NHMRC Public Consultations

Skip Navigation and go to Content
Visit NHMRC website

Review of the Australian Code for the Responsible Conduct of Research Submission

ID: 
51
Personal Details
This submission reflects the views of
Organisation Name: 
The Australian National University
Specific comments
Specific comments: 
Principles in Code

Principles

  1. The principles do not currently explicitly fully reflect all of the principles underpinning the National Statement on the Ethical Conduct of Human Research (merit and integrity, justice, beneficence and respect). Currently, integrity and respect are explicitly covered (P1 and P5, respectively), but the fairness principle is currently written in the broad context of “others” with exemplars basically constrained to fellow researchers (either explicitly or implicitly) – “others” should be broadened to include research participants and further to more clearly relate to the broad principle of justice.

  2. P2 pre-supposes that research is monolithic in the sense of having well-defined endpoints and outcomes. For example, sometimes research is hypothesis generating rather than definitive in terms of an explicit outcome, and such research needs to be allowed in the context of requiring “rigour”. Perhaps this is where the broader concept of “merit” could be employed as this description can marry concepts of worth and quality. The current wording of P2 is too narrow – “rigour in conducting high-quality research” begs the question of whether rigour is required within all research?

  3. P3: “Transparency” might interfere with commercial arrangements in place within research. There needs to be a provision within the principles to allow commercialisation of research without violating P3. For example, to accommodate delay in publication to allow commercial funding arrangements.

  4. P7 :”Researchers will consider the consequences and outcomes of research prior to its communication.”

What is the intention of this?  Is the expectation that a researcher is expected to consider how his research might be used by others and potential self-censure? We think this statement has no place in the code and should be removed or if the intention is otherwise it should be reworded or clarified.

This point would imply that institutions should have the ability to suppress the outcomes of research that they or other interest groups such as politicians might not agree with. This principle may have the consequence of stifling academic debate and freedom of ideas.

Responsibilities for Institutions in Code

Institutional responsibilities

    1. R3 should explicitly include research trainees.

    2. R9: It may not be possible to “resolve” all breaches of the Code – for example, where a breach results in a human death, there may be no reasonable resolution. Perhaps a better wording would be “investigate and resolve (where possible)”.

    3. R10: How does promoting detection of potential breaches of the Code relate to current whistle-blower legislation (which differs between State/Territory and Federal jurisdictions)?

Responsibilities for Researchers in Code

Researcher responsibilities:

The document should make clear that it is wholly and solely focussed on research misconduct and not general disciplinary matters. There needs to be recognition that research misconduct issues can arise from other issues.

 

  1. R14, R15 – these principles should apply broadly, not only to human research. Would it not be more straightforward to simply make acquiring relevant ethics approval (human, animal, rDNA) a responsibility of researchers?

  2. R19: What does it mean to disseminate research findings “broadly”, and why is this a responsibility? Some research findings are, by nature, narrow and technical and broad dissemination is neither practical nor reasonable. We would suggest deleting “and broadly”

  3. R20 - confidentiality should be a separate point

  4. R21-  insert “effectively”

  5. R24 Please amend to “Where appropriate monitor their conduct”.

  6. R25 This should be two separate points ensuring the obligation on the researcher to undertake training is clear

  7. R28: We feel that the document needs to be clear on Whistle Blower Legislation and the Public Interest Disclosure Act.

Definitions in Code

Definitions:

  1. “Breach” = “Behaviour”? We suggest it should be based on an action rather than behaviour.

  2.  “Researcher”: the examples are exclusively about universities. Suggest “Person who conducts research” with no examples.

Guide Section 3
  1. Preliminary assessment stage / minor breaches – this appears to operate as a committal process to make a determination as to whether to go forward to a full hearing but it’s also the process for determining outcomes of “minor breaches” (3.2, page 4) but without any proposal of how this might work or be determined. Also a lack of indication of what a “minor breach” is and who might make a decision etc.

     

  2. Box 1: “research and scholarly community” – is this meant to be interpreted broadly, or in a disciplinary context? If the latter (or both), it would be useful to perhaps include explicit reference to disciplinary norms (which can vary very widely between disciplines). The word “behaviour” in the last dot point might better be expressed as “breach”: maybe “whether the breach was inadvertent or intentional”.

     

We suggest that the section on mitigating circumstances has a separate table for clarity, it does not sit comfortably under Figure 1. Under mitigation it should also include whether it was done under direct supervision or not. There is also the possibility of a situation where a researcher has followed the ethics protocol but this has led to a breach elsewhere. This could be the result of an Ethics Committee mistake or a Legislative or Policy requirement.

 

  1. Examples of breaches (major - minor) Figure 1. Box 1.

    1. “Distortion of visual presentation of data” is SO subjective and open to argument. For example, using defaults in Excel OFTEN leads to distortions inadvertently!

