Of the 145 scientific journals with the 489 papers related to biodiversity conservation in the insular Caribbean for the period 2000–2015, eleven jointly accounted for 40% of all published papers. These were PLOS One (40 papers), Molecular Ecology (20), Molecular Phylogenetics and Evolution (20), Biological Conservation (18), Caribbean Journal of Science (18), Coral Reefs (18), Marine Ecology Progress Series (18), Conservation Biology (13), Revista de Biología Tropical (13), Conservation Genetics (9) and Zootaxa (9).

Overall, papers with only foreign-based authors accounted for 63.4% of the total number of papers, followed by papers with both Caribbean-based and foreign-based authors with 28.6%. Papers with only Caribbean-based authors accounted for 8.0% of the total number of papers. The total number of papers increased over time, going from 9 papers observed in 2000 to 54 in 2015 (Fig. 1a; Model A in Tables 1 and 2). There was also a significant increase in the number of authors per paper over the same period (from 1.7 authors per paper in 2000 to 4.4 in 2016; Fig. 1b; Model B in Tables 1 and 2).

Fig. 1.

Temporal trends (period 2000 to 2015) in peer-reviewed papers on biodiversity conservation in the insular Caribbean shown as (a) total number of papers (all authorship categories combined), (b) number of authors per paper, and (c) number of papers as a function of authorship category. Generalized linear model (quasi-Poisson) fits of the number of papers or authors as a function of year are shown as lines (along with corresponding 95% confidence bands).

(0.27MB).

Temporal trends in the number of papers differed significantly among authorship categories, as evidenced by a statistically significant interaction between year and authorship category (Model C in Tables 1 and 2). Papers by Caribbean-based authors did not increase significantly over time (Fig. 1c; Model C in Table 2). In contrast, papers by foreign-based authors and by both Caribbean- and foreign-based authors did increase over time (Fig. 1c), although only the latter category showed a rate of increase that was significantly larger than that of the Caribbean-based authors (Model C in Table 2).

The proportion of Caribbean-based authors per paper remained low over the entire 16-year period, with an overall average of 17.4% of all authors per paper per year (±8.0% standard deviation, n = 16) (Fig. 2a). The odds of having a Caribbean-based author in a paper did not change significantly over time, but decreased significantly as the number of authors per paper increased (Fig. 3a; Model D in Tables 1 and 2).

Fig. 2.

Temporal trends (period 2000 to 2015) in the contribution of insular Caribbean-based authors to the production of peer-reviewed papers on biodiversity conservation in the insular Caribbean shown as (a) average proportion of Caribbean-based authors per paper, (b) average proportion of papers including at least one Caribbean-based author, (c) average proportion of papers led by a Caribbean-based author, and (d) average proportion of papers with Caribbean-based authors also led by a Caribbean-based author.

(0.21MB).

Fig. 3.

Modelled effects of the number of authors on (a) the proportion of Caribbean-based authors per paper, and (b) the proportion of papers including at least one Caribbean-based authors between 2000 and 2015, using a quasi-binomial and a binomial GLM, respectively. Solid black lines and grey enveloping bands show model fits and corresponding 95% confidence intervals. Blue circles are the 489 papers (some jitter was added to their position to help visualization).

(0.15MB).

The proportions of papers including Caribbean-based authors and those led by Caribbean-based authors also remained low throughout the entire 16-year period (Fig. 2b and c), with an observed average of 31.9% of papers per year (±15.0% SD) and 22.0% of papers per year (±11.6%), respectively. The odds of having papers of either type did not significantly change over time (Models E and F in Table 1). In contrast, there was a significant and positive effect of the number of authors on the odds of having a paper including Caribbean-based authors (Fig. 3b; Model E in Tables 1 and 2), but not on the odds of having a paper led by Caribbean-based authors (Model F in Table 1).

Finally, when only papers including Caribbean-based authors were considered, the proportion of these papers that were also led by Caribbean-based authors decreased over time, going from 100% in 2000 to 62.1% in 2015 (Fig. 2d). The odds of having such a paper led by Caribbean-based authors significantly decreased over time, after accounting for the effect of number of authors (Fig. 4a; Model G in Tables 1 and 2). Moreover, these odds were also significantly and negatively related with the number of authors, after accounting for the effect of time (Fig. 4b; Model G in Tables 1 and 2).

