International scientific collaborative activities and barriers to them in eight societies

ABSTRACT

Scientific research increasingly requires international collaboration among scientists. Less is known, however, about the barriers that impede such collaboration. In this pioneering study, more than 9000 scientists from eight societies – the United States, the United Kingdom, India, Italy, Taiwan, Hong Kong, Turkey, and France – were surveyed to gauge scientists’ attitudes and experiences. While most scientists claimed international collaboration was important, their actual participation in such collaborations was much lower. We identified the prevalence rates of three types of barriers (political, logistical, and cultural) based on categories developed from previous work. In addition, we identified nine additional categories of barriers. Key barriers to collaboration that scientists identified included lack of funding for international work, restrictions on material and data sharing, and differences in academic standards. Respondents also complained about bias against scholars from emerging or developing countries. Our study highlights areas where efforts could be made to address policy issues, institutional barriers, and national biases to promote more productive collaboration in the global scientific community.

KEYWORDS:

• Collaboration
• policy
• research barriers
• survey
• bias

Introduction

Over the last two decades, research and development (R&D) funding and outputs have consistently increased around the world and across all science and engineering (S&E) disciplines (NSB, NSF 2019, 2020a; Wagner and Fukuyama 2009). Publication data demonstrate that S&E is a global enterprise. As the number of total S&E publications steadily rise, so do multiple international authorships (publication with authors from two or more countries) overall as well as a share of the total number of publication, which represent 23% of total publications in 2018 (up from 13.2% in 2000) (NSB, NSF 2019). The trend of increasing international collaborative research reflects the increase in R&D investments overall, from 722 billion USD to 2 USD.2 trillion in 2017, as well as improvements in information technology that allow scientists easy access to publications and to communicate more effectively through e-mail, video conferencing, and shareable online documents (NSB, NSF 2020a). The increase in coauthorship is not the sole indicator of international collaboration. It is indicative, however, of the strengthening of global, self-organizing scientific networks (Wagner and Leydesdorff 2005).

Scholars have suggested numerous reasons for the increased international collaboration. C.S. Wagner proposed five major motivations for engaging in international collaborations: (1) to increase visibility; (2) to share project costs; (3) to gain or share access to an expensive or unique physical resource; (4) to achieve greater leverage by sharing data; and (5) to increase creativity by exchanging ideas (2006). In addition, scientists seek international collaborations to expand the scope of their research, engage with a desirable collaborator, and/or come together to take on projects that are more global in nature (The Royal Society 2011; Wagner 2018). Additionally, surveys of publications found that financial resources, academic excellence, individual motivation, and active informal communications encourage international collaboration (Parker, Lortie, and Allesina 2010 and Wagner et al. 2015, Jeong, Choi, and Kim 2013). Furthermore, international collaboration increases the impact of research, as measured by citations (Sugimoto et al. 2017; Lariviere et al. 2015).

In the past, the network of global cooperation was limited to a small core group of countries, often because of linguistic and historical factors (Adams 2013; Leydesdorff and Wagner 2008). Recently, the growing openness of the global S&E network has made it possible for more scientists to participate in international collaborations (Adams 2013; Barrios et al. 2019; Gui, Lui, and Du 2019; Narin, Stevens, and Whitlow 1991). The United States, the United Kingdom, and France, along with Germany and Canada, have become prominent hubs for migrating scientists (Sugimoto et al. 2017).

Patterns of collaborations have changed over the past 20 years. For example in 1996, US scientists engaged with UK researchers the most, but by 2018 Chinese scientists were the top US collaborators (NSB, NSF 2019; Gui, Lui, and Du 2019). This shift is likely linked to China’s R&D infrastructure development (as R&D expenditures grew from 14 billion USD in 1996 to 496 billion USD in 2017) as well as in higher education, with Chinese students being the largest block of foreign science and engineering doctoral recipients in the United States (NSB, NSF 2019, 2020a).

