When are names named (those of the accuser and accused, and/or other persons involved in the investigation)? If no finding of misconduct is made, can the.

160 KB – 13 Pages

PAGE – 1 ============
Page 1 of 13 ORGANISATION FOR ECONOMIC CO – OPERATION AND DEVELOPMENT GLOBAL SCIENCE FORUM Best Practices for Ensuring Scientific Inte grity and Preventing Misconduct Contents 1. Rationale .. .. .. 1 2. Background .. .. . 2 3. The Varieties of Misconduct, and its Consequenc es .. 2 4. Options for Dealing with Misconduct Allegations .. 5 5. Responding to Misconduct Allegations .. .. 8 6. Investigating Misconduct .. .. 9 7. International Considerations .. .. .. 1 1 8. Causes, Contributing Factors, and Prevention .. . 11 1. Rationale Misconduct in resea rch ( for example, fabrication, falsification, and plagiarism) damages the scientific enterprise, is a misuse of public funds, and undermines the trust of citizens in science and in government. Misconduct is a special concern for governmental administrator s , who are the primary constituency of the OECD Global Science Forum . On behalf of the public, and to achieve societal benefits, they fund, oversee and evaluate research, much of which is conducted directly in public institutions or is otherwise sponsored by governments . At a time when scientific advances are considered to be critical in areas such as economic competitiveness, health, national security, and environmental protection, public officials are strongly motivated indeed obligated to ensure th e highest levels of integrity in research . W idespread attention has recently focussed on a few cases of misconduct in research . Their significance, the damage done, and po tential preventive measures are debate d by scientists, government officials, the p ress, and concerned members of the public. Recognising that the issue affects all of these stakeholder communities and that, like science itself, the problem has a major international dimension, the OECD Global Science Forum sponsor ed an international con sultation of government – designated officials and experts , based on an initiative from the Delegation s of Japan and Canada . On February 22 – 23, 2007, in Tokyo , the Global Science Forum and the Ministry of Education, Culture, Sports, Science and Technology o f Japan (MEXT) held the Workshop on Best Practices for Ensuring Scientific Integrity and Preventing Misconduct . The goal of the OECD Workshop was to d eepen the understanding of the underlying phenomena, to identify the range of possible solutions and, ba sed on experience, to enumerate the pros and cons of various practical measures, lessons learned and good practices. This report summarises the deliberations that took place in Tokyo. Its findings and conclusions pertain to all domains of basic and appli ed science : the physical and life sciences, social and behavioural sciences, and the humanities. Ensuring integrity in science is a complex, multifaceted task, touching upon education, publication, the functions of scientific and academic institutions, and the responsibilities of funding agencies. The present report refers to all of these, but its main focus is on the practical and administrative dimensions of dealing with allegations of misconduct. A number of cou ntries are currently creating, modifying, or reviewing their administrative mechanisms for dealing with such allegations . For these count ries, the Global Science Forum workshop and report should be particularly timely , by providing an opportunit y for international consultation , and for learning from the experiences of others. W orkshop participants address ed the issue of integrity in international collaborations, and they deliberated about possible new steps that might be needed to deal with special problems created by the differences in the ways that collaborating countries deal with allegations of misconduct. This matter will be the subject of follow – on work by the Global Science Forum.