Fig. 4.

Modelled effects of (a) year and (b) number of authors on the proportion of papers including Caribbean-based authors that were also led by Caribbean-based authors between 2000 and 2015 using a binomial GLM. Solid black lines and grey enveloping bands show model fits and corresponding 95% confidence intervals. Blue circles are the 489 papers (some jitter was added to their position to help visualization).

(0.13MB).

The median impact factor of papers with only foreign-based authors and of papers with both Caribbean- and foreign-based authors over the entire 16-year period was relatively similar at 2.275 (1st and 3rd quartiles: 1.589 and 3.916) and 2.844 (1st and 3rd quartiles: 0.245 and 3.299), respectively. In contrast, the median impact factor of papers with only Caribbean-based authors over the same period was considerably lower at 1.065 (1 st and 3rd quartiles: 0.4385 and 3.279). However, the impact factor of Caribbean-based papers steadily increased over time and reached values similar to those of the other two categories by the end of the study (Fig. 5a). This difference in trends in impact factor between authorship categories was confirmed by the model selection, which retained an interaction term between year and authorship category in addition to the number of authors as predictors (Model J in Tables 1 and 2). It also indicated that the log (impact factor) significantly increased with the number of authors, after accounting for the interactive effect of authorship category and year (Fig. 5b; Model J in Table 2).

Fig. 5.

Modelled effects of (a) year and authorship category and (b) number of authors in a paper on the (log-transformed) impact factor of 488 papers (circles) between 2000 and 2015 using a Gaussian GLM. The various lines and enveloping bands show model fits and corresponding 95% confidence intervals. Some jitter was added to the position of the papers (circles) to help visualization.

(0.18MB).

Focusing only on the papers that included Caribbean-based authors, but distinguishing among papers with authors based in SS versus those based in OT revealed marked differences in their respective contributions to the total number of papers published. Indeed, over the entire 16-year period the majority of the papers (58.1%) came exclusively from OT (OT only), followed by papers from SS (SS only; 38.0%), with a negligible contribution of papers from collaborations between authors based on both OT and SS (SS & OT; 3.9%). There were also marked differences in observed trends among these categories. SS only and OT only papers increased similarly overtime whereas SS & OT papers did not increase (Fig. 6). The difference in temporal trend between SS & OT papers and the other two categories was confirmed statistically by a significant interaction between year and author location (Model I in Table 1), with SS only and OT only papers exhibiting nearly identical slope estimates (Model I in Table 2). In fact, re-running the analysis after excluding SS & OT papers confirmed that rates of increase did not differ significantly between SS only and OT only (location × year interaction: p = 0.968), whereas there was a significant effect of both year (p < 0.001) and location (p = 0.0312) in the simplified model, with OT only consistently producing more papers than SS only at any given time (Fig. 6).

Fig. 6.

Modelled effects of year and Caribbean-based author location on the (square root transformed) number of papers (circles) that included at least one Caribbean-based author between 2000 and 2015 using a Gaussian GLM. The various lines and enveloping bands show model fits and corresponding 95% confidence intervals. Location is determined by whether authors in a paper were based on overseas territories (OT only), sovereign states (SS only), or a combination of both (SS & OT). Note that the response variable was back transformed for plotting.

(0.11MB).

Further comparing SS only and OT only papers (which jointly accounted for 96.1% of papers with Caribbean-based authors) indicated that both categories were largely dominated by papers involving collaborations with foreign-based authors (76.0% of papers for SS only and 82.4% for OT only). These proportions did not differ significantly between both categories over the study period (chi-square test: X2 = 0.652, d.f. = 1, p = 0.419). Moreover, there were no significant differences between both categories in log (impact factor) (Gaussian GLM: F = 2.02, d.f. = 1, 169, p = 0.1569) nor in the number of authors per paper (quasi-Poisson model: F = 0.36, d.f. = 1, 170, p = 0.5475) over the study period.

