Ethics regulation and sociology in France

Abstract

The French research ecosystem long resisted extending the ethics regulation processes established for biomedical science into the social sciences. This is now changing. This history of resistance is examined, together with the alternatives proposed. These include self-regulation by professional associations. Consideration is also given to the wider legal environment for French social sciences, particularly the laws on defamation and privacy, which also influence the opportunities for research and for the pursuit of grievances by participants. The introduction of ethics regulation, through IRB-style committees in universities and research organizations, reflects isomorphic pressures from the international research community, reinforced by the impact of the European Union’s General Data Protection Regulations (GDPR) on the application of national privacy law. Nevertheless, outside the specific domain of health-related research, ethics regulation still sits more lightly on French social science than in the anglosphere.

Introduction

There are many well-documented complaints from social scientists working in the anglosphere (US, UK, Canada, Australia, etc.) about the impact of systems of ethics regulation modeled on the Institutional Review Boards (IRBs) created in the USA from the 1950s to the 1970s to govern biomedical research with human subjects (Dingwall, 2012). The US system extended into the social sciences during the 1990s and took on more of the characteristics of a compliance bureaucracy (Babb, 2020; Schrag, 2010). It became a template for developments in Canada (van Den Hoonard, 2011), the UK (Dingwall, 2008b; Hedgecoe, 2008) and other nations within, or closely connected to, the anglosphere (Israel, 2015). Attempts to roll back this expansion have had little success (van Den Hoonaard, 2023; van Den Hoonard and Hamilton, 2016). Less attention has been given to the experience of countries that are more loosely coupled to the networks of English-language social science.

The French research ecosystem long resisted extending the model of the ethics regulation processes established for biomedical science into the social sciences but now appears to be yielding. This paper examines developments in France since 2006, when some members of the Association française de sociologie (AFS) first sought to develop a code of professional ethics, partly with a view to fending off the regulatory systems they had seen elsewhere. The AFS voted to reject the proposed code at its 2011 conference. This paper describes the debates within the association, based on internal documents and the accounts offered by some of those involved and traces the subsequent development of research ethics committees and data protection law as forms of governance.

French social science does not, of course, lack ethical values or escape ethics governance. However, these take different forms and balance interests in different ways. In particular, research and publication are subject to French law on defamation, privacy, intellectual property, and Republican secularism. Their impact will be examined through legal analysis, case reports and discussions among French social scientists and publishers. The French approach shows that alternatives to the IRB model are possible, although they come with their own problems and complaints. The influence of the IRB model is now growing, though, as a result of other changes in the socio-legal environment and an increasing orientation toward collaborations with anglosphere partners and publication in English-language journals.

The AFS and the ethics debate

Despite the high international esteem in which it is held, French sociology has a history of institutional fragility and internal fractiousness (Chenu, 2002; Dingwall, 2018). The impact of les évenements of May 1968 provoked quarrels and splits within the discipline that broke up existing social networks (Masson, 2008; Masson and Schrecker, 2016). The Société française de sociologie (1962–2002), which had attempted to unify and represent the whole community of sociologists in a similar way to the American and British associations, stagnated and, according to its last President, declined to around 50 members by 2001, when it lost its subsidy from the Centre national de la recherche scientifique (CNRS), the national research funder (Dufoix, 2023) However, the fall-out from the Teissier case in 2001, when an astrologer was awarded a doctorate in sociology by the Sorbonne, underlined the need for a more effective body to defend the discipline. The AFS (https://afs-socio.fr/) was launched as an independent, self-financing professional body in 2002 (Vannier, 2023). It currently has almost 1700 members. The AFS began discussing the development of a deontological code of research ethics in the spring of 2006, largely on the initiative of the association’s Comité d’action sociologie professionnelle (CASP), a group representing members from outside the academic sector. These ‘sociologues practiciens’, applied sociologists, were working for regional and local governments, public agencies, charities and, sometimes, private companies, alongside people from other professions where such codes were taken for granted (Pudal, 2014; Vassy, 2010; Vassy and Keller, 2008). This group had a set of questions about their moral obligations:

What should you do when the need to balance the budget of an agency, association or enterprise required you to work or intervene in certain fields, or to adopt approaches, that you found unacceptable? Should you accept any type of contract or instructions or were there some that should be refused – and on the basis of what principles? Could you, should you, allow yourself to be manipulated by a client? Taken together, these concerns called for, and emphasized, the need for a moral boundary around professional principles. Might it be necessary to develop a charter or a code of professional ethics for sociologists? (Vassy, 2010: 261) ¹