PAGE – 2 ============
Page 2 of 13 2. Background This report is based on the Tokyo workshop , and on information compiled during the preparatio ns for the event . A preliminary version of the report was provided as input to the that was held in Lisbon on September 16 – 19, 2007 1 . The document was subsequently modified, and reviewed by the Chair and Vice – Chai rs of the Global Science Forum. They approved and accepted this final version of the report on behalf of the Forum, and authorised its release to the general public. It is important to state explicitly that this report is of an informative and advisory n ature, without any attempt to instruct governments regarding what they should do in the matter of misconduct in research . The Global Science Forum, and the participants of the Tokyo workshop, have neither the authority nor the inclination to impose any pr escriptive measures on sovereign governments. In addition, it is recognised that there is not an all – embracing, one – size – fits – all global solution , due to the considerable diversity among countries in such variables as the overall structure of the science system, the roles of public and private institutions, the status of researchers (e.g., whether they are public servants), the legal system, and historical traditions and customs. Even so, and within the constrai nts imposed by these legitimate differences, b enefits could be derived from harm onising definitions, rules and procedures, sharing information internationally, and encouraging cooperation among officials and administrators who are responsible for promoting and enforcing inte grity i n research. The To kyo workshop was attended by over 50 government – appointed representatives of 23 countries, 2 invited experts, and the OECD secretariat. It was chaired by Professor Makoto Asashima of Japan and Dr. Nigel Lloyd of Canada. To supervise the workshop preparat ions, fourteen GSF delegations nominated members to the International Steering Committee (ISC). Delegations also designated national experts who were interviewed by the GSF secretariat. A detailed annotated agenda was prepared based on these interviews. The Global Science Forum of the Organisation for Economic Co – operation and Development is a venue for consultations among senior science policy officials of the OECD member and observer countries on matters relating to fundamental scientific research. The recommendations for actions by governments, international organisations, and the scientific community. until 2009 w as endorsed by ministers in February 2004. The Forum serves its member delegations by exploring opportunities for new or enhanced international co – operation in selected scientific areas; by defining international frameworks for national or regional scienc e policy decisions; and by addressing the scientific dimensions of issues of social concern. The Global Science Forum meets twice each year at OECD headquarters in Paris. At these meetings, selected subsidiary activities are reviewed and approved, based o n proposals from national governments. The activities may take the form of studies, working groups, task forces, and workshops. The normal duration of an activity is one or two years, and a public policy – report s are available at www.oecd.org/sti/gsf . 3 . The Varieties of Misconduct, and its Consequences A wide range of (mis)behaviours by scientists can be labelled . C larity and consistency i n defining mis conduct are p rerequisite s to establishing or evaluating an administrative system for processing misconduct allegations, and for understanding the underlying causes and effective remedies. A variety of administrative m echanisms and modalities (including pr evention and investigation/enforcement) may be needed to deal correctly with the diversity o f inappropriate behaviours. In particular, it is important to identify instances of misconduct that can be remediated via education, or that merit a full investiga ti on , including procedures for establishing innocence or guilt . 1 The report of the conference rapporteur, Dr. Peter Tindemans, can be found at www.esf.org/conferences/researchintegrity

PAGE – 3 ============
Page 3 of 13 During the course of preparing the OECD workshop, i nterviews with experts revealed a broad spectrum of ty pes of misconduct by scientists, as shown in the following table 2 : sconduct F abrication of data F alsification of data P lagiarism FFP normally includes: Selectively excluding data from analysis Misinterpreting data to obtain desired results (including inappropriate use of statistical methods) Doctoring images in pub lications Producing false data or results under pressure from a sponsor Research p ractice misconduct Using inappropriate (e.g., harmful or dangerous ) research methods Poor research design Experimental, analytical, computational errors Violation of human s ubject protocols Abuse of laboratory animals Data – related misconduct Not preserving primary data Bad data management, storage Withholding data from the scientific community NB: The above applies to physical research materials as well Publication – relate d m isconduct Claiming undeserved authorship Denying authorship to contributors Artificially proliferating publications – Failure to correct the publication record Personal m isconduct Inappropriate personal behaviour, harassment Inadequate leadership, mentoring, counselling of students Insensitivity to social or cultural norms Financial, and other m isconduct Peer review abuse e.g., non – disclosure of conflict of interest, unfairly holding up a Misrepresenting credentials or publication record Misuse of research funds for unauthorised purchases or for personal gain Making an unsubstantiated or malicious misconduct allegation At the core of the spectrum of inappropriate behaviours F abrication, Falsification and Plagiarism (FFP ). V arious definitions of these terms are possible. For example, t he United States government defines research misconduct in a way that has been adopted in some other countries 3 : Research misconduct is defined as fabrication, falsification, or plagiarism in proposing, performing, or reviewing research, or in reporting research results. Fabrication is making up results and recording or reporting them. 2 The chosen classification scheme is not intended to be exhaustive , or to constitute a un iver sally valid intellectual framework for theoretical studies of research misconduct. In this report, it is presented merely as a way to summarise the information distilled from the expert interviews . 3 The definition given above is not unique. It can, for example, be b roade ned to include . A lternative broad formulation s can be used, such as text has been adopted by the Committee on Publication Ethics.

