A few changes for 2010BMJ 2009; 339 doi: https://doi.org/10.1136/bmj.b5632 (Published 31 December 2009) Cite this as: BMJ 2009;339:b5632
All rapid responses
Bibliometric data are increasingly required in applications for
research funding. During a recent application, I noticed a large
discrepancy between the number of citations in two widely used databases
for a BMJ article published in 2010.1 In Scopus, both the print 'Pico' and
full-text online articles are separately indexed and the article has 53
citations in total. In the Science Citation Index (SCI), the online
article only is indexed and there are 15 citations. None of the 53 citing
articles in Scopus were duplicates, and all 15 citing articles in SCI were
also identified as citing articles in Scopus. Many of the citing articles
in Scopus that were not identified as citing articles in SCI are in
journals indexed by SCI.
Differences in number of citations between the two databases are
common for BMJ articles. In SCI, the 20 BMJ articles with the most
citations in 2010 were cited 541 times. Counting 'Pico' and online
articles as 1 article, the top 20 cited BMJ articles in Scopus had 41%
more citations (763 in total). All the top 20 cited articles in Scopus had
more citations in Scopus than SCI, (median ratio 1.4, range 1.1-11.0), and
8 had >80% more citations (Figure).
The differences between Scopus and SCI for other leading medical
journals seemed consistent with the different numbers of journals indexed
by each database. The top 20 cited Lancet articles in Scopus had 13% more
citations than the top 20 cited articles in SCI, and the top 20 cited NEJM
articles in Scopus had 21% more citations than the top 20 cited articles
in SCI. For both journals, only 1 of the top 20 cited articles in Scopus
had >50% more citations in Scopus than SCI, and none had >80%.
Thus, the issue seems primarily related to BMJ articles, and more
particularly to the indexing of the BMJ 'Pico'and online articles.
Individuals using BMJ bibliometric data such as the number of citations
for an article or the impact factor of the BMJ should be aware of the
differences between the databases. In particular, SCI undercounts
citations for a significant proportion of BMJ articles, presumably leading
to a lowered impact factor.
1. Bolland MJ, Avenell A, Baron JA, Grey A, Maclennan GS, Gamble GD,
et al. Effect of calcium supplements on risk of myocardial infarction and
cardiovascular events: meta-analysis. BMJ 2010;341:c3691.
Competing interests: No competing interests