A working group was established by the AFS, under the leadership of Abou Ndiaye (https://fr.wikipedia.org/wiki/Abou_Ndiaye), a sociologist working as a trainer and facilitator for a number of public and private organizations, Olivier Vasseur (linkedin.com/in/olivier-vasseur-6bab6a168), a physicist and sociologist working for a government research laboratory in the field of aerospace, and Catherine Déchamp-Leroux, a senior lecturer/maître de conférences in health sociology at the university of Paris 13. Ndiaye was a Vice-President of the AFS, one of the founders of CASP and a strong advocate for the code project. Vasseur was also a member of the Société française de statistique (Journal de la société française de statistique, 1997). He had already been involved in discussions within that society about professional ethics. Catherine Déchamp-Leroux was a member of the AFS Executive committee (comité exécutif) and was aware of the constraints imposed by medical ethics committees (Le Clainche-Piel, 2021). A proposal was presented to the AFS conference in September 2006 to draft a code for consideration at the next conference, in 2009. Vassy (linkedin.com/in/carine-vassy-818415a4) joined the group in 2007, on the basis of her experience as a qualitative medical sociologist with close connections to countries with established professional codes and governance regimes.

According to her account (Vassy, 2010), the committee made slow progress. The sociologists working outside higher education, who had pressed for the code, were unable to free themselves sufficiently from other commitments to be actively involved in drafting it. The academic sociologists, who dominated the AFS, did not understand the contexts within which the applied sociologists were working and the ways in which these generated the demand for a code. The lack of a budget meant that participation was effectively restricted to sociologists based in the Paris area. A workshop was organized by the AFS in October 2007, without attracting much interest or opposition (Vassy and Keller, 2008: 140). In February of the same year, the École des Hautes Études en Sciences Sociales (EHESS - School for Advanced Studies in the Social Sciences) in Paris had also convened a meeting to discuss ethics regulation (Fassin, 2008). Together with Richard Keller, a historian of medicine from the University of Wisconsin, Madison, Vassy argued that the negative impacts of externally-imposed ethics regulation, on the US model, could be avoided by at least some degree of self-regulation, which articulated the basic rules and principles considered to be relevant and acceptable by researchers. Their actual impact should not be overestimated but they would establish a degree of legitimacy for sociological research in the eyes of the media, public opinion and research funders (Vassy and Keller, 2008). At the same event, Dingwall (2008a, 2008b) presented a more critical view, arguing that ethics regulation for sociology was fundamentally damaging to the free inquiries necessary to inform the proper workings of democratic societies and lacked any justification comparable to the past abuses documented in the biomedical and psychological sciences. He did not, though, criticize the principle of self-regulation.

There was little other publicity for the group’s work, other than in the association’s electronic newsletter which was sent intermittently to members. As a result, the group was unable to build a wider constituency of support for its efforts, and to show that it had listened to a diversity of views within the profession. Nevertheless, an initial draft was produced and revised in the light of feedback from the association’s executive committee. This version was placed on the AFS website in spring 2009 with an electronic forum for comments, open for 1 month (https://web.archive.org/web/20130729022640/http://www.afs-socio.fr/sites/default/files/congres09). Only eight responses were received but the text was reviewed and revised before being distributed to participants in the April 2009 conference in Paris, where two sessions were scheduled for its discussion.

Although the group’s efforts met with some encouragement, they also encountered familiar objections. Vassy (2010) summarizes these. There was dissent from the opening statements that sociological research should be capable of being useful for society and that sociologists should seek to be as objective as possible in their work. The critics asserted the right of sociology to be carried out as an intellectual activity with no specific intention to bring about immediate change in French society and resisted the language of ‘objectivity’. Sociology could, and perhaps should, be a way to promote certain political or moral values. There was little or no disagreement about the sections that dealt with integrity, data sharing, dissemination and archiving. A section on the treatment of junior colleagues was thought inadequate in dealing with concerns about sexual harassment and discrimination against women. The draft’s reassertions of the importance of anonymizing data and of respecting confidentiality were broadly accepted. However, there was widespread resistance to the sections dealing with the information that researchers should give to those who were being researched. The original text stated:

Sociologists have the responsibility to explain clearly their research to those who are going to take part in it. In order to make a fully informed decision about their participation, research participants should be informed in terms they can understand about the following points: the subject of the research; its objectives; who is responsible for it; who is carrying it out; who is funding it; and how the results will be disseminated and used. (Vassy, 2010: 263)

Participants should also be told that they had a right to refuse to participate or to withdraw at any time.

