Page 161 - Becoming Metric Wise

152 Becoming Metric-Wise

Moreover, GS Metrics provides the following indicators for journals:
• The h5-index: this is the h-index restricted to publications in the
latest 5 complete calendar years.
• The h5-core: the core corresponding to the h5-index.
• The h5-median: the median citation value of the items in the
h5-core.

5.18 COMPARISONS

Many colleagues performed investigations comparing GS, Scopus and the
WoS. Because of unique features related to each of these databases a gen-
eral consensus, see e.g. (Meho & Yang, 2006) for one of the earliest stud-
ies, is that the three databases complement each other.
Vieira and Gomes (2009) compared Scopus and the WoS as to differ-
ences for some typical (Portuguese) universities. They agree with the gen-
eral consensus, mentioning that some high impact documents were found
in only one of the two databases. Yet, documents included in both data-
bases were generally the most-cited ones. Bar-Ilan (2010) studied a diffi-
cult case, namely a book, which is cited under many different forms. She
carefully searched for variants in all databases, removed duplicates and
made a thorough analysis of unique citations. She concluded that for this
particular case the coverage of the WoS and Scopus was quite comparable
and that GSs coverage was surprisingly good—it found the most unique
citations and was more accurate than she expected. More generally,
Kousha et al. (2011) assessed the citation impact of books studying
Google Books, GS and Scopus. They found that, especially in book-
oriented disciplines such as the arts and humanities online book citations
could better support peer review for research evaluation than Scopus cita-
tions. Based on their investigations Amara and Landry (2012) urge those
that perform assessments in Canadian business schools to complement
WoS data by GS data.
Meho and Yang (2007) compared the WoS, GS and Scopus to assess
the extent to which the results of a citation analysis depend upon the data
source, using the task of ranking the faculty of a library and information
science school. Their findings showed that GS was probably too difficult
to use for a large-scale citation analysis and that the other two gave similar
results overall. Yet, as mentioned above they came to the consensus view
that using all databases in conjunction gives the fairest results. Fields which
emphasize conferences, such as computer science and computational