Responsible research and innovation: hopes and fears in the scientific community in Europe

ABSTRACT

We conducted interviews among some 80 researchers and research executives on their understanding of and their attitudes toward ‘Responsible Research and Innovation’ (RRI). The relevant notion of RRI concerns granting influence on research projects to the general public and stakeholders (‘science with society’), for one, or focusing research on the common good (‘science for society’), for another. We identified a welcoming attitude regarding RRI in the scientific community, but also worries that should be heeded in any effort to implement RRI. Concerns arise with respect to fears of ignorance and bias of societal agents, the loss of the autonomy of science, the neglect of basic research by RRI procedures, the difficulty to anticipate research outcomes and their social impact as well as the additional expenditure required by engaging with societal actors. Such worries provide information about obstacles in the scientific community and thus suggest ways of structuring acceptable RRI schemes.

KEYWORDS:

• Responsible research and innovation
• public participation
• attitudes in the scientific community
• interview study

Introduction: the concept of responsible research and innovation

As society is increasingly shaped by the impact of scientific innovations, calls to regulate research for greater public accountability and benefit emerge (Jasanoff 2003; Guston 2004; Bozeman and Sarewitz 2011; Kitcher 2011; Stilgoe, Owen, and Macnaghten 2013). A frequently proposed way to align research with societal needs is including social agents in the research process. Granting the public influence on setting the research agenda and on evaluating the social impact of the results is assumed to ensure that research proceeds in favor of socially accepted ends. This approach is endorsed in the research policy of the European Union under the label of ‘Responsible Research and Innovation’ (RRI) (de Saille 2015). In this paper, we present empirical results from an interview study conducted among European researchers and research executives regarding their views on RRI. Our study is part of the NUCLEUS project (an acronym for ‘New Understanding of Communication, Learning and Engagement in Universities and Scientific Institutions’) that aims to implement RRI in higher education institutions across Europe. In particular, we sought to identify reservations on and obstacles to RRI. Each endeavor to implement RRI schemes will need the assent and cooperation of the scientific community. For this reason, a clearer view of the positions and feelings of scientists is worthwhile.

René von Schomberg (2013) has suggested a conceptual clarification of RRI, which is useful for defining our endeavor. RRI is a strategy of societal agents to become mutually responsive to one another and to warrant moral acceptability, sustainability and societal desirability of research and innovation outcomes. In this framework, von Schomberg distinguishes between RRI as a procedure of participation and a product meeting certain standards. Product-oriented RRI means research proceeding on behalf of the people (or science for society), while process-oriented RRI is research conducted in a dialog with the people (or science with society). The study focuses on these two items and explores what researchers and research executives think of science with and for society.

‘Science for society’ or ‘product-oriented RRI’ means that the direction of research is determined on social grounds such as practical urgency or societal desirability. The expected outcome of a research undertaking should be beneficial to society. ‘Science with society’ or ‘process-oriented RRI’ emphasizes public participation (Von Schomberg 2013, 59, 63–67). Stilgoe, Owen, and Macnaghten (2013) and Macnaghten (2016) have elaborated a more advanced scheme for structuring the dialogical process between science and society. They emphasize the reciprocal adjustment upstream between social agents and scientists. Public participation (or ‘inclusion’) is essential in their view as well, but they stress ‘reflexivity’ and ‘responsiveness’ in addition. Reflexivity means that scientists are aware of a particular framing of an issue and of the options for approaching the matter in alternative ways. Responsiveness indicates that scientists are prepared to adjust a research endeavor in the light of such considerations. RRI in this broader sense is claimed to go beyond committing science to certain social goals. The point rather is to enlarge the spectrum of items under debate and to put governance structures in place in whose framework scientists and the public interact to devise desirable futures.

The two social groups relevant for such public involvement are stakeholders and the general public. Stakeholders have a vested interest in certain results; relevant bodies range from economic companies to patient groups. The general public, by contrast, is assumed not to be directly affected by the research (they are no stakeholders) and is thus expected to evaluate issues from a more detached stance and without special interests. ¹ Science operating with society proceeds in interaction with both social bodies and thereby shows an enhanced responsiveness to social aspirations. Such research takes up needs, wants, and desires as defined by social agents and transforms them into research projects.

However, previous work has revealed that giving societal actors a say in matters of research is controversial within the scientific community. Our large-scale, qualitative study adds new insights by using a large and diverse sample, including research executives along with researchers, and by placing emphasis on attitudes toward science on behalf of society. We begin by briefly presenting what is known about the viewpoints on RRI as adopted in the scientific community. We move on to describe the study's design, methods and sample. Afterward, we present the results and depict general features and tendencies. In the discussion section, we consider our results in light of the literature. Our conclusion identifies opportunities and barriers to be taken into account for successful RRI implementation.

The state of the art: how RRI is viewed in the scientific community

The existing literature shows that RRI is widely endorsed by scholars and policymakers, to be sure, but that scientists adopt a more reserved stance. Researchers regard their own interaction with the public, or science with society, as a duty that arises from taking public money (Mathews, Kalfoglou, and Hudson 2005; Peters 2013; Burchell 2015; Hamlyn et al. 2015). However, they tend to perceive the general public as lacking the knowledge necessary for understanding research findings, let alone giving research endeavors guidance and direction. In a similar vein, the wider audience is assessed as irrational and exploitable by special interest groups (Small and Mallon 2007; Young and Matthews 2007; Besley and Nisbet 2013). Accordingly, the chief idea which researchers associate with public involvement is science education or the dissemination of information (Peters et al. 2009; Bührer et al. 2017; Loroño-Leturiondo and Davies 2018). By contrast, stakeholder involvement is viewed as enhancing relevance and impact of research (Smallman, Komme, and Faullimmel 2015; Bührer et al. 2017). However, the influence of industry and politics is seen as harmful for the freedom and disinterestedness of research (Bond and Paterson 2005; Small and Mallon 2007; Pew Research Center 2009; Tartari and Breschi 2012; Perkmann et al. 2013; van Hove and Wickson 2017).