    2. A couple of the items under the “major breach” column are perhaps misplaced. This is very confusing. For example, “Excellent support from institution” seems an incongruous entry under “major breach” (the extent described is really not about the gap between a minor or major breach but rather the extent to which an institution might be jointly culpable). Similarly, retractions of a paper might not be exclusively related to “major breaches”. For example, if an error is discovered in a paper, the paper might be retracted, but this may not reflect a breach at all but part of the famous “self-correcting” feature of academic inquiry. While it is appreciated that these are only examples, such tables tend to become templates for how people think, so it is important that they don’t overreach.

    3. A clear question that arises is how to deal with simple mistakes. Simple mistakes should never be regarded as a breach of the Code. Inadvertent mistakes in the current table would be captured as breaches which, we feel, should not be the intention. Perhaps a statement needs to be made covering this important (and common) circumstance. For example, a clerical error is not research misconduct, neither is an incorrect date in a lab book unless it was done with malicious intent or to deceive.

    4. The section on composition of the panel is constraining as minor breaches may be dealt with early by the management structure of the institution or the designated person and some investigations may not merit a panel. The proposal to convene a panel for even minor breaches adds complexity, delay and cost to any process.  Nor do we think to reference a large panel is useful. For example our present panel structure is 3 individuals who are at arm’s length from the institution, this, we feel, is more important than the size of the panel. The panel is also to make findings of fact and recommendations. Yet it is not a decision making body. If panels are used they should make findings of fact, not make recommendations or decisions.

Guide Section 4
  1. Figure 2

  1. This is unduly simplified and needs to reference a step for admittance of misconduct, for example.

  2. It would be useful to have inserted into the table that complaints are often unclear and need clarification before investigation.

  3. Confidentiality of how the complaint is dealt with needs to be clearer; in section 5 page 1 for example “a complaint should not be shared unless required” by whom? We suggest amend to “a complaint should not be shared beyond those with specific need to know in the context of the investigation”

Guide Section 5
  1. Information to the complainant – it is proposed that an essential part of the process is providing the complainant with a running commentary on what is happening – this could result in a breach of privacy, and potentially creates a sense of private prosecution, which creates a (further) conflict between complainant and respondent and the institution, setting up expectations that the complainant will be able to influence decisions and outcomes. It ignores the fundamentally private nature of the disciplinary process. The whole process needs to align with the relevant Privacy legislation and Whistleblower legislation.

Guide Section 7
  1. Section 7.4.3: “The institution is responsible for ensuring the public record is corrected”. Suggest insert “where possible and appropriate” into the above sentence.

