Introduction

Over the past decades the science systems of many countries have undergone rapid transformations and reconfigurations. Gibbons et al. (1994) observed in this regard the emergence of socially distributed knowledge systems, which means that scientific knowledge has increasingly become co-produced across sectors and national borders, often within the context of application, as a result of various collaborative arrangements (both formal and informal) between the individuals, organisations and countries. As a consequence of these and other developments, the boundaries of national research systems have become blurred, to the extent that the scientific output of individual countries is now better described in terms of global networks of scientific output. These networks are largely self-organised, in the sense that scientists sustain these networks through established disciplinary practices of co-producing and co-reporting research. However, governments are increasingly using science policy mechanisms to mediate and steer scientists’ international collaborations and communicative patterns towards meeting their socio-economic and political objectives, such as strengthening global inter-university and university–industry interaction and establishing regional integration.

Science policy makers, and particularly those from developing countries, can no longer ignore the size and structure of their countries’ international scientific collaborations. As Leydesdorff and Wagner (2008) show, the accumulative effect of global scientific collaboration has been the formation of a core group of collaborating countries that excludes developing countries. Exclusion from the core of collaboration and production means that developing countries, especially those located at the very periphery of global scientific collaboration (like most countries in sub-Saharan Africa), have limited capacity to efficiently absorb and implement the scientific knowledge that is produced elsewhere. Studies on the patterns of international research collaboration, and especially on the integration of developing countries with global and regional science networks, are therefore imperative. In addition, the benefits that are reportedly associated with international research collaboration (e.g. Katz and Martin 1997) and the validity of these claims need to be studied within the context of developing countries (not only by using bibliometrics but also qualitative research methods and case studies). Examples cited of the positive outcomes of international research collaboration include, for instance, the increased international visibility of a country’s domestic science base and the facilitation of knowledge and skills transfer.

In the scholarly literature the international research collaborations of developing countries are almost invariably discussed within the context of North–South relationships (e.g. Binka 2005; Costello and Zumla 2000; Gaillard 1994; Jentsch and Pilley 2003; Melon et al. 2009; Velho 2002). This is not surprising because for many decades the more affluent countries of the world (the political ‘North’) have been driving—or at least significantly influencing—the science and technology (S&T) and research agendas of poorer countries (the political ‘South’). The North’s impress on the S&T and research agendas of the South can be explained in terms of a number of factors, the most obvious being the North’s ability to unleash significant amounts of funding for global science development.

An aspect of developing-country collaboration that is almost never referred to in scholarly literature is the notion of South–South research collaboration, i.e. research collaboration between developing countries. Two examples from the few bibliometric studies available are those by Arunachalam and Viswanathan (2008) on collaboration between India and the People’s Republic of China, and by Boshoff (2009) on collaboration among the countries of Central Africa. Both studies analyse the co-authorship patterns of published research papers. In the study by Boshoff, the North emerged as a strong mediator of South–South research collaboration; European countries participated in 77% of the papers that were co-authored by two or more scientists from Central African countries.

The current paper also uses a bibliometric approach and explores the extent of South–South research collaboration in the Southern African Development Community (SADC). The latter is a regional economic community (REC) consisting of 15 countries: Angola, Botswana, the Democratic Republic of the Congo (DRC), Lesotho, Madagascar, Malawi, Mauritius, Mozambique, Namibia, the Seychelles, South Africa, Swaziland, Tanzania, Zambia and Zimbabwe. Three instances of South–South collaboration are examined:

  • Intra-regional research collaboration, i.e. collaboration among the 15 SADC countries;

  • Inter-regional research collaboration, i.e. collaboration between the SADC region and other RECs on the African content, namely the Economic Community of Central African States (ECCAS), the East African Community (EAC), the Economic Community of West African States (ECOWAS), and the Arab Maghreb Union (AMU) of North Africa; and

  • Continental research collaboration, i.e. collaboration between the SADC and the rest of Africa.

In the next section I first present the case for South–South research collaboration, before turning to a brief discussion of the SADC. This is followed by an overview of scientific production and collaboration in Africa, which describes the context in which South–South research collaboration in Africa takes place.

Why South–South research collaboration?

The harsh reality for many developing countries is that whatever research expertise they may have and whatever pockets of research excellence there may exist, these are largely invisible and insignificant within the context of globalisation and an ever-expanding knowledge base that is dominated by the North (Hassan 2000). However, the combined effort and strategic blending of the best of Southern research could increase the international visibility of science produced by developing regions, and strengthen the South’s participation in global science (Ohiorhenuan and Rath 2000). In addition, the weight of a combined South, according to Binka (2005, p. 208), could “give developing country scientists the requisite capacity to effectively negotiate with northern collaborators”.