Many sociologists thought that this statement failed to reflect the ways in which they worked, partly as a matter of practicality and partly as a matter of strategy. For example, they could not provide this information to every single person who passed through their research field. As such, the research would not be consequential for them and their consent was irrelevant. Researchers would not necessarily know at the time how their findings would be disseminated or used, especially if they were using an inductive methodology. More fundamentally, where the researcher did know what were the topic and objectives of the project, they might not want fully to disclose these in order to avoid biasing the results. The text on covert research came in for particular criticism. First, there was discontent that this type of research had been singled out for specific comment. Second, some were unhappy with the emphasis on the limited circumstances in which it might be justified and the need for special attention to the ethics of the research, particularly in the anonymization of participants. These were seen as threats to this type of research, verging on prohibitions. Why, critics asked, should they be prevented from doing things that were accepted in journalism, like concealing their identity or engaging in illegal activities alongside members of the groups being studied? The final section of the code, proposing an ethics committee to review problems presented by members of the association attracted no comments.

As a result of the criticism, the draft statement was rejected by the membership of the AFS and a new version was called for. The two prime movers, Olivier Vasseur and Abou Ndiaye, were not re-elected to the AFS executive committee, probably because of their work on the code. Both declined to take any further part in discussions. The group reconvened under the leadership of Catherine Déchamp-Le Roux. There were a few new members but the group remained small. The issues were, however, explored further in a workshop convened at the Université de Limoges in the autumn of 2009 (Laurens and Neyrat, 2010). This will be discussed in more detail later.

Reflecting on the failure of the first version, Vassy (2010) conceded that it was probably too influenced by anglosphere models. It enshrined a formal version of research where the researcher had a broad idea of the possible outcomes and was poorly adapted to the practices of ethnography and inductive reasoning. In particular, its version of informed consent did not sit well with these approaches. At the same time, the draft was also the victim of misunderstandings. It did not require a contractual approach to consent and did not seek to block covert research, merely to explain why such research required special care to avoid harming participants. While its force might be largely symbolic, it would be no less real as a collective statement of values. Like any statement of values, it would be open to debate, change and revision with the passage of time.

Nevertheless, the final version of the code, presented to the AFS conference in Grenoble in 2011 was rejected (Le Clainche-Piel and Fournier, 2020; Pudal, 2014).

Why did the AFS reject a code of ethics?

The tortured history of the attempt to introduce a code of ethics for members of the AFS reflects a combination of factors that are both internal to the association and of the wider socio-legal and institutional environment for French sociology.

From an internal point of view, it had been hard to build a coalition of support for the proposal, which had originated within one segment of the profession, working in applied settings and experiencing problems that might be better managed if a statement of professional ethics were available. This segment combined applied sociologists from outside the academic sector and sociologists working in health and medical sociology, like Déchamp-Le Roux and Vassy who could see at first hand the potential challenges of external ethics regulation as it spread across all research in biomedical settings. Support also came from organizational researchers working in the tradition of Michel Crozier (Berrebi-Hoffmann and Grémion, 2009; Masson, 2008). Crozier’s laboratory undertook both theoretical research on collective action and applied research to improve social relations and cooperation in organizations (Musselin, 2024 ) However, relatively few of its graduates became academic sociologists and AFS members. The majority of AFS members were insulated from these concerns through university and CNRS ² appointments, as observed by the President of the AFS at that time (Ferrand-Bechmann, 2020). They simply could not see the point of a code of ethics that might constrain their work.

There was also the unfortunate coincidence that one of the plenary speakers at the 2009 conference was Michael Burawoy, who was a prominent advocate of a politically-committed ‘public sociology’ in the US and elsewhere.

As mirror and conscience of society, sociology must define, promote and inform public debate about deepening class and racial inequalities, new gender regimes, environmental degradation, market fundamentalism, state and nonstate violence. (Burawoy, 2002)

Burawoy (1972) had been asked to speak about his study of class and race in Zambian copper mines and seems to have made a point of underlining that this work would not have been possible if he had been too overt about his objectives. As Pudal (2014) reports, this fed into a local concern for the possibility of research critical of major institutions or elite groups:

Would we be able to conduct research on institutional discrimination, corruption in politics, economics or journalism, power as it plays out in a minister’s private office, among managers or in the secluded social worlds of elusive elites if we had to comply with the research constraints imposed by such a code of conduct? The response is obvious: no.

