Is academic freedom at risk from internationalisation? Results from a 2020 survey of UK social scientists

ABSTRACT The question of the form that academic freedom takes and how it can be maintained in the context of the internationalisation of universities has become prominent in the UK in recent years. Both governmental and societal voices have raised concerns about perceived threats; however, much of the existing evidence is scattered and anecdotal. In October 2020, we distributed a survey in order to assess these issues. In this paper we report three main findings. First, UK social scientists express high levels of concern across a number of dimensions, from the effects of funding on research, to teaching content, to freedom of expression, and risks created by the online environment. Second, these concerns are somewhat greater in Politics, IR and Area Studies, suggesting that those disciplines which are most international in their content report greater risk. Finally, there appears to be demand for greater support. A majority of respondents did not know if guidelines existed in their department, and state that academic freedom was discussed infrequently or not at all. This suggests that institutional guidance and professional discourse have not kept pace with heightened concern. We find majority support for new legislation and even stronger support for a code of conduct.


Introduction
Is academic freedom at risk from internationalisation? This question may, until recently, have been counter-intuitive. 'University'from universitas magistrorum et scholariumetymologically denotes 'the whole' community of masters and scholars; a community which since its foundations in European territories has generally been international. The community of science and scholarship may, we assume, know no borders. Internationalisation is an increasingly important factor in the work of universities and other higher education (HE) providers. Over the last twenty years, the international dimension of HE has become more central to the agenda of international organisations and national governments as well of higher education providers (HEP) and their representative bodies. 1 And yet there remains uncertainty in internationalisation's definition. There are many different terms used in relation to the internationalisation of higher education. 2 Often the term has been used interchangeably or in conjuncture with globalisation. 3 As noted by Kehm and Teichler 4 this has caused an increasing 'fuzziness' or multidimensional character of the topic that is itself not neatly defined. Our approach to study internationalisation in HE is defined here as the 'international/intercultural dimension into the teaching, research and service functions of the institution'. 5 In recent years, the economic rationales and commercial manifestations of internationalisation have undermined the content of internationalisation and its humanist purposeas a driving force leading to justice and equity, peace and mutual understanding. 6 As pointed out by de Wit, the 'increasing competition in higher education and the commercialisation and crossborder delivery of higher education have challenged the value traditionally attached to cooperation, such as exchanges and partnerships'. 7 In light of these trends, the UK makes an extreme case to assess these trends with a HE sector which has long been internationalised but has seen a recent increase in exposure to the global market and an explicit policy agenda to increase exports. The British sector has historical status and linguistic advantage in a global academic environment whose lingua franca is English. In 2019, the UK government released an international education strategy 'to increase the value of our education exports to £35 billion per year, and to increase the number of international HE students hosted in the UK to 600,000 per year, both by 2030'. 8 This strategy made no reference to academic freedom and did not mention any risks or threats associated with internationalisation.
However, the political reality is that competition between states and non-state actors does create perceptions of threat. The status associated with the heights of the recentlyintroduced global university rankings may also be presented as a matter of 'soft power' and therefore national interest. While 'educational exports' are presented as a good, research which creates new intellectual property is to be closely guarded. President Trump's foreword to the 2017 National Security Strategy of the United States spoke of adversaries which 'steal and exploit our intellectual property'. 9 In the main body of the report, China was specified as the primary adversary in this regard, while Russia is mentioned alongside China as a 'revisionist power'. 10 It mentions coordination between government, the private sector and academia on four occasions. 11 The Russian national security strategy of 2015 prioritised 'ensuring national security in the area of culture' which entailed 'strengthening the coordination of the activity of interested federal organs of executive power and the Russian Academy of Sciences'. 12 While, for most academics, cross-border cooperation is an unmitigated good which is positively encouraged by one arm of government, the security agencies of the same governments found that such cooperation constituted a threat.
In the UK, a slew of reports which addressed threats or risks associated with internationalisation, including those to academic freedom, exposed an increasingly fraught environment. In 2019, the Foreign Affairs Committee released a report, A Cautious Embrace, which argued that universities had hitherto denied or dissembled in the face of evidence of authoritarian influence in UK universities. 13 In 2020, new proposals were issued for security management by Universities UK, while a draft model code of conduct was proposed by the Academic Freedom and Internationalisation Working Group (AFIWG, of which Heathershaw and Prelec are members). 14 However, despite this increased attention, there is little research on the nature of the problem. Much of the existing evidence of risks to academic freedom from internationalisation is scattered and anecdotal.