Secondly, relative to North–South collaboration, South–South research collaboration is considered more sensitive to the challenges and needs of developing countries, and hence better suited to finding appropriate solutions to common problems experienced by developing nations (Ohiorhenuan and Rath 2000). Southern partners, especially those within the same region, share similar environmental conditions and social burdens such as food insecurity and tropical diseases (Kane 2000), and have a better (almost intuitive) sense of the research actions required to address national and regional development priorities.

Thirdly, South–South research collaboration could contribute to the optimal and cost-effective use of limited resources in developing countries because research systems require a critical mass of human, material and financial resources to function efficiently (Kane 2000). Specifically, “[p]ooling of research resources would bring developing countries closer to meeting the critical minimum of investment required, as well as minimizing duplication of efforts in some areas” (Ohiorhenuan and Rath 2000, p. 12).

Lastly, South–South research collaboration could slow the current Diaspora of talented scientists and researchers from the South to the North, and transform it into increased South-to-South circulation. Regional centres of excellence in developing countries are important mechanisms of this circulation, as they provide the required platform and infrastructure for Southern researchers to collaborate on issues of regional importance, while also creating work and incentives for talented scientists to remain in their region.

SADC and regional S&T integration

The forerunner of the present SADC was the Southern African Development Co-ordination Conference (SADCC). SADCC was established in 1980 by the countries of Southern Africa (excluding South Africa) as a counter-strategy to the Constellation of Southern African States (COSAS), a strategy proposed by the South African government in 1979 (Gibb 2007). COSAS, being a South African initiative, sought to pursue friendly relations with the country’s neighbours by offering economic and technical assistance, in the expectation that it would reduce regional pressures for political change. At that stage South Africa was still firmly in the grip of apartheid and white minority rule. The assistance offered was presented within a spirit of co-prosperity but COSAS’ real aim was to strengthen regional economic dependence on South Africa.

SADCC was transformed into SADC in 1992, after the organisation changed its policy of balanced trade (through project co-ordination and large infrastructure projects) to that of free trade based on conventional neo-liberal principles (Gibb 2007). However, South Africa only joined the organisation in August 1994, 4 months after the country’s first democratic election. The objective of the integrated post-apartheid SADC, as formulated in the SADC Treaty, is (SADC 2001):

To promote sustainable and equitable economic growth and socio-economic development that will ensure poverty alleviation with the ultimate objective of its eradication, enhance the standard and quality of life of the people of Southern Africa and support the socially disadvantaged through regional integration” [Article 5, Objective 1.a].

Given the established belief that strengthening S&T produces long-term sustained economic growth and indirectly contributes to improvements in the general quality of life, regional S&T co-operation is included among the priority intervention areas in the SADC’s Regional Indicative Strategic Development Plan (RISDP) (SADC 2003). Moreover, in July 2007, the SADC officially adopted a Protocol on Science, Technology and Innovation (STI). The protocol expresses the commitment of member countries to collaborate towards fulfilling a number of objectives (DST 2007):

  • To establish institutional mechanisms to strengthen regional co-operation on and co-ordination of STI;

  • To institute management and co-ordination structures with clearly defined functions, which will facilitate the implementation of regional STI programmes;

  • To promote the development and harmony of STI in the region;

  • To pool resources for scientific research and technological development within the region;

  • To demystify STI by promoting the public understanding and awareness of and meaningful participation in these disciplines; and

  • To work towards the elimination of restrictions on the free movement of scientists, technologists and engineers for the purposes of education, research and participation in joint STI programmes.

Certain initiatives for contributing towards S&T research collaboration in SADC were already in place prior to the signing of the protocol. These initiatives had been introduced by the South African Department of Science and Technology (DST), or Department of Arts, Culture, Science and Technology (DACST) as it was known until mid-2002, and were being administered by the South African National Research Foundation (NRF). In 1999, for instance, the sum of 3 million South African Rand (then worth about 450 thousand US Dollars) was earmarked for the funding of large collaborative research projects that involved research teams from South Africa and other SADC countries. In 2000, SADC nationals who had recently obtained a doctoral degree were invited to apply for post-doctoral grants from the NRF to do research at South African institutions. Moreover, in the 2005/06 financial year the DST established a Regional S&T Fund to support small- to medium-scale projects that involved collaboration between researchers from at least two SADC countries. Project submissions to this end were invited in both 2006 and 2007. (For more details on these initiatives consult www.nrf.ac.za.)

Finally, as the SADC is part of the African continent, such regional integration also implies the alignment of SADC-specific S&T activities with continent-wide initiatives. The blueprint for Africa’s socio-economic renaissance is NEPAD, the New Partnership for Africa’s Development, a strategic framework that emerged from the activities of the African Union (AU) that was formally adopted in 2001. This blueprint aims to bridge existing structural gaps in various domains, including S&T. NEPAD’s S&T activities include linking up centres of excellence on the African continent, and strengthening S&T in select areas, such as biotechnology, geosciences, and product engineering (NEPAD 2001). NEPAD has since become institutionalised with a Secretariat in South Africa. In 2005 the NEPAD Secretariat (through its Office of Science and Technology), together with the AU Secretariat/Commission (through its Department for Human Resources, Science and Technology) produced a Consolidated Plan of Action (CPA) for Africa’s S&T (NEPAD 2006). The political and policy leadership for implementing the CPA lies with the African Ministerial Council on Science and Technology (AMCOST), a body that was established in 2003 under the auspices of NEPAD and the AU, to serve as a high-level platform for developing policies and setting priorities on African S&T (www.nepadst.org).