In many cases, the national politics and diplomatic relations between states have become less important in determining S&E collaborations, even as it continues to hinder collaboration between some scientific communities supported by opposing states (Jeanmonod and Firstenberg 2019; Sugimoto et al. 2017). Instead of politics, intrinsic factors, such as the availability of resources and equipment as well as the nature of the issues being addressed, have taken precedence (Jang and Ko 2019; Leydesdorff, Park, and Wagner 2014; Wagner 2018). For example, in the field of astronomy where large expense projects are common, international collaboration increased to 54% of articles, the highest of any field (NSB, NSF 2018; Coccia and Wang 2016). Increased collaboration through open national policies enhances team diversity and has been found to improve the quality of the work (Barjak and Robinson 2008; Leydesdorff, Bornmann, and Wagner 2019; Wagner and Jonkers 2017). In contrast, recent restrictive travel policies in the United States have been found to negatively affect collaborations (Jeanmonod and Firstenberg 2019; Sugimoto et al. 2017).

As more scientists collaborate across national borders, they often find barriers from their own governments and home institutions. Previous studies at the Baker Institute brought together scientists from the United States, China, Taiwan, Hong Kong and Singapore and identified a list of barriers researchers perceived affecting their work (Lane, Matthews, and Lewis 2009). While funding was the largest obstacle, participants also identified several cultural differences that were barriers including academic, fiscal, and cultural calendars as well as standards for peer review, research ethics, and accreditation. They also noted policy or regulatory barriers related to visas, intellectual property rights, and export control policies limiting material and data sharing.

Other scholars have also noted funding as well as other cultural and political barriers to international collaborations (Aarons et al. 2019; Hoekman et al. 2012; See 2018; Widmer, Widmer, and Lerman 2015). A 2017 workshop and survey of clinical researchers found that the majority of respondents described funding, time, academic differences (culture and calendars, for instance), and research capacities as barriers to international collaborations (Noonan et al. 2018). Other research points to language influencing researchers’ ability to engage and collaborate (Hwang 2013)

To better identify barriers and experiences of scientists, we surveyed 9,422 biologists and physicists in eight countries and regions – the United States, the United Kingdome, India, Italy, Taiwan, Hong Kong, Turkey, and France – which for the sake of simplicity we will refer to as societies due to the inclusion of Hong Kong and Taiwan (please note that this use of “societies” is different from scientific and medical associations and organizations). This survey, as discussed in detail below, was part of a larger project focused on religion and science that allowed us to ask questions related to international collaboration.

Survey participants were asked questions about the importance and extent of their international collaborations and about barriers impeding it. The results demonstrate that collaboration was important to the majority of scientists, although the actual rates of participation varied among the societies. Most scientists did not perceive strong barriers, but for those who did, funding was universally the chief concern in each society surveyed. Scientists also noted issues such as material and data sharing regulation and difference in academic standards as prominent obstacles. In addition, scientists described perceived biases against scholars in emerging and developing countries. With S&E international collaborations becoming increasingly important, by changing funding frameworks and instituting new guidelines key barriers to productive collaborations in the global S&E community could be reduced.

Materials and methods

The survey on which the article is based was part of a larger international study of 22,525 biologists and physicists working in universities and research institutes located in the United States, the United Kingdom (which included England, Northern Ireland, Scotland, and Wales), Italy, France, Turkey, India, Hong Kong and Taiwan. The primary focus of the broader study was on scientists’ attitudes toward religion, but it also included questions related to other social contexts of scientific work including ethics, family-life, gender, and international collaboration. It was this broader aim of understanding scientists’ views of religion, that directed the selection of locations to be surveyed.

The survey was conducted in eight countries and regions, which we will refer to as “societies,” due to the inclusion of Hong Kong and Taiwan. Scientists in Hong Kong and Taiwan were chosen to represent Sino-cultural views, as scientists in these societies could participate in a survey about these issues without facing potential political discrimination. The eight societies were selected in order to yield a sample with sufficient variation in dominant religious tradition (e.g., Catholic, Protestant, Hindu, Muslim, Buddhist) and diverse levels of religiosity (as measured by sociological indicators in surveys such as the World Values Survey). The societies also have various research and development (R&D) infrastructure (as measured by gross R&D expenditure (GERD) and GERD per gross domestic product (GDP)) (NSB, NSF 2020a). Detailed protocols for sampling and the full list of survey questions are available in Ecklund et al. (2019).

