European Science Foundation Policy Briefing Good scientific practice in research 10 and scholarship December 2000

Foreword ESF statement

Contents t a time when the need to build trust Good scientific practice in research and A between science and society is scholarship is essential for the integrity ESF statement becoming ever more important, it is vital of science. It sets internationally valid p. 1 that the conduct of science itself is based on benchmarks for quality assurance, which enable replication and further studies by Introduction the highest ethical considerations and that p. 4 misconduct within science itself can be other scientists. And it provides identified and dealt with in an open and safeguards against scientific dishonesty Good and fraud. Good practice, thus, nurtures scientific transparent manner. Several cases of trust within the scientific community practice misconduct have been reported over recent p. 6 and between science and society, both of years from across the World. This does not which are necessary for scientific Managing mean that there is an epidemic of such good advance. scientific cases but each one destroys trust both in the practice science system itself and between scientists. Several European Science Foundation p. 11 Most agencies concerned with science have (ESF) Member Organisations and some Investigating taken action to deal with these problems individual research institutions and allegations of and develop best practice. The ESF universities have already published scientific statement which follows and the report on guidelines, or codes, for good scientific misconduct practice across the full range of the p. 14 which it is based reflects the very large natural and social sciences, engineering amount of work which has already been Bibliography and the humanities. However, to be p. 16 undertaken by our Member Organisations. fully effective, such codes have to be Although we may have overlooked some Acknowl- more widely adopted by European edgements activities, nevertheless, I trust that this universities and research institutions, p. 16 report will help in the ongoing actions observed by all researchers and scholars necessary to develop and further improve and monitored for compliance. Both good research practice across Europe. This institutional and individual report is not the end of such efforts. Further commitment are prerequisites. developments in the way in which science is conducted are inevitable in a rapidly Procedures for investigating allegations of scientific misconduct complement changing world and there will always be a codes of good scientific practice. Such need to update and refine our approaches investigations are commonly carried out and this calls for ongoing action. ESF hopes at local (institutional) level, with gui- that this may be carried out in close dance and oversight by national bodies. The European partnership with other European Some countries, however, prefer to carry Science Foundation organisations representing institutions such out investigations at national level. acts as a catalyst as academies of sciences and the universities. for the development Finally, I should re-iterate that the public To achieve full compliance, and thus of science by must have confidence in the conduct of demonstrate effective self-regulation, the bringing together science. We in ESF and our Member various players – national academies and leading scientists research funding agencies, universities Organisations are determined that only the and funding and research institutions employing agencies to debate, highest standards should prevail. scientists and the scientists themselves, plan and implement each has distinctive advisory, managerial pan-European Enric Banda initiatives. ESF Secretary General or regulatory responsibilities. ESF, with its two sets of stakeholders monitoring of those standards and in firstly, (its membership drawn from developing transparent procedures for funding agencies, national research investigating allegations of scientific organisations and academies of sciences misconduct. Pan-European progress in and letters and, secondly, the research these areas would improve quality community at large) is uniquely placed to assurance, strengthen the self-regulation play a significant role in promoting the of science and help reinforce public trust highest levels of scientific integrity and in science. Therefore, ESF believes that better self-regulation across Europe. At a the following conclusions and strategic level, there is a need for more recommendations set out a basis for commonality in codes of good scientific further action at European level on this practice, in the effective managing and important topic:

1. Both the globalisation of science, with its extensive inter-organisational and international collaborations, and current public concerns about self- regulation underline the need to extend and harmonise codes of good scientific practice and procedures for investigating allegations of scientific fraud.

2. European scientific institutions are responding, though somewhat unevenly, to these pressures and are addressing the moral issues of scientific ethics and integrity and the more practical matters associated with self-regulation.

3. With its extensive membership in 23 countries, the ESF is uniquely placed to play a pan-European role in promoting common approaches amongst its Member Organisations for managing and regulating good scientific practice.

4. The current debate about a European Research Area introduces a favourable political dimension and creates a window of opportunity for action.

5. At a strategic level, there are several possible initiatives which need to be taken, at a European level, to strengthen approaches to scientific integrity and good scientific practice across Europe. Some of those listed below are purely advisory; others require a more active intervention. .ESF commits itself: to support and promote vigorously the concepts and principles of good . scientific practice in research and scholarship; and to promote the principle that the selection of scientists by academic insti- tutions should be transparent, based primarily on criteria of scientific quality, creativity and promise, without discrimination on grounds of sex, race, political opinions or cultural backgrounds.