PAGE – 4 ============
Page 4 of 13 Falsification is manipulating research, materials, equipment , or processes, or changing or omitting data or results such that the research is not accurately represented in the research record. without giving appropriate credit, including those obtained throu gh confidential review of other s research proposals and manuscripts. Research misconduct does not include honest error or honest differences of opinions. There is general agreement that credible allegations of research misco nduct (FFP) should be investigated , and that corrective actions should be undertaken if the investigation makes a positive finding. The same applies to financial misconduct, and harassment , which are in the province of accounting and administrative depart ments in research institutions and appropriate government agencies . At the other end of the spectrum are such phenomena as inadequate mentoring of students , or incompetence in performing research . For these, the internal mechanisms of the scientific comm unity can , in most cases, provide effective remedies, without the need for formal investigative actions . But there are also intermediate categories of misconduct where science administrations may need to be involved. The establishment of an op timal mappi ng between the offenc e and the method/venue for dealin g with it is difficult. It is c omplicated by the importance of determining whether an inappropriate action was deliberate, i.e., of establishing intent . This is notoriously difficult to do in an y inve stigation. F or example, if the validity of published results is questioned, the subsequent ina bility of the researcher to provide primary data m ay be the result of a genuine mistake or accident, but it could also be considered as prima facie proof of seri ous misconduct if it can be established that it was done deliberately to conceal an act of FFP. M isconduct in research damages science , but its consequences also extend into the broader societal sphere . The following general areas where negative impact oc curs were identified in the run – up to, and during, the OECD workshop : Harm to individuals and to society, if fraudulent research results in the release of an unsafe product or process (e.g., a drug or a therapy). Society may be harmed if false results bec ome widely known and believed . T he formal responsibility for protecting the public lies mostly outside the research administration system, and is assured by a well – developed structure of national laws, regulations, and institutions (e.g., the drug approva l process). Even so, research administrations must assume responsibility for not burdening the regulatory process. Direct damage to science itself, by creating false leads for other scientists to follow, and/or forcing others to waste time , effort and mon ey to reproduce fraudulent results. Fortunately, t he research record is inherently self – correcting , since re peatability, verifiability and consistency are hallmarks of the s cientific m ethod . H owever, incorrect results can persist and mislead for extende d periods of time. T he degradation of relations among scientists, between senior researchers and students, and between researchers and agency programme managers . D ability to foster and promote research in a competent and responsible manner. A possible consequence is a decline in the credibility of scientific analysis and advice on issues that have important implications for society. These issues (in such are as as health, environment, energy, national security) often have a major scientific component, and science – based laws and regulations may be needed to address them.