The vision of sociology as social criticism, as endorsed by Burawoy, had been a strong theme in French sociology in its reconstruction at the end of the Second World War, rivaling the more applied model that had generated the pressures for a code of ethics (Dingwall, 2018). Critical sociologists thought that a code of ethics was unnecessary and potentially constraining. Ethics in research was a personal matter, reflecting engagements in the world beyond the university or research institute. Debating a code in a context immediately adjacent to Burawoy’s modeling of his vision of public sociology would inevitably present a challenge for its authors and supporters. Pressing the point would jeopardize the AFS attempt to rebuild the community.

As several of the contributions to Laurens and Neyrat’s (2010) collection of papers from the Limoges workshop make clear, though, these concerns also have to be set into a context where researchers may have felt more vulnerable to external legal interventions than has generally been the case in the anglophone world.

Ethics regulation through the French courts

In contrast to the anglophone world, scientific publications in France are subject to the same legal regime as all other forms of publication, with the basic statute being the Loi du 29 juillet 1881 sur la liberté de la presse (Legifrance, 2021a) (It should be noted that the deference shown by anglosphere courts to scientific publications is more a question of custom and practice than of specific protection but, nevertheless, this type of communication tends to be left to self-policing rather than judicial intervention.). Article 29 specifies that:

Any allegation or imputation of an act which harms the honour or reputation of the person or entity to which the act is imputed is defamation. The direct publication or reproduction of this allegation or attribution is punishable, even if it is made in a tentative form or if it targets a person or entity not expressly named, but whose identification is made possible by the language used in speeches, chants, threats, written or printed matter, placards or incriminating posters.

Any outrageous expression, contempt or invective that does not contain the imputation of any fact is an insult.

These constitute criminal offenses and may lead to fines for authors, editors and publishers at a level depending upon the official status of the offended person or organization. Article 13 creates a right for offended persons or organizations to require publications to print responses or corrections with equal prominence to the original text and at no more than twice the original length.

Several of the papers presented in Limoges were case studies of encounters with French law. Torreiro and Sommier (2010) described their experiences of being prosecuted on the basis of a complaint from the Confédération Savoisienne, an association seeking to promote the independence of Savoy, a region of southeast France which had been an independent state until annexed in 1860. A publication had noted that some members of the association had advocated violence in support of this objective, something which contradicted the official public position of the leadership. The court of first instance fined the authors, editor and publisher €20,000, considering that two lines in a book displayed a lack of caution and balance. This judgment was overturned on appeal, although that decision rested on a technicality about the complainants’ delay in bringing the case to court. Naudier (2010) gives an account of a right of reply forced on a journal, Genèses, in response to an article she had published, using an author, Victoria Thérame, as a case study in the literary careers of women being published in the 1970s. This is accompanied by a chapter based on an interview with Oriot (2010), publisher of the collection, about the legal risks of this role. Oriot recounts being obliged to recall from bookstores and pulp a participant observation study of a writing workshop in the face of legal threats and shares similar experiences from other publishing houses.

Other social scientists have been sued for defamation in France (Atlani-Duault and Dufoix, 2014): Catherine Lutard-Tavard, a doctoral student in anthropology about her review of a book on Croatia and Serbia during the Yugoslavian war and Alain Garrigou, a professor in political science about an interview on opinion polls commissioned by the presidency of the French republic. According to Atlani-Duault and Dufoix, the accusation of defamation is used to question the impartiality of a researcher, but could also dissuade other scientists from expressing their thoughts.

Academic freedom and the French state

Unlike anglosphere governments, the ministry responsible for higher education research and innovation responded to concerns about the implications for academic research and established a commission chaired by an academic lawyer, Denis Mazeaud, to consider the implications for science and scholarship. This reported in April 2017 (Rapport sur les procédures bâillons, 2017). Its task was summarized thus:

For many years, researchers have been the target of law suits for defamation or denigration as a consequence of their scientific work published in academic journals or the press in general. These actions known as « procédures bâillons », have the clear object, or effect, of constraining their freedom of expression.