European Science Foundation Policy Briefing 2 ESF considers that a number of other actions are necessary. In taking action, it is vital that the approach is inclusive and sensitive to what has already been achieved by many of the ESF Member Organisations and other European organisations and by relevant international developments carried out by International Council of Scientific Unions (ICSU) and other similar bodies. Real progress will require linkages with these initiatives. And it is important that the goal of harmonising policies and procedures on the basis of best practice should be achieved without compromising the principle of subsidiarity in matters of executive action. .Therefore it is recommended that: ESF Member Organisations that are national academies should draw up national codes of good scientific practice in research and scholarship, . where these do not yet exist; and ESF Member Organisations that are national academies should initiate dis- cussions on the most appropriate national approach to procedures for investigating allegations of scientific misconduct (where this has not yet been done), whether by means of an independent national body (as in Denmark), formal procedures in each university and research institution, . or by other means. ESF Member Organisations that are research-funding agencies should consider ways of making an institution’s eligibility to apply for research grants conditional on that institution having adequate policies for good . scientific practice and procedures for investigating scientific misconduct. ESF Member Organisations that employ scientists should act as responsible employers with clear, fair and robust guidelines for good scientific practice, coupled with effective and transparent management procedures for implementing these guidelines and for investigating allegations of scientific misconduct.

Finally, it is important to consider whether there is a need for any pan- European structures to reinforce national arrangements, for example, by maintaining a college of eminent scientists who might serve on local or natio- nal committees investigating scientific misconduct, or by setting up an Om- budsman system to provide a third party for counselling “whistleblowers” in the scientific community. Consideration of such issues will need to involve not only ESF and its Member Organisations but also other relevant European or- ganisations, including those representing the universities.

European Science Foundation Policy Briefing 3 Introduction 5. Greater competition between scientists for scarce research and scholarship The nature of research and funds and the emphasis on publica- scholarship tions as measures of performance have put pressures on scientists to produce 1. Scientific research and scholarship are results quickly, in turn creating diverse and multifaceted activities temptations to short-cut proper embracing a wide range of intellectual procedures. Senior researchers and and practical endeavours. These scholars sometimes have insufficient include theoretical studies, time to involve themselves personally experimental work and surveys, as well in the day-to-day conduct of the as the verification, further analysis and various investigations they may be extension of earlier work. The objec- directing. The greater weight that tive is always to extend human some public funding agencies attach to knowledge and our understanding of the utilitarian value of science, too, has the physical, biological and social sharpened the focus on outputs, as well worlds. as challenging traditional academic 2. Progress in science depends on trust. values of freedom of thought and ac- Scientists must have confidence in the tion. results of other scientists. Also, society 6. The ethical issues always inherent in has to trust the honesty and motives of social science and clinical research, scientists and the integrity of their where people are the subjects, and results. Much of the current increasingly posed by advances in disillusionment with science in Europe biomedical and biotechnological is due to a loss of public trust. research, have added to the problem. 3. To regain and retain public trust, it is In today’s more inclusive society, these vital that the ethics and integrity of issues are now widely held to be too science are beyond question. Good important, at best, to be left to practices in the design, conduct, informal and private debate within the interpretation and reporting of scientific community, or, at worst, scientific research and scholarship are neglected by scientists. Last, but not the gatekeepers of integrity. They are least, self-regulation has been damaged the prerequisites of mutual trust by several well-publicised allegations within the global scientific community and some proven cases of scientific and of greater trust between scientists misconduct and fraud. and the public. Where there is a 7. All this has turned a spotlight on issues climate of trust, the results of science of scientific integrity and professional are more likely to be accepted, standards, and put pressure on the exploited or applied, for the benefit of scientific community to strengthen the humankind. process of self-regulation and make it more visible. Self-regulation 4. Science has had a tradition of informal Principles of scientific integrity self-regulation to ensure that the 8. Scientific integrity is at the heart of highest professional standards of the trust on which scientific communi- integrity are maintained. Over the cation and collaboration depend. past 20 years several trends in the Scientific integrity demands that those increasingly complex world of science, engaging in research and scholarship however, have strained the traditional, should at all times, and without excep- low-key approach to self-regulation. tion, adhere to the following basic principles:

European Science Foundation Policy Briefing 4 .highest professional standards in 10. To be effective, good scientific practices designing and conducting investiga- have to be made explicit in written tions guidance or codes. There also have to .a critical, open-minded approach in be managerial procedures for conducting research and scholarship implementing them and monitoring and in analysing data processes to ensure compliance. .frankness and fairness with regard to European universities and research the contributions of partners, institutions are increasingly competitors, and predecessors introducing these measures.

