Introduction

There are many ways in which the geographic dimension—and distinctions between local, regional, national and international realms—can affect the processes of scientific output and circulation (Marginson, 2006). Traditionally, from a bibliometric standpoint, because national scientific output is delimited as the contribution made by each country for the sake of furthering knowledge, the geographic origin or affiliation of researchers/authors is used as a demarcation criterion for publications. This conceptualization of national science is operational, a tracking device of sorts for National Research Systems; yet there are other aspects of “national output” or “domestic production” that have not been well explored. For instance, a given country may investigate some natural phenomenon or social trait inherent to its own land, e.g., diseases that are more widespread in certain countries, or flora and fauna whose habitat is geographically determined, regional geomorphological characteristics, or specific cultural heritages. There may also be global problems that have locally determined features, or particular historical, political, economic or cultural contexts, that could attract research efforts anchored in a specific country or region. In some disciplines, depending on the greater or narrower score of the topic of interest, the spatial realm is a consideration inherent to research design, especially when determinant of the place where empirical evidence will be sought or reconstructed to arrive at contextualized conclusions (Batthyány & Cabrera, 2011). This perspective of the analysis of geographic domains based on spatial delimitation is a novel element in scientific discourse, and may offer indications of the activity that takes place in given countries dealing with topics or problems that are geographically dependent. The notion of territory is understood here in a broad manner, embracing the classical meaning of the portion of surface land that belongs to a given sovereign State, or in reference to the natural, geographic, social or political setting, or other specific uses in the scientific discourse of diverse disciplinary fields (Capel, 2016). As geographic divisions do not necessarily reflect conventional administrative units (Grossetti et al., 2016; Matthiessen et al., 2010), the study of this dimension in the production of scientific knowledge, and the means of representing them in scientific literature, conform a challenging topic, largely unexplored to date.

There is a call to improve scientometric indicators to more properly evaluate global science, and topics potentially relevant for the resolution of grand challenges or societal problems have been put forth (Wilsdon et al., 2017). The main motivation and the contribution of the present study is the search for alternative ways to understand how knowledge production revolves around topics of local interest, and how this information can help in when designing research agendas or prioritizing research portfolios in terms of collaboration with partners.

In light of these considerations, our study proposes to operationalize research of a “national scope” (NS) as the portion of output of international circulation by a given country that makes geographic reference to the country itself in the fields established in bibliographic databases; this, in contrast with the output of the “global scope” (GS), is indicative of territorial problems and the research efforts/relationships on an international level in the output of national scope.

This focus on analysis of a comparative nature among countries is not common in the specialized literature. Our study distinguishes two groups of countries. The first, those in the region of Latin America and the Caribbean (LAC), considered to occupy the scientific periphery; the second, a group of countries at the forefront in developing mainstream science (WORLD). The proposed objectives are: (a) to determine how research on the national scope (NS) is represented and evolves in the greater realm of scientific output, both on a general level and by disciplinary field; (b) to explore whether there are similar patterns between output from the national scope (NS) and that of the global scope (GS) in terms of the language of publication and relationships of international collaboration. Then, (c) focusing solely on the group of LAC countries as a specific regional block, the collaboration among countries is characterized in view of the affinity index, so as to detect collaborating associates who may prove strategic.

This endeavor is meant to contribute to a corpus of new outlooks for identifying topics of local interest and their international connections; as well as for discerning patterns of collaborative communications taking on national and international dimensions, all in the context of knowledge production and circulation. To date these are scarcely contemplated aspects, yet they may prove helpful for the configuration of research agendas.

Related works

While few studies have focused on the geographic dimension as an element of delimiting national research topics or problems, one study looked at international co-authorship and the performance of research groups in Colombia. The underlying assumption was that local problems contribute ‘to the local stock of information necessary to increase local understanding and to produce new knowledge valuable to solve local intellectual, technical, or social issues’ (Ordoñez-Matamoros et al., 2010). Chavarro et al. (2014) adopt this approach to explore the relationship between interdisciplinarity and research topics related with local problems in Colombia. Aguirre-Ligüera and Fontans (2019) use the output expressed in curriculum vitae of researchers to identify local research topics in Uruguay. Other studies refer to Argentina, covering the output of national scope included in Scopus, comprising all areas of knowledge (Miguel et al., 2013; Miguel et al., 2015), the social sciences sciences (Castro-Torres & Alburez-Gutiérrez, 2022; Chinchilla-Rodriguez et al., 2015), agroindustry (González & Chinchilla-Rodriguez, 2020) and Latin American output (Miguel et al., 2019).