The commission was asked to collect evidence, in consultation with the organization representing university presidents, and to make recommendations that would reinforce the independence of the universities and protect them from litigation of this kind. The commission identified eight further cases involving academic lawyers, economists and political scientists, going back to 1986 but mostly between 2010 and 2017. As Atlani-Duault and Dufoix (2014) note, some academics might qualify for protection because of their status as civil servants under Article 11 of the Loi n° 83-634 du 13 juillet 1983 sur les droits et obligations des fonctionnaires (Legifrance, 2025c). This requires public employers to provide legal assistance to civil servants who are sued for acts directly connected to the performance of their professional duties. However, the provision is restricted, difficult to operate and not widely known. The Mazeaud commission took a different route in recommending various changes to the Loi du 29 juillet 1881 that would protect the right of free expression by academics and prevent lawsuits against universities, unless the ministry had consented to them. Article L952-2 of the Code de l’ education was subsequently modified by Loi n° 2020-1674 du 24 décembre 2020 de programmation de la recherche (Legifrance, 2021b) to state:

Teacher-researchers, teachers and researchers enjoy full independence and complete freedom of expression in the exercise of their teaching duties and research activities, subject to the reservations imposed on them by the principles of tolerance and objectivity, in accordance with university traditions and the provisions of the present code.

This seems to offer a degree of protection to communication in academic or scientific journals but it is not clear how far this would extend to commentary in the press or other media.

Another obstacle is the Loi n° 78-17 du 6 janvier 1978 relative à l'informatique, aux fichiers et aux libertés (Legifrance, 2021c). This greatly restricts the circumstances in which data can be generated, stored in computer files and used to study differences between people, which would be inconsistent with the Republican value of the absolute equality of all citizens. The starting point is article 8, 1:

It is forbidden to process personal data that reveals the supposed racial or ethnic origins, political opinions, religious or philosophical beliefs, or union membership of an individual person or to process genetic or biometric data in order to uniquely identify an individual, or personal data relating on an individual’s health, sexual life or sexual orientation.

This is followed by a series of exemptions, which still leave very little scope for social science research of the kind that might be taken for granted in the anglosphere, whether by the use of official statistics or by social surveys for specific purposes. There is little public evidence in the anglosphere of comparable legal pressure on scholars or publishers to tailor their work in particular directions, although there have been cases of conflict in US courts over the deontological obligation to protect informants taken by most social scientists (Scarce, 1995, 1999; https://bostoncollegesubpoena.wordpress.com/; Kara, 2022). Even the controversy over Goffman’s (2014) study of the impact of criminal law enforcement on a predominantly Black community in Philadelphia was carried on primarily at the interface between scholarship and journalism rather than in the courts (Lubet, 2018). Mention might be made of the 2014 English prosecution of the urban geographer, Bradley Garrett (Israel, 2022), for criminal damage in the course of breaking into closed properties in pursuit of ‘place-hacking’, although this attracted limited sympathy from other researchers (Oxford University Academic who Scaled Shard is Spared Jail Sentence, 2014).

It is difficult to assess the direct impact of this legal environment on the everyday practice of French social science research. The privacy legislation is certainly a constraint on the sort of large-scale surveys investigating ethnic or racial disparities that are routinely carried out by anglosphere sociologists and demographers. There are some exemptions but it is for the Commission Nationale Informatique et Libertés (CNIL - https://www.cnil.fr/), the agency responsible for implementing the statute to determine whether these apply to any particular research project, considering how the data are to be collected and managed. A large survey, Trajectoires et origines (TeO) was able to obtain a waiver allowing its authors to ask questions about respondents’ country at birth, nationality at birth, physical appearance and their parents’ country of birth (Beauchemin et al., 2016; Héran, 2016). These have, however, been controversial and criticized by civil society organizations representing minority ethnic groups and some researchers who adopt a traditional Republican stance (Héran, 2016; Le Bras, 2007; Perru, 2021). These constraints may have been important in sustaining qualitative research in France. If it is difficult to generate quantitative data on ethnic or racial disadvantage, it is possible to carry out qualitative studies of communities that have a high proportion of residents from particular ethnic or religious backgrounds. While qualitative studies may be important in highlighting social problems, they can be difficult to incorporate into policy processes dominated by planners and consultants from disciplines like economics.