.absolute honesty at all stages in 11. The main components of good scientific scientific enquiry, in particular, practice are described in paras. 18-50, avoiding: along with brief accounts of the present – any form of fraud, such as position in selected countries. fabricating or falsifying data or records; Scientific misconduct – piracy or plagiarism; 12. Allegations of scientific misconduct – sabotaging the work, records or and fraud first attracted major public protocols of other scientists; and political attention in the USA, – breach of confidence as a reviewer where there were several well- or supervisor, and publicised cases in the 1980s. Some of – complicity in such actions by these cases led to litigation. Although fellow scientists. a few prominent cases may have To retain professional and public trust, it is attracted disproportionate publicity, it vital that all scientists accept personal was difficult to deny the conclusion responsibility to uphold these fundamental that self-regulation of science, based values. on traditional approaches to instilling values of scientific integrity, was not Good scientific practice sufficiently meeting heightened public and political expectations. 9. Good scientific practice embraces all the procedures and practices that are 13. In response, the US National necessary for planning, conducting and Academies of Science established a reporting research and scholarship Panel on Scientific Responsibility and within a framework of scientific the Conduct of Research to review integrity. By providing a common scientific misconduct. In 1992-3 the currency, good practice facilitates the Panel published a defining two- vital, external processes of peer review, volume report Responsible Science: verification and repeatability. This Ensuring the Integrity of the Research enables other scientists to judge the Process.i Volume II contains guidelines validity of new contributions to for good research practice and for knowledge and understanding. Stan- handling allegations of scientific dard methodologies for collecting and misconduct. interpreting information also reduce the individual bias that might be Ethical and responsible science introduced, perhaps unwittingly, by a 14. The moral dimensions of the sciences scientist’s personal background and and the ethical and social responsibilities values. And the audit trail created by of scientists are themselves the subject good scientific practice provides of academic study and debate. These quality assurance and a valuable topics are developed, for example, in buttress against scientific misconduct two collections of essays published and fraud. under the patronage of the Confederation of Swiss Scientific Academies ii iii and

European Science Foundation Policy Briefing 5 books and articles by authorities such Good scientific practice as Professor John Ziman FRS.iv There is now an international peer-reviewed Introduction journal, founded in 1995, devoted to 18. In the late 1980s, biomedical research ethical issues of direct concern to witnessed some of the first initiatives scientists and engineers.1 These in codifying good scientific practice studies, however, go beyond the scope and establishing procedures for dealing of current review. with misconduct. By mid-2000 a good 15. The growing concern about the ethics deal of progress has been achieved by of science is also reflected in the individual European research organi- creation of high-level fora by the In- sations. Several of the more ternational Council of Scientific significant developments are Unions (ICSU) 2 and by UNESCO,3 by summarised below. the agenda of the joint ICSU/ 19. In the mid-1990s the UK Medical UNESCO World Conference on Research Council (MRC) produced a Science in Budapest in 1999, and more series of reports on the ethics of recently by a working group report on biomedical and clinical research, on the Misuse of Science, presented to the good scientific practice and on 50th Pugwash Conference in Cam- procedures for inquiring into bridge, UK in August 2000. allegations of scientific misconduct. The MRC has recently published up- The present position dated guidelines on good scientific 16. Generally speaking, the global practice that could serve as a general scientific community is responding model for biomedical research.v Other positively, though too slowly in the opi- UK research councils have now nion of some scientists, to the concerns produced their own guidelines of good and expectations of society. Questions scientific practice.vi of scientific integrity, high professional 20. Stimulated by a well-publicised case of standards and public trust are high on scientific misconduct in Germany – scientists’ agendas. Several national, the Herrmann/Brach affair, the European and international bodies 4 Deutsche Forschungsgemeinschaft have taken initiatives designed to raise 1 Science and Engi- (DFG) in Germany appointed an inter- awareness of the ethical dimension of neering Ethics. national Commission, Selbstkontrolle in science, to encourage good scientific 2 The Standing der Wissenschaft, to: practice, and to set procedures for . Committee on explore causes of dishonesty in the Responsibility and investigating allegations of scientific science system Ethics in Science misconduct. All this is helping to (SCRES; created in .discuss preventative measures strengthen the processes of self- 1996). .examine the existing mechanisms of regulation. 3 World Commission professional self-regulation in on the Ethics of 17. However, one senses a lack of science. Scientific Knowledge unanimity in the European scientific and Technology 21. The Commission presented a (created in 1998). community, at both institutional and comprehensive report in late-1997 4 individual levels, on the extent to For example the with an analysis of the issues along Comité d’éthique which new measures are needed and with 16 recommendations covering pour les sciences on the pace of their introduction. (Comets) of the Cen- principles and operation of good There is a clear opportunity for the tre National de Re- scientific practice as well as procedures cherche Scientifique ESF to exercise leadership by co- for investigating allegations of (CNRS) in France; the ordinating further developments in the All European scientific misconduct.vii These European arena. Academies (ALLEA) recommendations would provide an and the European Union in Europe; the excellent basis for developing a set of ICSU and UNESCO common European standards. internationally.