Among other things, the CPA outlines a number of flagship research and development clusters, each with specific programmes and projects. The SADC and the other African RECs are encouraged to promote the CPA programmes in their respective regions, in order to strengthen continental co-operation and integration. Thus, any study of the patterns of research collaboration of SADC would be incomplete without also examining research collaborations between SADC and the other RECs and the continent as a whole.

Scientific production and collaboration in Africa

The contribution of Africa—and specifically sub-Saharan Africa—to worldwide scientific production is declining. In 1987, sub-Saharan Africa’s share of world scientific papers was 1%, but since then the proportion gradually decreased to about 0.7% in 1996, where it stayed until 2004 (Tijssen 2007). It must be noted that the decrease is not in terms of the absolute numbers of papers, but in relation to worldwide growth rate. This means that sub-Saharan Africa’s growth in publications has fallen behind the worldwide growth rate. Africa’s share of world scientific papers is also markedly less than that of other developing regions. Between 2000 and 2004, Africa’s world share was only 1.8%, compared with 3.5% for Latin America (Pouris and Pouris 2009).

South Africa is the primary producer of science on the African continent, contributing 30% of Africa’s scientific papers, based on figures for 2000–2004 (Pouris and Pouris 2009). The notion of ‘African science’, however, is thus a misnomer because scientific production is concentrated in only a few countries (Tijssen 2007). South Africa, together with five other countries—Egypt, Morocco, Nigeria, Tunisia and Kenya—is responsible for 75% of Africa’s scientific papers. Tanzania is the only other SADC country, besides South Africa, with a notable share of scientific activity, producing approximately 2% of Africa’s scientific papers (Pouris and Pouris 2009).

The research strengths of South Africa (as measured by the numbers of papers in Thomson’s Essential Science Indicators database) are largely in Clinical Medicine and Plant and Animal Sciences, with a fair proportion also in Chemistry, Geosciences and Ecology/Environmental Sciences. Indications are, however, that South Africa is losing ground against Egypt and Nigeria in terms of the volume of output in selected fields (Jeenah and Pouris 2008).

The field concentrations of African science continent-wide reveal a significant over-representation of the medical and life sciences (with 61% of Africa’s scientific papers being published in this field against a worldwide average of 44%), and an under-representation of both the natural sciences and engineering sciences (only 25% of African papers are published in the natural sciences, compared with 37% worldwide; and the corresponding figure for the engineering sciences is 7 vs. 13%, based on the analyses by Tijssen 2007). According to Arvantis et al. (2000), the medical sciences are particularly well-represented in French-speaking African countries. The agricultural sciences, although also relatively under-represented, are better covered in the science profiles of English-speaking African countries. This reflects the diverse influences of colonialism on science production in Africa.

A variety of modes of research relations between African countries and their former colonial rulers emerged after independence. France, for instance, maintained close relations with its former colonies through the presence of specialized research institutes (e.g. the health-oriented Pasteur Institutes). These French institutes centralized research in the former colonies and this resulted in France playing a significant role in the steering and shaping of those countries’ research agendas. In the former British colonies, on the other hand, the foreign researchers were replaced by local researchers relatively soon after independence, meaning that the research agendas shifted to address local needs much sooner (Dahoun 1999).

The colonial legacy not only shaped the newly independent African countries’ research agendas but also their patterns of collaboration. Colonial ties, for instance, may explain why the United Kingdom accounts for 29% of all South Africa’s co-publications (Schubert and Sooryamoorthy 2010) and why France and Belgium, in the case of Central Africa, account respectively for 66 and 53% of the total scientific output of Chad and Burundi (Boshoff 2009). However, it should be noted that this trend is changing. The USA is fast becoming the main Northern collaborator for many African countries, and already ranks first in the case of South Africa, accounting for almost 32% of all South African co-publications (Schubert and Sooryamoorthy 2010).

Generally, the less productive a developing country is in terms of scientific output, the greater its dependence on international research collaboration. This is not unique to the African region as it also applies to other developing regions such as Latin America (Narváez-Berthelemot et al. 2002). Moreover, many researchers in African countries, when they do collaborate internationally, tend to participate in projects that have been conceptualised and designed by the partners in the North, meaning that the African-based researchers are just part of the execution. North–South relationships tend therefore to be highly unequal, and it is not uncommon for Southern collaborators to even be completely ignored as co-authors in the eventual publication (Dahdouh-Guebas et al. 2003). A partial explanation for this is that there is a lack of capacity to store and manipulate large sets of electronic data in Africa, which means that the analysis often occurs in the countries of the North, resulting in the African partners being estranged from the data and ensuing publications (Ogodo 2009).