The study focused on scientists from two disciplines – physics and biology – for a number of reasons. Biology and physics are considered classic natural science disciplines. Most prominent research universities have departments (or more than one department) in these areas. The two fields also have differing levels of female involvement (which may affect social factors such as religiosity and family-work balance a focus in the broader study). Further, they reflect important historical conflicts in the relationship between science and religion, which was the focus of the study. The term “scientists” in this article is used for linguistic simplicity, but the results are generalizable only to the population of biologists and physicists in the societies studied. We do not intend to generalize to all scientists, or to disciplines we did not include in our study (Ecklund et al. 2016, 2019).

The sampling frame for the survey was developed from publication data from 2001 to 2011 in the Thomas Reuter Web of Science (WOS), which was used to identify publishing biology and physics departments within each society. The goal was to include actively publishing scientists assuming they would also actively work on research (as opposed to retired or teaching faculty). From this sampling frame, a stratified, random sample of organizations was selected. Departmental websites were reviewed to confirm eligibility and data was extracted on all eligible scientists in each department and stratified first by discipline (physics and biology) and academic reputation (elite and non-elite based on country/region specific metrics). The distinction between elite and non-elite institutions was important to the broader project as well, because of important differences in the organizational contexts of these different types of scientific institutions (Hagstrom 1964; Hermanowicz 2009). To stratify our sampling frame in this manner, we triangulated three factors: research productivity (measured as number of times an institution appears as an author affiliation in a WOS journal article); insider opinions of scientists from each region in the study; and in-country ranking systems of educational institutions.

Second stage sampling included scientists eligible from sampled department stratified by gender (male and female) and three academic ranks: graduate students, post-PhD scientists not yet established in their careers (postdoctoral fellows, lecturers, assistant professors, etc.), and the most established scientists (associate professors, professors, etc.). A nonproportional number of scientists (generally a maximum of 20) were sampled from each stratum within each department. A random sample of scientists were invited from each organization to participate in our survey. We sampled only from physics and biology departments, without restricting or sampling on the range of subfields within these disciplines. Our analysis accounted for departmental clustering through sampling weights, which take into consideration the probabilities of any scientist or department being selected into our sample, thus minimizing sampling bias (Ecklund et al. 2016; Ecklund, Scheitle, and Peifer 2018; Ecklund et al. 2019).

Web-based surveys were administered across the societies between 2013 and 2015 that were fielded by two different survey firms (GtK NOP and Abt SRBI). A total of 9,422 survey responses were received, which constituted an overall response rate of 42% for the study. The survey response rates per country were: 57% (1986 responses) in the United States, 50% (1581 responses) in the United Kingdom, 46% (779 responses) in France, 57% (1411 responses) in Italy, 44% (1763 responses) in India, 40% (326 responses) in Hong Kong, 39% (892 responses) in Taiwan, and 39% (684 responses) in Turkey (Ecklund et al. 2016, 2019). This was a high response rate for a web-based survey of scientists, much higher than other prominent studies of scientists, such as the 2015 Pew AAAS Study of Scientists, which had a 19% response rate (Pew Research Center 2015).

The survey, protocols and consent forms used were reviewed and approved by the Institutional Review Board from Rice University (project ID 681517-2). Informed consent was obtained at the beginning of the web-based survey. For a more detailed description of the survey and methodology, see Ecklund et al. (2016) and (Ecklund et al. 2019).

For the purposes of this article, we focus on answers to six questions from the survey related to scientists’ collaboration with individuals in other countries:

1. Among collaborators you have worked with in the past 12 months, about what percentage of them were from institutions located in a different country? Your best estimate is fine. Choices were: I have not collaborated in the past 12 months; None of my collaborators were from institutions located in a different country; 1–25%; 26–50%; 51–75%; 76–100%; or Don’t know.

2. Please estimate what percentage of your publications is coauthored with scientists from institutions located in a different country? Choices were: I don’t have any publications; 0%; 1–10%; 11–25%; 26–50%; 51–75%; 76–100%; or Don’t know.

3. How important is it to your success as a scientist to try to collaborate with scientists from institutions located in a different country? Choices were: I’m not a researcher; Very important; Somewhat important; Not very important; Not at all important; or Don’t know.