European Science Foundation Policy Briefing 6 22. By mid-2000, 13 German universities Scope of codes of good had finalised their rules on good scientific practice scientific practice, in line with the 27. Though the details will vary to meet Commission’s recommendations, and the requirements of particular discipli- work was in progress in another 17 nes and national circumstances, universities. guidelines for good scientific practice 23. In France, a Working Group on should cover the following key areas, scientific integrity recommended in which are elaborated in the following 1998 that the Institut National de la .paragraphs: Santé et de la Recherche Médicale (IN- designing and conducting research SERM) should actively promote good and scholarship, including laboratory practice in its units, buil- documenting and analysing the data ding on existing legal requirements for .or findings clinical research. The Working Group accumulating, storing or archiving also recommended that the organisa- .data tion should have formal principles and publishing the results of research procedures for dealing with allegations .and scholarship of scientific fraud. protecting intellectual property (IP) .training, development and 24. The INSERM Working Group noted mentoring of young scientists that, to be most effective, good .appointing academics and other laboratory practices and procedures for researchers. investigating allegations of fraud require some harmonisation between 28. Several codes of good practice have the various universities and research been built around a core of legislative organisations in France, especially requirements for health and safety in when research was being undertaken the workplace, the use of human collaboratively. Looking wider, the beings and animals in research, Working Group envisaged that the ESF environmental protection, data protec- might seek to harmonise ethical codes tion and individual privacy. and good practices at a pan-European level. Design and methodologies 29. All research should be designed so that 25. On handling allegations of serious it has a clear objective, either misconduct, the Working Group answering a valid scientific question or, argued that there should be a degree of in scholarship, adding to the independence (from INSERM) understanding of an event, individual, amongst the investigating experts and concept or phenomenon. The design suggested that, longer term, the ESF of the study must be robust, the should create a college of scientists procedures proposed technically who could serve as independent ex- feasible and the intended methods of perts, on a case-by-case basis, in investi- analysis appropriate. gations of scientific fraud allegations anywhere in Europe. 30. Protocols and plans should, therefore, be written in clear and unambiguous 26. In accord with the Working Group terms. They should include specific recommendation, INSERM has now details of the aim, materials, methods, produced a comprehensive code of time schedules and analytical good practice in conducting biomedical approaches to be used. Unambiguous research. It is in an attractive format and fully documented protocols are not suitable for distribution to all its only necessary for those conducting the researchers.viii research, but also for those who may wish to assess or replicate the work at a later date. It is essential that all

European Science Foundation Policy Briefing 7 participants in the research accept surveys – a point that has been 5 At http:// responsibility for these crucial initial acknowledged by data protection www.ird.fr/fr/inst/ steps. commissioners. ird/debat/ en_remarq.shtml 31. Throughout the conduct of research 34. The UK Economic and Social Research all participants must keep clear and Council (ESRC) has produced a draft accurate records on a daily basis of the policy statement and guidelines for the procedures followed and the results social science community designed to obtained. Particular attention should maximise the benefit of social data to be paid to the completeness, integrity the community while protecting the and security of these records. Those interests of the data subjects.x conducting the research should 35. The French Institut de Recherche authenticate their findings by signing pour le Développement (IRD) has the records at the end of each day’s recently initiated a debate 5 on a work. These records must be kept professional code specifically for securely in paper or electronic format. development research. It is seeking to The aim is to provide a continuous and determine whether there are special verifiable record of good scientific factors concerning scientific conduct in practice. co-operation-based research for 32. Research in the humanities and social development, bearing in mind the sciences often involves interactions vulnerability of partner countries in with people. In these circumstances, terms of their political, social and cul- private citizens have a right to be tural characteristics, as well as their protected against unethical human and material resources. interference in their personal lives. The Swedish Council for Research in Data accumulation, handling and the Humanities and Social Sciences storage has published a code of ethical 36. Data are produced at all stages in principlesix, which sets our four key experimental research and in requirements for such research, as scholarship. Data sets are an impor- follow: tant resource, which enable later .to inform individuals about all as- verification of scientific interpretation pects of the proposed research and conclusions. They may also be the .to secure their voluntary agreement starting point for further studies. It is to participate – the principle of vital, therefore, that all primary and ‘informed consent’ secondary data are stored in a secure .to handle and store personal infor- and accessible form. mation under conditions of the highest possible confidentiality 37. Institutions must pay particular atten- .to use such information exclusively tion to documenting and archiving ori- for the purposes of the research. ginal research and scholarship data. Several codes of good practice 33. The balance between protecting the recommend a minimum period of 10 individual and allowing genuine years, longer in the case of especially researchers to access data is, however, a significant or sensitive data. National delicate one. Guidelines should be or regional discipline-based archives sufficiently flexible to allow legitimate should be considered where there are replication and even secondary analy- practical or other problems in storing ses of valuable (and costly) data sets to data at the institution where the address new, and quite possibly research was conducted. unforeseen, research questions. Such an approach actually reduces the need for new data collection and social