A major point of concern expressed by researchers is the time and effort needed for engaging the public and stakeholders (Kuhn et al. 2014; Burchell 2015; Hamlyn et al. 2015; Smallman, Komme, and Faullimmel 2015; Bührer et al. 2017; Porth et al. 2017; van Hove and Wickson 2017). Usually, such activities are not an asset in career progression or in receiving funding; therefore, incentives are missing. Also, a lack of institutional support and infrastructure as well as insufficient abilities of researchers present an obstacle.

Regarding science for society, the opinions of researchers are less well studied. They seem to value academic freedom (Small and Mallon 2007). External influences are accepted only with qualifications, for example, confined to allocation of resources at a general level (Mathews, Kalfoglou, and Hudson 2005, 163), research design (Lo 2016, 172) or applied science (van Hove and Wickson 2017, 223). Researchers fear intrusion in their work, especially by economic demand, and the imposition of short-term thinking. They expect this to result in a loss of creativity and scientific quality.

RRI in the sense of science for society seems to require some sort of anticipation of the future course of research, innovation, and their social impact. After all, such anticipation looks like the only option for assessing whether a novelty is likely to benefit society. However, such foresight knowl­edge is widely considered impossible among a large number of RRI scholars (see, e.g. Blok and Lemmens 2015, 26–28). Accordingly, judgments about the socially responsible nature of certain pathways of research or innovations are said to be doomed to failure. A ‘consequentialist’ approach to RRI founders on the tight limits set on predicting the development of knowl­edge and its social repercussions (Grinbaum and Groves 2013, 122–126; see also Nordmann 2014; Carrier 2019, section 2). However, on closer inspection such non-predictability claims turn out only to mean that technology development is not deterministic but subject to uncertainty and human choice. In this vein, many RRI scholars consider it possible to delineate several alternative scenarios of technological and social evolution, to assess their plausibility in qualitative terms, and to adjust research in light of such projections (Hansson 2011, 141–146; Stilgoe, Owen, and Macnaghten 2013, 1569–1571).

It is true that spelling out the notion of science for the common good does not necessarily invoke such foresight knowl­edge. If a technology is already available, debating about its introduction does not need to anticipate future pathways of research. Rather, under such conditions responsible innovation means to be responsive in that researchers are willing to adjust technological novelties to social demands. However, this process leaves the notion of ‘responsible research’ unaddressed. The latter notion still presupposes the ability to project lines of research into the future, while recognizing the uncertainty and malleability of any such projection. This state of the art suggests that the defeasible anticipation of lines of research and their social impact is part of various approaches to RRI. This is why we included relevant items in our questionnaires.

Few studies have been done on investigating the views of research executives on RRI. Lindner et al. (2016, 124) diagnosed only little awareness of RRI among European university administrators who instead focused on areas that are legally regulated (like gender equality or open access) or affect economic profit of their institution. Otherwise, the views of administrators and researchers mostly agree (Hartley, Pearce, and Taylor 2016). However, university managers saw the role of RRI rather as supporting the goals of their institution, e.g. acquisition of funding, while researchers regarded RRI as a civic responsibility that arises from public funding. Managers also feared the reputational risk for their institution from engagement done badly (Burns and Squires 2011).

The study: its design and context

In this section, we describe our interview study that was conducted among members of the European scientific community on their attitudes toward science with and for society. Our investigation adds to what is currently known about the attitudes of the scientific community in this regard in three ways.

First, previous qualitative studies mainly focus on researchers from one field, one institution or one country, and their sample often includes only 20–30 participants. The more large-scale quantitative surveys, on the other hand, are often unable to explore the grounds of the attitudes exhibited in the data. To assess the attitude toward RRI and its implementation, we combined a more diverse and larger sample with in-depth investigations. Second, current research often neglects a relevant academic group: research executives, such as deans and presidents of higher education institutions. The important role of administrators for RRI implementation becomes evident in researchers’ complaints about missing institutional infrastructure and support (Hamlyn et al. 2015, 37; Smallman, Komme, and Faullimmel 2015, 17, 55; van Hove and Wickson 2017, 221). Still, their views are under-examined (Burchell 2015, 39) or not studied independently of the views of researchers. We included them in our sample to compare their views to those of researchers and to illuminate the barriers to RRI within the institutional context of research. Third, there are only few studies on attitudes toward research driven by societal demand, a crucial but controversial RRI aspect. Clarifying underlying assumptions about, for example, freedom of research or innovation models is a prerequisite for addressing reasons behind reservations of scientists regarding RRI.

Outline of the study

We conducted qualitative interviews with over 50 researchers from different European countries and research fields. Their views were complemented and contrasted with the views of over 30 research executives, likewise from different European countries and institutions. The questions posed in the interviews aimed to capture attitudes toward RRI. For example, interviewees were asked about their views on involving social agents in the research process and on channeling research efforts into socially beneficial directions. Also, we were interested in the participants’ intuitions on what makes research responsible and asked for examples of what they view as instances of RRI. We focus on science with and for society as overarching RRI principles in order to obtain a general impression of attitudes and obstacles within the scientific community. We take this notion to include aspects such as reflexivity and responsiveness which specify possible ways to implement RRI. Therefore, we did not address the latter aspects separately in our questionnaires. We had originally planned to include RRI keys like gender equality, science education, ethics as well as open access (European Commission 2012). However, in accordance with previous research (Bührer et al. 2017; van Hove and Wickson 2017), we effected the selection of research items by appealing to social goals and granting influence to considerations from outside of science as the most controversial aspects of RRI. Gender equality, science education, ethics and open access seem to be generally approved, although their implementation often leaves a lot to be desired. Yet, whatever impediments remain to be overcome in this regard, they are unlikely to show up in interviews. It is not to be expected that interviewees openly distance themselves from such generally endorsed social goals.