These challenges aside, Tijssen’s (2007) analysis shows that African countries are increasingly co-authoring with foreign authors in international journals. In South Africa, for instance, the proportion of international research co-authorship increased from 41% in 2000 to 53% in 2005 (Sooryamoorthy 2009). Sooryamoorthy’s study also reveals that only about 1% of South Africa’s publications are co-authored with scientists from other African countries. This figure is supported by Jacobs and Pichappan (2006), who analysed the publications of the five most research-orientated universities in South Africa and found that only 3% of the universities’ publications involved co-authorship with other African countries. These extremely low figures speak directly to the current study and have implications for the establishment of South–South collaboration in the SADC region.

Methodology

The measurement of research collaboration

It has become standard practice to measure research collaboration through multi-author or multi-address papers that appear in journals that are indexed by international bibliometric databases, such as the ISI citation indexes by Thomson Reuters. The main advantage of using the ISI citation indexes is that they provide a complete list of all the authors and their affiliations (Luukkonen et al. 1993). Co-authored papers in the ISI citation indexes were thus also used in this study as a measure of collaboration. However, although the analysis of collaboration in terms of multi-author or multi-address papers is well established and regarded as a legitimate reflection of scientific collaboration, it is not without its shortcomings. I will therefore first highlight some of the shortcomings (that should be kept in mind when interpreting the results) before describing the data source and method.

The first shortcoming is that there is no conceptual clarity as to what constitutes research collaboration. This is an important consideration, given that the appropriate measurement of a construct is normally informed by its theoretical definition. In the case of research collaboration, however, a clear theoretical definition is lacking. According to Katz and Martin (1997, p. 16):

the notion of a research ‘collaboration’ is largely a matter of social convention among scientists. There is little consensus on where other, less formal links between scientists ‘end’ and collaboration ‘begins’. What some might deem a ‘collaboration’, others may merely regard as a loose grouping or a set of informal links. What constitutes a collaboration therefore varies across institutions, fields, sectors and countries, and very probably changes over time as well.

Moreover, research collaboration is not necessarily synonymous with research co-authorship, meaning that there is a difference between co-producing knowledge and co-reporting knowledge. In other words, not all instances of research collaboration will lead to a jointly authored paper, and not all co-authored papers imply that the authors listed have laboured together. One must therefore accept a certain level of uncertainty when relying on co-authorship analyses (Melin and Persson 1996).

Laudel (2002), in an attempt to pin down the uncertainty, investigated whether scientists who participated in four modes of research collaboration were listed as co-authors in the articles that emanated from their collaborations. Laudel’s finding was that, in the context of research collaboration of an industrialized country such as Germany, co-authorship is largely awarded to collaborators who creatively contribute to the research process, i.e. who contribute towards the theoretical-conceptual and experimental tasks involved. Collaborators who perform routine work, or who provide access to research equipment, or who only convey special knowledge relevant to the successful accomplishment of a research process, are not listed as co-authors.

Studies that investigate the correspondence between collaboration and co-authorship in the context of developing countries are virtually non-existent. In one of the few studies available, Boshoff (2009) surveyed 82 reprint authors from France, the United Kingdom and the USA, enquiring about the nature of the research contributions made by co-authors from Cameroon, a developing African country. Approximately 80% of reprint authors said that the Cameroonian co-authors helped with the fieldwork or data collection, and 60% said that they helped with the interpretation of results. Thus, in co-publications that include developing country participation, the contributions by authors from developing countries largely concern the collection of location-specific data for Northern partners, as well as the contextualization of the findings obtained.

As previously stated, there may be instances where scientists from developing countries are excluded as co-authors in papers written by their counterparts in the developed world—even though both groups contributed to the research—because of neo-colonial dynamics that are at play (Dahdouh-Guebas et al. 2003). Also, scientists from developing countries often study or work overseas for periods of time, and when they do publish their research they tend to list the foreign affiliation as their address (Kim 2006). All of these mean that the true collaborative research capacity of developing countries is probably underestimated through the analysis of co-authorships alone. It is important to keep this reservation in mind.

Data and method

The bibliometric data source used is the three citation indexes (Arts & Humanities, Social Sciences, and Science-expanded) from the on-line version of the Web of Science (WoS) in the ISI Web of Knowledge by Thomson Reuters. The list of SADC papers published between 1975 and 2008 were downloaded from the WoS during April 2009. (A SADC paper was taken to mean an article, letter or review with at least one author address located in any of the 15 member countries of SADC.)