European Science Foundation Policy Briefing 8 Publishing the results of research and 41. Many in the scientific community scholarship share their ideas and data freely with 38. Publication in a peer-reviewed journal colleagues as their thinking proceeds - or as a scholarly book is an important through discussion, correspondence or stage in the scientific process, marking at scientific meetings. Any subsequent the point when data, theories, exploitation of information gained interpretations and paradigms formally through these informal contacts, enter the public domain. The right to without the direct involvement or the authorship of publications derives explicit approval of the originator of solely from a creative contribution to the ideas, amounts to infringement of the work in question. In the case of the proprietary rights of the scientist joint authors, each should have made a concerned. significant contribution to the creative or analytical process and each has to Protection of IP accept shared responsibility for the 42. Research workers have a duty to ensure content of the resulting article or book. that intellectual property arising from The practice of honorary, or “ghost”, their work is properly safeguarded. authorships is inconsistent with these This requires them to keep thorough, principles and with good scientific accurate and contemporaneous records practice. of the steps leading to their discovery. It is important they understand that 39. Authorship brings further their records may have to stand up to responsibilities. In particular, authors legal challenge. It is also vital that need to provide accounts of the they avoid public disclosure before pa- materials and methods and of any tent protection is achieved. Laws on analytical and statistical techniques disclosure vary significantly between they used in sufficient detail to enable Europe and the USA. the reader to judge the validity of the approaches adopted and, if so desired, 43. Scientists have a further duty to to replicate the analysis. Authors must ensure, insofar as is possible, that their also be honest and frank in referring to research and scholarship should be earlier work, acknowledging the developed for the benefit of the intellectual contributions of other community. This may involve scientists and declaring any potential assigning or licensing the IP to conflicts of interest. industry or commerce if a product needs to be developed and marketed. 40. Scientific journals, too, have responsibilities. They should make it Training, development and clear in their guidelines that they are mentoring of young scientists committed to best international publishing 44. The training and development of practice. Generally accepted rules have young researchers is an important been drawn up and the majority of responsibility of all those in science. high quality publications adhere to These activities should not be limited them. In particular, reviewers and to providing the technical skills members of editorial boards should be necessary to enable them to conduct required to declare actual or potential their research and become conflicts of interest. Moreover, the independent investigators. Training membership of such boards and the must also inculcate the core ethical names of those who serve as expert standards and norms of science, as well referees should be published on a as principles of best scientific practice. regular basis. Many publishers have also issued clear guidelines for authors. 45. In the past, young scientists have learned these values and norms informally, by working alongside

European Science Foundation Policy Briefing 9 senior scientists and by mentoring. Academic and other scientific Such approaches were supplemented appointments by occasional publications that offered 49. Advances in science are the result of general advice; for instance Sir Peter free, creative thinking by individual Medawar’s book Advice to a Young scientists. When recruiting to Scientist. xi scientific posts, academic and related institutions should put a high 46. With the pressures of today’s world, premium on scientific excellence, greater formality is needed to help creativity and potential as selection young scientists understand the impor- criteria. tance of scientific integrity and to adopt good scientific practices as early 50. Appointment procedures for scientific as possible in their careers. Some positions should be transparent, with universities now routinely provide the selection criteria clearly publicised short courses on these issues for their in advance and adhered to during the graduate students. In 1989 the US Na- selection process. The procedures tional Academy of Science published a should also be socially inclusive, booklet On Being a Scientist: aiming to address deficits of under- Responsible Conduct in Research, represented social groups. Under no which described the ethical foundations circumstances should political or any of scientific practice and some of the other external influence be applied to professional issues and dilemmas that press the appointment of particular scientists might encounter. It was candidates. addressed to junior research workers, and some 200,000 copies were distributed to graduate and undergraduate students.

47. An expanded second edition was published in 1995, jointly by the Na- tional Academy of Sciences, the Institute of Medicine and the National Academy of Engineering.xii Although it is written for American readers, the principles and values it describes are universally valid. The booklet could be of value throughout Europe.

48. On the question of mentorship, the DFG Commission advised that it is good practice for graduate students to be associated with two experienced scientists in addition to their formal supervisor, one of whom should be chosen by the student. This arrange- ment would create a safety valve for mediating in any conflict situations, on scientific practice or other matters, which might arise.