We used semi-structured questionnaires that addressed the topics listed but also allowed us to inject additional questions when appropriate. Questionnaires differed between researchers and research executives, but their structure and the majority of questions agreed. While questionnaires for researchers focused more on their own research and research in general, questionnaires for executives put more emphasis on the institutional context. We also posed biographical questions about position, age, and disciplinary background to both groups of interviewees.

The interviews were conducted in 2016 and 2017. Most of the interview partners were suggested by NUCLEUS consortium members and were from consortium partner universities. Potential participants were approached via email, sending them information about the research as well as an invitation to participate. Informed consent was obtained from all participants and interviews were conducted mainly via video calls, while some were carried out face to face. After transcription of the recorded interviews, they were analyzed using the program ATLAS.ti. We employed qualitative content analysis following Meuser and Nagel (1991). Relevant sections from the interviews were paraphrased and then grouped under common themes that emerged from the interview material. Themes were assigned, attitudes were identified and their frequencies recorded jointly by both authors.

The sample

Overall, we interviewed 54 researchers and 32 executives from eleven countries. The largest group of interviewees was from Germany, owing to the fact that both project management and several consortium partners were located in Germany. Large groups of interviewees also came from the United Kingdom, the Netherlands and Serbia (see Table 1).

Table 1. Interview partners per country.

Country	Number of Interview Partners
France	3
Georgia	7
Germany	24
Greece	1
Ireland	4
Italy	1
Malta	5
Netherlands	15
Serbia	10
United Kingdom	16

Seventy interview partners were from consortium partner universities. About one third of all interview partners came from three institutions, namely, the Mathematical Institute of the Serbian Academy of Sciences and Arts, Delft University of Technology and Ruhr-University Bochum. Overall, members of 25 institutions were interviewed. Most of the researchers we interviewed worked in the natural sciences or in technology and engineering (see Table 2). Social scientists and researchers from health and medicine were also prominent in our sample.

Table 2. Research fields of researchers.

Field of Research	Number of Interview Partners
Architecture & Design	4
Computer Sciences & Mathematics	4
Education	1
Health Sciences & Medicine	7
Humanities	2
Natural Sciences	15
Social Sciences	7
Technology & Engineering	12
Other	2

Eighteen female and 36 male researchers were interviewed. Their mean age was 50 years, ranging from 30 to 73 years. Most of them have worked in research for longer than 10 years after obtaining their PhD. The executives we interviewed held various positions at their institutions, such as dean, principal or head of department. 16 of them were female and 16 male. Their mean age was 50 years, with ages ranging between 38 and 64 years. Most of them have held their current office for less than 10 years.

Study results and obstacles to implementation ²

The general attitude of researchers and research executives

Most of the researchers and research executives interviewed welcomed RRI in the double sense of interacting with society (both with stakeholders and the wider audience) and of selecting research topics according to social urgency. They were eager to serve society and explained that societal challenges play an important role in shaping their research agenda. When researchers were asked about the role of science in society, they stressed social progress, human well-being and overcoming social challenges, for example, via ‘Innovation, new technologies, new treatment modalities […] Better cars, better television sets, better internet progress’ (#005, 2). One respondent suggested that the orientation toward social challenges gives justification to research:

[Y]ou relate your research towards something which is of current, high interest for the society as well, not only for science. So, you are somehow spending tax payers’ money in an appropriate way, not just to enjoy, let’s say, science as an art or something which is of potential interest in the future […] I think it is closer related to applications. (#014, 2)

Yet, many participants also included basic research in the sense of creating understanding. Most executives defined their institution’s role in society as education and/or research. Many also mentioned contributions to societal benefit and generating economic impact.

Our interviewees appreciated social input for identifying pressing problems that they could set out to solve. Thus, engaging the public and identifying their demands is understood as a win-win scheme in that it stimulates research, on the one hand, and increases the public interest in science and the dissemination of scientific ideas, as well as public support for science, on the other. However, the preferred direction of interaction went from science to society. That is to say, interviewees mostly regarded the general public as recipients of information about science. Dissemination activities, such as lectures, open days, festivals, school programs and science shops, featured as the primary goal. These activities are tied to aspirations to enhance public acceptance of science. Scientific results are hoped to be taken up more readily by the wider audience and to be understood more thoroughly so that public support for science can be expected to increase.

The role of providing input to the scientific community was rather assigned to stakeholders. Stakeholders were frequently seen as competent sources of social demand. ‘They are the users of the potential product and therefore can define the conditions best. In this way we can consider far more realistic conditions than if we carried on without consulting them’ (#079, 5). That is, interaction with stakeholders was hoped to illuminate actual practical demand and thus to direct the attention of researchers to such problems. One respondent described the development of a health technology for doing rehabilitation exercises at home: ‘So, trying to create a technology that in some way can mimic [the patients’ social environment] is very difficult but it is a challenge that we would never have been aware of had we not spoken to them’ (#086[RE], 2). To use stakeholders as a source for research funding is another hope mentioned several times, especially by executives. Accordingly, many of the executives’ institutions had ties to industry partners as well as non-industry partners, such as NGOs and governments. They were regarded as an important source of research funding and many saw an expansion of these relationships over the past. Exchanges between science and industry were also seen as offering opportunities for non-academic careers to graduates. Several interviewees mentioned plans to expand their relationships with industry in the future.

In spite of their general endorsement, interviewees qualified their willingness to practice RRI. A proviso frequently mentioned was that RRI makes more sense in some fields than in others; it was said to be more suited to practice-driven fields than to basic research, more suited to environmental research than to elementary particle physics. Executives saw natural science, economics and technology as more apt for industry engagement, while they considered social sciences and humanities as better suited for non-industrial engagement. Relevance of a research field for society and interest expressed by the public was assumed as a precondition for engagement. Regarding types of institutions, technological and entrepreneurial universities as well as universities of applied science were seen as more appropriate for industry contacts than institutions doing predominantly basic research.