The continuing advance of ethics regulation

What has happened since 2011? One source might be the graduate-level textbook on research ethics in sociology, by Genard and Roca I Escoda (2019). This devotes a chapter to the growth of ethics regulation in the francophone world but gives little new detail. The authors review the critical literature, mainly from the anglosphere. This includes discussions of the implications of bureaucratizing informed consent on a narrowly individualized model, and of the insistence on degrees of transparency in fieldwork that are incompatible with rigorous research. The book reviews the legal context of French sociology, but still relies heavily on Laurens and Neyrat (2010) and on Atlani-Duault and Dufoix (2014), who also draw on the Limoges papers and several of the anglosphere cases discussed above. Apart from this, the content would be familiar to anglosphere readers, although perhaps with a little more emphasis on moral philosophy. This is encouraged by the close relationship between sociology and philosophy in France (Dingwall, 2018): Genard and Roca I Escoda (2019) provide ample citation of deontological concern among French-speaking sociologists. To the extent, however, that it leaves research ethics to individual judgments, it may also leave the profession vulnerable to allegations that research participants are being treated as means to pre-determined ends, that data are being cherry-picked to support a prior argument rather than to test it. Perhaps the most interesting observation, following Gagnon (2010), is that the movement toward ethics regulation in the francophone world began in Quebec and only then spread into France, Switzerland and Belgium (Genard and Roca I Escoda, 2019: 150). This may suggest a need to consider the drivers for ethics regulation in Canada (van Den Hoonard, 2011; van Den Hoonard and Hamilton, 2016) as one means by which ethics regulation penetrated the linguistic and cultural barriers between the global anglophone and francophone academic communities.

The spread of ethics regulation across the francophone world has been gathering pace through the organizational processes identified by DiMaggio and Powell (1983) in their account of institutional isomorphism. This suggests that organizations operating in the same field will tend to converge on the same structures, cultures and outputs, regardless of their actual suitability to the business of the organization. This convergence is accomplished through three processes: coercive isomorphism where organizations are forced by political or legal pressures to adopt particular forms in order to be treated as legitimate; mimetic isomorphism where organizations are simply copying others that they perceive as more successful rather than taking on the uncertain task of devising their own forms; and normative isomorphism where they are competing for credentialled workers (professionals) who have their own values and expectations of what the organization should look like.

French biomedical research has experienced coercive isomorphism to adopt an IRB model with the 1988 passage of the Loi Huriet (Loi du 20 décembre 1988 relative à la protection des personnes qui se prêtent à des recherches biomédicales) (Legifrance, 2025a) and subsequent revisions, like the Loi Jardé (Loin° 2012-300 du 5 mars 2012 relative aux recherches impliquant la personne humaine) in 2012 (Legifrance, 2025b). This seems to have been driven mostly by external regulatory demands, particularly from the European Union and by the desire to participate in international scientific networks, where anglophone countries have imposed IRBs as a condition of collaboration. Hedgecoe (2009) describes how the IRB model was exported from the US to the UK and Sweden by similar means in the late 1960s.

Social science research has also experienced a pressure for isomorphism, but it is more recent. An increasing number of universities and research institutions have established ethics review processes, in response to both coercive and mimetic pressures. Genard and Roca I Escoda (2019: 149) cite cases of French researchers struggling with anglosphere ethics regulation in the absence of similar processes in their own country (Fassin, 2006). There are difficulties in meeting the coercive ethics requirements of European research funding streams, like the Horizon programs and the mimetic pressures of publishing in anglosphere journals that have voluntarily adopted ethics compliance certification as a condition of publication. This has been widely institutionalized by publishers as a normative expectation through a mimetic spillover from the medical field via the Committee on Publication Ethics, originally created by a self-appointed group of editors from anglosphere biomedical science journals. For those French social scientists who aspire to publish in English it is necessary to show at least ceremonial compliance (Meyer and Rowan, 1977) with this expectation. While it is still possible to be recruited to an academic position by a French university, collaborating exclusively with French speakers and publishing only in French-language journals, promotion depends increasingly on wider international recognition, which implies some kind of engagement with the anglosphere.

Nevertheless, in France it is still not obligatory to obtain approval from an ethics committee before starting a social science research project. It is, though, becoming more difficult to do research on health and health care without such clearance. Some hospital managers and clinical directors are requiring ethics committee approval before allowing sociologists to interview patients or their carers: the position about interviewing employees is less clear.