4. Have you ever encountered any barriers to collaborating with scientists from institutions located in a different country? Choices were: Yes or No [if no, skip remaining questions.]

5. Have you encountered any of the following barriers to international collaboration? Choices were Yes, No or Don’t Know.

   1. Visa–Getting visas for myself or collaborators

   2. Language

   3. Funding–Funding for collaborative research

   4. Calendar–My academic calendar is inconsistent with my collaborators

   5. Separation from family–Separation from family and friends prevents international travel

   6. Political beliefs–Differences in political beliefs among collaborators makes collaboration difficult

   7. Gender discrimination

   8. Sexual orientation – Discrimination because of my sexual orientation

   9. Religious differences–Religious differences among collaborators

   10. Ethical Standards – Differences in ethical standards for scientific work

   11. Intellectual property–Differences in enforcement of intellectual property rights

   12. Time zone–Difficulty finding a time to communicate because of time zone differences

6. Have you experienced any other barriers to international collaboration? Choices were Yes [Specify – what were the barriers?] or No.

The survey questions were designed to assess the importance of international collaboration and to identify social, cultural or political barriers that impeded scientists from working with collaborators in other countries. Following the close-ended survey prompts (questions 1–5), which asked respondents about a range of specific barriers, the survey also provided respondents with a text-box in which they could enter any additional barriers to international collaboration that they encountered (question 6). These responses were reviewed to determine if the responses fit into categories presented previously as barriers, which served as our initial coding framework (question 5, based on Lane, Matthews, and Lewis 2009). Remaining responses were coded inductively to identify additional themes. Because these open-ended responses were quite straightforward and our aim was simply descriptive categorization rather than theory-development, these responses were coded by a single coder, who classified them into emergent categories (EY). They were subsequently reviewed independently by KM to ensure the codes correctly fit the new coding categories. Responses varied from one word to full paragraphs referencing multiple issues. As a result, responses were coded in one or multiple categories as needed.

Results

The survey assessed scientists’ perception of international S&E collaboration. Biologists and physicists from eight societies – the United States, the United Kingdom, India, Italy, Taiwan, Hong Kong, Turkey, and France – were asked about the importance of international collaboration. In all eight societies, the majority of scientists believed international collaboration was important (Figure 1). Italy and India had the highest approval rates (95% and 94%, respectively), while the United States had the lowest rate (68%).

Figure 1. Importance of international collaboration. The majority of scientists in each society identified international collaboration as important. Abbreviations: FRA = France, HK = Hong Kong, IND = India, ITL = Italy, TAI = Taiwan, TUR = Turkey, UK = United Kingdom and US = United States.

The survey also asked scientists how much they collaborated with international partners. First, it asked respondents what percentage of their collaborators during the past year were from foreign countries (Figure 2(a), Table 1). Here, in almost all societies the majority of scientists indicated that they participated in some form of international collaboration. The exception was India, where only 48% of respondents indicated they had an international collaborator. In contrast, highest collaborators where in the United Kingdom and Hong Kong. In these societies, 14% (United Kingdom) and 12% (Hong Kong) of respondents reported 76–100% of their collaborators were from other countries.

Figure 2. Participation in international collaborations. Scientists were asked to estimate (a) the percentage of collaborators from other countries that they worked with in the past year and (b) the percentage of their publications that involved authors from other countries (for percentages in each category see Tables 1 and 2). Answers were in categories: 0%, 1–25%, 26–50%, 51–75%, and 76–100% publications with international collaborations. The more a society participated in collaboration, the darker the color on the graph. Abbreviations: FRA = France, HK = Hong Kong, IND = India, ITL = Italy, TAI = Taiwan, TUR = Turkey, UK = United Kingdom and US = United States.

Table 1. Percentages of Scientists’ International Collaboration. Scientists were asked to estimate the percentage of collaborators from other countries they worked with in the past year. Answers were in categories: 0%, 1–25%, 26–50%, 51–75%, and 76–100% publications with international collaborations.