PAGE – 5 ============
Page 5 of 13 Conclusion A : Instances of misconduct in research are regrettable, but real, o c currences within the scientific enterprise. Scientists, like all professionals, are subject to pressures and temptations, and they are no more n o r less likely than others to behave badly. The prevalence of misconduct is difficult to measure but, when it occurs, th e damage to science, and to the way it is perceived and utilised , can be severe. Misconduct by scientists can take many forms, each affecting differently the stakeholders, such as researchers, institutions, government agencies, publishers, and the public. A well – designed strategy for promoting integrity should take this diversity into account by identifying the most appropriate methods and venues for dealing with each category of misconduct. As in other instances where society confronts individual wrongdo ing , a n optimal response contains elements of both prevention and enforcement. However, it is always better to prevent bad behaviour than to be forced to deal with its consequences. Accordingly, an optima l strategy consists of actively promoting integrit y and deterring misconduct within all of the components of the scientific enterprise: universities and other research institutions, funding agencies, professional organizations (unions, academies, etc.), the publishing establishment, and in for a w here scie ntists and the public interact. 4 . Options for Dealing with Research Misconduct Allegations Sections 4 – 7 of this report concern the practical, administrative aspects of dealing with misconduct allegations. The focus is on underlying principles and acti onable procedures. It is worth repeating that no attempt is being made to devise a universal prescription for a system that all governments should put in place. Rather, the enumeration and analysis of the selected topics are meant to constitute a kind of of issues that should be given consider ation when creating or fine – tuning a system of national or local principles, rules, and procedures. Given the inter – governmental status of the OECD Global Science Forum, this material is presented primari ly for consideration by resp onsible public officials. Their role is a special one, and it is sometimes underappreciated. Th ere is a body of opinion claiming that all matters pertaining to integrity should be handled exclusively within the scientific comm unity, and in the context of the corresponding institutional frameworks (for example , academic departments at univ ersities). However , g overnment officials have certain responsibilities that t h ey cannot delegate: They are formally accountable for the prope r spending of public funds. In particular, they manage the granting process (including reviews of applications and monitoring of progress) which cannot function properly if it becomes compromised by dishonesty. As described f urther in Section 8 , the gran ting procedures that agencies establish may have an effect on the p revalence of certain forms of misconduct (i.e., they can have a corrupting effect on susceptible individuals). They are responsible for public safety , which can be compromised by the conseq uences of misconduct in research . They fund ( and are otherwise involved with) the education and training of researchers – activities that are vital for promoting integrity and preventing misconduct . On a practical level, t hey are sometimes the only agents who have the means to conduct especially complex or difficult investigations , or ones that transcend national borders .

PAGE – 6 ============
Page 6 of 13 Based on the information gathered during the preparations for th e Tokyo workshop, and at the workshop itself, it appears that dealing wi th misconduct in research is a shared responsibility of public officials, scientists and institutional administrators. The division of roles differs from country to country but, in general, t hree generic ways of handling misconduct cases can be identified : a) A d – hoc committees establishe d to deal with specific cases. Such committees are often composed of prestigious individuals, possibly under the aegis of existing university – based ethics committees. The advantage of this approach is that ethics committees already exist at many institutions, although t hey are often associated chiefly with the life/medical sciences, and handle matters relating to human experimental subjects and patients. While th e work of these bodies is vital, it ca nnot be assumed that the y can handle al l case s of misconduct in research . Ethical issues (i.e., questions of right and wrong and fairness ) underlie the very concept of misconduct , but the practical adjudication of concrete cases revolves more around the determination of facts and the careful analysis of events, documents and other data records . This, in turn, can be difficult to do without specialised expertise, as well as special – purpose rules, regulations and precedents. All ad – hoc processes suffer, to some extent, from a defi cit of consistency, since the functioning of each individual committee depends critically on its makeup, i.e. , on the preferences, opinions and experiences of its members. For misconduct investigations, fairness and consistency are critical attributes, wh ich can be difficult to ensure in an ad – hoc process. E xtending the mandate of ethics committees to handling cases of misconduct in research should be accompanied by careful analysis and, if nee d ed, modifications of existing rules and procedures. b) Standing committees in research institutions. Some countries rely on standing entities (offices, officers, committees) and corresponding procedures , at the level of the institution (e.g., university, large laboratory) where the misconduct occurs. These can be res ponsible for receiving allegations, processing them (including conducting investigations), and recommending outcomes. Typically, these entities are not entirely autonomous : there is a measure of interaction with a government – mandated central national auth ority , for example, a funding agency . A system of this kind generally benefits from good acceptance by scientists, who prefer to put their trust in local arrangements that operate under terms and conditions that they can observe and understand. Acceptanc e by the community is a vital attribute of any misconduct – processing system . Scientists are understandably protective of their reputations and careers, which can be seriously damaged by allegations , or even mere rumours, of misconduct. A purely local ar rangement ensures consistency of procedures throughout a given organisation (for instance, a large university) but does not necessarily do so at the national level a source of potential problems when allegations of misconduct involve more than one instit ution. T he cost , workload and administrative overhead s of maintaining standing bodies must also be considered . Furthermore, there may be an inherent conflict of interest that could lead to the unjustified suppression of cases, based on a desire to avoid unfavourable publicity for the local institution. c) One or more dedicated committee (s) at the n ational level . This variant may be preferred by countries whose scientific communities are small, and where it may be difficult to establish committee s of imparti al scient ists , free of personal conflicts of interest . Members of permanent national committees can be selected so as to represent a wide spectrum of relevant expertise (for ex ample, detailed legal experience ), drawing on extensive human resources. A nat ional committee c an establish a consistent track record of cases, and there are benefits from having a stable support staff, consistent long – term relations with funding agencies , and independence from the vagaries of changing national governments . A commi ttee of this kind can play a major role in reviewing and fine – tuning its own procedures, in advising the government on misconduct – related policies , in maintain ing a permanent record of misconduct – related information, and in coordinating with similar commit tees in other countries.