All possible variants of country names were considered when extracting data from the WoS—for example, Madagascar was formerly known as the Malagasy Republic, Namibia as South West Africa, Zimbabwe as Rhodesia, and South African included four homelands during the apartheid era (Bophuthatswana, Ciskei, Transkei and Venda). The identification of papers by authors from the DRC presented a challenge because the WoS allocates papers from both the DRC and the Congo Republic (which is not an SADC country) to a single country category of ‘Congo’. Some papers produced by scientists in the DRC were also assigned to the category of ‘Zaire’, which was the country’s previous name. Thus, institutional and city names had to be used to correctly classify papers produced by the DRC in order to avoid confusing them with papers produced by the Congo Republic.

The extracted papers were imported into a Microsoft Access database, where after the country names were standardized and the data manipulated for the bibliometric analysis. At this stage it was discovered that the WoS had erroneously assigned a few papers to South Africa, where the addresses provided referred to institutions in countries such as South Korea, Saudi Arabia and Australia. These were manually corrected in the database. Also, as Namibia was known prior to 1990 as South West Africa (SWA), it was administered by South Africa, resulting in some SWA authors listing South Africa as their country affiliation. Such country affiliations were regressively changed to Namibia in the database.

A total of 137,632 papers were identified as belonging to the SADC. The SADC papers were categorized into scientific fields using the journal subject categories of the WoS. The Centre for Research on Science and Technology (CREST) at Stellenbosch University has developed a structure according to which these journal subject categories can be classified into four broad fields: agricultural sciences; natural and engineering sciences; health sciences; and social sciences and humanities. (The field classification framework is displayed in the Online Supplement.)

In the analysis, all papers with a single author were regarded as single-authored papers, irrespective of the number of institutional or country affiliations listed for the author. International co-authored papers were those that listed at least two countries (and at least two authors). Also, absolute measures of international research collaboration were used because the focus of the study is not on the relative strength of the links between countries but on the extent of collaboration among countries in the SADC, and between the SADC and the rest of Africa. Absolute measures, according to Luukkonen et al. (1993, p. 16), can answer questions such as “whether collaborative links reveal a centre–periphery relationship, or in absolute size, which countries are the most important collaborative partners of another country”.

Lastly, the classification of African countries into RECs (e.g. ECCAS or ECOWAS) was done using information provided on the websites of the relevant institutions. High-income countries refer to industrialized countries that were classified as ‘high-income’ by the World Bank in its list of economies for July 2008.

Results

Intra-regional research collaboration

Table 1 provides a detailed breakdown of the pattern of co-authorship of each SADC country for the period 2005–2008 and, as baseline, also for the period 1975–1978. (It must be remembered though that the WoS databases have been expanded substantially in the past as well as more recently, meaning that year-to-year fluctuations in output can also be explained in terms of new journal additions.) The table includes four kinds of co-authored papers that represent intra-regional or SADC collaboration: (1) papers where all foreign co-authors are from other SADC countries (category 3); (2) papers where the foreign co-authors are from other SADC countries and from other African countries (category 6); (3) papers where the foreign co-authors are from other SADC countries and from countries outside Africa (category 7); and (4) papers where the foreign co-authors are from other SADC countries, from other African countries, and from countries outside Africa (category 9). Thus, intra-regional research collaboration represents papers where at least one other SADC country appears in the list of foreign author addresses, irrespective of whether or not other African or non-African countries are listed.

Table 1 Pattern of co-authorship of research papers of individual SADC countries, 1975–1978 and 2005–2008
Full size table

The lowest proportions of intra-regional research collaboration (for 2005–2008) can be observed for South Africa (3%) and the three French-speaking SADC countries (Mauritius—4%; the DRC—6%; Madagascar—7%). In the French-speaking countries the dominant practice seems to be co-publication with non-African countries (the DRC—86%; Madagascar—85%; Mauritius—52%). Although co-authorship with non-African countries is also substantial in the case of South Africa (43%), South Africa has the largest share of ‘self-sufficiency’ in its publications (55% in 2005–2008) of all the SADC countries. ‘Self-sufficiency’ is composed of two indicators: the share of single-authored papers and the share of co-authored papers listing only own-country addresses.

The highest proportions of intra-regional research collaboration in 2005–2008 can be observed for Lesotho (54%) and Namibia (52%). More than half of the papers produced by scientists in these countries are co-authored with at least one other SADC country (54 and 52%). Although the proportions for Lesotho and Namibia are markedly similar in size, the structure of co-authorship of these two countries differs. For Namibia, the 52% of intra-regional co-authored papers is composed of 34% of papers co-authored with scientists outside Africa. For Lesotho, however, co-authorship with scientists outside Africa comprises only 19% of the 54% of intra-regional co-authored papers.

Table 1 clearly shows that South Africa dominates the production of research papers in the SADC. In 2005–2008 South Africa produced 21,277 papers, almost 14 times more than the 1,563 papers generated by Tanzania (the second largest knowledge producer in SADC) over the same period. Although not shown in Table 1, it was calculated that from 2005 to 2008 South Africa produced 81% of the total pool of SADC papers, and this figure deviates only marginally from the 83% in 1975–1978. Moreover, 78% of all intra-regional or SADC co-authored papers in the 2005–2008 period included authors based in South Africa.