	US	UK	FRA	ITL	IND	HK	TAI	TUR
0%	33%	19%	16%	21%	52%	12%	35%	49%
1–25%	37%	29%	38%	35%	28%	37%	42%	34%
26–50%	18%	20%	23%	21%	9%	20%	13%	9%
51–75%	8%	19%	13%	14%	6%	18%	6%	4%
76–100%	4%	14%	9%	9%	4%	12%	5%	3%

Abbreviations: FRA = France, HK = Hong Kong, IND = India, ITL = Italy, TAI = Taiwan, TUR = Turkey, UK = United Kingdom and US = United States.

In addition, scientists were asked to estimate the overall percentage of their publications that included an international collaboration (Figure 2(b), Table 2). Here, Italy had 23% of respondents reporting that 76–100% of their publications had internationally coauthors.

Table 2. Percentages of Scientists’ International Collaborative Publications. Scientists were asked to estimate the percentage of their publications that involved authors from other countries. Answers were in categories: 0%, 1–25%, 26–50%, 51–75%, and 76–100% publications with international collaborations.

	US	UK	FRA	ITL	IND	HK	TAI	TUR
0%	27%	13%	7%	13%	35%	9%	30%	36%
1–25%	42%	32%	42%	39%	44%	42%	46%	40%
26–50%	13%	18%	19%	15%	11%	23%	11%	10%
51–75%	7%	14%	15%	10%	5%	13%	7%	6%
76–100%	11%	22%	17%	23%	6%	13%	6%	8%

Abbreviations: FRA = France, HK = Hong Kong, IND = India, ITL = Italy, TAI = Taiwan, TUR = Turkey, UK = United Kingdom and US = United States.

Scientists from India and Turkey reported the lowest rates of collaboration, both in terms of perceived collaboration as well as estimated publication: 52% and 49% respectively reported that 0% of their collaborators in the past year were from institutions in other countries. Furthermore, 36% of Indian scientists and 35% of Turkish scientists reported 0% of their publications had international coauthors. Another 44% (India) and 40% (Turkey) said they had 1–25% of their publications authored with individuals from another country. This data seems particularly interesting given that the majority (84% India and 94% Turkey) indicated international collaboration was important to their success as scientists (Figure 1).

To understand why scientists might be collaborating less than they desired, participants were asked if they encountered any barriers. In all eight societies, the majority of scientists indicated that they did not perceive any barriers (Figure 3). Nevertheless, a significant minority in each country or region was able to identify obstacles: United States (30%), United Kingdom (29%), France (24%), Italy (38%), India (21%), Hong Kong (34%), Taiwan (29%), and Turkey (33%). These respondents were asked about the impact of 12 specific issues, in three categories – political, logistical and cultural issues – previously identified by researchers in the United States and Asia (Table 3) (Lane, Matthews, and Lewis 2009). Of these 12 issues, the majority of respondents identified funding as a major issue, with other issues varying based on society.

Figure 3. Barriers to international collaborations. Scientists were asked if they perceived any barriers to international collaborations. Approximately one-third of scientists from each country and region indicated that they encountered barriers to international collaboration. Abbreviations: FRA = France, HK = Hong Kong, IND = India, ITL = Italy, TAI = Taiwan, TUR = Turkey, UK = United Kingdom and US = United States.

Table 3. Scientists’ perspectives of specific barriers to international collaboration. The percent of “yes” responses to each barrier surveyed.

	US	UK	FRA	ITL	IND	HK	TAI	TUR
Political Issues
Funding	68%	77%	81%	79%	85%	73%	72%	84%
Visas	41%	38%	18%	26%	43%	27%	18%	37%
Intellectual Property Rights	20%	21%	16%	17%	25%	20%	25%	13%
Politics	5.7%	7.0%	5.3%	6.9%	11%	6.1%	10%	7.4%
Logistical Issues
Time Zones	55%	48%	20%	21%	24%	36%	24%	13%
Family Separation	41%	42%	31%	36%	35%	36%	45%	45%
Language	41%	33%	32%	42%	17%	20%	49%	42%
Calendar	29%	25%	26%	35%	22%	36%	32%	36%
Cultural Issues
Ethics	16%	18%	15%	17%	20%	20%	20%	18%
Gender	9.3%	9.1%	7.2%	11%	5%	0.0%	2.4%	6.5%
Religion	1.7%	4.2%	2.3%	3.4%	4.7%	1.2%	3.1%	3.6%
Sexual Orientation	1.7%	0.7%	0.7%	0.4%	1.2%	0.0%	2.4%	1.5%

Abbreviations: FRA = France, HK = Hong Kong, IND = India, ITL = Italy, TAI = Taiwan, TUR = Turkey, UK = United Kingdom and US = United States.