Some studies employ natural language processing techniques to derive a useful bibliographic corpus of reference by detecting local research topics from place names (González & Varela, 2019; González et al., 2017; González et al., 2018; González et al., 2019). The coverage and overlap of different sources have been elucidated by comparing SciELO and Scopus (Hidalgo et al., 2019) or undertaking studies located in geographic areas for micro-analysis (Arias & González, 2021). At the global level, and in the case of social sciences, Castro-Torres and Alburez-Gutiérrez (2022) find that the names of countries in the title and abstract of bibliographic registers are more frequent in the output of peripheral countries as opposed to mainstream ones.

In the case of Argentina, and for all the disciplinary areas contained in Scopus, Miguel et al. (2015) reveal that some 25% of the country´s scientific output is oriented to the study of some aspect of national relevance. The figure differs according to discipline, and in the Social Sciences, Economics, Agricultural and Biological Sciences, Earth Sciences and Environmental studies, the percentage of nationally-focused output amounts to nearly 50%. Just 28% of the papers were published in Argentinian or Latin American journals; most were in extra-regional journals. Work without collaboration or in national collaboration prevails, though under international collaboration the links with the United States and European countries (Spain, United Kingdom, Germany, France, Italy) stand out. Relations with other LAC countries are incipient and less intense.

Such findings shed new light on certain assumptions about how disciplines oriented toward the study of local or regional phenomena would be published in national journals and in the local language, beyond the circuit of international circulation (Beigel, 2013a; Castro-Torres & Alburez-Gutierrez, 2022). Accordingly, the scientific systems of peripheral countries would lack the visibility and prestige of the consolidated knowledge producers (Gómez et al., 2022), mainly because of the low level of internationalization of their systems of knowledge output and divulgation (Vessuri et al., 2014), generating in the periphery a tension between publishing globally but perishing locally, versus the opposite, perishing globally and publishing locally (Demeter, 2018; Hanafi, 2011).

Another important factor to bear in mind is the language of publication. It plays a key role in the comparative evaluation of national scientific systems. Publication in a language other than English reduces readership, regardless of the topic and its local or global orientation. In general, lingua franca publications attain an impact level much lower than those put out in English (Egghe & Rousseau, 2000). This trend in communication—not an isolated case—is certainly is not unique of Latin America, affecting non-English-speaking countries overall (Van Leeuwen et al., 2001). In view of previous studies, we assume that in research on national topics in LAC there is a predominance of publishing in some language other than English, unlike the pattern observed for output that does not have a geographic/local focus (Chinchilla-Rodriguez et al., 2015), but it can also be seen in countries of the group WORLD where English is not the maternal language.

Meanwhile, international collaboration is a determinant factor in the modes of knowledge production and diffusion (Adams, 2012). One might argue that the national production centered on topics of local interest is less prone to be undertaken by international research teams. However, studies show that some areas reflect a high concentration of international collaboration in the output of local interest, (McManus, 2020; Plutniak, 2018). Other studies reveal asymmetries from the perspective of relations of scientific collaboration, where the strongest links continue to run North–North, or else in vertical, North–South; but relations within peripheral circuits are still very weak (Beigel, 2013b). Some previous authors find that, despite an increase in intra-regional relationships, the main collaborators partners of the countries in the LAC region are still the US and the top Western European producers of science (Chinchilla-Rodríguez et al., 2012; Santa & Herrero Solana, 2010). Another stream of research looks into the asymmetries in collaboration, working from the premise that scientific collaboration could be reciprocal but not symmetric, as countries differ in size (number of publications) and scientific capacity (economic strength, level of development) (Chinchilla-Rodriguez et al., 2019). Thus, the use of the percentage of documents in international collaboration as the unique indicator to analyze international collaboration can provide unrealistic information about collaborating associates (Melin, 1999), which has led to the use of different indexes to analyze such asymmetries (Chinchilla-Rodríguez et al., 2021; Eck & Waltman, 2009; Fuchs et al., 2021; Leclerc & Gagné, 1994; Zitt et al., 2000). Identifying the strategic associates present in the scientific agenda of many nations can therefore prove useful (Chinchilla-Rodríguez et al., 2021) when defining agendas around local research topics.

Material and methods

Data were retrieved from Scopus database (www.scopus.com), of Elsevier B. V., it being one of the most widely used databases for international bibliometric studies since its appearance in 2004; it moreover affords better coverage than the Web of Science regarding the output from peripheral countries (Moya-Anegón et al., 2007; Singh et al., 2021). This study analyzes output in the period 2009–2018 from the 20 most prolific countries according to the Scopus database when data were retrieved, broken down into two groups: LAC (Brazil, Mexico, Argentina, Chile, Colombia, Cuba, Venezuela, Peru, Ecuador and Uruguay), and WORLD (United States, China, United Kingdom, Germany, Japan, France, India, Italy, Canada and Spain).