The USA overseas territory of Puerto Rico accounted alone for nearly 23.7% of all paper authorships (Fig. 7; Fig. S1 in Supplementary material). This was followed by the sovereign state of Cuba, accounting for nearly 12.8% of all paper authorships (Fig. 7; Fig. S1). Other important OT contributors were the USA Virgin Islands and Curacao (Netherlands OT) with 9.5% and 8.5% of all paper authorships, respectively (Fig. 7; Fig. S1). Other important sovereign state contributors were Jamaica (5.7%) and the Dominican Republic (5.2%; Fig. 7; Fig. S1). Comparatively, on the lower end, two SS (Haiti and Antigua & Barbuda) and one British OT (Anguilla) each accounted for only 0.6% of all paper authorships (Fig. 7; Fig. S1).

Fig. 7.

Connectivity network showing paper collaborations between 2000 and 2015 among locations (circles) that were Sovereign States (grey fill) and Overseas Territories (British: green fill; French: yellow; USA: orange; Netherlands: light blue) as well as the relative contribution of each location to total number of papers published (which is proportional to circle diameter). The width of the line connecting pairs of locations is proportional to the number of paper collaborations between locations.

(0.17MB).

A closer look into the level of collaboration among Caribbean-based authors from different insular Caribbean locations (whether SS or OT) showed that, overall, 58.3% of all paper authorships involved collaborations with foreign-based authors, but not with Caribbean-based authors from other locations. This was followed by 19.9% of paper authorships that included both foreign-based as well as Caribbean-based authors from other locations, and by 15.6% of paper authorships that did not involve collaborations with any authors from other locations (whether foreign- or Caribbean-based). Finally, only 6.2% of paper authorships involved collaborations with Caribbean-based authors from other locations, but not with foreign-based authors.

The connectivity network showing paper collaborations among Caribbean locations (whether SS or OT) revealed a few small collaboration networks over the entire 16-year period (Fig. 7). The French OT (Guadeloupe and Martinique) had collaborated between themselves, but not with other locations in the region (Fig. 7). Most of collaborations of the Netherlands OT (Saba, Curacao, St Eustasius, St Martin and Bonaire) had taken place among themselves (Fig. 7). The OT of the USA (Puerto Rico and US Virgin Islands) shared a relatively large number of intra-collaborations (Fig. 7). Another small network of collaborations involved Barbados, St Kitts and Nevis, Grenada, the Bahamas, and the British Virgin Islands (Fig. 7). The last conspicuous network involved the Greater Antilles (Cuba, the Dominican Republic, Puerto Rico, Jamaica and, to a lesser extent, Haiti) (Fig. 7). In contrast, a couple of SS (Dominica and Antigua & Barbuda) and British OT (Monserrat and Anguilla) lacked collaborations with other locations during the study period (Fig. 7). Finally, Puerto Rico, in addition to being the major contributor to the total number of papers published, was the location most likely to be involved in collaborations with other locations across the region, whether SS or OT, thus helping connect most research networks (Fig. 7).

Our survey of the peer-reviewed literature on biodiversity conservation in the insular Caribbean region between 2000 and 2015 has revealed a low contribution of Caribbean-based authors as measured by a variety of metrics. We found that both the total number of papers published and number of authors per paper increased over time and this translated into multi-author papers with at least one Caribbean-based author being the most rapidly growing contribution to the literature. However, the latter came with no concomitant increases over time in the relative contribution of Caribbean-based authors to paper authorship, whether measured as the proportion of authors that were Caribbean-based or the proportion of papers that included at least one Caribbean-based author or which were led by a Caribbean-based author. In fact, when we focused only on papers that included at least one Caribbean-based author, we found that the proportion of such papers that were also led by a Caribbean-based author decreased over time.

Independently of time, we found contrasting effects of the number of authors in a paper on the contribution of Caribbean-based authors to paper authorship. Whereas increases in the number of authors led to higher odds of having at least one Caribbean-based author in a paper, it also led to lower odds of having additional Caribbean-based authors in the authorship, and had no effect on the odds of having a paper led by Caribbean-based authors. Our results are thus in line with a global pattern of increased inclusion of authors from the Global South via multi-author collaborations, but without necessarily reflecting any improvement in the relative contribution of such authors to lead (and/or corresponding) paper authorships (Gonzalez-Alcaide et al., 2017; Malhado et al., 2014; Mammides et al., 2016). This is in line with recent studies showing a persistent under-representation of the Global South in biodiversity conservation science (Campos-Arceiz et al., 2018; Espin et al., 2017; Gaillard, 2010; Livingston et al., 2016; Melles et al., 2019; Stocks et al., 2008; Tydecks et al., 2018).

