Scholarly Inbreeding in Latin American Academically Managed Journals

From a bibliometric perspective, scholarly inbreeding has been identified in journals through the excessive use of both author and journal self-citations. However, editorial bias toward researchers from the same institution as the editorial management team has seldom been considered. According to the Spanish Foundation for Science and Technology, this occurs when a journal publishes more than 20% of documents authored by researchers affiliated with the same institution as the editorial management team. The purpose of this study is to establish the extent to which 81 Latin American journals managed by universities publish intramural documents (defined as those published by its own faculties). Results revealed that 56% of Brazilian journals were not compliant with the 20% benchmark as well as 44% of Colombian journals, 50% of Chilean journals, and 71% of Mexican journals. Interestingly, one third of these journals published the majority of the documents in English. By examining the documents published by these journals and subsequent citations to these articles, it was established that the intramural documents of some journals registered a higher ratio of citations per document in comparison with extramural documents published in the same journals. The results presented in this study provide evidence of inbreeding in some academically managed journals from Latin America. Although no one specific reason can account for this phenomenon, plausible explanations are given that may contribute to its understanding.


Introduction
From a biological perspective, inbreeding has been defined as preferential breeding between close relatives. In this context, Falconer and Mackay (1996) argued that inbreeding may impact any trait under selection, such as the physiological efficiency of an organism. Although in academia, inbreeding refers to the employment of doctorates by the same institutions that formed them (Horta et al., 2010;Inanc & Turner, 2011), from a bibliometric perspective, inbreeding may be understood as the excessive use of author self-citations (Glanzel et al., 2004;Seeber et al., 2019) or journal self-citations (Yu & Wang, 2007;Campanario, 2017). However, editorial bias toward researchers from the same institution as the editorial management team has seldom been considered (what will be referred to as intramural publication). According to the Spanish Foundation for Science and Technology, this occurs when a journal publishes more than 20% of documents authored by researchers from the same institution as the editorial management team (Fuentes et al., 2013). Though this practice may seem harmless to most, it may actually have an effect on university rankings. Among the many parameters assessed by these rankings is research performance, which, in some cases (such as the SCImago Institutional Ranking), constitutes 50% of the overall score (SCImago, 2019). Likewise, other university rankings, such as QS World University Ranking and Academic Ranking of World Universities, assign a lower but still significant weight (20%) to research performance. Thus, inadvertently, a perverse stimulus may have been created for universities to use their own journals to improve this parameter, as a higher ranking position will facilitate university branding. From a journal's perspective, publishing more than 20% of intramural documents will limit the scope of intellectual coverage, diminishing their geographical reach and provenance. Furthermore, because peer-review is a subjective process, questions may be raised about the validity of the process. For instance, journals with a single-blind peer-review process are prone to biased peer reviews because the names of the authors are known by the reviewers (Helmer et al., 2017;Seeber & Bacchelli, 2017).
One of the early studies on editorial practices identified favoritism to friends and personal associates as one of the least ethical practices (Sherrel et al., 1989). A latter study showed that almost 25% of articles published by 28 top economic journals in 1984 were characterized by an author-editor connection that may have been through a degree-granting institution or by current affiliation (Laband & Piette, 1994). A subsequent study on economy journals (Colussi, 2015) established that researchers from the same institution as the editor tend to increase the number of published articles by approximately one article per year. Similarly, a study showed that law journals publish more articles from members from their own institutions than from faculty affiliated with other law schools (Yoon, 2013). In a seminal work, Reingerwertz and Lutmar (2018) not only established the existence of academic bias in four leading international relations journals during the years 2000-2015 but also showed that these manuscripts received significantly fewer citations.
Because all of these studies have been performed using journals mainly from Europe and North America, it became of interest to determine whether this academic in-group bias was a usual practice in Latin American journals. Thus, this study aims to establish the extent to which journals managed by universities publish manuscripts from researchers belonging to the same institution and compare with those manuscripts written by researchers from other institutions. To achieve this, all the documents published between 2008 and 2015 among 81 journals from five Latin American countries were analyzed.