Respondents were allowed to fill out an open-ended response asking for “other barriers.” Here, responses were garnered from 769 scientists: 227 United States, 79 United Kingdom, 78 France, 68 Italy, 164 India, 21 Hong Kong, 40 from Taiwan and 48 Turkey. Comments varied in length from one word, such as “regulation,” to more detailed descriptions mentioning multiple issues. Of the 944 issues, one-in-three (313) reiterated one of the 12 issues from Table 3.

The remaining 632 issues were categorized into nine additional issues (Table 4). These categories included: (1) academic standards; (2) material and data sharing restrictions; (3) bureaucracy; (4) communication challenges; (5) lack of institutional support; (6) lack of opportunities; (7) nationalism; (8) geographical challenges; and (9) insufficient time (in contrast to calendars which was limited to conflicting fiscal, cultural, and holiday calendars). The description of these categories and their prevalence in our data are presented in Table 4.

Table 4. Additional barriers to international collaborations. The number of responses to the open-ended question regarding “other barriers” to international collaborations. Responses for previously identified barriers (Table 3) were excluded. The remaining responses were coded into distinct categories based on ideas described.

Category and Definition	#	US	UK	FRA	ITL	IND	HK	TAI	TUR
Academic Standards: Disparities in competitiveness, intensity of work, and general academic climates.	120	24	7	14	17	36	7	14	1
Material and Data Sharing: Controls prohibit the sharing of data, samples, and equipment.	119	77	10	7	5	8	3	9	0
Bureaucracy: Hierarchies and “dirty politics” in the administration systems of local and federal governing bodies.	101	23	5	12	14	33	4	1	9
Communication: Difficulty in communicating effectively via e-mail, phone, or video. Different standards in each country in terms of response times and different comfort levels in networking with other scientists.	67	29	8	3	4	6	2	14	1
Lack of Institutional Support: Lack of support or even opposition to participation in international collaborations by investigators or institutions	62	2	2	6	9	27	0	6	10
Lack of Opportunities: Lack of scientific activities (seminars or conferences) and invitations to present research at international platforms.	60	7	7	8	2	17	1	10	8
Nationalism: A reported or perceived bias against or toward working with a specific region/country.	53	13	3	4	3	16	1	5	8
Geography: Long distances between collaborators.	29	6	8	4	1	1	2	5	2
Time: Busy schedules.	21	5	3	8	0	3	0	1	1
Total	632	186	53	66	44	160	20	29	40

Abbreviations: FRA = France, HK = Hong Kong, IND = India, ITL = Italy, TAI = Taiwan, TUR = Turkey, UK = United Kingdom and US = United States.

The top new barriers respondents mentioned in the open-response were academic standards and regulations on sharing material and data. Scientists mentioned differences in the competitive cultures within and between research institutions, which we defined as “academic standards.” Example responses included an Italian scientist mentioning differences in “the method of approaching solutions to problems” and a Hong Kong scientist remarking on differences in “understanding the scope or definition of a collaboration.” Several others mention lower standards in other countries (of note, these responses were also coded within the “academic standard” category as well).

US scientists frequently noted data and material sharing regulations as barriers. Several specifically mentioned the International Traffic in Arms Regulations (ITAR), a series of US laws that strictly regulate and control equipment and technical data that might be related to defense or military technology (Blount 2008; Culhane and Worms 2001). ITAR limits foreign involvement or exporting of materials in research projects that the government deems related to national security (Abbey and Lane 2009). One US scientist remarked, “The FBI visits me about once a year concerning my collaborations in China,” and another said, “NASA discourages collaborations with Chinese [researchers].”

