Emerging economies practicing excessive self-citation draw attention to the effects of isolation
If a person were to say that China, Brazil or Iran are islands, we would say that he is demonstrating little or no familiarity with basic geography. But in the field of scientometrics—which studies quantitative elements of scientific output—his statement would not sound strange. A number of countries, including the three mentioned, have become islands in a sea of academic publications and citations. This phenomenon, known as scientific insularity, occurs when a country’s rate of national self-citation is higher than average—in other words, when many of the citations come from researchers within that country. The causes of this phenomenon range from a scientific agenda focused on more localized issues, to insufficient training in methodological practices at the university level. A research study published last year in the journal Scientometrics takes a new look at the problem and suggests that political and geographic factors such as a country’s size exert a strong influence on the frequency of national self-citation.
“Large countries have more opportunities to develop research traditions and independent academic networks, particularly when there are a large number of potentially interconnected groups,” explains Richard Ladle, principal author of the study and senior research associate at Oxford University. Scientists from small countries end up being forced to collaborate and establish a dialogue with colleagues in other countries, which increases the number of foreign citations. To arrive at this assessment, Ladle and his team, which includes his Brazilian wife Ana Malhado of the Institute of Biological Sciences and Health at the Federal University of Alagoas (Ufal), analyzed data on numbers of publications and citations by country. The data, for the 1996 to 2010 period, were made available by SCImago, a portal of scientific indicators supported by the Scopus database.
The study revealed that the leading self-citers include large countries such as the BRIC nations (Brazil, Russia, India and China), those with historically closed political systems such as Iran and Cuba, and former Soviet bloc states (Serbia, Ukraine and Poland). Countries with lower levels of self-citation are mostly small in size but have a high proportion of post-graduate programs taught in English, such as Israel, Denmark, Singapore and the United Arab Emirates. While China registered a 51% self-citation rate in the period analyzed, only 15% of the total citations received for articles from Israel came from authors within that country.
The influence of a country’s size on national self-citation is linked to other factors. High rates of national self-citation in large developed countries like the United States, with 47%, are to be expected. That country produces the largest number of research papers in the world, generally having a high impact factor, so one would expect its papers to be heavily cited. Brazil, with a high and growing rate of national self-citation, is also a large country but less developed, which means that the causes differ from those that exist in the United States.
Explaining further, Ladle compares his native country of England with Brazil, where he has worked since 2009 as a visiting professor in the same department as his wife at Ufal. “One difference I’ve noticed is the low number of foreign professors and students here, and the considerable number of academic institutions in England,” he says. While a professor at Oxford, Ladle had as many as 70% foreign students in his program, and 50% of his department colleagues were from various parts of the world. “English scientists write mainly for international journals, and therefore they are obligated to “sell” their work to the international community from the very beginning of their career,” he adds.
This preference for publications with international prestige over those from one’s own country is a common characteristic of nations that attach great importance to cutting the umbilical cord that connects a paper to the mother country. From 2006 to 2011, articles from Sweden were cited more than six million times, and only 1.06 million were self-citations. The difference between Sweden and China lies in what may be called the “miracle of proliferation” of scientific publications. China has 537 publications listed in Scopus, and Sweden is content with 47 publications in the database.
“The more scientific journals a country has, the more self-citations it will receive,” explains Rogério Meneghini, scientific coordinator of the virtual library Scientific Electronic Library Online (SciELO), who also calls attention to the differences between the systems adopted by the most widely recognized international databases. The Web of Knowledge provided by Thomson Reuters, the most traditional of these, has about 12,000 indexed publications. Its competitor Scopus, from Elsevier, has 23,000 publications, indicating that it is less selective. The market competition between the two creates favorable conditions for publications with low international prestige to be accepted quickly.
The emergence of new publications in some countries is driven by policy and market considerations. Contrary to the global trend, in which a national focus is decreasingly the principal factor used to rate quality, developing countries are being treated like veritable “islands of lost treasure” by the large scientific publishing houses. “Nowadays the ones paying attention to the scientific journal market are the commercial publishers like Elsevier and Springer, whose profit margin is as much as 45%,” notes Meneghini, who points out that the major investment targets have been China and Brazil, among others. “Sweden regards it as more important to have researchers serving as managing editors of major international publications and dictating the direction of science in the world, than to simply have a lot of national publications,” he comments.