Method
An initial query was made on Journal Citation Reports 2015 to establish all the Latin American journals indexed. This search resulted in 247 journals from Argentina, Brazil, Chile, Colombia, Costa Rica, Ecuador, Mexico, Uruguay, and Venezuela. All journals were evaluated to determine their publisher, selecting only journals that were managed by universities. In total, 84 journals were identified. However, Boletin de la Sociedad Geologica Mexicana (from Mexico) was excluded because it was indexed by Web of Science from 2012 onward. Furthermore, the journals Revista de Biologia Tropical (from Costa Rica) and Revista Cientifica-Facultad de Ciencias Veterinarias (from Venezuela) were not considered in this study because they were the only academically managed journals from each country. This exclusion was made due to the fact that very small samples would undermine the validity of study. Thus, 81 journals were used for this study: four from Argentina, 32 from Brazil, 22 from Chile, nine from Colombia, and 14 from Mexico.
To analyze the journals managed by universities, the following variables were selected: language of publication, total number of documents published by the journal, number of documents published by researchers from the same country, number of documents published by researchers from the same institution (intramural documents), number of documents published by researchers from other institutions (extramural documents), citations received by intramural documents, citations received by extramural documents, average citations per intramural document and average citations per extramural document. Table 1 summarizes the total number of documents published by each journal managed by a university and the total number of documents published by researchers from the same country and institution as the editorial management team. The data indicates that for 56 journals (two from Argentina, six from Colombia, 31 from Brazil, 10 from Chile, and seven from Mexico), more than 50% of their documents were authored by researchers from the same country as the journal. It is worth noting that Brazilian journals exhibited the highest ratio of documents published by local researchers, ranging from 74.7% (for Journal of Applied Oral Science) to 99.2% (Ciencia Rural). On the contrary, only three journals (Journal of Applied Economics, Latin American Journal of Solids and Structures, and Electronic Journal of Biotechnology) published less than 20% of documents written by authors from the same country.

Results
More interesting results arise when the proportion of documents that were authored by researchers affiliated to the same institution that managed the journal was established. The majority (55.6%) of these journals accepted and published intramural manuscripts well above the 20% threshold. Whereas Argentinian journals were more compliant to the rule (only one journal exceeded the 20% threshold), 56% of Brazilian journals were not compliant, as well as 44% of Colombian journals, 50% of Chilean journals, and 71% of Mexican journals. In fact, five journals (Caldasia, Medicina Veterinaria Recife, Geofisica Internacional, Revista Mexicana de Astronomia y Astrofisica, Veterinaria Mexico) published more than 50% of intramural manuscripts during the period assessed. It is worth noticing that only one journal, Journal of Applied Economics (from Argentina), did not publish any manuscripts from researchers that belonged to the same university.
Considering the possibility that researchers might be publishing in their local language due to a language barrier, the language of publications of the journals that exceeded the 20% threshold of intramural documents was determined. As Table 2 depicts, 18 of the 45 journals from Argentina, Colombia, Chile, and Mexico published more than 50% of their documents in Spanish. In the case of Brazilian journals, 12 used Portuguese as the main language of publication. The other 15 journals published more than 50% of their documents in English. An interesting finding was that three of these journals (Archives of Clinical Psychiatry, Geofisica Internacional, and Revista Mexicana de Astronomia y Astrofisica) published more than 50% of intramural manuscripts.   The next question was to determine if the high proportion of intramural documents had an effect on the journal citation rate. To answer this, the intramural documents for each journal were disaggregated and their citations estimated to later compare with those from documents authored by researchers from other institutions. Table 3 shows the ratio of citation counts per document for each of the journals that published more than 20% of intramural manuscripts. The 44 journals were grouped by country to facilitate data interpretation. In the case of the only Argentinian journal (Revista de la Facultad de Ciencias Agrarias), the intramural publications were cited 27.7% more often than those authored by researchers from other institutions. Contrarily, we established for two Colombian journals (Revista Colombiana de Ciencias Pecuarias and Ingenieria e Investigacion) that the publications written by researchers from other institutions were slightly more cited. The rest of the Colombian journals did not display differences in the ratio of citations per document. Interesting results emerged when Brazilian journals were analyzed. First, the intramural manuscripts published in 12 of the 18 journals received a higher ratio of citations per document than those authored by researchers from other institutions. Surprisingly, two of these journals published almost, if not all, the manuscripts in English (94.5% for Revista Latino-Americana de Enfermagem and 100% for Scientia Agricola). Secondly, the manuscripts from five of the journals received similar citations per document, independent of the main language of publication (two used English and three used Portuguese). Only one journal (Journal of Applied Oral Science) presented a higher ratio of citations per document for manuscripts written by researchers from a different institution.