A decrease in the contribution of Caribbean-based authors to both paper authorship and lead authorship as the number of authors increases might simply faithfully reflect the respective contributions of all authors, since foreign-based authors will often have greater access to the funding and cutting-edge technology (and associated know-how) (Dangles et al., 2016; Habel et al., 2016; Malhado et al., 2014); this would facilitate foreign-based authors leading the research effort. However, in other instances, this might result from questionable practices. For example, it might reflect an obligation of foreign-based authors to associate local Caribbean-based authors to their projects to either be eligible for cooperation programs with the Global South or to obtain relevant permits from local authorities (Barber et al., 2014; Habel et al., 2016), even though the Caribbean-based authors made negligible contributions. Independently of the reasons for this underrepresentation, persistent non-lead authorships could ultimately be detrimental to the careers and international recognition of Caribbean-based authors. This might also stall the much-needed increase in the representation of biodiversity-rich regions among editors and reviewers of conservation journals (Espin et al., 2017; Primack et al., 2017).

We found that the impact factor of a paper increased with the number of authors per paper, and this was the case irrespective of authorship origin, which is qualitatively consistent with the observation that collaborations tend to yield papers of higher global significance in the natural and social sciences (Larivière et al., 2015). More interestingly, we found that, over the entire study period, papers by only Caribbean-based authors were associated with a lower impact factor than those that included foreign-based authors. However, our findings also indicated that the impact factor of papers published by only Caribbean-based authors increased steadily over the 16-year period, reaching levels similar to those of the two other authorship categories. Although this may indicate an improvement in the global significance of local research capacity, such improvement was not matched by a concomitant increase over time in the total number of papers produced solely by Caribbean-based authors. Thus, without collaboration with foreign-based authors, important limitations to the production of papers in journals with high impact factor persist within the region.

Different non-mutually exclusive factors can help explain the lower number of papers published by Caribbean-based authors, particularly in journals with high impact factor. These factors include among others: research mainly focused on addressing locally relevant problems; a pervasive lack of access to research funding and cutting-edge research technology; an insufficient number of well-equipped biological stations across the region; high teaching loads for academic staff; limited opportunities for training and conference attendance; low access to international journals; language barriers; low incentives for peer-reviewed publications by post-graduate students; and low interest in pursuing careers in conservation (see Table S1 in Supplementary material for further details). Many of these factors apply to other biodiversity-rich regions of the Global South (e.g. Barber et al., 2014) and operate under a persistent background of “brain drain” emigration (Gaillard, 2010; Mishra, 2006). We do, however, acknowledge that the relative importance of these diverse factors is likely to vary considerably within the insular Caribbean region due to differences in historical context (see below).

Caribbean-based authors in OT consistently produced more papers than those from SS, even though both groups were equally likely to engage in collaborations with foreign-based authors. However, there is a high variability in the number of papers within both OT and SS. Indeed, the USA OT of Puerto Rico singlehandedly accounted for about ¼ of all Caribbean-based publications, in spite of its population size being only about 1/3 of that of larger locations such as Cuba, the Dominican Republic and Haiti (UN, 2019), making this OT by far the major Caribbean-based contributor to the number of papers published.