In October 2020, we distributed a survey to gather data to investigate perceived threats to academic freedom in higher education. In this paper we report three main findings from the survey. First, UK social scientists express high levels of concern about threats to academic freedom in their institutions across a number of dimensions from the effects of funding on research, to teaching content, to freedom of expression, and particular risks created by the online environment. Second, these concerns are somewhat greater in Politics & International Relations and Area Studies, suggesting that those disciplines with the most explicitly international contentindeed which are defined as internationalreport greater risk. Finally, there appears to be demand for greater support. A majority of respondents state they didn't know if guidelines exist in their department and that academic freedom was discussed infrequently or not at all at their institutions, suggesting that institutional guidance and professional discourse has not kept pace with heightened concern. It is therefore perhaps unsurprising that we find a majority support for new legislation and even stronger support for a code of conduct.
The paper proceeds as follows. First, we introduce the concept of academic freedom, making a distinction between negative (freedom from) and positive (freedom to) aspects. Second, we relate academic freedom to internationalisation via a survey of the limited academic evidence and febrile policy context. Third, we outline the difficulties involved in fielding a survey in this environment and the steps we took to avoid acting as advocates despite the fact that two of our members (Prelec and Heathershaw) are members of a group engaging in advocacy in this area. Fourth, we report our results and the three main findings stated above. We conclude by highlighting the need for further research in certain areas, including the causes of risks to academic freedom and the responses of staff and students to these risks.

Academic freedom IN comparative perspective
Academic freedom is at the very core of higher education and of the mission of the university: that much is undisputed. Most scholars argue that a properly developed HE system is impossible to achieve without it. Academic freedom is widely acknowledged to be central to both teaching and research. Having first taken shape in the specifically European setting of the early universities, the concept of academic freedom has expanded beyond a limited professional right 'to cover faculty members in a great variety of institutions 'beyond the high school,' and to protect the liberty to participate in extramural as well as intramural activities.' 15 In its basic form, academic freedom is a central value of HE that 'affects the academic profession in all aspects of academic work'. 16 There are, however, different understandings of this concept. Discourse around it has only begun to pay attention to changes taking place in HE and in society at large. 17 It has been only recently that scholars began to carefully examine the concept of academic freedom and the idea that lies behind it. For instance, the work of Traianou 18 investigates academic freedom in terms of erosion of autonomy that universities have from governments, and the autonomy that academics have within universities. 19 Others have examined the impact of government research evaluations and neoliberal market driven policies on academic freedom. 20 Greater clarity can perhaps be attained by distinguishing between negative and positive dimensions of academic freedom. 21 Considering that critical thinking is at the centre of what scholars should pursue in their research and stimulate in their interaction with students, it is elemental that academics usually operate, by and large, without strict supervision or direction in their teaching and research. This understanding of academic freedom is negative as it emphasises freedom from interference in one's actions. Negative freedom is the dominate strain in the UK policy debate, perhaps since the 1986 Education Act and through to the 2021 policy paper by the Department of Education (DfE). This latter paper invokes 'freedom of speech' in railing against "the rise of intolerance and 'cancel culture' upon our campuses', introducing and advocating a new Free Speech and Academic Freedom Champion within the UK's regulator The Office for Students (OfS), as well as a statutory tort against institutions which breach their free speech duty. 22 The paper recognises that academic freedom and freedom of speech are 'related but distinct concepts' in so far as the former is a professional right while the latter is a general right. However, it conflates the two throughout the paper and its policy proposals, completely failing to address the specific sources and conditions of academic freedom.
It is unsurprising that this narrow and confused account of academic freedom has been received with scepticism by many respondents to the DfE's consultation. 23 The discourse on academic freedom in Western societies in the twenty-first century largely neglects some aspects of it that no longer chime with the reality of HE in the twentyfirst century. In defining academic freedom,   24 writes that it includes 'the tenure rights of faculty members, which are conferred after a period of probation, bestow economic security as well as forestall restrictions on freedom that might stem from the power to dismiss.' 25 Such a concept of academic freedom is positive insofar as it entails freedom to act. Positive academic freedom requires security of employment, the time and stable funding for research, and the ability to participate in governance. The 1997 UNESCO Recommendationthe nearest thing to a global standard for academic freedomspecifies freedoms to 'take part in the governing bodies', and 'the policy of participation of all concerned in internal decision making structures and practices, and the development of consultative mechanisms'. 26 But the increasingly neoliberal understanding of universities as businesses and the precarity of a large portion of staff, especially in junior roles, has normalised the lack of job security to a large extent. For example, Exeter, the university which employs two of us permanently and has employed a further two of us casually, omits these key elements from its 2009 agreement on academic freedom and adopts an entirely negative definition of freedom from interference. 27 Therefore, and in spite of the very high positions many UK universities hold in international academic rankings, academic freedom does not enjoy great protection in Britain in comparison to its neighbours. Analysing comparatively the legal protections for academic freedom in an array of European countries, Karran 28 argued in 2007 that the UK was 'the sick man of Europe' in terms of academic freedom, explaining: in the UK, there is no constitutional protection for either freedom of speech or academic freedom, the law on academic freedom is designed to ensure 'just cause' for employment termination, the academic staff have only a minor input in the decision-making process, the Rector is an external appointment over which they have no rights, and academic tenure exists for only a few staff, who are dwindling in number as retirement beckons. 29 Despite these concerns, which remain in place at the time of writing, the UK nevertheless sits in the upper echelon of states worldwide with respect to academic freedom in studies relying on expert surveys. It is graded A in the recently-introduced Academic Freedom Index alongside other European, North American and a scattering of other states including Argentina, Mongolia and Nigeria. By contrast, at the opposite end of the spectrum at grade E we see China alongside several former Soviet, Middle Eastern and other states. 30 Thus, while the UK's historical legacies, liberal culture and advanced economy makes it a relatively strong example of academic freedom, it is far from being a shining beacon on a hill for the rest of the world.