Because of these specific requirements where research relates to health, the Institut national de la santé et de la recherche médicale (INSERM) created an ethics committee in the early 2000s to examine projects from the humanities and social sciences. INSERM is the national biomedical research center, similar to the National Institutes of Health in the US. The Comité d’Evaluation Ethique (CEEI) de l’Inserm (https://lorier.inserm.fr/en/ethics/ethics-evaluation/) reviews the ethical aspects of research involving humans which is not biomedical in nature. CEEI is registered with the NIH Office for Human Research Protections (OHRP) in the United States, which gives it international recognition. CEEI has jurisdiction over humanities and social science research in the health sector, but this only seems to be enacted in relation to studies involving patients: a recent PhD, for example, examined a medical innovation through documents and interviews with professionals but did not follow through with patients in order to avoid CEEI delays. The CEEI looks particularly at the ways in which research participants are recruited, the information that they are given about the research and the ways in which their consent to participate is obtained and recorded.

Ethics regulation is, though, increasingly extending to research in the humanities and social sciences that is not concerned with health matters. In 2024, the CNRS created its own ethics committee, which is also registered with the Office for Human Research Protections. The problems created by these committees for humanities and social science research – in both CNRS and universities – are the same as those long recognized in the anglosphere. The committees are multidisciplinary and do not always include representation from sociology, partly because sociologists refuse to be appointed. The committee members do not, then, understand the purposes and methods of inductive methods, which are fundamental to qualitative research in both sociology and anthropology (Derbez, 2023; Desprès, 2020). This can result in lengthy correspondence and delays before researchers can obtain clearance

New constraints from European law

The spread of ethics regulation in France is also driven by recent European law on data protection. Coercive pressures have been reinforced by the EU General Data Protection Regulation (GDPR, 2016) which requires organizations to designate a Data Protection Officer (DPO) to oversee plans for the protection of personal data. In the French context, this regulation extends the 1978 law, described earlier. It deals with the relevance and proportionality of data collected by private or public organizations, data security and protection, data retention periods and ways of informing individuals about their rights in relation to their personal data. This regulation is mainly intended to protect European citizens from the misuse of their data by business but some clauses also affect the collection and curation of identifiable data in scientific research, including that collected by social science researchers, whether in the public or private sectors.

Some research ethics committees in France, like those for INSERM and the CNRS, require clearance from a DPO, certifying compliance with the GDPR before they will approve a project. Every university and research organization now has a named DPO, who is responsible for implementing the regulation. They are accountable to the CNIL, which enforces both French and European legislation in this area, including its application to research organizations. In practice, that agency has mainly focused on sanctioning businesses for misuses of data, such as collecting online user data without consent.

The GDPR covers all kinds of, directly or indirectly, identifiable personal data, however it is collected or stored – on paper, audio or video recordings, etc. Leaders of scientific research projects must define the purpose of their data processing; explain the legal basis for the processing; specify the period for which data will be retained; show that the collection of data is limited to that which is strictly necessary for the project; describe the measures that will be implemented to ensure the security of the data; and set out how they will inform research participants about the purpose of the project and who is responsible for it, how the data will be used, how they can exercise their rights to access their data, have it deleted or corrected and how their consent to participation will be documented.

It is this last point that particularly concerns social scientists. The regulations about informing research participants and obtaining their consent can vary according to the legal basis of a project, but they are still difficult to implement where a research project requires a degree of informality in the relationship between researcher and researched. Another troubling aspect of these regulations is the specific concern for ‘sensitive’ data. These include data on health, racial origin, religious, philosophical and political opinions, sexual life and sexual orientation. If sensitive data are to be collected,, which includes all research dealing with health, the project must be more thoroughly scrutinized by the DPO, analyzing the impact on private life through a Data Protection Impact Assessment, also known as a Privacy Impact Assessment. This involves more exchanges between the DPO and the research leader, before the project can start, and while it is running, to ensure continuing compliance, introducing new sources of delay and distraction. An impact assessment is also mandatory if the research meets two criteria in a predefined list of nine. These include the collection of data on vulnerable populations (children, elderly people, etc.). The GDPR does provide exceptions for research carried out by public bodies. If, for example, a study involves sensitive data, and the consent of participants is impossible or would be likely seriously to jeopardize the investigation, it is possible to obtain a derogation from the CNIL by proposing other ways to protect participant data. This, however, is an additional step that is likely to introduce a 6-month delay (InSHS, 2021; Roux, 2022).