The disproportionate overall contribution of Puerto-Rico to the number of papers published in the region has been corroborated independently and goes beyond the life sciences (Ortiz-Rivera et al., 2000). This likely partly reflects the fact that Puerto Rico has one of the best records globally in education development over the past 50 years, leading to a relatively high proportion of college-educated adults (Ladd and Rivera-Batiz, 2006) and to a large base of adequately-trained human resource capacity. As such, the important relative contribution of Cuba for the sovereign states might better reflect the fact that Cuba not only has the largest human population of the insular Caribbean, but that it also has one of the highest proportions of university-educated adults in the region (Roser and Ortiz-Ospina, 2013). Given the contrasting historical, socio-economic and political contexts of Puerto Rico and Cuba, with the latter suffering from the embargo imposed by the USA, our findings suggest that, despite the cultural, linguistic and socio-economic heterogeneity that exists within the insular Caribbean, investment in tertiary-level or higher education remains one necessary condition to increase local research capacity in biodiversity conservation. In that regard, and despite a significant development of the insular Caribbean university system over the last 50 years (Cobley, 2000), several SS still rely on a few small universities with limited funding (Obino et al., 2017) and interdisciplinary research capacity (CDKN, 2012). In any case, there is great need for a more historically-informed understanding of how local institutions (universities, research stations, government agencies, NGOs) across the region have embraced (or not) research in disciplines related to biodiversity conservation to help guide the way forward; gaining such understanding will likely require directed surveys (Gaillard, 2010), more interdisciplinary communication, and input from historians of science in the region.

Our results provide insights into intra-regional cooperation in research on biodiversity in the insular Caribbean. Only about one quarter of all paper authorships involved collaborations among Caribbean-based authors from other locations within the insular Caribbean region (whether SS or OT). In fact, when authors from outside the Caribbean (foreign-based) were not involved, Caribbean-based authors were more likely to publish on their own (or with authors from the same Caribbean location) than via collaborations with authors from other Caribbean locations (whether SS or OT) . Our analysis also revealed a fragmented landscape with just a few small research networks, most of which appeared to reflect geopolitical legacies (notably the OT clusters of France, the Netherlands and the USA), supporting a particularly important role of cultural and historical post-colonial connections among locations in scientific production (Cayuela et al., 2018). In contrast, only one network, that of the Greater Antilles, seemed to capture high strategic importance in biodiversity conservation. Thus, despite sharing a large fraction of biodiversity and a number of threats to such biodiversity, such as biological invasions (e.g. Betancur-R et al., 2011), we found little evidence of strong collaboration. Some of the factors that likely hinder such collaborations include the relatively high cost of travelling between locations within the region and language barriers. In relation to the latter, the official bilingual status of Puerto Rico (English and Spanish) might help explain why it was also the most inter-connected location.

The insular Caribbean is one of the most rewarding areas for research on biodiversity and evolution in the world and is also one of the most threatened (Brooks et al., 2002; Myers et al., 2000). We also agree with the contention that tackling global environmental problems will require the rapid development of ecology science in the Global South (Livingston et al., 2016). As such, one would expect the numerous research programs implemented there by foreign-based research teams to contribute to increasing the productivity of Caribbean-based researchers through direct cooperation, training and capacity building. Indeed, global increases in scientific cooperation between countries of the Global South and the Global North in tropical ecology and conservation in recent years have been reported (Dangles et al., 2016; Malhado et al., 2014; Perez and Hogan, 2018). However, these same reports also acknowledge that more needs to be done to integrate underrepresented countries of the Global South. Our results for the insular Caribbean are qualitatively consistent with this assertion.

We recognize that the present study is based on a conservative sampling of the peer-reviewed journal literature pertaining to biodiversity conservation research. By definition, it necessarily ignores the production of unpublished academic research (e.g. graduate and undergraduate theses) and grey literature (e.g. technical reports) by Caribbean-based researchers, because these publications are difficult to rigorously quantify and will have a limited reach at the global level. It also excludes peer-reviewed journals not listed in the Web of Science Core collection, which might introduce a bias against regional and/or non-English journals of the Global South (Alonso and Fernández-Juricic, 2002; Holmgren and Schnitzer, 2004; Nuñez and Amano, 2021). This diversity of scientific contributions certainly deserves to be examined in future studies to provide a more complete picture. Still, in the context of global visibility and relevance, we believe that insights from this preliminary assessment provide a valuable starting point for taking positive action.