In an initial stage, searches strategies were carried out for each country in the study period specified, without restrictions in terms of document type, using the name of the country in the field AFFILCOUNTRY considering abbreviations and various language versions. In this way we recovered the full set of records having at least one signing author who mentioned the country´s name under institutional affiliation. As evidences suggesting that authors select the words in the title strategically (Hartley, 2005; Rosner, 1990), and that the most relevant indexed keywords are chosen by the providers of content and standardized according to a controlled vocabulary, to normalize synonyms, spelling variations, and plural forms (Muñoz-Ecija et al., 2022). In the second stage and for each set of records recovered by country, we searched for output containing a place name and/or names of country in the fields of title or indexed keywords (See Annex-Table 1 for the search strategies applied). Each subset was denominated as output of national scope (NS), and the remainder was identified as output of global scope (GS) (Fig. 1). Percentages of national (NS) and global scope (GS) documents are calculated respect to the total number of documents published by each country.

Fig. 1
figure 1

Example of Argentinean documents of national scope with place name/country name in the title

Full size image

Excluded was the Abstract field, due to the fact that the database for this field includes Copyright information, which usually contains country names, and together with homonyms creates noise in information retrieval, for instance, © Akadèmiai Kiadó, Budapest, Hungary 2014. Such eliminations slightly reduced the volume of relevant documents. (See Annex-Table 2 for the sample of relevant documents from searches for place names/adjectives in the fields of Title and Keywords, as opposed to field Abstract).

Given the resulting corpus of over 18 million documents, we proceeded to determine the total volume of documents from each country and the volume of NS documents. Likewise, for each country the percentage of NS documents over total output—at a general level and by discipline— was determined, and for the classification by discipline, a scheme of 27 categories was used; it is offered by SCOPUS, ASJC (All Science Journal Classification), and categorizes the journals as well as the articles they contain (See Annex-Table 3). These categories were, in turn, broken down into vast areas of knowledge, according to the classification proposed in the Manual of Frascati (OECD, 2015) to appraise the activities within science, technology and innovation. The index of mean annual variation (AGR) served to measure the evolution of output within the study period: \(AGR=\left(\frac{\left(\frac{E}{B}\right)1}{N}-1\right)\times 100\), where N is the period of years; E is the volume at the end of the period; B is the volume at the beginning of the period. Language of publication was determined by the percentage of documents published in English and the native language of each country (if not English).

For the indicators of international collaboration, full counting at the country level was carried out and used. This implies that one point is accredited to the country mentioned in the affiliation of each one of the papers coauthors (Leydesdorff, 1988; Park et al., 2016). On this basis, the international collaboration index of a country would be the percentage of documents signed in conjunction by authors from another country or other countries with respect to the total number of documents (OECD, 2016).

Then, the affinity index (AFI) served to calibrate the relative importance and the asymmetries among countries, measuring the number of collaborative articles published in conjunction, and the total number of international collaborations of each country. To this end, an asymmetric matrix of 20 × 20 was constructed to incorporate the collaborations of the countries studied here. The number of countries was designated as N, and the number of publications in international collaboration between country i and country j was expressed as nij. The values of the diagonal of the matrix correspond to the total national output. Given a dataset containing N countries, the total number of co-authorships in the matrix, referred to as n(…) and n(n), would represent the total of internationally collaborative documents published by the country i (j). To measure the asymmetry in collaboration patterns, we use a pair of inclusion indexes: AFI (i, j), and counterpart, AFI(j, i) with one-way normalization (Zitt et al., 2000). AFI is a measure of the links between a given country (i) with another country (j), compared to the total country links (j) with the entire world during the same period. AFI (j, i) is calculated as: AFI (i, j) = n(i, j)/n(i), where n(i, j) is the volume of links between countries i and j, and n(i) = ∑jn(i, j), total co-authorship linkage of i. This index is size-dependent given that the preference for partner j in AFI (i, j)is influenced by the global size of j (conversely, preference is dependent upon the size of i for AFI (j, i). This index highlights the important partners in terms of quantity and demonstrates asymmetry in partnerships (Chinchilla-Rodriguez et al., 2018b).