In addition, especially in Taiwan, India, and Turkey, several respondents mentioned bureaucratic impediments and lack of institutional support or opportunities (Table 4). One Indian scientist wrote “Ignorance of scientists in developed countries about institutions in India. They know of only a few institutions in India. It is extremely difficult to get travel funding. This prevents people with similar/complimentary backgrounds from interacting.”

Another notable issue was bias against scientists from specific countries, sometimes named but often in general terms of “other countries,” which we categorized as “nationalism”. For instance, a scientist from Taiwan said, “Potential collaborators don’t trust me due to my non-USA affiliation.” One Turkish respondent stated, “Coming from Turkey is perceived as a negative.” An Indian scientist mentioned, “a lot of Western countries … look down/are suspicious of Indian science. Unfortunately, Indian scientists behave similarly with other Third World countries.”

Discussions

Over the past two decades, S&E R&D funding has increased globally, contributing to tremendous growth in global S&E publication output (NSB, NSF 2019, 2020a). The pursuit of S&E R&D can be framed as a competition between different nations, which at times has been a driving force in promoting policy change and encouraging increased national expenditures in science and technology R&D (Amer Acad 2014; NASEM 2007). However, many scientists are approaching S&E with an increasingly collaborative outlook. While the United States still maintains its position as the top R&D contributor, China has steadily increased expenditures and publication outputs leading the world in total publication since 2016 (NSB, NSF 2019). This increase in research has built a worldwide network of scientists participating in intra-national and international collaboration.

In the present study, drawing on nationally representative surveys of physicists and biologists in eight societies, we attempted to assess the importance of international collaborations to scientists, the extent of their participation in such collaborations, and the barriers they perceive to such collaborations. Overall, scientists in this study saw value in international collaboration, although levels of participation varied greatly.

Among the countries in our sample, India and Turkey in particular had lower levels of participation in international collaboration. Fifty-two percent of Indian scientists indicated that they had not worked with any international partners in the past 12 months, despite 84% saying it was important. Similarly, 94% of Turkish scientists believed international collaboration was important, but 49% indicated they had no partnerships in the last year. These results match data demonstrating that approximately 18% of Indian and 25% of Turkish publications have international coauthors (NSB, NSF 2019).

Of the societies surveyed, Turkey and India have less developed science infrastructures, which likely limits their ability to engage in international collaborations. Both countries have lower research intensity, defined as the percentage of GERD per GDP, which is 0.62% for India and 0.96% for Turkey (NSB, NSF 2020b). These rates are lower than the other socieites surveyed: France (2.19%), Hong Kong (China: 2.15%), Italy (1.35%), Taiwan (3.3%), United Kingdom (1.66%), and United States (2.81%) (NSB, NSF 2020b). Furthermore, India and Turkey represent a lower portion of the top 10%-cited publications than the other societies, despite the fact that India ranks third in the world in terms of the number of publications, behind only China and the United States (NSB, NSF 2019). However, at least in India, the percentage of publications in biology and physics trend lower than the global average for these fields, perhaps indicating why collaborations are lower as well (NSB, NSF 2019).

The gap between Indian and Turkish scientists’ research preferences and their actual practice underscores the need to better understand obstacles to international collaboration in developing economies with limited R&D resources. Indian and Turkish scientists described barriers in their open-responses, in which they specifically mentioned issues such as onerous paperwork, bureaucratic delays or denials, and a lack of institutional support for international collaboration. Also of concern were the accounts of national, racial and ethnic biases experienced by Indian and Turkish scientists. They thought that their work was perceived as lesser quality than that of scientists from the United States and Europe. This in turns affects their ability to publish as well as to be invited to present at conferences, which is one of the key ways to meet collaborators.

Overall, most of the researchers in the survey did not experience major barriers to international collaborations. These results are similar to a 2018 survey of scientists on international mobility, which determined that many scientists did not face travel obstacles when relocating for work (McInroy et al. 2018). The most common barrier across all countries and regions surveyed was funding, consistent with previous research (Aarons et al. 2019; Hoekman et al. 2012; Noonan et al. 2018; See 2018; Wagner 2006).