In Brazil, the reason for the high output of publications is simple: they serve as a channel for scientific output that does not find a foothold in international journals. “We have an extremely large number of publications, something like 5,000—there is no exact count—so there is always a demand for new articles,” Meneghini says. He believes the problem lies in the fact that many institutions were never prepared to publish articles, but since they want to establish post-graduate programs, publishing a journal might earn them points in evaluations by the Coordinating Agency for the Improvement of Higher Education Personnel (Capes).
“A publication regarded here as a benchmark and presenting high-quality articles abroad may be accorded less importance when its national citation rate is taken into account,” says Gilson Volpato, professor at the Institute of Biosciences at the Universidade Estadual Paulista (Unesp) in Botucatu and author of books on scientific writing. One example he cites is the Revista Brasileira de Zootecnia, published by the Brazilian Zootechnical Society. Although it was highly rated by the Qualis evaluation system of Capes, the publication was virtually frozen out of Journal Citation Reports (JCR), linked to the Web of Knowledge, because it had very high rates of national self-citation. The group of Brazilian publications that were suppressed from JCR’s Journal Impact Factor in 2011 also included Planta Daninha, Revista Brasileira de Farmacognosia and Revista Ciência Agronômica, because they were shown to have anomalous citation patterns that distorted their impact factors.
The self-citation rate of CA, A Cancer Journal for Clinicians, a major medical journal published by the American Cancer Society, is 0.029%, and its impact factor according to JCR is 101. The well-known journals Nature and Science have self-citation rates of 0.158% and 0.136% and impact factors of 36 and 31, respectively.
Some examples of mechanisms for according greater prominence to research studies conducted on the home campus can be found at the University of São Paulo (USP) and the State University of Campinas (Unicamp). In the past three years, both of these institutions have begun to invest in workshops to discuss tools for writing theses and papers, and methodologies for presenting data and results. “We think about how to improve the writing in English and the citation practices, and we create strategies for internationalizing our publications,” says Sueli Mara Soares Pinto Ferreira, coordinator of the Integrated Library System of the University of São Paulo (SIBi-USP), which includes 104 publications produced at USP.
Since 2008, Unicamp has maintained a Writing Space for the purpose of supporting the translation of scientific articles into other languages. The first workshop organized at Unicamp, held in 2010, had only 57 participants. In 2012, the number had soared to 800. There is also a project underway to create a repository for the 44 publications produced by the university. “The repository should begin to receive the first publications in the first six months of this year, and in the future it will be linked to a larger repository hosted on the website of the Council of São Paulo State University Presidents (Cruesp),” explains Ronaldo Pilli, Provost for Research at Unicamp.
“When I did my doctorate at Oxford, my initial project to describe leaf morphology patterns in the Amazon was completely deconstructed,” says Ana Malhado, a co-author of the study. She had to develop hypotheses and adopt a new approach that was much more attuned to the world literature. Many students and researchers seem to prefer writing a descriptive article for a national publication over using the same data to address globally important topics through generalization. Malhado’s comments underscore how insularity can often make life difficult for Brazilians who plan to study abroad.
“Every country has a level of national self-citation that it regards as healthy, because their scientists have to address national interests,” she comments. Malhado believes that the main point to emphasize is that public policymakers should be more mindful of insularity and seek to refine research agendas at all levels.
In Brazil, agriculture and public health are good examples of sectors that focus on local issues but, at the same time, produce articles that receive national recognition. “Many countries are interested in knowing how Brazil has dealt with a climate- or soil-related situation to grow a particular crop. It’s a matter of using local data and then generalizing,” Meneghini says. As an example, Volpato mentions the educator Paulo Freire (1921-1997), who studied adult literacy and was internationally recognized for developing global concepts using local data. “To fail to do so is a sign of weakness in the scientific standing of our researchers,” he concludes.
Another way to avoid insularity is to hire foreign professors such as Ladle. The couple’s research is an example of how a partnership between researchers from different countries can lend their work a desirable global character and, in some cases, show that overcoming national boundaries can also bear fruit aside from scientific interests. During one chaotic week, first-time parents Ladle and Malhado granted an interview to Pesquisa FAPESP soon after the birth of their daughter Jasmine.
LADLE, R. J. et al. Assessing insularity in global science. Scientometrics. 28 Mar. 2012.