PAGE – 8 ============
Page 8 of 13 Conclusion B : There is no universal optima l system for dealing with misconduct in research . A dministrations are free to design and implement the system that meet s their needs and is consistent with the way research is managed in a give n country or institution, and is compatible with local laws and traditions. Nonetheless, interested administrations are encouraged to implement such a system and to publicise it in the relevant communities . Wide – ranging consultation s with principal stake h olders are desirable : with researchers , funding agencies , scientific publishers, representa tives of the public. Section s 4 – 7 of this report can 5 . Responding to Misconduct Allegations Misconduct allegations most often arise in a spontaneous, unsolicited way ; for example, when a graduate student ( or other collaborator of the accused scientist ) suspects that data have been fabricated. A researcher in the same domai n may become suspicious when unable to reproduce a measurement , o r evidence may emerge from a computerised search for plagiarised text , or evidence may be discovered by potential employers during verification of claims made in a CV . All too often , the po tential accuser has no idea where to turn with the suspicions, and such uncertainty can be a powerful disincentive to taking action . T hose seeking to create, review, or modify a system for dealing with misconduct would benefit from seeking answers to the following questions regarding the all – import a Who is the first person/organisation to turn to with an allegation or suspicion? Is there a special office/officer located near the same venue as the person who susp ects misconduct? If so, does the person receiving the allegation have special expertise or training? Is the receiving office/officer someone whose elevated standing (e.g., dean of an academic faculty, high – level official of a science ministry) could dis courage a student or other person who is in the lower ranks of the hierarchy? Does the allegation have to be presented to a person who has authority over the accuser (for example, a departmental chairperson vis – à – vis a graduate student) ? Is adequate infor mation available to the potential accuser? Is there generally accessible information on a web site, for instance, or via an anonymous hotline? Is there someone to consult when merely a suspicion exists, without certainty or definitive evidence? Are there requirements/restrictions on who can be accused (and be an accuser)? Can anyone come academic field, work not published in a peer – – type work)? Does suspect work need to be published, versus presented in a conference, or mentioned in a conversation? Are anonymous allegations accepted ? for misconduct allegations? How does the syst em deal with frivolous or malicious accusations? Does bringing forward a false accusation itself constitute actionable misconduct? decide the merits of the allegation?

