Researchers from the University of São Paulo (USP), the Federal University of the ABC (UFABC), and the Federal Technological University of Paraná (UTFPR) studied a gigantic volume of information—the bibliographic output of about 1.1 million Brazilian researchers and students who are registered on the Lattes Curriculum Vitae (CV) platform—and obtained an unprecedented overview of collaborations among them, based on co-authorship of scientific articles, books and book chapters. The article, published in January in the Journal of the Association for information Science and Technology, showed that the number of collaborations between Brazilian researchers has increased notably over the last two decades. A little more than 1 million collaborations were recorded between 2008 and 2010, contrasted with the 63,944 observed between 1990 and 1992. This advance greatly exceeds the growth of the scientific community itself during the period.
The authors analyzed the evolution of eight major areas of knowledge (see figure). Health Sciences were one of the highlights: nearly 42% of the 272,783 researchers established collaborations with colleagues in the same area. At the other extreme is linguistics, languages, literature and the arts: only 15% of the nearly 100,000 CVs reviewed involved co-authorship. Despite this, researchers in linguistics, literature and the arts have a peculiarity: they established the most partnerships with colleagues in other fields, especially the humanities. “In relative terms, linguistics, languages, literature and the arts are the areas with the highest degree of interdisciplinarity. However, in absolute numbers, the biological sciences were responsible for the greatest number of interactions with other areas,” says Jesús Pascual Mena-Chalco, a professor at the UFABC Center for Mathematics, Computer Science and Cognition and the principal author of the study.
The data relating to collaborations in the humanities surprised Samile Vanz, professor at the School of Library Science at the Federal University of Rio Grande do Sul. Collaborations reached 169,869 in the area of the humanities, and 102,058 in the applied social sciences between 1990 and 2010. In other studies, she observes, the academic production in these areas appears more timid. “This is because studies normally use the Web of Science or Scopus databases, which primarily index scientific articles. Researchers from the humanities traditionally publish mostly books and book chapters, which are more visible in the Lattes data,” says Vanz.
JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGYThe study’s contribution is not restricted to characterizing collaborations among Brazilian academics. It also advanced computer science and information technology by developing mathematical models and algorithms capable of refining the organization of large volumes of data. “What we did was look at the universe that is recorded on the Lattes Platform and, using new algorithms, identified the collaborations between authors,” says Mena-Chalco. “Using this methodological tool, sociologists and managers will be able to perform more in-depth analyses,” he suggests. “It is a pioneering study,” says Rogério Meneghini, scientific coordinator of the SciELO Brazil virtual library. “The authors were able to extract data from the Lattes Platform and, to that end, developed their own methodology. It is still difficult to extract information from this platform.” Currently the platform, maintained by the National Council for Scientific and Technological Development (CNPq), has more than 2.7 million academic CVs and, due to its collection of information from almost all Brazilian researchers, is considered to be unique; there is nothing like it anywhere in the world.
Big data
According to Roberto Marcondes Cesar Júnior, professor at the USP Institute of Mathematics and Statistics and another co-author of the study, the research is directly related to an emerging issue, Big Data—an information technology concept based on the storage and interpretation of gigantic volumes of data. Recently, Marcondes—who is assistant coordinator of the FAPESP Sciences and Engineering Division—has focused intensely on the subject. This study by his group is part of a more ambitious line of research within the Centers of Excellence Support Program (Pronex), the result of an agreement between FAPESP and CNPq. “One of the problems we studied was how to create a mathematical model that would allow us to ask questions such as: if I hire a researcher today, what is the expected contribution of his or her work to my department in the future? The advent of Big Data comes from the realization that we have more data to analyze than we are humanly capable of processing,” says Marcondes.
Journal of The Association For Information Science And TechnologyAccording to Jacqueline Leta, professor at the Federal University of Rio de Janeiro, the study’s results are helping to confirm some hypotheses. “The fact that the biological and health sciences have the greatest number of collaborations makes sense, as they have been areas of excellence in Brazil since the foundation of the first institutions, such as the Oswaldo Cruz Institute and the Butantan Institute, which practically established Brazilian science,” she says. The health and biological sciences also involve high costs, says Leta, and one of the reasons for partnerships between researchers is the desire to share infrastructure.
To understand the impact of the results of the study, one must understand certain limits on the raw data. For example, a distinction must be made between collaborations involving individuals and those between institutions. “The study shows an increase in collaborations between people, but that does not necessarily mean that research institutions are collaborating more,” says Samile Vanz. According to her, collaboration between institutions is especially important, because it requires greater sharing of teams, resources and equipment. “Brazil is stabilizing at the level of individual and co-authorship and might be going through a new phase of expansion of cooperation between institutions, but this will be seen only in future studies,” says the researcher, who in 2009 defended her doctoral dissertation on Brazilian participation in international research networks (see Pesquisa FAPESP Issue nº 169).
Caveats
Mena-Chalca explains that the cross-referencing of data concentrated on identifying co-authorship between researchers registered with the Lattes Platform, without looking at their institutions. One example: the collaboration between two Brazilians with Lattes CVs, one at USP and the other at Harvard, would be an international collaboration. “However, since the study only looks at individuals, we cannot identify which collaborations are international and which are domestic. Our goal was to see how Brazilian researchers are working with each other,” explains the UFABC Professor. Another limitation is related to the lack of CV updating. If two researchers wrote an article together, but only one of them recorded it in their Lattes CV, the lack of information from the other author prevents identification of the collaboration.