Also, interviewees named several preconditions for successful engagement. First, researchers wanted to see their commitment pay off in terms of their own career. Others mentioned prerequisites that enable individual scientists to encourage and take up social input. Most prominent was the demand to incentivize engagement, e.g. via career opportunities and calls for proposals. Respondents felt that they need assistance if they are supposed to productively interact with society in terms of time, reward, money, and training. A third condition that was brought up frequently was that science as well as external parties need to make an effort to enter into a genuine dialog. This requires willingness and openness on the side of the researchers as well as interest in scientific matters on the part of the public and stakeholders.

Some interview partners expressed reservations: they felt uncertain about how to implement RRI in practice, or viewed RRI only as a fashionable policy concept with a short life-span. One respondent bluntly stated: ‘I think it's nonsense’ (#067, 7). But the general attitude toward reaching out to society and being open to social input was positive. Interacting with the general public and with stakeholders was welcomed both as a source of funding and support as well as of ideas about useful pathways of research.

Examples of RRI provided by respondents

We asked interviewees to give positive examples of RRI. This was supposed to ascertain that the understanding of RRI among them roughly agreed and to capture their intuitions about responsible research. The most common feature of positive examples was fulfilling societal needs through research. Many of these examples came from medical research, environmental protection and sustainability. Other societal demands taken up in the examples were safety and security, technology assessment, development aid, prevention of conflicts and violence, gender equality and the creation of jobs. In addition, many respondents referred to demands that are urgent in their local context, such as supporting local businesses and improving working conditions, reducing traffic and noise. The involvement of societal groups in research was seen as enabling these kinds of socially beneficial research. However, several researchers emphasized in their examples that RRI should not exclusively focus on practical use but should also pursue basic research aimed at understanding.

As counterexamples to RRI, researchers mentioned, e.g., weapons research but also the use of mathematical models by banks in the financial crisis of 2008. One researcher called financial mathematics a weapon of mass destruction. Another type of negative examples was unrecognized harmful side effects of research, such as neglecting ethical issues when developing self-driving cars. Other negative examples given by researchers focused on biased interests influencing research. This mostly concerned profit-seeking companies who design clinical trials in such a way that desired results are more likely to show up or who even block publication of unfavorable studies. One interviewee worried about research funded by the pharmaceutical industry:

You might end up in a situation where you find out that a particular drug is not working or something, and that these results are not published, because they are not favorable to the party that is funding the research. (#002, 4)

In addition, many respondents brought up negative examples for RRI that refer to a lack of research ethics and scientific integrity. In this understanding, it is a major part of the responsibility of science to produce trustworthy and unbiased results that have been gained in morally impeccable ways. Misconduct like plagiarism and fabrication of results was cited as irresponsible research. A specific example that was brought up several times was research falsely linking MMR vaccines and neurological disorders in children. The irresponsible character of the relevant study was attributed to its poor scientific quality, but also to negligent peer review that had enabled its publication in the first place.

Worries associated with RRI

Worries respondents uttered with respect to RRI had chiefly to do with the fear of ignorance and bias. Public engagement was said to suffer from lack of knowl­edge on scientific matters. The general public was considered to be not sufficiently competent to provide useful input. One interviewee noted that non-scientists might not be familiar with the state of the art in science:

So if you ask society about: ‘What should we do on environmental problems?,’ they come up with things like: ‘Well, let's separate paper from glass or so’ and we already do that. So what worries us, then, is that, you see, there is a long way to go from what people in general think about a topic and what actually scientists are able to do. (#033, 4)

Another interviewee stressed that non-scientists need to familiarize themselves with topics in order to avoid unrealistic expectations: ‘The problem is that [citizens, policymakers, third sector organizations] tend to have unrealistic expectations and do not want to get the background knowledge to make informed decisions’ (#079, 14). Due to its lack of familiarity with research matters, the general public is likely to suggest problems that are either banal or beyond the reach of present-day science.

Researchers often viewed the influence of stakeholders like economic companies and policymakers as a source of bias on the selection of research topics. Political forces and businesses were feared to pursue their vested interests and not to act in favor of the common good. Concerns were articulated, in particular, regarding the impact of pharmaceutical companies on medical research. One respondent complained that the pharmaceutical industry would not necessarily care about curing patients. Stakeholder bias could alienate science from its social responsibility. In addition, the financial pressure that stakeholders exercise could infringe on the freedom of research.

Research executives, by contrast, apparently harbored no such doubts, but indiscriminately welcomed stakeholder funding and involvement for their institution. Among the few significant differences between research executives and researchers was this more blanket acceptance of stakeholder influence on the part of the executives. They had no qualms about stakeholder influence on research; they were mainly afraid that such influence would make science appear biased and less objective in the eyes of the public, even though this is not actually the case. The involvement of companies was feared to diminish the public credibility of science.

A major worry of respondents concerned the freedom of scientific research and the creativity of research that is thought to flow from the unfettered pursuit of approaches. They laid stress on the requirement that the influence of the general public and the engagement of researchers with stakeholders must not impair their freedom of research. Interviewees worried that a dialog with society could encroach on the autonomy of research and accordingly demanded that not all research be directed at societal aims. Especially in its early stages, research needs to develop unconstrained to allow for unusual and risky approaches that may lead to unexpected findings.

And for me, part of doing research is also being creative, coming to new ideas that maybe other people would find nonsense at first. I think that's very important for doing good research that you have the room to combine things that maybe people don't like in the beginning or don't see as promising. So I would really see it as a risk, if RRI means this, everything you do has to be discussed with everyone from the start because I think that’s going to lead to bad research if you do it that way. (#020, 8)

The anticipation of research and innovation

We asked researchers whether they think that research outcomes and their social impact can be anticipated. Most of them thought it possible to predict research outcomes to a certain extent, while the rest felt that foresight depends on the details of the project and the researcher. In particular, anticipating the development of a project was only regarded possible for moderate modifications of existing knowledge and for well-defined aims. Moreover, foreseeing outcomes was assumed to be easier in applied research than in basic research: ‘In engineering, it is possible to anticipate where the research will lead to. For [chemistry], it is more difficult to say’ (#044, 12). Basic research was often perceived as bearing a higher risk of failure. Several interview partners considered it impossible outright to anticipate research outcomes because of inherent uncertainties.