Specific consultation questions
Question 1: Do you like the new approach to the Code, namely the principles-based document being supported by several guides that provide advice on implementation?: 
Yes, we like the new approach and, in particular the principles-based document. We do have some concerns regarding how the Guide might be positioned. It is extremely important that the Guides are only providing advice and context, whilst the instruments by which institutions will implement the Code takes into account institutional policies and procedures, Enterprise Agreements and Student regulations. It is necessary for any disciplinary consequences that these are done through the approved institutional processes. The preamble to the Guide needs to be much more explicit regarding the advisory nature of the document with cross-references made to individual institutional agreements. Otherwise it would indicate that a University (or another employer) could run the whole research misconduct procedure and once a finding of a 'breach' has occurred, then run a whole other procedure pursuant to the relevant employment contract or enterprise agreement. In addition, to the above it is also important to recognise that the institutions also function in a legal environment which is a mix of Commonwealth and State/Territory Legislation which requires referencing in the Guide as well.
Question 2:The draft Code is intended to be used by all research disciplines. Do the principles adequately capture the expectations for responsible research across all research disciplines?: 
It is almost impossible to capture expectations across all disciplines. Authorship and publication practices, for example, vary very widely across disciplines. Acknowledging disciplinary differences is extremely important and should be noted in the text. Reference to AIATSIS’s Guidelines for Ethical Research in Australian Indigenous Studies should be included and possibly to the UNDRIP (United Nations Declaration on the Rights of Indigenous Peoples). We strongly recommend that “will be consulted and involved in decisions about research” be replaced with ‘”Involved in decisions about research according to the principle of free prior and informed consent.” Researchers should be aware of the need to promote Australian Indigenous values (either local or nation-wide) about language, land (homelands, outstations, acknowledging the connection to country, etc.), traditions, culture and cultural heritage (protect, preserve, promote) and of the need for cultural training. The above would also relate to R15. The principles also need to make reference to the importance of informed consent and confidentiality when undertaking research with human subjects.
Question 3: The draft Guide refers to breaches of the Code rather than providing a definition of research misconduct, and states that institutions can decide whether or not to use the term research misconduct in their own processes.: 
The word “incident” might be a better alternative, with the word “breach” being used only when a preliminary investigation suggests a case to answer. It seems inappropriate to regard someone making a mistake in a publication as a “breach” if that mistake was completely inadvertent, and particularly so when the mistake is later discovered and corrected by the researcher without any involvement from the institution at all. This is where it is exceptionally important that the Guide is just that. It cannot be considered to be the model which everyone must follow as that is dependent on State/Territory legislation and institutional policies and procedures. On a more specific point the definition of breaches of the Code do not refer to the treatment of participants, including potentially vulnerable participants. The only reference to the National Statement is the need for ethics approval - whereas the emphasis should be on ethical practice.
Question 4: Do you think the process described for investigating and managing potential breaches of the Code is clearly described and practical?: 
This comes to the same point as made in the answers above. Each institution will need to implement the principles through their own policies and procedures. The code for investigation is a Guide only and as stated it has no legal status. It would appear that the Guide assumes it possible to run a whole investigation, make not only factual findings, but findings that misconduct (namely a Code breach) has occurred, but somehow be outside and able to be dis-connected from employment disciplinary processes which then need to be put in train. In R11 in the Draft Code (and at 7.3 in the Draft Guide) we are concerned with the use of the term “Balance of Probabilities” which has a specific meaning in law and could lead to such interpretations being applied. We recommend that the definition of “balance of probabilities” is amended to no longer refer to whether something is ‘more probable than not’; this may lead to a mathematical calculation of whether a breach occurred that was warned against in Briginshaw v Briginshaw (1938) 60 CLR 336. The definition should recognise that the strength of the evidence necessary to be satisfied of a matter on the balance of probabilities may vary according to the nature or gravity of the fact to be proved. The Civil standard of proof as outlined in Bringinshaw v Briginshaw is described as ‘The court should consider the seriousness of the allegation, the inherent unlikelihood of the occurrence, and the gravity of the consequences flowing from a particular finding, before deciding whether the issues has been proved to the court’s reasonable satisfaction’ (Laws of Australia [16.2.360]. This has been incorporated in legislation, including the Evidence Act 2011 (ACT) section 140. We recommend that R11 of the code is amended so that findings are to be made “to the satisfaction of the decision maker(s) on the balance of probabilities’”. This clarifies that the decision maker must be satisfied that a breach has occurred. The proposed document also says, p. 2: “This Guide should be used to inform the application of existing institutional processes”. This is at odds with the statement on p.1 “This Guide is not a legally binding document. Its purpose is to assist institutions to fulfil their responsibility under the Code to investigate and manage concerns or complaints about potential breaches of the Code.” The role of the Head of School/Dean also needs to be recognized in the process, as they should deal with minor issues as they arise and will often be involved in any subsequent actions.
Question 5: The Code Review Committee and working group are considering what additional resources should be developed to support implementation of the Code and Guide.: 
Yes. A great example of this kind of guidance in practice is COPE (Committee on Publishing Ethics) http://publicationethics.org/ - see the Cases, in particular. The flowcharts they offer are also extremely instructive. Case studies may be helpful but should be brief and their aim clearly stated. Appropriate case studies could be particularly useful for training purposes for example when it comes to dealing with an Indigenous scholar.
Question 6: Are the mechanisms for review of an investigation clearly and correctly described in Section 7.6 of the Guide? If not, where are the inaccuracies?: 
Section 7.6 is very brief and simply states that processes for appeal must exist and be appropriately documented in outward-facing guidance. There is a restriction on appeal to matters of procedural fairness. Noting the potential for appeal to higher authorities (where, presumably, no such restriction on the basis for appeal exists), it may be worth considering whether it is unreasonable to limit appeals to only procedural fairness grounds. Perhaps not well considered here is that the NHMRC has no power to regulate employment contracts that might overtake the provisions/jurisdiction of the Fair Work Act. While the proposed Guide says that it will not be legally binding, it also states that if there is a complaint about an investigation/breach there is an implication that the ARIC will use the Guide as the benchmark for investigation processes: The Australian Research Integrity Committee (ARIC) provides a review system of institutional processes responding to allegations of breaches of the Code for institutions that are in receipt of funding from the National Health and Medical Research Council (NHMRC) or the Australian Research Council (ARC). Following a review, ARIC advises the Chief Executive Officer of NHMRC or ARC of the outcome of the review and recommendations for further action (7.6, pg 19) In our view the ARIC may review institutional processes but must judge these against the Principles and not the advice given in the Guide. The ‘existence’ of two separate processes, one through institutional procedures and one through the Guide could result in adverse actions. There is an ability for the respondent to respond in writing to the allegation and have access to the evidence (see 7.3), but there is no specific review once the panel has determined whether the respondent has breached the code. We recommend a right of response before the decision is final.
Question 7: Please comment on which three topics you would nominate as being the highest priority and why.: 
Providing the status of the Guides is resolved effectively as noted above then Guides on Authorship and Data Management are the highest priorities (as the Committee has deemed) as these are issues that are perennially at the heart of potential breaches. These would be one and two in terms of ordered priority. A guide on Supervision of research trainees would be welcome – this is a difficult space for institutions since supervision is so broadly distributed across the institution and occurs at so many levels. Also, the separation of the responsibilities of the supervisor and the person supervised is a critical area that needs exploration and explanation.
General comments
Comments: 

As noted elsewhere in our submission a primary issue which is worth underlining is the status of the Guide and ensuring this is providing advice to be taken into consideration by institutional processes and that any review of insitiutional processes is judged against the principles and not against the detailed information in the guide. We are happy to amplify further if required.

Page reviewed: 17 September, 2018