Internationalisation and academic freedom
Since the 1980s, a gradual increase in the internationalisation of HE has increasingly exposed it to international political and global market dynamics. This move 'from margin to core' has occurred as a consequence of a series of developments, such as the increasing importance of research and education for economic development, but also 'the rapidly growing demand for higher education in the world, the end of the Cold War, and regional cooperation in higher education, the latter particularly in Europe'. 31 In this sense, academia, and the attempt of some to advance academic freedoms, has indeed become 'a real part of the globalisation process'. 32 That is not, per se, surprising, as academic freedom has long been recognised to rest on both cultural and institutional factors and to change 'from time to time and from place to place'. 33 However, the move from national to international of the past four decades has been specific and, in many ways, distinct from the 'globalisation' paradigm, narrowly construed as a homogenising and flattening of products and procedures on a global scale. As argued by Hans de Wit, the processes of globalisation flows in our economies and societies have expanded and influenced the manner in which internationalisation has been implemented. 34 In this sense, internationalisation of HE is one of the ways in which a country responds to the impact of globalisation, yet at the same time safeguards the individuality of a nation. 35 Internationalisation is thus a process more readily steerable by governments than is globalisation. 36 The literature on the international dimension of HE is complex and draws from research of a broad range of disciplines and thematic areas. 37 Existing research on internationalisation in HE has received considerable attention in the last two decades and has largely pursued questions that relate to institutional and organisational challenges 38 , benefits, and implications. 39 These include 'commercial advantage, knowledge and language acquisition, enhancing the curriculum with international content' among other objectives and drivers. The dimension of cultural exchange, while not as prominently discussed, is no less relevant: universities have engaged in a process of 'integrating an international/intercultural dimension into the teaching, research and service functions of the institution''. 40 As a result, cross-border cooperation projects are becoming increasingly common. The establishment of campuses abroad, cross-border collaborative arrangements, programmes for international students and the creation of Englishlanguage programmes are some of them. In this view, studies on internationalisation of HE do not solely address student mobility or internationalisation policies of HE but rather, as note by Kehm and Teichler, 'various links between internationally oriented activities embedding institutions, people, and knowledge'. 41 Assessments of the implications of internationalisation in HE differ widely. On one hand, the advent of internationalisation in HE had been regarded as something positive and important. 42 The benefits of internationalisation are frequently associated with funding resources, research collaboration and innovation, international staff and student recruitment, and intercultural exchanges among others. 43 In the European context, as noted by Wihlborg and Robson,44 internationalisation has become a key topic of HE policy debates with concerns of meeting European labour market needs and strengthening research and innovation capacity.
At the same time, the many aspects and the sheer complexity of internationalisation also raise various challenges for policy makers and the HE sector. Hénard and colleagues, for instance, argue that internationalisation introduces alternative ways of thinking about education that impact the governance and the management of the HE sector, which will vary according to a country's social or political development. 45 Kehm and Teichler also write that internationalisation accentuates existing international inequality between nations and world regions. 46 Taking a similar view, van der Welde argues that global competition has the potential to enhance vertical differentiation between research intensive HEPs and others. Such processes could further increase structural inequities between institutions and departments, which could further impact on the quality of education and its diversity. 47 As further stressed by Bamberger and colleagues: '[this process] serves to normalise inequalities, transforming internationalisation into a part of the meritocratic global race as nations seek to compete in the global knowledge economy; which deflects attention away from the embedded inequalities within the system'. 48 Among other key concerns of internationalisation are ways to sustain and enhance the quality of learning programmes and institutional credentials 49 as well as the neoliberal practices in HE (the deregulation of the education sector, market competition, rational choice as well as the global knowledge economy) which have been criticised for undermining professional and intellectual enquiry. 50 This trend towards internationalisation has been said to follow a paradigm of competition, rather than of cooperation: 'Competition for students, for scholars, for talents for the knowledge economy, for funding of complex research, for access to the top 500 in global rankings, and for access to high impact publications'. 51 Especially since the second half of the 1990s, this trend has increasedto the benefit of a select few institutions and at the cost of a large majority of tertiary education institutions, and entailing pressures such as revenue generation, competition for talents, and branding and reputation. In the UK specifically, the trend of 'competitive' internationalisation has been flanked by an equally clear process of marketisation, which has left UK universities compelled to operate within a fraught environment characterised by market pressures. 