European Science Foundation Policy Briefing 10 implement their codes of good Managing good scientific 6 A good, recent practice scientific practice, including example of an mechanisms for: institutional code is .delegating responsibilities for direc- the Code of Good Introduction Scientific Practice (in tion, supervision, conflict resolution the field of Health 51. Scientists have a moral duty to and quality assurance within their and Life Sciences) maintain the highest standards of management structures, taking into published jointly in integrity without the imposition of account the size of each scientific July 2000 by the external controls and the threat of Universitat Pompeu unit Fabra and the Institut sanctions. Nevertheless, in today’s .maintaining an effective manage- Municipal world the sensitivity of the issues ment audit trail to verify these d’Investigació Mècida, Barcelona. involved in scientific integrity procedures underline the need for the scientific .appointing mediators to whom community to be seen to be regulating scientists can turn in conflict situa- itself. Hence the importance of tions, including cases of suspected scientific institutions having formal scientific misconduct and transparent procedures for .investigating allegations of scientific managing and monitoring their misconduct policies of good scientific practice. .incorporating the principles and rules of good scientific practice into Responsibilities of institutions teaching curricula and the education 52. It is primarily the responsibility of of young scientists and scholars. individual universities and research institutions to develop practical rules Responsibilities of research for good scientific practice for the groups scientists they employ. The need to 56. Institutions should delegate to establish clear and robust institutional individual departments, research policies is a central recommendation of laboratories and groups the most recent reviews of scientific responsibility to adopt good scientific integrity – for example, the DFG Com- practice and to operate within mission. institutional policy frameworks at all 53. Institutional policies for good research times. These operational groups must practice must incorporate and reinforce develop mechanisms, appropriate to any existing civil legislation or codes of their particular discipline and situa- practice concerning, for example, the tion, for ensuring compliance with use of animals in scientific good practice. In particular, there need experimentation, human patients in to be mechanisms for monitoring biomedical research and the use of methodologies, data records and surveys in the social sciences. notebooks and checking the integrity of audit trails. Responsibility for 54. The rules will affect individual compliance monitoring is best assigned scientists and it is important that to an experienced member of each universities and research institutions research group. formulate their rules of good scientific practice in a democratic manner that 57. At a more philosophical level, there is involves their professional members of a responsibility on the heads of staff. Once agreed, these rules should schools, departments and research be widely publicised 6 and made groups and their senior colleagues to binding on all members of an institu- create a climate in their groups or tion, if necessary through terms and units that encourages all to aspire to conditions of employment. the highest professional standards in the conduct of their research and 55. Universities and research institutions scholarship. need to have appropriate management structures and procedures to

European Science Foundation Policy Briefing 11 Responsibilities of individual research and scholarship and to make 7 For example, the scientists the results and data collections Czech Academy of Sciences, the UK 58. available, for example by archiving. Recently, there have been suggestions Medical Research that the intrinsic moral responsibility Some research funding agencies have Council, and the of scientists to work with absolute already gone further. In the USA, as a Swedish Research precondition for accepting research Council for integrity might be reinforced if Engineering students were to make pledges at their grant applications, the National Sciences. graduation, along the lines of the Science Foundation (NSF) and the Na- Hippocratic Oath in the case of tional Institutes of Health (NIH) medical graduates. Some professional require all submitting universities and bodies and institutions already do this other research institutions not only to for their members. The idea would be have in place rules for good scientific to extend the approach to the practice, but also procedures for generality of scientists, irrespective of handling allegations of scientific their discipline, at the time of gradua- misconduct. The DFG Commission tion. The support and co-operation of recommended a similar approach for universities would clearly be essential. Germany.

59. Such a proposal was discussed at the 62. By the same token, funding agencies, Budapest World Conference on Science research councils and foundations have in 1999 and is now being followed up a duty to set an example by the probity in Europe and in the USA. Some of their research appraisal processes. It members of the scientific community, is essential that their operating policies however, strongly oppose the idea as and practices be characterised by impractical. equity, integrity, confidentiality and transparency. Some have published guidelines.7 Leadership by national academies 63. Confidentiality requires that all those 60. National academies are well placed to who assess or administer applications provide leadership in the pursuit of for research funds should not pass scientific integrity and good practice. privileged information to others and They are often the most appropriate should take all necessary steps to independent body to establish and sup- ensure that it is stored securely. They port a national committee for scientific must be required to treat the research ethics and to nominate independent proposals they review confidentially experts to panels investigating cases of and to disclose any conflicts of interest. alleged scientific misconduct. Those This extends to those who contribute to academies that employ scientists have the review process by acting as the added responsibility of external referees. formulating and managing their own 64. Considerations of openness or guidelines and codes of practice. transparency require that the procedures used by research funding The role of research funding agencies should be published, agencies including the criteria that peer 61. Research funding agencies have a reviewers will apply. The names of particular opportunity to demonstrate their advisory committees, as well as leadership in promoting high stan- those who carry managerial and admi- dards of scientific integrity. As a con- nistrative responsibilities, should also dition of their research grants they can be publicly available. oblige institutions and principal investigators to adhere to good scientific practice in the conduct of

European Science Foundation Policy Briefing 12 The contributions of learned should be compared with a view to 8 Section 7.2 and professional societies alignment around common principles, 65. Learned and professional societies in while respecting differences in science have traditionally prepared sensitivities and opinions between guidelines of professional standards for member states. The paper concludes their members, particularly in areas that the process of spreading best with obvious ethical considerations. practice would be enhanced if the The DFG Commission encouraged various national committees included scientific learned societies to be more experts from other European countries. active in this area. 70. Though primarily concerned with scientific ethics, these observations are Contract research equally relevant to the policies and 66. The guidance provided by codes of management practices needed to good scientific practice is equally appli- achieve more uniform standards of cable to contract research funded by good scientific practice across Europe. commercial sponsors, governments or official agencies. Certain tensions do, however, arise from time to time when research projects are carried out under contract. These frequently relate to the ownership and exploitation of intellectual property and to publication arrangements, which should be clearly addressed and agreed before a contract is finalised.