The pressure for conformity is increasing as leaders of research groups are held personally responsible for every project’s compliance with the GDPR. Although Article 9 of the GDPR offers significant flexibility for research much of this rests on the discretion of the national agencies, like the CNIL, who are expected to enforce the regulations (Mondschein and Monda, 2018). For self-protection, research leaders and DPOs tend to pressure all their colleagues toward a risk-averse interpretation of the rules. Despite leaving the European Union, the UK is still bound by the GDPR but advice to researchers tends to be more pragmatic and place more emphasis on the available exemptions.

Discussions with French colleagues suggest that DPO approval is relatively straightforward where data are not sensitive and do not involve vulnerable populations. Where these conditions do not apply, social science researchers experience the same difficulties with DPOs as with some ethics committees. Approval requires a considerable amount of time-consuming paperwork, which cannot be delegated to administrative staff because of the level of scientific content required. The regulations regarding prohibitions and exemptions are complex. Moreover, the DPOs frequently have a legal background and a poor understanding of the methods and objectives of the social sciences. Their offices are often under-staffed, resulting in a rapid turnover of personnel and little organizational memory. It is often necessary to wait for several months for a DPO decision. This may be negative simply because the DPO has not understood the social science approach and has refused to engage with it in a legally-detailed fashion. When added to the delays introduced by ethics committees, the time required for regulatory approval may make it impossible to carry out the project at all.

Conclusion

The French case shows that sociology can function adequately in a bounded linguistic and institutional context without either an explicit statement of professional values or a specific structure of ethics regulation. This position is, however, very hard to sustain once its practitioners choose, or are pressed, to engage with the anglosphere and its institutional models. There are powerful isomorphic forces. Transnational legal pressures seem to have coerced French biomedicine to adopt at least the forms of ethics regulation (DeVries et al., 2009), providing an organizational base for extensions into the social sciences. European policy choices, as interpreted at a national level, on citizen privacy have also had an impact on the institutional environment for social research. Mimetic pressures, particularly from journal editors and publishers, have added to this. While French laws on privacy and publication, and Republican values of citizenship and academic freedom might have offered an alternative way of addressing ethical issues in social science research, they seem to have delayed rather than prevented concessions to anglosphere isomorphism.

These are early days to determine whether France will see the same clashes between researchers and regulators as persist in the anglosphere where this continues to be a running sore.

It is no exaggeration to speak of placing the craft of the researcher under administrative control, of subjecting it to bureaucratic logics. . .Researchers at the beginning of the 21^st century must deal with regular surveillance from a very diverse set of actors. . .All sorts of authorities instrumentalize different circumstances to legitimate their interventions to limit or control the autonomy of researchers. They pretend to be defenders of research through arguments about ethics, explicit consent and the protection of personal data, with the consequence of establishing their capacity to prevent social science research (Aldrin et al., 2022: 46–47).

Ethics committees in France have claimed the right to evaluate both the scientific quality and the ethics of research in all non-medical disciplines (Fagard et al., 2023). French researchers may just resign themselves to designing projects that minimize the likelihood of conflict by avoiding particular topics and methods, as already happens in the anglosphere (Bledsoe et al., 2007). Maro (2022) draws on her experience with an Australian university’s research ethics committee to identify this as a risk for French researchers.

As Dingwall (2016: 38) has observed, the result would be ‘systematic areas of ignorance about social conditions’. It is, though, unlikely that the politicians will come to their aid as they did in 2020: relations between the government and the academic community have become increasingly fraught over allegations of ‘islamo-gauchisme’ (left-wing sympathy for Islamic causes) and ministers have signaled interest in restricting rather than enlarging the space for free academic speech (Louati, 2021).

It is difficult to assess what the AFS gained by refusing to put in place an ethical charter. It is possible that the association gained 15 years of freedom for sociologists, without a professional body being able to call them to account. But it can also be argued that sociologists lost time and leverage because of the lack of internal debate and awareness of regulatory approaches that were spreading elsewhere. France may now be getting the worst of both worlds, with anglophone regulatory institutions being added to the existing legal constraints, in an environment where resource limitations and political agendas are being expressed through increased levels of managerial control over research and researchers. Many sociologists, especially those without secure employment, are increasingly apprehensive about the consequences and the continued possibility of their national tradition of free inquiry.