Results

General analysis

Even though the output volumes registered for the 10 countries of the group WORLD are much superior to those of the 10 LAC countries —as to be expected— the relative weight of NS research in the LAC group is substantially higher (mean 21.5%) with respect to the main producers of science in the world (mean of 7.8%). This reveals a differential profile in the output of LAC countries, where the studies that are geographically delimited (to self-study of some form) are much more noticeable. Peru and Ecuador stand out in the first group, and India and Spain in the second (Fig. 2).

Fig. 2
figure 2

Volume of total output (blue) and output of a National Scope (orange circles). The tones of blue reflect the weight of total output in the 20 countries, and the orange circles the % of NS production NS

Full size image

Notwithstanding, there is a strong positive correlation (0.95) between the absolute number of NS publications with respect to the total volume of each country´s overall output. Within the LAC group, the value of average growth rate (AGR) in the NS output over the decade of reference is greater (6.3%) than that of the group WORLD (3.4%); and it is furthermore greater that the rest of output (3.4% among LAC and 1.4% in the group WORLD). Only a handful of countries —Mexico, Cuba, Venezuela, the US and Canada—showed an inverse relationship, whereby the NS output grew less than GS research; even in Venezuela and the United States, the value is negative. The LAC countries of study that registered the highest growth rates in NS scientific output were Ecuador (18%), Peru (11%), and Colombia (10%); among the WORLD group they were China (8%) and India (6%) (Fig. 3).

Fig. 3
figure 3

Average growth rate (AGR) in national scope (NS) and global scope (GS) production

Full size image

Disciplinary fields

Both in the LAC group and in the WORLD group, the scientific fields of study exhibiting most NS output are: Agriculture & Biological Sciences, Veterinary Sciences, Earth and Environmental Sciences, Social Sciences, Economics and Business, Psychology, Arts and Humanities. In the area of Medical and Health Sciences, deserving notice are Nursing, and Clinical and Immunological Medicine (only in the group LAC). As seen at the general level, LAC presents more relative weight in terms of NS output under the disciplinary breakdown, and again Peru and Ecuador stand out. Within the WORLD group, however, China is the country whose disciplinary profile most resembles that of the LAC countries (Fig. 4).

Fig. 4
figure 4

Percentage of NS output in the 27 subject areas ASJC (Scopus) grouped into Frascati areas

Full size image

Language of publication

Both in the LAC countries and in the WORLD group (Fig. 5) we find that the English language predominates within the corpus of national output. This is true both of NS documents and GS ones. Still, the LAC countries show a percentage of documents in their native language that is superior to English in the NS output. In the group denominated WORLD, a similar behavior is seen only for Germany, France, Italy and Spain. In Japan, both NS and GS output equal that of native language documents; in China there is an inverse relationship (percentage-wise higher in GS than in NS). In the United States and the United Kingdom, only English-language articles were considered, so that the percentage was nearly 100%; and in Canada, English predominates over French, as seen in the GS output. In India, the Hindi and Bengali languages are very rarely used for publication, English being the prevailing means of scientific communication.

Fig. 5
figure 5

English vs. lingua franca in LAC (left) and WORLD countries (right) for local (%NS) vs. global scope research (%GS)

Full size image

International collaboration in local vs global output

Overall, in the NS documents, the degree of international collaboration is lower than that of GS in all countries except China. Within the LAC group, the proportion of documents involving international collaboration in both groups (NS and GS) is higher than among the group WORLD, especially for countries of lower output: Ecuador, Peru and Uruguay. In contrast, Brazil presents levels of collaboration that match those of the WORLD group. As for global output, these countries —together with Chile—appear to depend greatly on international collaboration, the exception being India. This may reflect patterns of collaboration of countries with limited output who look for international associates to develop their research lines. In the case of Brazil, previous studies have detected its low rate of international collaboration in comparison with its regional peers (Grácio et al., 2020).

The ratio between international collaboration rates in NS output (as opposed to GS) is lower in the LAC countries (as opposed to WORLD). On average, LAC countries show 1.3 documents involving international collaboration of GS for each document of NS; Peru and Ecuador have the least difference (1.1: 1) and Argentina the greatest (1.6: 1). In turn, the WORLD group (excepting China), gives an average ratio of 2.1 to 1. Noteworthy are the highest difference (India, 3 to 1) and the lowest (Japan, 1.6 to 1) (Table 1).