52 The backdrop is one of substantial changes taking place in the UK's HE sector, caused by both internal and external factors. As EU research funding to the UK has declined, the British government has sought to step in with national funds. However, in March 2021, the announcement of its new strategy for Global Britain coincided with announcement of drastic cuts to research funding from UK Aid. All the while, the amount of private donations, both domestic and international, has skyrocketed, almost tripling from 2009 (£0.5bn) to 2018 (£1.3bn) in the UK and Ireland. 53 While there are very strong indications that a large portion of this private funding comes from overseas, including from authoritarian regimes, the means to assess this problem in earnest are lacking due to the lack of transparency in reporting by UK HE institutions. 54 The long-standing question of the effects of the market dynamics of internationalisation on academic freedom has recently been joined by a second major concern that has hitherto been disregarded by the internationalisation literature: security. In recent policy reports, a debate has emerged between those emphasising the opportunities of the market and others highlighting threats to the security of universities, particularly with regard to China. 55 Some have also raised concerns about transnational education and the offshoring of campuses and university services (such as conducting degree programmes abroad) to countries that are not perceived as democracies. Wilkins 56 points out ethical issues such as lack of academic freedom and civil liberties which might arise in establishing international branch campuses overseas, particularly in countries with authoritarian regimes. Romanowski and Nasser 57 observe that in countries like Qatar or Gulf nations, expatriate faculty adapt their research and teaching materials to the country's socio-political and cultural conditions. The work of these scholars provides important insights on the challenges stemming from the internationalisation of higher education. Our study aims to complement their findings and explore the effects of internationalisation on academic freedom from the perspectives of British social science academics and consider potential tensions that can emerge between strategic imperatives to internationalise HE and the fundamental prerequisite of academic freedom.
In recent years, the question of the threat to academic freedom has gained increasing attention from government and parliament. A 2019 UK Parliament inquiry about the influences of autocratic regimes considered, among other issues, whether and how political leaders and/or business interests from non-democratic countries could pose a threat to institutional autonomy and academic freedom. Since then, a slew of policy reports have raised alarms about internationalisation. These debates are taking place against the backdrop of more widespread concern regarding freedom of speech and threats emerging from China (particularly among groups within the governing Conservative Party). It is not the place of this article to assess either the debate between 'cancel culture' and 'safe spaces' or the foreign policy of the UK towards China. However, it must be noted that interventions in this debate since 2019 are exclusively conceived in terms of a negative understanding of academic freedom and tend to be preoccupied with China. This includes concerns about Chinese interference in UK defence research, over-reliance on student fee income from China, major Chinese technology companies 'buying' influence at elite UK institutions, and even arguments about 'decoupling' from Chinese universities entirely. 58 Regardless of these published reports, few have engaged with the question of academic freedom and internationalisation in the policy realm. The UK's University and College Union (UCU) has focused on casualisation and insecure employment rights and conditions of academic staff compared to their colleagues in Europe 59 as well as raising awareness on the 'dehumanising effect' casualisation has in the UK, with fixed-term and casual contracts disproportionately likely to be held by women and BME staff. 60 In 2019, UCU has also submitted an official complaint to the UNESCO/ILO Committee of Experts on the Application of the Recommendations concerning Teaching Personnel, arguing that the law on academic freedom in the UK failed to offer sufficient protection for academic freedom. 61 Civil society organisations working to monitor and support academic freedom internationally are either small, emergent or both. The Global Public Policy Institute (GPPI), based in Germany, has created a new index which was published for the first time in 2020, while the US-based Scholars At Risk (SAR) has only recently expanded its networks into Europe. Meanwhile, the UK's Council for at-risk academics (Cara) is much older, originating in its aid to Jews fleeing Nazi Germany in the 1930s. Both Cara and SAR have seen big increases in demand from at-risk academics in recent years. SAR has received 5,000 applications since it began in 2000 with 500 of these in the academic year 2019/20. 62 SAR's Free To Think 2020 report analyses 341 attacks on HE communities 63 in 58 countries. 64 At the time of writing, Cara was processing over 100 applications and reported that they were 'still certainly busier than at any time since the 1930s, and there's no let up'. 65 Both organisations remain small and reliant on universities to host and fund at-risk scholars under their fellowship programmes.
It is in this context that the Academic Freedom and Internationalisation Working Group (AFIWG), with whom we collaborated on this survey, was established in September 2019 under the auspices of the Parliamentary Human Rights Group (PHRG). As of September 2021, the group is composed of twelve academicsa mix of genders, origins, career-stages and disciplines of the Humanities and Social Sciences. 66 Two of the present authors (Prelec and Heathershaw) are members of the group, which also includes named supporters from UCU, Cara, SAR and PHRG. In contrast to the aforementioned policy literature, the group conceives academic freedom in explicitly positive and global terms.