The same picture came out regarding the anticipation of the social consequences of technological innovations. Only a minority of interviewees deemed the anticipation of societal impact possible, at least to a large extent, whereas most respondents were more cautious. They restricted any such advanced knowledge to cases in which similar technologies already existed. Also, such judgment was seen as precarious early in research, while it was thought to get easier as research matured. Some interviewees rejected any possibility of anticipating societal consequences of technological innovations. One statement highlighted a previous false estimate:

I mean, a good example is the building of the bank of England, they built the biggest building, a very big building in London in order to put the biggest computers there, so it was in the 70s. And even they could not predict that computers were going to be smaller and smaller, so nowadays, you can put a computer in a … (laughing and showing a mobile phone). So, this kind of thing cannot be predicted. (#060, 7)

Accordingly, the surge of computational power and the breath-taking shrinkage of the pertinent hardware had escaped the attention of observers, along with its social impact, namely, the ubiquity of computational facilities. What was thought to be limited to large-scale machines is now available in portable devices everywhere.

The role of basic research

A concern regarding RRI articulated by several respondents was the insufficient inclusion of basic research that is aimed at understanding the world. Some interviewees thought that research released from the pressure of practice was a prerequisite of research bearing practical fruit in the long run. Basic research was supposed to create and fill up a repository that future practical innovations can tap. That is, basic research was considered essential for making practice-driven research sustainable. In this vein, a strong emphasis on RRI was feared to marginalize basic research and thus to undermine the future practical relevance of science. Research often starts with ideas that are vague and exploratory and whose relation to practical use is tenuous.

[W]e also run formats that are more exploratory in nature where we have sometimes a vague idea that a certain form of mechanism can be useful. … And then we have a greater chance of arriving at unexpected results which then in the longer term become useful in another project. (#041, 2)

Other respondents highlighted a genuine, non-instrumental role for understanding in scientific knowledge production. Science fitting the RRI scheme should be complemented by science creating an understanding of the world.

There is a role of science to make sense of the world around us, and there is a role for […] more application-oriented science and research to solve problems, to create solutions, to find new ways of looking at things. (#066, 2)

Although some respondents criticized any social restrictions on research, the majority welcomed RRI aspirations but sought to limit the research that falls under this heading. Many respondents thought it to be vital to maintain at least a corridor for research that is not shaped by societal needs, because future social benefits are hard to gauge, innovations cannot be planned or anticipated, or the search for understanding is worthwhile in itself.

Types of responses

While the positions described so far featured prominently in our sample and were widely shared across countries and disciplines, we could also discern more specific clusters of views on RRI. We identified four groups that varied in their intensity and kind of support of demand-driven research and societal engagement. We recognized one group that was clearly positive toward science with and for society as well as one that was clearly negative. The largest group by far (about one third of all participants) was ambivalent about RRI. A fourth group saw their national context as decisive for determining whether RRI should be implemented. Especially among those ambivalent about RRI, a few considered reflexivity an element of responsible research by being ready to call their own assumptions into question in light of social feedback.

The respondents clearly negative toward RRI advocated autonomous research that is undisturbed by societal influence. They emphasized lack of knowledge on the part of the public and biased interests on the part of stakeholders. For them, responsible research mainly had to do with scientific integrity and good scientific practice, for example, respecting ethical requirements. Furthermore, they were skeptical about the predictability of research outcomes and their social impact. Members of this group often came from Germany and worked in fields of basic research.

The largest group we distinguished was made up of respondents that were ambivalent about RRI. While they welcomed demand-driven science as well as involvement of societal actors, they also advocated protecting some areas of research from external influences. The worries prominent in the first group regarding public and stakeholder influence were present to a lesser extent. This group was also skeptical about the anticipation of research outcomes but more optimistic about the predictability of social impact. Its members often came from Germany, the Netherlands or the United Kingdom and often worked in applied research.

The third, clearly positive group focused on research intended to fulfill social demand and welcomed engagement, often regarding all groups, all stages of the research process and all fields. When asked to give examples of responsible research, they frequently mentioned instances of science with and for society. Members often worked in applied research or were research executives, many of them came from the UK.

In the fourth group, the pertinent national context dominated views on RRI. Members stressed the unfavorable environment they operate in as constraining their opportunities to do research in general, for example, instable economy, lack of funding and international exchange as well as brain drain. One respondent stated that scarce public budgets in his country give priority to things other than research, let alone RRI: ‘Science is important, but if you need to decide what is more important, to have lunch or to invest in science, of course you must provide food first, otherwise no science at all’ (#014, 13). Accordingly, the group demanded improvements for their research in general, independently of any concerns for RRI. Still, many thought research should be focused on improving the living conditions in their country and not on generating knowledge for its own sake. Also, they often lacked experience with stakeholder involvement and focused on public engagement in terms of dissemination and education. These interviewees came from Serbia, Georgia, Malta and Greece.

Some interviewees showed awareness of reflexivity. This means that scientists through the interaction with social agents become aware of the uncertainty of scientific knowl­edge, the relevance of a particular framing of an issue, and the possibility of alternative approaches (Stilgoe, Owen, and Macnaghten 2013, 1571). Indeed, some interviewees, answered that they were prepared to reconsider and revise their own viewpoints. For example, a researcher recounted an experience of discussing with members of the public and seeing ‘how my perspectives of research can be felt as enthusiastic or, on the contrary, pretty frightening’ (#011, 3). This indicates that this scientist was able to perceive his own position as seen from the outside. Other interviewees responded that scientists should participate in critical debates and be willing to learn from social communities. They recommended a process of reciprocal adjustment of expectations and actions. The importance of non-scientific aspects in framing topics, such as ethical considerations, came up repeatedly. Some interviewees also showed reflexivity by critically assessing the kinds of engagement activities exercised at their home institutions. They disapproved of engagement that proceeds top-down and suffers from lack of time or lack of impact on research. They also criticized a focus on affluent people as well as engagement practices that are conducted as part of public relations. Reflexivity was primarily found among interviewees working in fields like philosophy, science and technology studies, or social design.

