with or without a middle name, a woman’s maiden name and her
married name, different transliterations of nonwestern languages and
misspellings. A famous example in the field of informetrics is given by
the different forms of Derek J. de Solla Price’s name. This problem as
well can be solved by using ORCID or a similar unique identifier.
• Synonyms do not only exist for persons but also for journals or even
countries (Congo vs Zaire; Russia, the Russian Federation and the
larger (former) Soviet Union). In addition to variations in the abbrevi-
ated form of a given journal title, journals merge, split, change titles
and may appear in translations (Garfield, 1975).
• Which sources should one use: only journals(andwhich type(s)ofjour-
nal publications), or does one include books and/or conference proceed-
ings? As the type of used sources can influence the results of a citation
analysis (especially in the social sciences and the humanities) the choice
and the number of sources should depend on the purpose of the analysis.
• Implicit citations. Most citation analyses only consider explicit cita-
tions for the simple reason that most databases only give information
on this kind of citations. Yet, the A&HCI includes some implicit cita-
tions too. Implicit citations are frequently found in the form of epo-
nyms, discussed further on.
• Which periods to use? When counting publications, one needs to
determine a time period called the publication window: only publica-
tions published in this window are counted. When counting citations
to publications, we need a publication window as well as a citation
window, i.e., the time period during which citing publications are
published. As there may be large variations in citation counts from
one year to another, citation data should not be too restricted in time.
• Variations in the number of references, and hence the number of cita-
tions per field. Publications and referencing practices depend strongly
on research specialties. As a consequence variations between citation
counts also strongly depend on research fields. This may lead to diffi-
culties in cross-disciplinary comparisons.