Table 1 International collaboration output of National Scope vs. the rest of output. NS: total number of “local scope” documents in international collaboration; % NS: percentage of “local scope” documents in international collaboration; % GS: total “global scope” documents in international collaboration; % NS: percentage of “global scope” documents in international collaboration
Full size table

Main collaborators in LAC countries

Here, we seek to focus on the international relationships of the LAC countries as a case study. Taking as the unit of analysis the proportion of documents involving international collaboration within the NS output in group LAC, we find that in 8 of the 10 countries, the United States is the main collaborator, with a more substantial presence in Peru (32%) and Ecuador (25%), followed by Mexico (17%), Colombia (13%), Chile and Venezuela (12%), Argentina (10%) and Brazil (8%). In the case of Uruguay and Cuba, collaboration with the US occupies the third place in our list. But LAC shows a noteworthy level of collaboration with Western Europe, especially the UK, Germany, Spain and France. Portugal is mostly linked to Brazil; Italy with Brazil and Argentina.

Despite the fact that the United States tends to be the main collaborator partner of most countries, tconsidering only the percentage of collaboration of each country, this indicator does not reveal other potential partners with whom stronger relationships are established, taking into account all possible partners.To illustrate this point, we analyze the percentages of documents in international collaboration among countries and the Affinity Index (AFI), which allows us to derive the principal partners of each country taking into account all the possible associates. These indicators are represented as % GS and AFI_GS for the global output, and % NS, and AFI_NS for national output. For instance, Brazil´s main partner would be the US, with whom it produces 8.2% of its NS output, whereas Brazil contributes only 0.4% to the NS output of the United States. In contrast, if we take the AFI into account, its main collaborator partner is Uruguay. As for GS output, the US represents 10% of Brazil´s production, while Brazil itself contributes only in 1% of what the US published. Uruguay continues to stand out as the main scientific associate of Brazil in its global output.

Identifying these asymmetries in the production of a national scope and that of a global scope can help to discover partners that may be relevant, despite not being the ones with the most documents in collaboration. Figure 6 shows the percentages of international collaboration for output of a local scope (% NS) or global one (% GS) in LAC countries, along with the affinity index values (AFI).

Fig. 6
figure 6

Percentage of national scope (% NS) and global scope (%GS) documents in international collaboration along with the value of the affinity index for national scope (AFI_NS) and global scope (AFI_GS)

Full size image

Brazil produces 4% of its NS output in conjunction with 20 LAC countries (out of a total of 157 countries altogether). Its main associates lie outside its geographic region, Argentina, Mexico, Colombia and Chile respectively occupying positions 9, 12, 14 and 16 in the ranking of countries with the highest percentages of international collaboration. It is a noteworthy collaboration partner for almost all the LAC countries, yet particularly for Uruguay and Peru (% NS). Its collaboration with Mexico is quite balanced, even though Mexico ranks 12 among the regional collaborative efforts (% NS). Uruguay and Colombia, and Spain and France, are the closest extra-regional associates, according to the AFI. In terms of worldwide output, the contribution of the LAC countries is amplified, and among the WORLD group, we see a more balanced picture of collaboration with Spain than with the other countries, even though Spain is 6th in the total ranking of collaborators in output. If we do not only bear in mind the bilateral relations (percentages of international collaboration) but rather all those given among the 20 countries of study, Brazil is seen to present a very different portfolio in terms of international relations. The strongest associates according to this index would include all the LAC countries, with Argentina and Chile at the lead; and as main extra-regional associates, Canada and China loom into view.

In the case of Mexico, nearly 40% of its national output is published in conjunction with institutions lying outside the LAC region, especially with North American institutions (20%). In percentages of international collaboration (% NS), the main LAC associates are Brazil, Argentina and Colombia, and Mexico has an important relationship with Cuba (% NS). The US does not have strong research ties with Mexico, despite showing higher NS values, but Cuba and Venezuela do. For global output, 16% of Mexico´s international research efforts involve LAC countries, with Brazil, Colombia, Chile and Argentina as the countries with greatest % NS (10% of the total). A comparison of the percentages of collaboration and the AFI reveals that the main LAC associates are Cuba, Ecuador and Peru; among the WORLD group, Spain and France.

Argentina collaborates with 142 countries on the world level, and with 20 on the LAC level. This constitutes 13% of its output of national scope. The countries with highest values in NS are the US, Spain, Brazil and Germany; meanwhile, the AFI indicates they are Uruguay, Chile and Peru (among the LAC region) and at a much greater distance, Spain and Canada as closest collaborators. Bearing in mind the collaboration of Argentina in the LAC realm, Uruguay stands out because of its roughly 15% of output in conjunction with Argentina (% GS). In contrast, we see that Chile shows more production in collaboration, this asymmetry evidencing the key role of Argentina in Uruguay´s output of national scope, while Chile shows values in the context of Argentina´s output that resemble the AFI. It collaborates with 20 LAC countries, on 25% of its global output: the US, Spain and Brazil occupy the top spots as collaborators (% GS), but according to the AFI the main associates would be Brazil, Chile, and Colombia among the LAC, and Canada, China and France among the WORLD.