Funding for research can be obtained from governments, intergovernmental groups (such as the United Nations), nongovernmental organizations, or philanthropic groups. These entities, however, have their own priorities and missions with international collaborations often only a small portion of their overall budget (Wagner 2006, 2018). For example, the US National Institutes of Health’s (NIH) Fogarty International Center has a mission to support and facilitate global health research, but the center’s budget is only approximately 80 USD million annually, less than 0.2% of the 40 billion USD annual budget for NIH (www.nih.gov).

Based on additional information given in the open-responses, funding obstacles scientists faced include lack of funding programs for international collaboration, no funding for travel associated with collaborations, and grants limiting money to only one country in the collaboration. Collaborations across borders requires higher costs than similar work over shorter distances (Wagner 2006). When multiple parties participate in research, it often requires travel to validate and share data and techniques, as well as having access in both locations to the same materials and equipment. The current system often requires each collaborator to find funding independently instead of working together on one large grant (Wagner 2006).

Policy changes associated with government funding can alleviate some of the barriers associated with international collaboration. Existing funding sources could be modified to allow shared resources across countries or new sources developed between nations to fund international collaborations. This would provide the infrastructure for collaborative international projects including staffing, equipment, and travel funding. For example, research in the European Commission Framework Programmes was created to encourage members from different countries to collaborate. The funding resulted in new partnerships in addition to rewarding existing collaborations (Hoekman et al. 2012).

Other policy adjustments could help promote international collaborations. To alleviate the burden of travel logistics, governments can “fast-track” visa applications for scientists and engineers for education, science conferences, and research collaborations, just as many countries do with entrepreneurs providing investment capital (Lane, Matthews, and Lewis 2009).

Additionally, the new movement to require publications to be open access would help provide access to all scientists without having to pay for journal subscription charges (Registry of Open Access Repository Mandates and Policies: http://roarmap.eprints.org/). Government policies promoting open access allow the dissemination of research more broadly and funding to shift away from journal fees and toward other expenses (Wagner 2018).

Furthermore, scientific associations or organizations would play a more enhanced role in developing and promoting international collaborations. Scientific conferences are an important part of communicating science as well as engaging with colleagues. They offer an opportunity for researchers from different parts of the world to engage with each other and potentially collaborate in the future (Wagner 2018). Scientific organizations could promote international collaborations through prioritizing speaker invites and travel awards to resource poor countries as well as scheduling sessions to include geographic diversity.

Although some barriers are government-related, other issues are institutional or local. These obstacles include bureaucratic paperwork and a lack of opportunities and support within departments or local universities where scientists are located. While some universities and research institutions promote international collaborations, some scientists still noted that this work is discouraged at their institutions. One Indian scientist remarked, “Top scientists who evaluate my research think international collaboration is a bad idea.” Without support from universities or department leadership, international collaboration rates will likely stagnate.

While this survey identified the importance of collaboration, the data was limited in its ability fully describe the barriers, nor could the data be generalized to other areas of science such as chemistry and engineering. Further work is still needed to fully understand the barriers, the motivation behind collaborations, and the factors that contribute to successful long-term collaboration. In-depth interviews with scientists would help elucidate the nature of the barriers they perceived or experienced, including whether issues are more personal or whether they are structural, and thus can be ameliorated by policy changes. Previous research on US scientists hinted that successful collaborations often involves women, mentorship relationships and favorable views of industry (Bozeman and Corley 2004).Other work points to international collaborations being interdisciplinary and performed by elite scholars (Jones, Wuchty, and Uzzi 2008). Interviews would help determine if these factors are relevant to successful international collaborations or if alternative aspects are more significant. Future research could also identify if there are nationalistic attitudes or biases that hinder collaboration regardless of opportunities. By addressing these issues, international collaboration can help improve the scientific enterprise necessary for global efforts to address problems facing all of humanity.

Acknowledgments

The authors would like to thank the students and faculty at Rice University’s Religion and Public Life program who helped with creation and implementation of the survey, especially the program director Elaine Howard Ecklund. This work was supported by the Templeton World Charity Foundation under Grant No. TWCF0033/AB14 and the Rice University’s Baker Institute for Public Policy Civic Scientist Program.

Disclosure statement

The authors declare no conflicts.

Additional information

Funding

This work was supported by the Templeton World Charity Foundation [TWCF0033/AB14].