You are on page 1of 4

Cardiovasc Intervent Radiol

https://doi.org/10.1007/s00270-020-02535-0

LETTER TO THE EDITOR

Disparities in Citation Metrics Amongst Web of Science, Scopus,


and Google Scholar for Interventional Radiology Journals
Vibhor Wadhwa1 • George K. Vilanilam1 • Jeffrey Forris Beecham Chick2

Received: 17 April 2020 / Accepted: 25 May 2020


 Springer Science+Business Media, LLC, part of Springer Nature and the Cardiovascular and Interventional Radiological Society of Europe
(CIRSE) 2020

Editor, CVIR, and JET in 2016 and 2017, were obtained from the
The scientific community relies on citation metrics to WoS database and included in this study. Citation counts
gauge the reputation of a journal. Academic institutions were recorded for each article on April 10th, 2020 from
also utilize citation metrics as a criteria for promotion and WoS, Scopus, and GS. Means and standard deviations were
tenure. Although there are many available citation metrics, calculated for citations obtained from all three sources for
the most important, for scientific journals, is the journal the journals. Paired sample t test was used to determine
impact factor (JIF) published by Clarivate (Clarivate; statistical differences in citations between the three sour-
Philadelphia, PA). The calculation of JIF relies on citation ces. Additionally, the Altmetric score (Holtzbrinck Pub-
counts of the articles published in the journal and uses the lishing Group; Stuttgart, Germany) for each article was
Web of Science (WoS) database to count the number of recorded and the Spearman correlation test was used to
citations. The other well-known citation indexing services determine correlation between citation counts and Alt-
used by the scientific community include Scopus (Elsevier metric scores. Two-tailed p value of \ 0.05 was considered
Inc; Amsterdam, Netherlands) and Google scholar (GS) significant.
(Alphabet Inc; Mountain View, CA). It has been shown A total of 150 articles were included in this study
that disparities exist amongst these services in terms of (JVIR = 50; CVIR = 50; JET = 50). The mean number of
citations metrics for other medical specialties. For the field citations obtained from WoS, Scopus, and GS for JVIR was
of interventional radiology (IR), however, no analysis has 24.64 ± 10.43, 27.5 ± 12.43, and 36.8 ± 16.15, respec-
reported the differences between these citation indexes. tively; for CVIR were 25.62 ± 16.25, 27.52 ± 16.73, and
The purpose of this study was to evaluate the citation 37.32 ± 22.11, respectively; and for JET were
counts obtained from WoS, Scopus, and GS for three well- 22.90 ± 8.69, 24.60 ± 9.24 and 29.38 ± 11.69, respec-
known IR journals—Journal of Vascular and Interven- tively (Fig. 1). For JVIR, Scopus found 11.61% more
tional Radiology (JVIR), CardioVascular and Interven- citations than WoS and GS found 49.35% more citations
tional Radiology (CVIR), and Journal of Endovascular than WoS, both differences were statistically significant
Therapy (JET). (p \ 0.001). Similarly, for CVIR, Scopus found 7.42%
Institutional review board approval was not required for more citations than WoS and GS found 45.67% more
this study. The 50 most cited articles, published in JVIR, citations than WoS, both differences were again statisti-
cally significant (p \ 0.001). For JET, Scopus found
7.42% more citations than WoS and GS found 28.29%
& Vibhor Wadhwa more citations than WoS (p \ 0.001). The mean Altmetric
vibhorwadhwa90@gmail.com scores for the journals were: JVIR 13.4 ± 16.27, CVIR
1
Department of Radiology, University of Arkansas for 5.08 ± 5.22, and JET 2.20 ± 2.74. There was no statistical
Medical Sciences, Little Rock, AR, USA correlation between the Altmetric scores and the citation
2
Division of Interventional Radiology, Department of counts obtained from WoS (p = 0.323), Scopus
Radiology, University of Washington, Seattle, WA, USA (p = 0.627), or GS (p = 0.066).

123
V. Wadhwa et al.: Disparities in Citation Metrics Amongst Web of Science, Scopus, and…

Fig. 1 Mean citations for 50


most cited articles published in
JVIR, CVIR, and JET journals
in 2016 and 2017 obtained from
Web of Science, Scopus, and
Google Scholar