Mexico
As for six of the 11 Chilean journals, their intramural manuscripts received a higher ratio of citations per document than their counterparts. Peculiarly, two of these journals (Ciencia e Investigacion Agraria and Estudios de Economia) published the majority of the manuscripts in English. Likewise, English was the main language used by four other Chilean journals (Archivos de Medicina Veterinaria, Revista de Geografia Norte Grande, Estudios Filologicos, and Revista Chilena de Derecho) that received a similar ratio of citations per document, independent from the fact that the author was from the same institution as the journal.
The analysis also revealed that intramural manuscripts published by four Mexican journals (Revista Mexicana de Astronomia y Astrofisica, Revista Mexicana de Ingenieria Quimica, Geofisica Internacional, and Revista Chapingo Serie Ciencias Forestales) received a higher ratio of citations per document in comparison to documents authored by researchers from other institutions. Again, it seems that the language barrier was not an issue because two of these journals published all of their documents in English, whereas the other two used English as the main publication language.

Discussion
Just as livestock breeders justify inbreeding on the genetic value of their own stock, some editors may prefer researchers from their own institutions. In fact, Medoff (2003) established that editors of some economic journals used personal ties and institutional connections to simplify the search for high-quality articles. Colussi (2015) suggested that preexisting ties may reduce the cost of communication and increase collaboration, hence improving the quality of the manuscript. For instance, articles authored by former doctoral students of an editor increase citations more than 27% when the same editor is in charge of the journal (Colussi, 2015). Nonetheless, it is important not to generalize because social dynamics are different for each journal. It is important to keep in mind that not all universities offer a doctoral program in the same discipline as the scope of the journal being editorially managed.
One question that might arise is the rationality behind the 20% institutional inbreeding threshold set for scholarly journals by the Spanish Foundation for Science and Technology. Although this value was defined as one of the criteria to assess the quality of scholarly journals, no explanation was provided in the source regarding its origin. After a systematic search of the literature, no studies examining this issue of intramural publication were found. However, various studies investigating journal self-citation rate could provide a plausible explanation for this value. Already in 1963, Garfield and Sher (1963) established that 20% of the citations received by a journal corresponded to journal self-citations. Later studies considered this percentage as the standard during journal assessments (Campanario & Molina, 2009;Krauss, 2007;Tighe et al., 2011). Likewise, Thomson-Reuters (2002) published an essay stating that 82% of the journals listed on the 2002 Journal Citations Report had journal self-citation rates of 20% or less. Perhaps the Spanish Foundation for Science and Technology considered that 20% was a reasonable threshold to avoid the excessive use of Spanish scholarly journals by researchers from the same institution responsible for their editorial management. Clearly, such a behavior would have a confounding effect on the international visibility of Spanish scholarly journals. After all, journals are considered the primary mode of communication and record for scientific research (Brody et al., 2006). Perhaps some years from now, the 20% threshold will be commonly used as a standard just as the 5% p value, which was set almost 100 years ago, is used in statistical analysis (Fisher, 1926).
The results of this study showed that 45 Latin American journals published more than 20% of documents authored by researchers affiliated to the same institution as the editorial management team. Although the most plausible explanation may be a language barrier, one third of the journals published the majority of the documents in English. It is possible that the scope of the journal affects the language of publication. According to the data collected from Journal Citation Reports, the categories of the 30 journals that published mainly in Portuguese or Spanish were the following: agriculture, anthropology, biodiversity and conservation, construction and building technology, forestry, geography, history, information science and library science, law, linguistics, nursing, plant sciences, psychiatry, public/environmental and occupational health, social sciences, sociology, veterinary sciences, and zoology. An analysis of the total publication output between 1996 and 2011 revealed that researchers publishing in Spanish and Portuguese tend to publish their work most in fields related to health sciences, social sciences, and arts and humanities (Van Weijen, 2012). Although some coincidence is observed, other reasons may account for this result.