67. Some government and commercial research customers now expect research organisations to have acquired formal accreditation, such as ISO 9000, as a measure of quality assurance, and may restrict their competitive tendering processes to accredited orga- nisations.

A pan-European approach 68. In early 2000 the European Commis- sion adopted a policy paper by the Commissioner for Research, Training and Development entitled Towards a European research area.xiii The paper was designed to stimulate a debate about the need for, and ways of achieving, a better overall framework for research in Europe.

69. On scientific ethics, Towards a European research area 8 argues that there should be stronger links between ethics committees established at natio- nal and European levels. The ethical criteria and rules adopted in national and in European research programmes

European Science Foundation Policy Briefing 13 Investigating allegations of the DFG Commission, should be of scientific misconduct included in all procedures for investigating allegations of misconduct: Introduction .a definition of categories of action 71. Major incidents of scientific that seriously deviate from good dishonesty are uncommon, but they do scientific practice and which are cause considerable concern when they held to constitute scientific do occur. They not only call into ques- misconduct tion the data reported, but also .jurisdiction, rules of procedure undermine public confidence in (including rules for the burden of science and the mutual trust between proof), and time limits for scientists. The Herrmann/Brach preliminary and substantive investi- affair is still reverberating in Germany. gations designed to ascertain the facts And as recently as August 2000, a prin- .the rights of the involved parties to cipal author had to retract a molecular be heard, and rules for the exclusion biology paper already published in of conflicts of interest Science, after peer review, because a co- .the confidentiality of investigations, author had admitted altering gel re- though if there is conflict between xiv cords and other data. the need for confidentiality and the need to seek the truth, the latter Formal procedures must prevail .the range of available sanctions, 72. The primary responsibility for establishing a procedure for which should be related to the investigating allegations of seriousness of any proven malpractice rests with each university .misconduct and research institute where research the jurisdiction for determining is carried out. Preliminary enquiries sanctions. should normally be carried out in that institution. It is also the responsibility Local and national investigations of each institution to ensure that its 75. In the USA, the two main public entire staff are aware of what research-funding agencies, the NSF constitutes misconduct and that its and the NIH, have established perma- investigating procedures are properly nent offices to maintain the integrity publicised. of their science programmes - the NSF 73. Whatever the source of an allegation Office of Inspector General (OIG) and of scientific misconduct, it is essential the NIH Office of Research Integrity to ensure that justice is done, and is (ORI). However, the primary seen to be done, to the complainant responsibility for dealing with and to the accused. Bona fide allegations of scientific misconduct complaints must be pursued with rests with institutions. The federal bo- integrity, in confidence and without dies, OIG and ORI, are there to detriment to the complainant. Equally, provide policy guidance and technical staff who are the subject of such assistance to those institutions and to allegations are entitled to expect that perform a review and oversight their work will be regarded as honest function. Since it was formed in 1992, unless proved to be otherwise, and that ORI has logged more than 1,500 they will be protected against ill- allegations of misconduct in public founded, frivolous, mischievous or health and biomedical research. About malicious allegations. 20% required a formal inquiry. Misconduct has been proved in about 74. With these considerations in mind, the 100 cases – about 6% of the original following general requirements, allegations. largely based on the recommendations

European Science Foundation Policy Briefing 14 76. Apart from in Scandinavia, the Appeals 9 Including the European approach is also to Medical Research investigate allegations locally – by the 80. Natural justice requires that Council (UK) and, in university or research institute where arbitration and appeal arrangements Germany, the Max are available. Responsibility for Planck Gesellschaft the alleged misconduct took place. In and the Herman von most countries this is done without the establishing such a facility might be Helmholtz oversight of national bodies like the undertaken by national funding Gemeinschaft. ORI in the USA. Several research bo- agencies and/or professional bodies. dies 9 have issued regulations for Arrangements for access to arbitration dealing with allegations of scientific and appeal mechanisms must be misconduct in their research institu- available equally to employing tions. authorities, complainants and those who have been the subject of 77. The approach in Scandinavia is the main departure from the prevalent allegations of misconduct. European practice of investigations being undertaken at institutional level. Ombudsman The Danish Medical Research Council founded the Danish Committee on 81. To address the dilemma facing Scientific Dishonesty in 1992, initially scientists who have doubts about the to investigate allegations of fraud in conduct of other, possibly senior, biomedical research. The Committee scientists, the DFG Commission now works under the Danish Research recommended the creation of a natio- Ministry and covers the full range of nal Ombudsman (or a small scientific disciplines. The Danes committee). Its mandate should be to believe that centralising the investiga- advise and assist scientists and scholars tion introduces an important in questions of good scientific practice independent element at the outset and and its impairment through scientific overcomes any inhibitions that dishonesty, and to give an annual pu- universities might have in investigating blic report on its work. one of their own scientists. 82. A mediating person or committee of 78. Other Nordic countries have largely this sort could become a trusted third followed the Danish model of an party to whom scientists would turn independent committee of investiga- with their problems. It would alleviate tion, but they generally allow local ins- the isolation that potential “whistle titutions to conduct preliminary investigations. For example, the blowers” experience and could provide Research Council of Norway wise counsel not available locally. The established a National Committee for Ombudsman could take up matters the Evaluation of Dishonesty in judged to be of serious concern with Health Research in 1994. As well as the university or research institute in investigating cases of alleged scientific question. But he or she would not have dishonesty, the Committee promotes a mandate to investigate alleged measures to prevent dishonesty in misconduct. In addition to its intrinsic health research. To date the Committee benefits, setting up a mediating has investigated nine cases.xv authority along these lines would send a clear message to the public and 79. In early 1999, a Parliamentary Com- politicians that science is taking self- mission in Sweden made wide ranging regulation seriously. recommendations in a report Good Practice in Research designed to 83. The DFG Ombudsman, in fact a three- increase public oversight of the person committee, has been active for a research system, including setting up a year or so and has made its first report national commission to deal with cases to the DFG Senate. of alleged scientific fraud.