Chile produces 13% of its national output with 25 LAC countries, and the rest with 122 extra-regional countries. The main LAC collaborators are Argentina and Brazil (% NS), and Chile itself is an important collaborator for Peru and Uruguay (% NS). The AFI values displace the positions of the US and Brazil (% NS) and show Uruguay, Argentina and Peru to be the most significant LAC associates, while extra-regionally, Spain and Italy. On topics of a global nature, it publishes 26% of its scientific texts with 26 regional neighbors, most notably Brazil, Argentina and Colombia, though it is a relevant associate for Peru and Uruguay as well (% GS). In view of the AFI, Peru, Uruguay and Argentina would be the strongest LAC contributors, and in the WORLD, context described here, Spain, France and Germany prevail.

Colombia collaborates with 157 countries worldwide, 21 of them LAC countries, with which it accumulates 18% of its output of national scope. Its main collaborators are Brazil, Mexico, Chile and Argentina (summing up 12% of their national scope output,—% NS) and in the group WORLD, the US and Spain. Colombia appears to be an important partner for Venezuela and Ecuador (% NS) and the AFI values confirm that the main associates are Venezuela, Ecuador and Peru.

In its output of global scope, Colombia collaborates largely (25%) with 17 LAC countries, and on a greater proportion of documents (% NS) with the US, Spain, Brazil and the UK. According to the AFI, it is a relevant associate for Ecuador, Venezuela and Peru; and among the extra-regional lands, Spain, France, Italy and Canada. Among the main collaborators —whether national or global— we find mostly Spanish-speaking countries.

Cuba presents a pattern of collaboration slightly different from the rest of the LAC countries. In its output of national scope, it collaborates with 25 LAC countries, on 22% of its production, while the rest is produced with 70 countries beyond the region. Cuba´s collaborators are, in terms of % NS, Spain followed by Mexico and the US; it is moreover an important associate for Ecuador and Venezuela (% NS and AFI) and, among the WORLD group, for Spain, India, Italy and China. In its output of world scope, it collaborates to an extent of 38% of output with 32 LAC countries. Spain, Mexico and Brazil are the predominant producers in collaboration with Cuba, but its strategic associates are Ecuador, Peru, Venezuela, Spain, France, Germany and Italy (AFI).

Venezuela publishes 28% of its output of national scope in conjunction with 20 LAC countries, and the rest with 74 from outside the region. Among its outstanding collaborators we find five LAC countries; in its global output the pattern is repeated, with Colombia and Mexico as main regional partners. A different portfolio is revealed by the AFI, with Ecuador, Colombia and Cuba appearing as its main LAC associates, and Spain, France and India standing out in the group WORLD. In its output of global scope, the % NS values show the US, Spain, and Colombia to be the countries with which most documents are published, although the relations of collaboration are strongest with Ecuador, Peru and Colombia, or with Spain, France, and Italy in the WORLD group.

Peru publishes 22% of its local research results in conjunction with 21 countries of the LAC region, out of the 144 countries of collaboration at the world level. The associates with whom it publishes a greater proportion of documents (% NS) are the US, the United Kingdom and Spain. In turn, Brazil, Chile, Argentina and Mexico are its outstanding LAC associates. Still, when all the potential associates are considered (AFI), the preferences change and Ecuador, Venezuela and Chile appear as the main LAC allies, whereas Spain, India and the USA do so at the world level. Peru would appear to be a key partner for Ecuador. In its output of a global scope, it publishes 70% together with 22 LAC countries, most notably the US, Brazil and Spain (% NS), having as strategic LAC associates Ecuador, Cuba, Venezuela and WORLD collaborators Spain, France and USA. Peru presents a distinctive pattern among the LAC countries, displaying a strong regional component within its output of global scope.

Ecuador collaborates with 23 counties in the LAC region, giving rise to 29% of its local output. The rest is produced with more than 100 extra-regional countries. Its main associates are the US, Spain and Germany (% NS); among the LAC most documents in international collaboration involve: Mexico, Brazil and Colombia. According to the AFI, however, the most solid collaborators among the 20 countries are Venezuela, Cuba and Peru within the LAC, and Spain, Germany and the US beyond. In terms of global scope output, the strategic allies (AFI) are Venezuela, Peru and Cuba, or (WORLD) Spain, France and Italy.