Discussion of our results

Summarizing and condensing the messages contained in our data yield three kinds of hopes and four sorts of worries associated with science with and for society in the scientific community. Regarding the hopes, first, the public and especially stakeholders are assumed to be a source of worthwhile information about which kinds of research could produce socially beneficial or socially detrimental impact. As a result, their involvement is granted a constructive role in shaping the direction of research. Second, a claimed asset of bringing in stakeholders was obtaining additional funds from industry and policymakers. Third, other interviewees thought of interacting with the general public primarily in terms of science communication. Respondents hoped that a wider distribution of research results increases the sense of relevance of the projects in question and would thus contribute to support and additional funding. All in all, judging from the interviews, a friendly and welcoming atmosphere toward RRI prevails in the scientific community. Accordingly, the examples of responsible research given by respondents frequently featured involving societal actors and doing socially beneficial research.

These positive sentiments match the findings of Smallman, Komme, and Faullimmel (2015, 78) while, in contrast, van Hove and Wickson (2017) found resistance toward stakeholder involvement as well as orientation of research toward social utility and concluded that central features of RRI are seen as harmful to good science. However, it needs to be borne in mind that this positive attitude might be the result of selection bias. The general support for RRI, as expressed in the interviews, might be due to our sampling procedure rather than accurately representing the views held in the scientific community. Respondents who are more positive toward RRI might be more likely to agree to be interviewed about it. Additionally, most of our interviewees were recruited by NUCLEUS consortium partners which suggests a preferred and non-random selection of the people we talked to. This form of selection bias is common in studies investigating attitudes of researchers because they rely on voluntary participation (see also Peters et al. 2009, 58–56; Pew Research Center 2009, 58; Besley and Nisbet 2013, 646–647; Hamlyn et al. 2015, 10). Therefore, we grant that our findings might show greater enthusiasm about RRI than actually prevails in the scientific community. However, given that our respondents tend to approve RRI, their worries should be taken all the more seriously since less welcoming members of the scientific community can be expected to have such reservations to an even greater extent.

We identified four kinds of reservations about RRI. They confirm and further illuminate the impediments emerging from the literature (see section 2) and also point to possible remedies. The first and most broadly voiced source of reluctance was the felt loss of autonomy. Many respondents expressed their concern that non-scientists are not familiar enough with the issues in question to make a useful input possible. They rather feared that an uninformed lay audience could encroach on the freedom of research and distort fruitful avenues of research. Part of the unease we spotted in the scientific community can be reconstructed to the effect that scientists consider input from the people as possibly misleading in that the goals ranked on top are impossible to accomplish. It is not helpful to simply demand that science cure cancer. This is too general to be manageable and runs the risk of overburdening science. A more sensible question would be: explore the odds of immune therapy of skin cancer. But framing such more detailed questions and comparing their merits is beyond the grasp of the general public. The contrary worry was that the general public was not sufficiently imaginative and would ask for research that had already been completed. As a result, scientists feared that the public input would be in part overambitious and in part underambitious, wavering between what is unrealizable and what has long been realized.

Loss of autonomy with respect to stakeholders does not arise from assumed ignorance but one-sidedness. It is feared that stakeholders promote their own interests rather than the common good. Funding provided by stakeholders is welcomed, but the flipside is that economic and political influences could control the research agenda and thus take over science. Accordingly, respondents frequently emphasized evenhandedness when giving their own example of responsible research. We did not ask explicitly for such epistemic qualities, but bias was mentioned spontaneously as a problematic feature. In addition, making sure that results are novel, reliable and reproducible was seen as an essential feature of responsible research. Thus, researchers welcome a research system that bestows a limited influence on social agents but preserves a leeway of discretion for researchers. In this judgment, RRI should leave room for conceiving and carrying out new ideas that do not exhibit any social relevance. According to our data, this critical reservation among some of the researchers regarding stakeholder influence on their work is not shared by research executives. They were merely worried that the public might perceive research done in collaboration with stakeholders as biased.

The problem of assumed lack of information on the side of the public could be mitigated by including some sort of lay education. Pertinent methods are tested by James S. Fishkin in his Center of Deliberative Polling (http://cdd.stanford.edu/). Citizens who are willing to take part in a guided deliberative process receive relevant information material and discuss that information in various formats with experts or politicians (depending on the nature of the case). Our data suggest that having public intervention preceded by a phase of science education enhances the odds of acceptance by the scientific community. The other wing of public involvement is bringing in stakeholders, and the unease emerging here is due to one-sidedness. Balancing out such bias could be achieved by projects growing out of basic research, as proposed by some scientists. But an alternative could be to make sure that stakeholders with contrasting commitments are granted influence on the selection of research items pursued. Stakeholder bias would be kept at bay by involving a variety of differently biased stakeholders (Carrier 2013, 2558–2559).

The second and the third obstacle to implementing RRI procedures are connected to the appreciation of basic or epistemic research and the assumed difficulty of anticipating research outcome. These hurdles surface in the science-for-society wing of RRI and, to the best of our knowl­edge, have not been explored in detail in the RRI literature. As to the first item, respondents largely confined the usefulness of social input on science to applied fields, while they thought that basic research needs to proceed freely and without outside intervention. In their view, basic research is, as a rule, not socially relevant and should not be judged by standards of social relevance. They felt that such research frequently lays the foundation for important innovations further downstream but that the distance between research and practical novelties is too large to allow for a meaningful input from society. This assumed unidirectional dependence of innovation on basic research mirrors the linear model of innovation which seems to linger on in the scientific community, although its validity has been widely called into question (Rosenberg 1991; Stokes 1997; Carrier 2010, section 5). These findings suggest that RRI faces stronger resistance in basic than applied research fields.