We believe that the complexity of socio-political systems in the insular Caribbean, their colonial histories, and consequent combination of SS and OT, may call for a reconsideration of some of the solutions that have been advocated so far to increase the representation of the Global South in biodiversity conservation science. Adding Caribbean-based editors to the editorial boards of ecology and conservation journals and/or waiving publication costs for authors from the Global South (Campos-Arceiz et al., 2018; Livingston et al., 2016; Mammides et al., 2016; Nuñez et al., 2019) is important and may help increase the relative contribution of Caribbean-based authors, but only to a point. Indeed, like elsewhere, the ability of the Caribbean-based scientific community to produce data of sufficient quality to match the standards of international journals critically depends upon the availability of local qualified scientists. Our data showing an increase in impact factor over the 16-year period for papers involving only Caribbean-based authors support that highly qualified Caribbean-based researchers already exist within the region. However, the lack of an increase in the number of papers published over time in the absence of foreign-based authors, coupled with the marked differences among locations in paper production, supports that more are needed.

Therefore, we suggest that the first priority should be building research capacity in conservation science in the insular Caribbean. In that respect, although we do not question an important role for short-term internships in helping increase interest and in providing some training in technical skills in biodiversity research (e.g. Latta and Faaborg, 2009), we argue that this alone is not enough. Short term internships come with the risk of maintaining young Caribbean-based naturalists in a subordinate role, relegating them to a technical task force supporting research and monitoring programs designed and led by foreign-based researchers.

We propose that the ultimate goal of capacity building in the region should be the development of a collective scientific expertise, i.e. the constitution of a culturally-diverse network of multilingual Caribbean-based scientists with diverse and complementary skills, sharing a common and strong theoretical background in ecology, evolution and conservation biology, and forming a cohesive academic force capable of orienting conservation strategies and informing policies. Ideally, such a network would help improve the current poor record of intra-regional collaborations and re-configure the current map of collaboration networks, aligning it with research priorities shared within the region, improving the understanding of impacts, and enhancing the sharing of information and facilities (Strigl, 2003). This approach aligns with the recent calls for more South-South cooperation with the Caribbean encouraging academic exchange and mobility among recent graduates (CCHE, 2010). Importantly, such networks could also help ensure that, when collaborating with foreign-based researchers bringing most of the funding, projects in conservation are also locally relevant (Baker et al., 2019; Malhado et al., 2014; Watkins and Donnelly, 2005). In parallel to this effort, it will be important to develop a regional agenda of strategic research priorities for the insular Caribbean – taking into account the important knowledge gaps in biodiversity and the need for technology transfer and scientific training – to help tackle shared threats to biodiversity while helping increase the global significance of locally-driven research.

We fully acknowledge that capacity building of this level will be a slow and expensive process (Heitor et al., 2014; Mammides et al., 2016), but we contend that it is essential to achieve long-term effectiveness and sustainability of conservation efforts (Mammides et al., 2016). That said, short-term returns are also possible. For example, recent research supported by the Caribaea Initiative – a Caribbean-based NGO that funds and facilitates postgraduate training and networking opportunities for Caribbean conservation biology students – has already yielded valuable discoveries (Atherley et al., 2020; Cambrone et al., 2021; del Castillo Domínguez et al., 2021; Exantus et al., 2021; Rodriguez-Silva et al., 2020; Zhao et al., 2020). This locally-based approach is not a new one in the global arena (Mammides et al., 2016). It also aligns with previous findings that even small increases in investment in local research capacity in the Global South can provide quick and efficient returns (Holmgren and Schnitzer, 2004), particularly if such investments are in subdisciplines underpinning biodiversity conservation (Livingston et al., 2016). In the short- to medium-term, an important fraction of this highly skilled Caribbean-based workforce could be absorbed by local environmental NGOs and consultancy firms assisting local and regional governments in operationalizing international agreements and policy instruments protecting biodiversity (e.g. Aichi Biodiversity Targets) and promoting sustainable development (e.g. 2030 Agenda for Sustainable Development).

Our results are a call to action to change the status quo in biodiversity research in one of the most biodiversity-rich regions of the world. This will require explicit diversity, equity and inclusion policies by international funding bodies, research teams, and scientific publishers and substantial funding (see Table S2 in Supplementary material for further details and additional proposed measures). We suggest that such funding should come from diverse sources, including, notably, Caribbean governments, which need to be more policy-committed towards these efforts (e.g. see Brazil in Barber et al., 2014), the private sector in the region (in particular the nature-based private sector that so heavily relies on ecosystem services), local and international NGOs, international research funding agencies, and the countries of the Global North, particularly those with a colonial past in the region.