Uruguay shows a remarkable pattern of collaboration, with Brazil and Argentina at the forefront (28%), in its national scope output. It collaborates with 20 LAC countries on 45% of local topics production, and the remainder with 63 countries outside the region. Its strategic associates are Argentina, Venezuela, Chile and Spain, France and Canada, both at the local and at the world level. Uruguay´s output involves 20 LAC countries (60%), Brazil and Argentina playing a relevant role, providing similar proportions to those of the US in % GS, though they lag well behind when the value of reference is AFI_GS.

Concluding remarks

This study looks at knowledge production by focusing on local and global topics of interest within a group of countries in the Latin American and Caribbean region (LAC), considered to be on the scientific periphery, as compared to countries at the forefront of WORLD mainstream science (including non-English speaking countries). We demonstrate that LAC countries present a particular pattern of knowledge production, characterized by a strong tendency to deal with issues of local interest published in lingua franca. It may be seen as a “brand of identity” that differentiates these countries from the countries of the global north. This notion would be supported by the results of a Social Science study reported recently (Castro-Torres & Alburez-Gutierrez, 2022).

At the same time, we explored whether similar or different patterns arise between the two groups of countries at the disciplinary level. Our findings are not surprising in that Arts and Humanities, as well as Social Sciences in general, and Economics and Business in particular, are the disciplines having a stronger component of national-oriented research—i.e., documents with geographical reference to the country (NS)—followed by Agricultural, Veterinary, Earth and Environmental Sciences (Fig. 3). Although the proportion of NS research is uneven among countries, Nursing, Clinical Medicine and Immunology stand out within the health sciences, in some LAC countries (Ecuador, Peru and Venezuela) and in the World group (China, India and Japan). This suggests that while areas such as Humanities, Social Sciences and Agricultural Sciences cope with problems inherent to a country, international collaboration may help broaden the knowledge generated; and because the main concepts are universal, they thus contribute to global science (McManus et al., 2020).

Our results show research co-authorship to be useful for combining different skill sets, especially for specific problems at the global level, but also for national-oriented topics. LAC countries collaborate as much on local output as for global output, with a similar number of collaborators countries, and a somewhat higher proportion of documents in output of a global scope. Important aspects can be underlined in this sense. The first, that even when relationships in peripheral circuits are still very weak (Beigel, 2013b), the LAC countries collaborate among themselves to a greater extent on work of a national nature rather than on global topics. The second point is, that in all cases, the percentage of output of a national scope produced with countries from outside the region is much greater than for LAC countries, which suggest that there may be scientific communities interested in the same key research topics. However, as collaboration is driven by multiple factors, this interest in local topics might be explained from different perspectives. First, internationally collaborative publications with a particular country may not necessarily stem from the interest of foreign co-authors in national issues but rather from the use of infrastructure located in a specific territory (e.g., observatories in Chile or in South Africa), or by the necessity of natural resources in tropical countries to study biodiversity (Stefanoudis et al., 2021); or the case in which researchers initiate short-terms collaboration in peripheral countries “to get the quota” in international projects where diversity and inclusiveness are a mantra in research agendas (Gewin, 2023). In some cases, these international collaborations might not address local research needs and research is completed without any further effective communication or engagement with others from that nation, which have been coined as parachute science (De Vos, 2022) or helicopter research (Adame, 2021).

To better understanding of the implications of establishing relationships in topics at the local and/or global level, this study presents a case study of the application of asymmetry measures to locate strategic partners. Using a set of the most prolific countries in the LAC region and at the worldwide level, we present an egocentric view of each country involved, in order to analyze the strength and reciprocity of the relationships of each of the various countries. Here we demonstrate an asymmetrical relationship between each of these countries and the United States —the US was found to be a dominant country in terms of collaboration for most countries under study, though representing a small fraction of the share of scientific relationships stemming from the US. Identifying these asymmetries in the production of a national scope and that of a global scope can help to discover partners that may be relevant, despite not being the ones with the most documents in collaboration.

As the internationalization of science affects the development and research performance of countries and, to some degree, depends on the attractiveness of a partner in the global network, (infrastructures, number of qualified researchers, economic wealth, natural resources, etc.) the country with which the scientific relationship is established is important for indicator construction, given the unequal magnitude of contributions between partners (Chinchilla-Rodríguez et al., 2018a). The results of this study indicate that, despite the low volume of international publications for many countries, the number of countries reached is relatively high and that different collaboration strategies can be designed to address topics of local and/or global interest. Therefore, this internationalization implies policy challenges and opportunities, particularly for peripheral countries, when designing research agendas or mobility programs. It also suggests a need to enhance the attractiveness and scientific capacities of certain countries in the global scientific system.