Another possible explanation may be institutional building (Yoon, 2013). Because the majority of local and global rankings consider institutional research output, it seems logical that institutions may benefit from intramural documents published in their own journals. To corroborate this theory, all the journals' websites were accessed to establish the type of peer review used to evaluate a manuscript. Regrettably, most of the journals do not specify any information about the peer-review process. Undoubtedly, if all the journals indicated whether a double-blinded or an open peer-review process was implemented, this hypothesis could be more clearly tested.
One should also consider the possibility that some of these peer-reviewed journals may be used to convey manuscripts written by doctoral students during their final year of training. Currently, doctoral programs encourage students to publish their research in a peer-reviewed journal as a way of ensuring they have developed logical thinking and the ability to communicate their scientific arguments accordingly (Glew et al., 2014). However, writing a manuscript for publication differs greatly from writing an essay for a university course (Jalongo et al., 2014;Kirkpatrick, 2019). Unfortunately, it is not uncommon for a doctoral student to graduate without having formal training on how to write a manuscript. In this context, mentors have a major influence on what doctoral students learn about the formal and informal rules of their discipline (Barnes et al., 2012). Most mentors contribute in two major aspects: socially, by serving as an interface between the student and the department/faculty, and careerwise, by relating to activities that boost career development Brill et al., 2014). Nevertheless, in research-oriented disciplines, mentors engage in what has been defined as "research mentorship" (Abedin et al., 2012). Among the many skills that these doctoral students need to attain are how to conduct research, how to write proposals to obtain funds, and how to write an article using the results of their research project. Although mentors usually share their advice on how to structure an article and how to choose a suitable journal, the doctoral student may suggest other journals based on their fast publication process. Some Latin American doctoral programs include a "minimum publication requirement" to obtain a doctoral degree, similar to other countries such as China (Li, 2006). Thus, a doctoral student may be driven to submit initially to a journal whose editors are linked to the home university and that is indexed by a global citation database such as Web of Science. Additionally, because this study considered only Latin American universities, a lack of English proficiency among these students can be assumed.
Another aspect that needs to be taken into consideration is the fact that countries such as Brazil and Mexico oblige researchers to form part of a national researchers' registry to apply for research funds, which involves publishing their research findings in indexed journals. In the case of Chile, the National Fund for Scientific and Technological Development not only requires researchers a minimum number of articles published in journals indexed by Scopus or Web of Science to apply for a research grant, but if the results of an awarded grant are not published in similarly indexed journals, the researcher may be requested a full refund of the allocated funds (Krauskopf, 2018).

Limitations
This study was confined to journals managed by universities from five Latin American countries, so it is possible that the inclusion from journals from North American countries may yield different results. Likewise, it covered an 8-year time period that ended in 2015 to allow enough time for documents to be cited.
Although the results of this study show that the consequences of inbreeding may not seem so detrimental, at least in terms of citations, it is important to consider that citation data could be manipulated through the use of self-citing (Krauskopf, 2013). Unfortunately, the origin of these citations was not available for analysis, so it was impossible to establish self-citation rate for each journal. However, a recent study (Seeber et al., 2019) established that the tendency for researchers to self-cite not only varies among disciplines, but it also relates to their institutional affiliation. On the other hand, some of these journals publish manuscripts that focus on areas of research that are of particular interest to local researchers. For instance, the majority of the articles published by the Brazilian journal Revista de Saude Publica relate to public health issues. Lastly, the scope of these journals converges to fields related to life sciences. Perhaps in these fields, the concentration of intramural documents on each of these journals generates a crowding effect that increases their chances of being cited (Yoon, 2013).