In fact, the justification of the worry about basic research depends on how universal the commitment to RRI requirements is supposed to be. In other words, does the failure to be socially relevant make research irresponsible or merely non-responsible? In the former case, research failing to address any sort of societal demand and being only intended to understand the cosmic order (particle physics, astrophysics) would not be respectable and thus unworthy of public funding. In the latter case, slots for decent research would be acknowledged outside of RRI. In particular, being socially responsible would not be necessary in order to obtain public funding.

Third, many scientists consider it difficult to anticipate the future development of a research field and to assess its social impact. Implementing RRI seems to require anticipating future technologies and their impact on society. Some researchers considered such foresight to be out of the question most of the time; innovative novelties are inherently unpredictable. Researchers involved in basic research emphasized contingency and serendipity. The general line among respondents was that only the consequences of small-scale changes were foreseeable but that innovative leaps were beyond the reach of imagination. This was assumed to limit the possible domain of RRI considerations.

It is debated controversially what amount of foresight knowl­edge is necessary for passing sensible judgments about RRI (see the section on the state of the art above). What looks feasible in any event is targeting downstream phases of research that are close to practical implementation and address concrete devices or procedures. Such considerations could directly take up social aspirations and concerns, and no severe problems of anticipation would emerge. Take as an example efforts to measure blood sugar levels of diabetes patients in the sweat on their skin rather than in their bloodstream. Such devices avoid pricking one’s finger for determining blood sugar and are clearly a worthwhile alleviation of the relevant procedures (Young 2017). Such small-scale technological advances are predictable to a fairly high degree. However, RRI judgments on such options would be directed at the use to which existing knowl­edge is put and concern a stage when the relevant research has been completed to a large extent.

Other approaches to RRI could well proceed without reliable foresight knowledge. Rather than assessing what might come up, one could focus on avoiding ‘irresponsible research’ in the sense of neglecting ethical constraints and precautionary measures, or one could aim at the social conditions under which a technology is implemented. Who benefits and who suffers from a new technology? Have the terms for introducing the technology been negotiated in a fair and transparent procedure in which all relevant parties have been represented? (Carrier 2019, section 5). Furthermore, assessing the potential practical impact of a research line typically does not require precise estimates of the development. Think of climate change. It does not matter much whether the global surface temperature would rise by 3 or 6 degrees centigrade by the end of this century. Both surges would make the earth an inhospitable place. Ballpark figures are often sufficient for RRI-considerations.

The fourth worry had to do with the expenditure required for RRI. Interacting with society was feared to create yet another administrative burden and take time and resources away from research. Furthermore, respondents felt that such interaction is a demanding task that needs preparation and resources. They emphasized that RRI claims a lot of effort which needs to be supported or offset by suitable resources. In particular, researchers were afraid that institutionalizing RRI would mean imposing an additional bureaucratic superstructure on them. RRI activities are demanded to remain a voluntary effort and to be recognized in terms of funding or career opportunities. This suggests that the complement to preparing the general public for RRI challenges is training the scientists for interacting productively with lay people. RRI activities require some sort of engagement literacy. Scientists supposed to enter into a fruitful exchange with the general public and to take up and process proposals for the research agenda should be able to rely on a specific training and additional resources, and their efforts should be acknowledged in terms of career opportunities.

As there is little research on the views of executives on RRI, our sample included them alongside researchers. The overall result is that the views held by researchers and executives agree well. A major hope for RRI is to garner support for science, while ignorance of the public and required expenditure are feared. Some minor differences emerged. For example, executives often focused on their institution’s mission rather than on science in general. They emphasized education and research and considered the generation of societal and economic impact as secondary to this goal. Furthermore, they often had a better overview of disciplines when talking about the suitability of RRI to different fields. The only significant discrepancy was an unreservedly positive appreciation of stakeholder funding by executives. They were worried about the public reception rather than the epistemic quality of research done in collaboration with industry. This is in line with the concern of higher education managers for the reputation of their institution and its economic success found by Burns and Squires (2011). One of the reasons for this harmony certainly is that executives in their overwhelming majority have been scientists or are scientists on temporal leave. Their mostly transitory occupation with the administrative problems of institutions of research and higher education has not altered their fundamental outlook on research. We take this null result to be significant.

Although the mentioned hopes and worries were widely shared, four more specific groups emerged within our sample. These are aligned to some extent with research fields and nationalities. For instance, the group most negative toward RRI was predominantly made up of researchers working in basic fields. This result coheres well with findings that such researchers are less open to RRI aspects than those in applied fields (Bührer et al. 2017, 8). RRI is also implemented to different degrees in different countries which might account for part of the emergent grouping as well. For example, public funding in Germany places strong emphasis on scientific excellence and basic research while in the UK universities actively stimulate RRI (Lindner et al. 2016, 122). In our sample, the respondents most negative toward RRI often came from Germany, while respondents from the UK were often the most positive, among them many executives. However, respondents from Germany and the UK alongside those from other countries were also frequently ambivalent about RRI. Repeatedly respondents from Serbia, Georgia and Malta highlighted scientific efforts to solve social problems rather than producing knowledge for its own sake. Bührer et al. (2017, 28–31) likewise found expected benefits from RRI to be slightly higher in the south and southeast of Europe than in other parts. Still, our interviewees often lacked experience with stakeholder and public involvement beyond dissemination. While other findings on the degree of engagement in the southern and eastern parts of Europe differ (Kuhn et al. 2014, 48; Bührer et al. 2017, 29), our data suggest that RRI is not always a primary concern for researchers and executives in Serbia, Georgia and Malta who are focused on improving the tight national conditions for research. The groups we identified further emphasize field-specific barriers for RRI and also indicate that the challenges for RRI we pointed out might vary among countries.