Limitations and future work

The first limitation to be acknowledged here is the diffuse distinction between local vs global scope (Marginson, 2022). Adopting geography as the framework for exploring research endeavors on topics/problems grounded in their own land can create difficulties for translating informational needs into a search equation to be used in a bibliographic database, especially when working on a global scale. The case at hand called for retrieving records of publications where the countries were the object of study or the geographic setting of the topic. This approach is certainly not the only one possible among all the terminological options with reference to studies about local interests or those of a national scope. Still, it did allow for a set of documents based on objective criteria to be retrieved, being reproducible for any posterior study or pertinent adaptation to other geographic domains (Miguel et al., 2013). To analyze a single country, one could descend to more specific levels of aggregation, including the place names of regions and provinces, administrative units, historical names of territories within the search strategy (González & Chinchilla-Rodríguez, 2020); include the field “Abstract”, discarding irrelevant documents, to adapt this methodology to study specific countries (Castro-Torres & Alburez-Gutiérrez, 2022; Miguel, et al., 2015) or use Natural Language Processing (NLP) software’s such as the Stanford Named Entity Recognizer (NER) to training sequence models on labeled data (Finkel et al., 2005; González & Varela, 2019; González et al., 2018). In this study, we are aware of the limitation regarding search’s specificity, accuracy and comprehensiveness. In future work, we will combine and explore approaches to refine local vs globally oriented outputs in retrieving information retrieval.

Then again, this study uses the Scopus database, which has limitations despite its better geographical coverage with respect to the Web of Science. There is empirical evidence that these two mainstream databases strongly under-represent the scientific production of developing countries, and over-represent production by industrialized countries (Rafols et al., 2015). The potential biases introduced by the use of Scopus should be considered to avoid underestimating domestic capabilities and research agendas that are more attuned to local needs, for example in the global south. Regional databases such as Scielo could serve as complementary data sources for analyzing LAC countries in further studies, to compensate for the fact that local science and journals are not perceived as participating in mainstream science.

The AFI is not calculated for the countries of the group WORLD simply because that lies beyond the focus of this attempt to elucidate the relations of scientific strength among developed countries. The main objective is to capture science of a national scope within the LAC region, making comparisons with the countries of the WORLD functional as a framework of reference. The results of the study raise a series of questions to be addressed in future research efforts. As evidences shown that collaboration between high-and low-income countries fail to establish long term and equitable collaborations (Odeny & Bosurgi, 2022), we ask whether are we dealing with sporadic or consolidated associations over this time period? Are the collaborations open research lines that could be shared with these or other countries? In which areas?

Although there is ample empirical evidence of disciplinary differences in terms of co-authorship, collaborations are believed to be more important in the natural sciences (Lewis et al., 2012). Social sciences and humanities scholars also experiment with collaboration as an effective research strategy, but it is comparatively rare in their home fields (Graham-Bertolini et al., 2019). It may be that national output regarding topics of local interest is hardly undertaken by international teams, even though some studies show the contrary is true in certain areas (McManus et al., 2020; Plutniak, 2018). Local topics under study might be of secondary importance or not valid from the national point of view because they might be imposed by high-income countries, with the periphery serving as an exemplification. In other words, articles based on cases from different countries do not necessarily mean that foreign partners are interested in issues inherent to a particular country but rather seek confirmation of global theories or trends by comparing case studies. Therefore, analyzing disciplines in which collaboration occurs would facilitate interpreting the results obtained. The disciplinary differences in addressing national issues should be the subject of future research to deepen the knowledge that the geographical dimension acquires in scientific research, either to gain better understanding about whether the research agendas are set by the international partners that provide financing, or whether they are focus on local issues of national interest that can lead to reinforce economic and/or scientific capacities.

Finally, the different roles played by collaborators might tell us something more about the engagement of researchers from peripheral/scientifically advanced countries in international collaborations. In this sense, key authorship positions in most disciplines (except in those dominated to alphabetical order such as economics) are an indicator of the contributions made to the research for each one of the authors involved in publications, and are correlated with access to fundings, promotions and academic tenure (Akudinobi & Kilmarx, 2022). Evidences show that authorship hierarchies in which local authors are by default assigned middle positions instead of key positions (first, last or corresponding author) are perpetuating scholarly and geographic inequalities (González-Alcaide et al., 2017; Odeny & Bosurgi, 2022). Authors providing funding usually adopt the corresponding author role leaving other positions for the rest of collaborators (Willems & Plume, 2021; Zhang et al., 2022) Therefore, the study of authorship positions and contributorships (Robinson-García et al., 2020) might help to shed light on the different roles that authors from peripheral countries play in international collaborations.