PAGE – 9 ============
Page 9 of 13 Conclusion C : Well – intentioned persons who have a legitimate suspicion that misconduct may have occurred should have access to local information and assistance . Recognising that suspicions of misconduct place both accused and accuser in vul nerable positions, the first administrative response should be characterised by sensitivity, confidentiality, objectivity, and fairness. Persons receiving a suspicion or allegation should have the appropriate competence, training and mandate (including li nks to higher – level authorities, should they be needed). If possible, these persons should have the authority to resolve conflicts that do not merit a full investigative proceeding. 6 . Investigating Misconduct The rules and procedures for mis conduct inve stigation s should explicitly address the following issues and questions : For an ad – hoc local body, are there guidelines regarding such matters as: o The number of members and their affiliation (from inside/outside the institution where misconduct is alleged) o The areas of expertise that committee members need (including professional/ judicial/procedural) o Avoidance of conflict of interest (and how conflict of interest is defined), including potential bias by local – level committees towards protecting the reputati on of a home institution. How and under what authority does the investigatory body obtain the cooperation of the parties, especially those who are not themselves accused? Can they compel collaborators to provide data or testimony? What if they need coope ration from outside the institution? Does lack of cooperation with an investigation itself constitute misconduct? In cases involving collaboration among scientists from two or more institutions, questions of jurisdiction naturally arise, since parallel du plicative investigations are undesirable. One possible practical solution is to assign the task to the institution providing the greater share of the research funds. Can an investigation be enlarged as new evidence is discovered? Can it be extended to ot her institutions? If so, should additional committees be established? Is there a time limit on investigations? What happens if the accused resigns, stops the work, etc.? What happens if regulatory or criminal violations are uncovered? Does the miscond uct investigation continue in parallel with other process es that may be triggered ? Are there limits on the power/authority of the investigators? How much new work can they require (for example, repeating an entire series of experiments)? What is the sourc e of funds for conducting an investigation ? Do funding agencies provide any support? Questions of fairness are particularly important when dealing with misconduct, because the investigation process is a quasi – legal one; that is, it has many of the attribu tes of criminal or civil procedures, but is reduced in complexity and is meant to function more quickly . Moreover, the Precise definitions, policies and procedu res for misconduct investigations are needed to prevent the perception (or , worse, the are contrived to persecute an individual, based on personal conflicts, or the unpopularity of a particular line o f

PAGE – 10 ============
Page 10 of 13 research. At the OECD workshop, attention was called to the undesirability of labelling as pursuit of research that is merely outside of the mainstream scientific consensus a hazard that could be linked to overly – broad definitions of misconduct. Accordingly, when constructing inves t igative procedures, answers can be sought to the fo llowing questions: o What are the conditions and rules of confidentiality for accuser and accused? Can be given anonymity and protected fro m retaliation, without generating spurious/frivolous allegations? o What is the standard of proof in a misconduct investigation (e.g., preponderance of evidence , proof beyond a reasonable doubt)? Is there a presumption of innocence? How can the validity of the proceedings be ensured, given that the investigators may be prominent the hostility of colleagues? In cases where intentional misconduct is hard to distingui sh from unintentional carelessness in carrying out research , how do the investigators establish intent? o How can the accused defend him/herself? Does he/she have access to documents, testimony? Can the accused confront accusers and witnesses? Can the acc used be assisted by a lawyer during the proceedings? Does the accused have a right to question the composition of the investigating body ? Can one set of allegations give rise to more than one investigation In general, how do the rig hts of the accused compare to those in a criminal or civil proceeding? o What are the rights of appeal and review (by accuser or accused) at each step of the investigation? o Who is notified of the progress of the investigation, and when? How much detail is p rovided (e.g., to the funding agency)? Can the agency provide feedback, suggestions, information? Can it play a more active role during the investigation? o What are the conditions of access by journalists and the public to the outcomes and records of inve stigations ? When are names named (those of the accuser and accused, and/or other persons involved in the investigation)? If no finding of misconduct is made, c an the exonerated scientist require that a formal exoneration be published? How do requests fo r – of – information – type laws? Is it feasible to Concluding a misconduct investigation o Can disciplinary measures begin dur ing the investigation (e.g., suspension of the research, withholding of a grant) ? o Is there a shadings possible? Is there a reasonable and consistently applied relationship between the seriousness of the misconduct and the severity of the imposed punishment? Does the investigating body just make findings, or can it also recommend corrective actions (including the punishment of guilty individuals, retraction of tainted public ations, and other measures to protect science and the public interest) ? Can action be taken with regard to persons who should have exercised better supervision, even if they ha ve not actively committed misconduct? o What specific steps can be taken to resto re a damaged reputation , and to restore a project that may have been delayed or disrupted during an investigation ? o Is there any provision for protecti ng projects may be terminated even if their work ha d nothing to do with the misconduct committed by the principal investigator?

160 KB – 13 Pages