Challenges to academic freedom may result from repressive government practices. They may also arise or be aggravated by marketisation and an opportunistic approach to building global ties within the higher education sector. Some of these factors may in turn exacerbate the casualisation of academic work and prompt the creation of funding structures that make universities vulnerable to interference by donors. 67 The AFIWG has drafted and promoted a Model Code of Conduct to protect academic freedom in the context of internationalisation with common standards of transparency, accountability, and support to academics and students. 68 Previous research and policy reports have identified four specific areas of academic life in which internationalisation may incur on freedoms. First, academic freedom is at risk in international partnerships, both in transnational education and research. For example, academics who work at UK overseas campuses in UAE 69 , China 70 and other authoritarian states have reported censorship and education which is technocratic rather than promoting critical thinking. 71 Second, academic freedom is at risk in fieldwork abroad. The well-known cases of Matthew Hedges 72 , Giulio Regeni 73 and others demonstrate the real risks to UK researchers while threats to research associates, assistants and participants based in the field often go unreported. Third, academic freedom is at-risk for expatriate faculty and students in the UK. Previous research has found evidence of the surveillance, or at least data insecurity, of students from Saudi Arabia, Kazakhstan and China, and expat academics in the UK admitting that their research is constrained due to the threat of retaliation. 74 Fourth, in donations and other overseas funding. In contrast to the US, there is a lack of transparency in the UK. Most of what we know from scandals that have eruptedsuch as the Woolf Report into the LSE/Ghaddafi case 75and Freedom of Information requests. For example, there was a hundred-fold increase in donations from Middle Eastern states to Oxford from 2001-2014. 76 These findings suggest that the exercise of academic freedom and the conditions of internationalisation are entwined. We still know little about the extent of the risks to academic freedom and the perception of academics regarding the problem and its potential solutions. Our contribution here is to explore the views of UK-based social science academics on the matter in a more systematic way.

The survey
The objective of the survey was to assess academic perceptions on internationalisation, marketisation, and academic freedom in the UK HE sector. Prior to the fielding the survey, a pilot study was administered to a convenience sample of academics working in the social sciences across the UK. Both surveys were submitted for review and approved by the ethics committee of the College of Social Sciences of the University of Exeter. The pilot study included open-ended questions. Based on the results from the pilot study and research objectives, we created a questionnaire. We included questions on the draft model code of conduct (DMCC) of the AFIWG after consulting with that group on the wording and seeking agreement from its members to proceed. Similarly, we asked a question about a possible law to protect academic freedom because this was part of the political and policy conversation in 2020.
To distribute the survey, we collected public-domain email addresses of academic staff working in social science and humanities departments from the university webpages using the automated web scraping software Parsehub. We hand-coded those websites that we were unable to scrape. 77 Such method, however, didn't filter the staff that did not have student contact (e.g. administrative roles). To remedy this issue, we have asked in our survey the academic position of the respondent. We collected email addresses from 94 universities, reaching out to academics working in social science and humanities departments (broadly defined) with student contact in the last 4-5 years. The survey was distributed to 25,000 faculty staff in the UK HE sector working in these departments. A total of 1,500 academics of all ranks working in HEPs throughout the UK took the survey. The survey was administered via Qualtrics. The sample was composed of 27% Professors, 38% Senior Lecturers/Associate Professors, 25% Lecturers/ Assistant Professors, 5% researchers on fixed-term contracts, 4% employed on fixed-term teaching contracts, and 1% PhD students. 78 In terms of gender distribution, 56% of participants were men and 38% were women. Following informed consent, participants were invited to complete the questionnaire.
The survey included open-and closed-ended questions that asked faculty employed in social sciences departments about the state of academic freedom in universities and whether academic freedom was discussed in their working environment. We asked participants questions regarding specific aspects of academic freedom such as: conducting research and teaching as well as academic exchange and dissemination, institutional autonomy and research integrity. The questionnaire also included demographic data on participants' gender, academic position, department, regional specialisation, and political orientation. Questions were not obligatory; respondents were free to skip one or more, if they so wished. The survey was fielded in October-November 2020. This was a particularly challenging time as UK social scientists had recently begun an academic year teaching online or hybrid (online and face-to-face) due to the COVID −19 pandemic.
This exploratory study 79 constitutes an important first step in learning more about how UK academics understand the effects of internationalisation and marketisation on academic freedom. As such, we do not have pre-determined hypotheses, but instead present descriptive statistics here. We report the findings below, including both those which provide evidence that academic freedom is at risk from internationalisation, and those which cast doubt on this link. We hope these findings are useful in providing a snapshot of faculty perceptions on academic freedom and shed light on potential challenges to academic freedom stemming from internationalisation and marketisation. Although the response-rate is relatively low due (approx. 6%) to the aforementioned challenges to distribution, the number and distribution of responses suggest that they are representative.