In a similar vein, the area of research seems to affect how RRI is conceptualized. Primarily, interviewees from fields which reflect on science and/or society regarded reflexivity as a feature of RRI. They appear to be more ready to question their own approaches than scientists who often view socially beneficial research primarily as a matter of responsiveness to societal needs. However, our result is a snapshot of community attitudes and fails to capture the evolution of such attitudes as it might be produced through continued interaction with society. Our study design is less suited to provide access to features such as reflexivity than long-term studies such as the SPICE project (Stilgoe, Owen, and Macnaghten 2013).

Rather than finding a dichotomy of supporters and critics of RRI, our study paints a more nuanced picture of the European scientific community. While we identified small unquestionably positive and negative groups, the better part of our interviewees was ambivalent about RRI. While generally welcoming it, they also shared the mentioned worries and advocated protecting certain areas of research from social influence. This contrasts with Hartley, Pearce, and Taylor (2016) who found the staff of a UK university to either acknowledge their responsibility toward society or to reject RRI as an illegitimate politicization of research. This difference might be due to our larger sample capturing more of the diversity present in the community.

Nevertheless, it should be kept in mind that our sample focused only on some European countries (mostly Germany, UK and Ireland, the Netherlands and Serbia). As nationality seems to play a role in views on RRI, the lack of participants from Eastern and Southern Europe as well as Scandinavia might have affected the results. Furthermore, some institutions were represented strongly within our sample and RRI implementation might vary with organizational and institutional culture (Forsberg et al. 2018). Also, most of the researchers we interviewed worked in applied fields and the average age of our interviewees was 50 years. Younger researchers and executives might hold different views, for example, they might be less inclined toward RRI (Bührer et al. 2017, 40).

Conclusion

A chief result of our study is that many scientists appreciate initiatives taken by social agents and are willing to let the public contribute ideas and stimuli to their own research. Many scientists are happy to use the demand articulated by society as a guide for selecting research topics. Our study revealed a lot of good will in the scientific community to engage with RRI as a bottom-up process. This attitude includes the appreciation of an interaction with society and goes beyond the mere willingness to let the public discuss science-generated scenarios. In view of the selection bias mentioned, the markedly positive and highly welcoming attitude toward RRI needs to be qualified. Still, the general tendency we recorded points so clearly in the direction of open-mindedness that it would be surprising if the scientific community rejected to enter into a dialog with society. On the contrary, judging from our data, there is plenty of room for implementing RRI. However, it should also be acknowledged that this needs to be done with caution so as not to backfire. In essence, scientists had qualms about the fruitfulness and feasibility of proposals from the wider audience and wanted to keep basic research, understood as the search for understanding, at their own discretion. Our results further entail that the barriers facing RRI vary between research contexts. Medical problems are thought to be more amenable to social choice than cosmology. Kuzma and Roberts (2018, 7) likewise advocate tailoring RRI to specific contexts in order to prevent immediate resistance in the scientific community. This calls for context-sensitive implementation as well as the need to consider whether RRI might demand different procedures in different settings. While our interviewees agreed that science should serve society, they feared that RRI might not always be the best way to achieve this. They warned that too much outside influence would bias science or make it ineffective. While part of this resistance might be due to the fear of losing power and autonomy, their uncertainty about how to design the science-society relationship justly and effectively is of high significance. Whatever the justification of the concerns expressed by researchers is, they provide key information about the prospects of implementing RRI. Worries present in the scientific community need to be addressed and overcome in order to garner support from researchers. Researchers are supposed to lend their ear to society. Conversely, society should lend its ear to researchers as well and take their concerns and worries seriously. RRI cannot successfully be introduced when resistance in the scientific community is strong. There is good will, and it should not be spoiled.

Acknowledgments

The study was part of the NUCLEUS Project that involved 24 institutions from Europe as well as Georgia, China and South Africa and was supported by the European Commission under the Horizon 2020 framework (GA-Nr.: 664932). The authors wish to thank Ellen Böger and Stefan Fuchs for organizing and conducting most of the interviews and Ellen Böger for preparing the analysis. They also wish to thank Caitríona Mordan and Anne Dijkstra for contributing to organizing and carrying out interviews and Peter Broks for providing advice on the analysis. Furthermore, they would like to express their gratitude to Alexander Gerber (Rhine-Waal University of Applied Sciences) for leading the project as well as to the whole NUCLEUS Consortium for input, feedback and support. The authors would also like to thank all interview partners for their contribution.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributor

Martin Carrier is professor of philosophy at Bielefeld University. His research presently deals with methodological characteristics of research that is subject to economic and political demands, with heuristic strategies suited to make research practically beneficial, responsible research and innovation, and the public credibility of science. He is a member of the German Academy of Science Leopoldina and various other German and European Academies. He was awarded the Leibniz Prize of the German Research Association (DFG) for 2008, the Blaise-Pascal-Medal for Social Sciences and Humanities for 2015 by the European Academy of Science, the Canadian Diefenbaker Award for 2016, and the Werner-Heisenberg Medal of the Alexander-von-Humboldt Foundation 2018.

Additional information

Funding

This work was supported by Horizon 2020: [grant number 664932].

Notes

1 Many authors use the term ‘stakeholder’ to include all sorts of social agents ranging from companies to the general public (see, e.g., Blok and Lemmens 2015). But in view of the heterogeneity of the groups subsumed under this heading, such an all-encompassing terminology is misleading.

2 For reasons of readability, we do not include detailed references for the answers we received. Upon request, we provide anonymized excerpts of the interviews. References are given for direct quotations by indicating the number of the interview (assigned internally) and the page of the interview transcript. Interviews with research executives are distinguished by the prefix ‘RE’; the remainder refers to interviews with researchers.