Some error in ESI (acknowledged, but not detailed by Thomson)
pislyakov at HSE.RU
Wed Oct 12 10:00:24 EDT 2011
While performing research using Highly Cited Papers section of Essential
Science Indicators, we with my colleague noticed that there are 11160 of
them published in 2008 and only 9311 ones dated by 2009, 8% less. (This was
in June, now the numbers are 10726 and 9094 respectively, the same ratio).
The science is continuously growing, and WoS shows at least 3% growth in the
number of papers for these years, notwithstanding which version of WoS you use.
As, by definition, the highly cited paper is every 100th one for a fixed
year, there is a clear inconsistency. We contacted Thomson support and,
after 2.5 months of investigation, in August, they answered "...upon
thorough review we have identified issues with the programming which
affected relative records that we left from being added to the database. I
sincere apologize for the inconvenience this has caused you".
It seemed to me that this is an important message not for me personally, but
for all scientometric society, as many of us use ESI in our studies.
Unfortunately, until now no public explanation appeared as to what records
were missed, how this influences the integrity of the data, when the correct
version of ESI will appear for public use.
I hope this will be explained in detail for the public. If someone has their
own versions/observations, they are most welcome.
Assistant Director for Electronic Resources Management
Higher School of Economics Library
20 Myasnitskaya street
Tel.: +7 (495) 6213785
Fax: +7 (495) 6287931
E-mail: pislyakov at hse.ru
More information about the SIGMETRICS