Worried, but not sure why: perceptions of threats to academic freedom in the UK higher education sector
The data suggest that academic freedom is perceived to be under threat by a substantial majority of UK social scientists. As shown in Table 1, over two thirds (67%) of respondents indicate that academic freedom was under threat in higher education. Looking at differences among academic disciplines 80 , the perception of academic freedom being under threat is highest among Politics & International Relations (IR) scholars and lowest among Humanities scholars. 81 The Economics, Business and Law departments fare second after Politics and International Relations: 70% of them think that academic freedom is under threat in the UK. The perception of academic freedom being under threat is felt across disciplines, regardless of the region of specialisation.
When asked if academic freedom was discussed in universities, 74% of teaching and research staff would answer in the affirmative, as shown in Table 2 (14% talk about it 'often' and 59% talk about it 'not very often'). Considering that Politics & IR scholars are those who perceive threats to academic freedom most acutely, it is unsurprising that they are also those who discuss it the most (78% in total). Academics working in Russell Group universities are more likely to talk about academic freedom (77% overall) than those in other UK universities (69%), as shown in Table 2; and academics who talk about academic freedom often are almost twice more likely to be in Russell Group universities than in others (62% versus 38%), as per Table 3. 82 However, this frequency of discussion is in contrast with the clear finding that the majority of the people working in UK HE are uninformed about academic freedom guidelines. As illustrated in Table 4, almost two thirds (65%) of respondents indicate that they do not know if their department provides guidelines on academic freedom. Only 14% of academics answered that their department has academic freedom guidelines; and Economics, Business & Law academics are those who are nominally more aware about this (17%). These figures reflect research findings by Karran and Mallinson (2017) 83 , who note that academic freedom is a neglected subject and that UK academics have a scant knowledge of the concept. 84

Old and new threats: the move to online teaching
With regard to teaching, we asked specific questions about teaching online and teaching international students from autocracies. The issue has become more acute during the Covid-19 pandemic as more of these students were studying online from their home countries. According to our data, 44% of respondents' report that they self-censor when teaching online, while 34% said they didn't. Although these figures might not have a direct connection with the pandemic, or international education, they however raise concerns about academic freedom in online education more generally. The latest report published by Scholars at Risk (2020) 85 notes that, as a result of the COVID-19 pandemic, new threats to academic freedom have emerged. Most notable among these are the increased opportunities for the surveillance of research, teaching and discourse, as well as sanctions, restrictions, self-censorship, and isolation. 86 However, despite these emerging concerns, our data also indicate that most UK social scientists do not have additional concerns when teaching classes to students from authoritarian states. About 73% of respondents state they do not self-censor when teaching students from autocratic states in the UK. A majority (58%) of respondents state that the nationality of their students does not constrain class content with 23% saying that it does.  Research funding was also addressed in the survey. Academics participating in the survey were asked whether academics in UK universities should accept funding from foreign entities or governments that do not respect human rights. Three quarters (75%) of respondents think that academics in UK universities should not accept funding from such entities or governments. But do academics in the UK feel pressed to work with non-democratic governments? Again, the data suggest not. A majority of academics (59%) do not feel pressured to collaborate with non-democratic partners in the aftermath of Brexit, although a minority (10%) said that they do.
Perceptions about funding are taking place against the backdrop of changes in the national/international funding landscape. Our findings are consistent with other research showing the UK's exit from the EU presents a longer-term challenge in relation to continued access to research funding from EU sources. According to the latest Universities UK report in 2020, the proportion of UK research funding from overseas sources has increased in comparison to previous years with almost 24% of total UK research funding coming from international or EU sources. 87 EU funding constitutes more than half of the UK's international funding, and has increased by 7%, while non-EU other funding has increased at a faster rate, by over 20%, since 2017-18. Research collaboration with non-EU countries has increased considerably, with the number of co-authored publications with China growing by over 41% in the period of 2016-19. 88 However, heightened concerns in the UK regarding China in 2020, including the application of the National Security Law in Hong Kong, might have affected some respondents. In our survey, 28% agreed they would have serious concerns conducting joint research with academics based in universities in Hong Kong while 38% disagreed with this statement and a large minority answered don't know.

Under pressure: politics, IR & area studies scholars are most impacted
A similar trend to what was noted in Table 1 above, with Politics and IR scholars perceiving greater threat to academic freedom, can be found across several other areas. These include those outlined in Table 5, i.e.: freedom to select teaching content (highest perception of this being under threat, i.e. 56%); freedom to conduct research (Politics and IR feel most threatened, 50%, whereas Economics, Business & Law feel least threatened, 41%); and institutional censorship (in relative terms, Politics and IR perceive this as a bigger problem, 39%, whereas Economics, Business & Law the least, with 33%). As suggested by Table 6, similar trends can be observed when it comes to the erosion of institutional autonomy over the past few years. Overall, we find that Politics and IR scholars have a higher perception of academic freedom being under threat than other categories of social scientists, whereas Humanities scholars seem to be least impacted. There is, however, an interesting exception in terms of the freedom to express potentially controversial views on campuses, for which Economics, Business and Law scholars, in particular, reporting higher levels of concern.