European Science Foundation Policy Briefing 15 Bibliography Acknowledgements i Responsible Science: Ensuring the Integrity This report is largely a synthesis of of the Research Process. Report of a Panel on Scientific Responsibility and the Conduct information publicly available in print of Research. (Two volumes.) National and on the WWW and was prepared by Academy Press, Washington D C, USA. 1992 (Volume I) and 1993 (Volume II). Dr. Brian Jamieson of Brian Jamieson & Associates, on behalf of the ESF. It has ii Scientists and their responsibilities. Eds. W Shea & B Sitter-Liver. Watson Publishing benefited from an initial study by Dr. International. Canton, Maine, USA. 1989. David Evered, comments by Dr. Christoph iii The Responsible Scholar. Ethical Schneider (Deutsche Forschungs- considerations in the Humanities and Social Sciences. Eds. G Berthould & B Sitter-Liver. gemeinschaft, Germany) and Dr. Gérard Watson Publishing International. Canton, Toulouse (Ecole Normale Supérieure, Maine, USA. 1996. Paris, France) and by an ESF Reference iv Why must scientists become more ethically Group of representatives from the ESF sensitive than they used to be? J M Ziman FRS. Science, Vol. 282; p. 1813 [4 Dec] Scientific Standing Committees (Dr. Ruth 1998. Barrington, Health Research Board, v Principles of Good Research Practice. Dublin, Ireland; Professor Michael Laver, Medical Research Council (MRC). London. Trinity College, Dublin, Ireland; Dr. 1999. Victor de Lorenzo, CSIC, Madrid, Spain; vi Safeguarding Good Scientific Practice. Biotechnology and Biological Sciences Professor Gretty Mirdal, University of Research Council (BBSRC). Swindon, UK. Copenhagen, Denmark; Dr. Ekkehard 1998. Mochmann, University of Köln, vii Recommendations of the Commission on Germany; and Sir Peter Swinnerton-Dyer, Professional Self-Regulation in Science. Deutsche Forschungsgemeinschaft (DFG). Isaac Newton Institute, Cambridge, UK). Wiley–VCH. 1997. viii Les bonnes pratiques de Laboratoire. Institut national de la santé et de la recherche médicale (INSERM). Paris. 1999. ix Ethics. Swedish Council for Research in the Humanities and Social Sciences. Stockholm. 1990. x Copyright and Confidentiality. A report to European Science Foundation Policy Briefings are the Economic and Social Research Council published by the European Science Foundation (ESF). They by Allen & Overy, London. 1998. address selected science policy issues of key concern to the Foundation’s Member Organisations and the wider scientific xi Advice to a Young Scientist. Sir Peter community. Medawar FRS. Harper & Row, New York, By drawing on the advice and expertise of the ESF’s member- 1979. ship, the briefings aim both to provide information and to xii On Being a Scientist: Responsible promote discussion. Conduct in Research. Second edition. Further information on the ESF’s scientific and science policy National Academy Press, Washington D C, activities is available from the Communication and USA. 1995. Information Unit, xiii Towards a European research area. European Science Foundation, P Busquin. Office for Official Publications of 1 quai Lezay-Marnésia, the European Union, Luxembourg. 2000. 67080 Strasbourg Cedex, France Tel: +33 (0)3 88 76 71 25 xiv Reflections on a Retraction. Donald Ken- Fax: +33 (0)3 88 37 05 32 nedy. Science, Vol. 289; p. 1137 [18 Aug] Email: [email protected] 2000. or from our web site at: http://www.esf.org xv Dishonesty in Health Research. Report of the National Committee’s work 1998. The Research Council of Norway. Oslo. 1998. ISRN ESF-SPB-00-10-FR+ENG © European Science Foundation

European Science Foundation Policy Briefing 16