The proportion of intra-regional co-authored papers of all SADC papers is extremely low. Figure 1 plots the proportions for the period 1975–2008. In 2007/2008 the proportion was 3.2%, compared to 0.4% in 1975/1976. Figure 1 also shows the proportion of South African papers that is jointly produced with other SADC countries, and what proportion of papers from authors based in other SADC countries as a group (South Africa excluded) is the result of co-authorship between at least two of those 14 SADC countries. Intra-regional co-authorship is marginally higher for papers produced by the cluster of 14 SADC countries than that for papers produced by South Africa (3.7 vs. 3.1% in 2007/2008). However, in all instances, Fig. 1 shows that intra-regional co-authorship only really starts to increase from 1995/1996. This means that the change to democracy in South Africa in 1994 had a positive effect on SADC research integration and co-authorship in general, as the other SADC countries also started to collaborate more with each other since that date. Nevertheless, the overall proportion of intra-regional research collaboration is still critically low.

Fig. 1
figure 1

Intra-regional co-authored SADC papers as a proportion of all SADC papers, 1975–2008

Full size image

Moreover, it is really a case of South Africa being a significant research collaborator for other SADC countries rather than it is the other way around (i.e. the other countries being significant research collaborators of South Africa). This can be seen from the comparison of the proportions reported in Figs. 2 and 3. South African scientists co-produce about 49% of Namibia’s research papers (Fig. 2), and are also responsible for 44 and 34% respectively of Lesotho’s and Swaziland’s research papers. This is not surprising because Lesotho and Swaziland are land-locked within South Africa, and Namibia was under South African administration until its independence in 1990. On the other hand, author collaboration with these three countries—and indeed any of the other SADC countries—account for less than 1% of South Africa’s total research output (Fig. 3).

Fig. 2
figure 2

Proportion of SADC country papers that are co-authored with South Africa, by individual country, 2005–2008

Full size image
Fig. 3
figure 3

Proportion of South African papers that are co-authored with rest of SADC, by SADC country, 2005–2008

Full size image

Table 2 further examines the low proportions of intra-regional research collaboration by disaggregating the proportions in terms of field, and also by comparing the proportions with those of SADC co-authorship with high-income countries. High-income country collaboration is used here as a measure of North–South collaboration. As can be seen, intra-regional research collaboration increased over time in all four broad fields. The highest incidence of such collaboration in 2005–2008 is associated with the agricultural sciences (5.5%), and the lowest with the social sciences and humanities (1.3%). The need for incentives to increase the proportions of intra-regional collaboration is particularly evident in the light of significantly larger proportions of co-authorship with high-income countries (47% overall and even as high as 61% in the health sciences for the period 2005–2008).

Table 2 Proportion of SADC papers that are co-authored with at least one other SADC country, by broad field, 1975–2008 (and, as benchmark, the proportion of SADC papers that are co-authored with at least one high-income country)
Full size table

Up to now we have focused on the proportions of SADC papers that involve regional co-authorship. The analysis now shifts to the proportions of regionally co-authored papers that were also jointly authored with the rich economies of the world. In other words, the focus moves to the global North’s contribution to regionally co-authored papers. This is examined in Table 3, where the results are presented by broad field. As can be seen, high-income countries are listed as co-authors in 60% of all SADC co-authored papers produced during 2005–2008, and the proportions seem to be growing—overall, as well by broad field. The proportion is highest for health sciences (84%). Arguably, joint intra-regional and international collaboration is highest in health sciences because of the complex nature of diseases such as HIV and malaria, whose combat requires large, internationally-funded research consortia that comprise both regional and non-regional partners.

Table 3 Proportion of intra-regional co-authored papers (i.e. papers that are jointly authored by at least two SADC countries) that also list a high-income country as co-author, 1975–2008
Full size table

Inter-regional research collaboration

The extent of inter-regional research collaboration is highlighted in Table 4. This mode of collaboration is measured as the proportion of SADC papers that are co-authored with scientists based in one or more countries in another REC. In other words, what is reported is the extent of collaboration between the SADC and each of the four African RECs that represent the other main regions of Africa (central, east, west and north). Figures are presented for the total SADC, as well as for South Africa individually and for the other 14 SADC countries as a group. Table 5 shows that collaboration between the SADC and the other African RECs is virtually non-existent (less than 2% of the total SADC output). Even if South Africa (which dominates the pool of SADC papers) is excluded, the figures for 2005–2008 only marginally improve to 5% of papers that are co-authored with ECCAS and 3% that are co-authored with ECOWAS. Cameroon is the main collaborating country in ECCAS (central African region) and Nigeria the main collaborating country in ECOWAS (western African region).