These findings invite interpretation. One possibility is that Politics, IR, Business and Law schools are some of those who have often expanded most rapidly, for both domestic and international students, perhaps creating an impression among staff that market demand trumps the maintenance of standards and academic freedom. A further possibility is that the higher number of international students in these fields means that these staff are more exposed to sensitivities arising when teaching students from and conducting research in autocracies. 89 However, further research, including interviews, is required to explore the findings from these scholars.
A clearer and perhaps more indicative finding is that greater academic freedom concerns are held by area specialists than by academics who do not report a specialisation in a specific region. 90 The most significant contrast comes in relation to academics who report to be self-censoring when teaching students coming from authoritarian regimes: this value is considerably higher among scholars specialising in China (41%) and Africa (39%) than those who specialise in European states (33%), and all are substantially higher than the average of all respondents (20%)as shown in Table 7.
Another considerable difference concerns academic freedom worries in terms of selecting teaching content: scholars working on Europe (51%) and Africa (52%) report a higher   concern than the overall average of 42%as per Table 8. The percentage of those self-censoring in reporting fieldwork findings is slightly higher in case of scholars specialising in Africa (26%) and China (22%) than those working on Europe (19%), as shown in Table 9.

Moving forward
The picture emerging above is complex and begs as many questions as answers. As mentioned above, the data suggest that the conversation about academic freedom is infrequent among social scientists. While clear majorities state that academic freedom and institutional autonomy are at risk, less than 15% state that academic freedom is discussed 'often' in their universities, with almost 27% saying it is not discussed at all (see Table 2). Similarly, over 65% answer 'don't know' when asked if their department has guidelines on academic freedom (Table 4). This absence of regular debate and uncertainty about guidelines is accompanied by preferences among a majority for action beyond the university itself. As shown in Table 10, 55% of respondents think that the UK government should introduce a law to protect academic freedom in international partnerships, with 28% saying 'maybe' and almost 17% opposed. Finally, survey respondents were asked about their opinion on a bottom-up Draft Model Code of Conduct (DMCC) discussed by academics and implemented by HE institutions. 91 Over 60% of social scientist supported adapting a model of conduct such as the DMCC, with just 2% opposed, as per Table 11. 92

Conclusions
Our findings constitute the results from what is, to our knowledge, the very first survey of UK academics exploring the link between internationalisation and academic freedom. They are basic and exploratory and raise as many questions as answers. These findings demonstrate a high level of concern about academic freedom, with a clear majority stating that academic freedom on campus is at risk. Our data also show that those UK social scientists reporting greater concerns tended to be located in disciplines where internationalisation is inherent to the field (Politics & International Relations, Business) or areas where authoritarian rule is preponderant (e.g. Chinese Studies, African Studies). Yet, there are some inconsistencies in these findings and only a minority of social  scientists report that factors related to internationalisation have caused them to selfcensor or that their academic freedom is compromised. It is important, therefore, not to overstate the certainty of these findings or the degree of risk to academic freedom which they demonstrate.
There are several areas of limitation and uncertainty about which further research is needed. We highlight two here. First, we understand very little about the relationship between what we identify as positive and negative academic freedom. Do academics who lack security of contract feel more or less concerned about academic freedom? Logically, we may expect early-career academics in International Relations and Area Studies who may be engaged in projects with authoritarian state partners or seeking jobs in a global marketplace to report greater risk. However, a survey is a blunt instrument to gather data as the size of these sub-groups is low and the consequent findings commensurably less robust. Second, it is very hard to disentangle the relationship between the domestic and international causes of risk to academic freedom. Our survey asks about academic freedom in general and the risks associated with internationalisation in particular but it does not ask about the freedom of expression concerns surrounding socalled 'cancel culture' raised by government and civil society. We simply do not know if reported concern around foreign authoritarian states correlates with reported concerns over domestic risks to academic freedom, or whether these are two quite different population groups.