The differences in citation counts for the same articles one study estimating 389 million indexed records [2]. GS
indicate the differing sourcing strategies of these services. includes [ 90% found by WoS and Scopus, with the
WoS has 129 journals under the subject subcategory additional citations including theses and dissertations,
‘‘Radiology, Nuclear Medicine, and Medical Imaging,’’ books or book chapters, conference proceedings, unpub-
and ranks journals using JIF, of which JVIR ranks 44 and lished materials, and other document types [3]. Another
CVIR ranks 79. Under the subject subcategory ‘‘Peripheral possible reason for GS finding more citations is the
Vascular Disease’’ on WoS, JET ranks 25 and JVIR ranks inclusion of non-English literature while WoS and Scopus
27. On the other hand, Scopus lists 272 journals under the predominantly index English literature. Some studies have
subject subcategory ‘‘Radiology, Nuclear Medicine, and shown that GS may return duplicate citations results, pos-
Imaging,’’ and ranks journals based upon CiteScore metric, sibly another reason for higher citation counts [4]. It is
of which CVIR ranks 102 and JVIR ranks 112. There is no worth noting; however, that differences between citation
distinct subject category for IR in either of these indices. databases were consistent across all three journals, despite
The CiteScore metric uses Scopus citations, rather than the different strategies.
WoS citations, and is calculated by including citations and Altmetric is a commercial organization which developed
publications from the three previous years (rather than the a social media-based ‘‘alternative metrics’’ scoring system
2 years for JIF). JIF and CiteScore trends of JVIR, CVIR, and publishes a composite Altmetric Attention Score for
and JET over last 5 years are shown in Fig. 2. Note that each article to aggregate and track online mentions. It is
while the JIF of JVIR has consistently been higher than calculated as a weighted total of the mentions of an article
CVIR, the CiteScore of both journals is comparable. Also, by various online platforms including social media, with
note that while the JIF for JET is comparable with JVIR in the weights reflecting the relative reach of each online
2018, the CiteScore for JET is higher than JVIR. source. Importantly, however, conventional citations in
The present study shows that Scopus and GS find sig- biomedical journals are not included within the score,
nificantly more citations than WoS, with the difference making it a different metric altogether. It has been previ-
between GS and WoS being larger for all three journals. ously shown that there is a weak correlation between
Findings of this current study are similar to previous citations and Altmetric score for radiology articles [5]. This
reports in other specialties, such as cardiology [1], and is consistent with findings of the current study, which
highlight the disparities in citation indexing within these showed no significant correlation between Altmetric score
databases. The Journal Citation Report (JCR), of Clarivate, and citation counts. Despite these findings, such novel
indexes almost 12,000 journals while Scopus indexes close scoring systems are likely to garner more attention with the
to 23,000 active journals; therefore Scopus potentially has continued growth of social media.
a broader library of citation sources. The size of GS Limitations of this study are inclusion of only 50 most
repository; however, is unknown at the present time, with cited manuscripts from the journals, which was purposeful,

123
V. Wadhwa et al.: Disparities in Citation Metrics Amongst Web of Science, Scopus, and…

Fig. 2 Journal impact factor and CiteScore trends for JVIR, CVIR, and JET from 2014 to 2018

123
V. Wadhwa et al.: Disparities in Citation Metrics Amongst Web of Science, Scopus, and…

since as highly cited articles are more likely to show dif- References
ferences between the various sources. Also, IR studies
published in other journals were not included, since this 1. Anker MS, Hadzibegovic S, Lena A, Haverkamp W. The
difference in referencing in Web of Science, Scopus, and Google
study was targeted towards IR-centric journals.
Scholar. ESC Heart Fail. 2019;6(6):1291–312. https://doi.org/10.
In conclusion, this study demonstrates that Scopus and 1002/ehf2.12583.
GS have a higher mean citation counts than WoS for IR 2. Gusenbauer M. Google Scholar to overshadow them all? Com-
journals and the differences between WoS and GS are paring the sizes of 12 academic search engines and bibliographic
databases. Scientometrics. 2019;118(1):177–214. https://doi.org/
much larger than WoS and Scopus. Given these differ-
10.1007/s11192-018-2958-5.
ences, individual researchers, academic institutions, and 3. Martı́n-Martı́n A, Orduna-Malea E, Thelwall M, López-Cózar ED.
journals are more likely to rely on multiple sources to Google Scholar, Web of Science, and Scopus: a systematic
gather complete citation data. Larger studies need to be comparison of citations in 252 subject categories. SocArXiv. 2018.
https://doi.org/10.31235/osf.io/42nkm.
performed to determine the quality of these citation
4. Adriaanse LS, Rensleigh C. Web of science, scopus and google
indices. scholar: a content comprehensiveness comparison. Electron Libr.
2013;31(6):727–44. https://doi.org/10.1108/el-12-2011-0174.
5. Rosenkrantz AB, Ayoola A, Singh K, Duszak R. Alternative
Funding None. metrics (‘‘Altmetrics’’) for assessing article impact in popular
general radiology journals. Acad Radiol. 2017;24(7):891–7.
Compliance with Ethical Standards https://doi.org/10.1016/j.acra.2016.11.019.

Conflict of interest The authors declare that they have no conflict of Publisher’s Note Springer Nature remains neutral with regard to
interest. jurisdictional claims in published maps and institutional affiliations.

123

You might also like