Table 4 Proportion of SADC papers that are co-authored with other African RECs, 1975–2008
Full size table
Table 5 Proportion of SADC papers that are co-authored with the rest of Africa, by broad field, 1975–2008
Full size table

Continental research collaboration

The four RECs mentioned above do not cover the non-SADC African region comprehensively—for instance, Egypt, Ethiopia and Sudan, are not included in any of these RECs. Hence, continental research collaboration was considered, which means collaboration between SADC and the rest of the African countries outside SADC (Table 5). The relevant proportions are broken down by broad field and reported for the total SADC as well as for South Africa separately and the other 14 SADC countries combined. As can be seen, in terms of research the 14 SADC countries (as a group) are significantly better integrated than South Africa is with the rest of the continent. Only 3.5% of papers produced by South Africa between 2005 and 2008 included a non-SADC African affiliation, whereas the corresponding figure for the other 14 SADC countries was 10.2%. However, given the vast differences in the total number of papers produced, the 3.5% of papers for South Africa translate into a significantly larger paper count than the 10.2% of the other 14 SADC countries.

Table 5 further shows that research collaboration between SADC and the rest of Africa is highest for the agricultural sciences (9% overall in 2005–2008, and 15.5% for the 14 SADC countries that exclude South Africa). Continental co-authorship, specifically as far as South Africa is concerned, only started to emerge during the last period listed (2005–2008), for which the overall figure is 3.5% (compared with only 0.6% in 1995–1998). The other SADC countries, however, had already achieved a higher level of continental research co-authorship (5.8%) during the period 1995–1998.

Lastly, Fig. 4 highlights the contribution of the global North towards continental research collaboration between SADC and the rest of Africa. Between 2005 and 2008, high-income countries participated in almost 60% of all continental co-authored papers. This substantially large proportion is no recent phenomenon, as an equally large proportion (61%) can also be observed for the earlier period of 1995–1998.

Fig. 4
figure 4

Proportion of continental co-authored papers (i.e. papers jointly authored by SADC and the rest of Africa) that also list a high-income country as co-author, 1975–2008

Full size image

Discussion

South–South collaboration has become a political trend, irrespective of whether it is in terms of trade, health, agriculture, transport, security or research. The notion of a regional research system, in particular, is politically desirable for many reasons. In the Middle East, for example, it is believed that regional scientific collaboration between Israeli and Palestinian scientists should be pursued amidst political instability in the region, as some argue that such collaboration could become the driving force for peace (Green 2008). In the SADC region, to use another example, Zimbabwean scientists are turning towards regional collaboration, as well as other forms of international collaboration, in their efforts to reignite the S&T system of Zimbabwe that was almost destroyed by the socio-economic and political turmoil of the last few years (Ogodo 2008).

This study focused on regional research collaboration as an instance of South–South collaboration, and specifically on collaboration among the 15 countries of the SADC, as well as between the SADC and the rest of Africa. The results sparked a number of insights, which I will explain below.

Firstly, research collaboration in the developing world is primarily a case of North–South collaboration and such collaboration is most likely to continue to be the dominant mode in years to come, even though South–South collaboration has entered the political agenda and is portrayed as an alternative to North–South collaboration. Two sets of findings from the current study support this conclusion.

  • SADC countries collaborate in terms of research significantly more with the rich economies of the world than with their regional counterparts. Only about 3% of papers produced by SADC during 2005–2008 were co-authored by scientists from two or more SADC countries, and only about 5% of SADC papers were jointly authored with scientists based in African countries outside the SADC. In contrast, 47% of SADC papers during that period were co-authored with scientists from high-income countries. (Similar figures were reported by Narváez-Berthelemot et al. (1999) for the Mercosur countries in Latin America. Between 1980 and 1990, when the Mercosur countries included Argentina, Brazil, Paraguay and Uruguay, respectively 40 and 41% of Mercosur papers were co-authored with countries from the European Union and the North America Free Trade Agreement. However, only 4% of Mercosur papers were co-authored with other Latin American countries.)

  • The few instances of intra-regional and continental research collaboration in the SADC are largely the product of North–South collaboration, on account of facilitation or mediation of the South–South collaborations of the SADC by the global North. In 2005–2008, 60 and 59% respectively of intra-regional and continental co-authored papers involved at least one high-income country.

Secondly, regional research collaboration that includes a country that is significantly closer to the core of world science than the other countries in that region does not represent, in my opinion, South–South collaboration. It is a highly unbalanced and unequal partnership that can best be described as a variant of North–South collaboration, where the scientific giant in the South plays the same role as the ‘political North’. In SADC, about 81% of all papers and 78% of all intra-regional co-authored papers are produced by South Africa. This gravitation of SADC publications towards South Africa is only one indicator of the centrality of South African science in the SADC. Other factors that set a scientific core apart from its surrounding periphery are also present in the South Africa-SADC context, such as the core’s ability to attract skilled human resources for S&T from the periphery. A recent report by the South African Council on Higher Education (CHE 2009), for instance, reported that, in 2005, 45 and 32% respectively of all foreign masters and doctoral graduates at South African universities were from other SADC countries. Schubert and Sooryamoorthy (2010) describe South Africa’s position as that of a semi-periphery because, in terms of global science, South African scientists find themselves at the periphery. However, in terms of regional and continental science, South Africa fulfills the role of a scientific core.