These are some of the avenues for further research this paper has raised. 93 A further set of questions relates to the comparative aspect of these challenges. What are the perceptions of academics working in other countries, both democratic and authoritarian, and how do they compare to those outlined here for UK social scientists? Does 'internationalisation' have a different meaning, and a different impact, in other educational systems? In which ways does transnational repression play out in the academe today? In a globalised and digitalised context in which academic freedom is increasingly under attack, while also beingas this paper has shownwidely misunderstood and therefore prone to manipulations, such questions appear more urgent than ever.   88. Universities UK. 89. Furstenberg,Prelec,and Heathershaw. 90. As with the categorisation by academic department (see above at note 83), we have proceeded in a similar way to analyse regions of interest. We have asked respondents to indicate whether they have a regional specialisation and, if so, which one (either from a drop-down menu of 10 choices, or with a write-in answer), and have subsequently coded the write-in answers. We have then grouped some of the country specialisms to draw more robust findings, while being wary of not oversimplifying the answers given by survey participants. The three groupings analysed here, which were chosen because they contained most responses and were therefore most representative, are as follows: European states (107 respondents), African states (86 respondents), and China (69 respondents); we created a dummy variable for each of these categories. Respondents could fall into more than one of these categories. 91. The question was explained as follows: 'The Draft Model Code of Conduct (DMCC) of the Academic Freedom and Internationalisation Working Group (AFIWG) provides a common set of standards and responsibilities to protect academic freedom for UK HE institutions which sign up to the code. The responsibilities include to provide training and support to staff and students, reporting mechanisms in cases where academic freedom is being violated, transparency about adherence to the code and accountability of the institution to its staff and students. The DMCC covers international partnerships, fieldwork and overseas travel, foreign students and staff in the UK, and the receipt of foreign gifts. It was drafted by academics as a model which staff and students can adapt and adopt in their own institutions. Do you favour the adoption of such a code of conduct within your institution?' 92. The far lower number of respondents answering 'no' to the DMCC as opposed to the law, invites the interpretation that opposition to a code is far lower than that to a law. However, the DMCC is a concrete proposal which was explained in the questionnaire while a law was a hypothetical proposal that invited speculation. Comparison of these two findings is therefore difficult. 93. Our research in this area proceeds in two directions. First, we are conducting analysis of survey experiments based on internationalisation scenarios from the 2020 survey and will report these in a subsequent paper. In these, we seek to test the effect on responses to risks to academic freedom in three different scenarios with a different treatment offered in each case. In one, where a colleague engages in self-censorship, we test peer influence. In a second, where a dean mentions the importance of the external partner, we assess the influence of senior management. In a third, where the UK Home Office refuses a visa to a foreign speaker, we assess the effect of domestic migration regulations on the exercise of academic freedom. In each case, our focus is how factors internal to the UK condition how academics respond to a perceived external threat. Our second direction of research takes this further, exploring via interviews and focus groups what internal resources and constraints UK academics face when assessing risks to academic freedom arising from internationalisation. 94. The difference in perception of academic freedom being under threat between respondents from different disciplines is statistically significant (χ2 = 27.3306, p-value = 0.001). 95. The difference in perception of academic freedom being discussed between respondents from different disciplines is statistically significant (χ2 = 20.5649, p-value = 0.008). The difference in perception of academic freedom being discussed between respondents from Russell Group universities versus non-Russell Group ones is also statistically significant (χ2 = 12.3497, p-value = 0.002). 96. The difference in knowledge about policy guidelines or procedures regarding academic freedom is not statistically significant between respondents from different disciplines (χ2 = 9.7029, p-value = 0.286). 97. The data in this table summarises the question whose formulation was as follows: 'From the list of aspects of academic freedom below, which (if any) do you consider to be currently at risk in UK universities? Please select all that apply.' Options included in the table: Freedom to select teaching content; Freedom to conduct research without commercial or political interference; Freedom from institutional censorship; Freedom to have potentially controversial views expressed on university campuses. For all these items the difference in responses among academics from different disciplines is statistically significant (teaching content item χ2 = 93.0946, p-value = 0.000; research without interference item χ2 = 73.3630, p-value = 0.000; institutional censorship item χ2 = 50.3549, p-value = 0.000; expressing controversial views item χ2 = 155.6609, p-value = 0.000). 98. The difference in considering the autonomy of institutions having eroded over the past few years is not statistically significant between respondents from different disciplines (χ2 = 7.8179, p-value = 0.451). 99. For all these dummy variables the difference in responses between those coded as belonging to one of these regional groups as those who were not are statistically significant (for European states item χ2 = 12.1936, p-value = 0.002; for African states item χ2 = 20.5353, p-value = 0.000; for China item χ2 = 19.8230, p-value = 0.000). 100. For the European and African states dummy variables the difference in responses between those coded as belonging to one of these regional groups as those who were not are statistically significant (for European states item χ2 = 4.1707, p-value = 0.041; for African states item χ2 = 3.9846, p-value = 0.046). This is not the case when comparing those who specialise in China relative to those who do not (χ2 = 0.2452, p-value = 0.620). 101. The difference in responses between those coded as specialising in Africa and those who do not is statistically significant (χ2 = 11.1120, p-value = 0.004). This is not the case when comparing those who specialise in China relative to those who do not (χ2 = 4.5834, p-value = 0.101), or when comparing responses of those who specialise in Europe to those who do not have that specialisation (χ2 = 2.4988, p-value = 0.287).