Google Scholar Citations versus Web of Science and ResearchGate: The Spanish Library and Information Science

[Emilio Delgado López-Cózar]

Dear colleagues,

Following the research front the EC3 Research Group initiated to 
discover the inner depths of Google Scholar and test its suitability as 
a tool for research evaluation, this time we have turned our efforts to 
investigate new uses for Google Scholar Citations (sometimes also known 
as Google Scholar Profiles). We present a new procedure to learn about 
the impact of a whole scientific discipline (in this case, Library and 
Information Science), a specific scientific and professional community 
(the Spanish community), and the main agents that are a part of it 
(scientists, professionals, the documents they produce, and the journals 
and publishers which print these documents). From the scientific 
production of the members of the Spanish Library and Information Science 
community who have made public their profile on Google Scholar 
Citations, we can develop a picture of this discipline.

Moreover, we want to compare this picture with the one offered by the 
Web of Science (Core Collection) and ResearchGate. This way, it is 
possible to present a new perspective of the discipline, and, most 
important of all (at least as regards our line of research on Google 
Scholar), we present an easy and intuitive comparative chart for all 
these products and the realities they portray.
The experimental nature of this action is also aimed at verifying the 
degree of acceptance of products of this kind in the Humanities and 
Social Sciences communities, which are not as of yet familiar with this 
sort of bibliometric practices. And the best way to do this is to 
display the results obtained in each area of study in order to analyze 
and assess the reactions that these products may provoke. From the 
ranking of 336 Spanish Library and Information Science authors who have 
made their Google Scholar Citations profile public, and the 492 most 
cited documents in those profiles (Top 10%), three additional rankings 
have been developed: a journal ranking, a publisher ranking, and an 
institution ranking according to the number of citations received.

We believe the data speak for themselves, and they confirm to the 
letter what we claimed recently in our work about the 64,000 most cited 
documents in Google Scholar 
(http://googlescholardigest.blogspot.com.es/2014/10/the-most-cited-documents-in-google.html). 
It demonstrated, among other things, and using the largest sample to 
date in a work of this nature, that 91.6% of the documents received more 
citations in GS than in WoS, and that GS has been able to find twice the 
number of citations per document than WoS finds. In this product we are 
now presenting, these figures are even more pronounced: 99.8% percent of 
the documents had received more citations in GS than in WoS, and 
documents in GS had received, on average, three times more citations in 
GS (47) than they had received according to WoS (15). Moreover, while 
126 authors in our list (37.5%) don’t have a single WoS document, and 43 
of them only one (that is, more than half of these authors are invisible 
in WoS), all the authors in our list have at least one document in their 
GSC profile.

Therefore, we can report again what seems ever more clear: Google 
Scholar is able to measure the most remote corners of knowledge… those 
that can’t be found in the shelves of Thomson Reuters (especially for 
the Humanities and Social Sciences). Furthermore, when comparing authors 
or documents that are present both in GS and WoS, GS portrays a very 
similar picture to the one that can be observed in WoS. Pearson 
correlation for the number of citations (244 documents) is 0.87 
(Confident level 95%, P-value <2.2e−16). If we compare authors’ h-index 
according to these two sources, the correlation is also high: 0.84. 
Lastly, it is also of note that ResearchGate is beginning to offer an 
interesting data set, with its own peculiarities but also yielding 
similar results when comparing its indicators with those of GS and WoS. 
Pearson correlation for the RG Score and GS h-index (for the 158 authors 
who have a profile in both services) is 0.75, while if we compare the RG 
Score with the WoS h-index (for 205 authors, that is, all those that 
have created a profile in ResearchGate in our authors sample), the 
correlation is 0.82, which suggests the two measures are also highly 
related.

Below you can find the link to the product, where a detailed 
explanation of the methods used to create it can be found (in Spanish).

http://www.biblioteconomia-documentacion-española.infoec3.es/

Emilio Delgado López-Cózar
EC3 Research Group: Evaluación de la Ciencia y de la Comunicación 
Científica
Facultad de Comunicación y Documentación
Universidad de Granada
http://googlescholardigest.blogspot.com.es/
http://scholar.google.com/citations?hl=es&user=kyTHOh0AAAAJ