Thirdly, and related to the above, it stands to reason that a semi-periphery such as South Africa has more to gain from collaborating with the global core (North) than with the global periphery (South), which also explains the limited collaboration between South Africa and the rest of the SADC. A case study on South African–German collaboration (Schubert and Sooryamoorthy 2010) shows that South African scientists prefer to collaborate with researchers or groups that are regarded as core players in their respective fields, i.e. with highly reputed international research partners. The decision for South Africans to collaborate internationally is thus more often a strategic one, taken to overcome factors related to their marginality, such as limited domestic funding opportunities. One can therefore play devil’s advocate and ask why South African scientists would willingly engage in bilateral collaborative activities with their even more marginalized SADC counterparts, as this would not reduce but in fact accentuate their own marginality. Multi-lateral collaborative activities between South African and other SADC scientists, that include Northern partners, seem more plausible, given that the general association with the scientific core is a positive one and would be advantageous to both the semi-periphery and periphery. However, one must caution against broad-based statements and generalisations, as pockets of scientific strength do exist in other SADC countries, although these are limited. At the same time, one must acknowledge and address the barriers that work against regional and continental scientific collaboration. These relate to language and cultural divisions (for instance, the SADC is composed of anglophone, francophone and lucophone countries) as well as to the ‘de-institutionalised’ (Mouton 2008) nature of the science systems of many sub-Saharan African countries, including some SADC countries. This refers to a state of affairs where the scientific institutions of the country are fragile, fragmented and constantly under-resourced, and where scientific inquiry serves to build the careers of individual scientists rather than strengthening the scientific institutions. Mouton ascribes this to a number of factors such as the legacy of colonial science; political instability and civil wars; and international research and funding agencies that—through well-intended support—have created better opportunities for talented African scientists outside their region of origin.

Fourthly, any criteria or guidelines for successful North–South collaborations should also be extended to include South–South collaborations that comprise highly unequal partners, such as collaboration between South Africa and the other SADC countries. Gaillard (1994), for instance, who questioned whether collaboration is possible between unequal partners, proposed the introduction of a checklist of complementary collaborative practices as a solution to the inequality. Gaillard suggested that the guidelines be used as a charter of collaborating partners, to ensure that each partner is associated as equally as possible with the different stages of the joint research. Examples from the charter include a strong mutual interest in the research, joint conceptualization and drafting of project proposals, and the inclusion of a substantial number of researchers from each collaborating group. The criterion of mutual benefit, in particular, is seen as vital for successful collaboration within the context of inequality. However, for joint research activity to be mutually beneficial to all those involved, partner countries that are marginalized and situated at the periphery of world science would need to ensure that they:

have a minimum of capacity, and thus that they undertake some minimum of national activities in science and technology. Areas where individual national effort is required include the integration of S&T into national development plans, with carefully selected priorities backed by adequate resources; increased spending on R&D from current levels; according higher priority to all educational activities; placing greater stress on education in basic sciences and effective systems of research … (Ohiorhenuan and Rath 2000, p. 24).

The building of a minimum of scientific expertise through strategic government intervention is important, as it will provide peripheral countries with the required capacity to absorb leading edge research. Without such absorptive capacity, knowledge and skills transfer within regional and other international partnerships will not be optimal and collaboration will remain unequal and asymmetrical.

Lastly, co-authored research papers are only a partial measure of collaborative research activity and the extent to which co-authored papers measure research collaboration in one country of the South may differ significantly from another country of the South, especially if an international data source such as the WoS is used. In 2008, South Africa had 52 domestic journals (out of a pool of about 300 government-accredited domestic journals) indexed by the WoS, compared to Malawi which had only one domestic journal in the WoS, and the other SADC countries which had none. Thus, with the exception of South Africa, SADC countries at the periphery of world science are largely invisible as far as papers in the WoS are concerned. Moreover, if one assumes that these marginalized SADC countries publish their regionally co-authored papers (to the extent that they do exist) in domestic (non-WoS) journals, some instances of SADC collaboration will go undetected. This clearly illustrates that additional research is required in order to clarify the extent to which an analysis of co-authored research papers in the WoS actually reflects research collaboration by the low-income countries of the SADC. For instance, is international visibility (which is achieved through publication in international sources) valued to the same degree as the dissemination of location-specific research to local scholars and the intended project beneficiaries? Also, do SADC researchers from scientific peripheral countries regard joint publication as the final stage of joint research and, if they do, in what sources do they publish the joint research (in local or international peer reviewed journals or in other, non-peer reviewed sources), and why? These are important questions for further studies that could help to contextualise and strengthen the measurement of research collaboration in developing countries.