Issues of Research Integrity and Open Science
Issues of Research Integrity and Open Science
Challenges of research integrity and open science focusing on research culture, research procedures, data practices and management, publishing and dissemination.
The European Code of Conduct for Research Integrity (ECoC) is the most important guidance document on the EU level in the research integrity realm. It outlines four fundamental principles of research integrity – reliability, honesty, respect and accountability – and describes good research practices in eight contexts: 1) research environment, 2) training, supervision and mentoring, 3) research procedures, 4) safeguards, 5) data practices and management, 6) collaborative working, 7) publication and dissemination and 8) reviewing, evaluating and editing. Since the transition to open science affects the entire research system, each of the eight contexts deserves closer scrutiny. In the interviews conductedwith stakeholders within the framework of ROSiE, many issues directly and indirectly related to research integrity were addressed, although in general interviewees consider open science mostly, if not entirely, conducive to research integrity because it increases transparency and has the potential to mitigate the reproducibility crisis experienced by several fields of research in recent years.
Changes in the research environment were mentioned as a crucial precondition for a successful transition to open science by most interviewees. Throughout many interviews various references were made to the necessity to establish a research culture that endorses and rewards open science. Interviewees strongly emphasised that incentives to follow open science practices need to be created, for example in research and researcher assessment and funding schemes. This clearly shows that in their view the transition to open science will only succeed if open science is aligned to incentives. Moreover, several interviewees underlined the need to create proper infrastructures for data management, although in general technical aspects of infrastructure development were not named as a major concern because existing infrastructural developments tare perceived to be on the right track. Also, several interviewees anticipate that technological progress and investments in platforms such as the European Open Science Cloud (EOSC) will decrease technological barriers further.
All interviewees consider training in responsible open science desirable. Several explicitly argued that in their view open science should be integrated in trainings in responsible research and good scientific practice. Two interviewees suggested that integrating open science into such trainings also could potentially decrease reluctance among researchers to participate in trainings and endorse open science practices because responsible research and good scientific practice are less contested terms than open science. Furthermore, interviewees largely agreed that open science trainings should be hands-on and practice-oriented rather than theoretical and general.
Major challenges related to mentoring one interviewee brought up are potentially detrimental socialisation effects. In his experience, especially older supervisors who are less aware or more critical towards open science often maintain cultures of closed science in their settings (such as their labs, for example). Thus, their younger mentees are socialised with research practices that hamper the transition to open science.
With regard to research procedures, the most important tensions mentioned in the interviews related to interests in intellectual property protection and the fear of being scooped if research procedures are opened up before studies have been completed and results published. A further challenge related to opening research procedures is that doing so requires significant effort and thus presupposes the availability of sufficient resources. In lab-based disciplines electronic lab notebooks were identified as a potentially helpful by an interviewee who, however, also cautioned that implementing them on a broader scale would be a long-term development rather than something that could easily be established over a short time-period.
As expected, data practices and management were discussed extensively in all interviews, and many discussions centred on the necessity to create research environments that reward good data practices and management. Consequently, in the open science transition research environment and data practices and management are closely related and cannot easily be analysed separately. Essentially all aspects of open science related to open data are inherently linked to data practices. A major specific challenge related to data management some interviewees hinted to and strongly emphasised by a policymaker is that clear guidance is necessary to ensure that research data management becomes an integral component of the entire research processes.
Another crucial aspect related to responsible data management extensively discussed by one interviewee is the question under what conditions access restrictions are justified and how access to data could be managed. He outlined that curating data is costly and that data is highly valuable to, for example, tech companies and insurers. Therefore, he argued restricting access is justifiable if access conditions are clearly specified and transparent. In this way, open science in his view can also mean creating legitimate yet transparent access restrictions that recognise the value of data. By extension, this also implies that open science should mean transparency about why some data is not or cannot be made open.
Implementing open science practices can create challenges in research collaborations, particularly in multi-centre research projects with partners from different countries. As an interviewee elaborated, the extent to which data can actually be opened still differs between countries, even in the EU. Thus, researchers moving from one country to another sometimes are surprised about regulatory heterogeneity, while international consortia often need advice on which infrastructure to use for storing publications and data in a manner both compliant with pertinent regulation and conducive to open science.
The move to open science also has created new challenges when it comes to publishing and disseminating research, albeit seemingly with some notable differences between different disciplines. A first major challenge identified by an interviewee is that following open science practices sometimes is not possible, even if authors of a publication would generally like to make it openly accessible. She illustrated this point by citing an example where she was invited by a publisher to contribute a chapter to a large edited volume. Accepting the invitation was only possible by agreeing to the publisher’s terms on access. Another issue particularly relevant to disciplines where book publications are common is related to the problematic effects creative commons licences, such as CC-BY, can have. Publications licenced under CC-BY, as a policymaker from arts and humanities field explained, can be republished in inadequate formats without the consent of authors or original publishers, as long as the text corpus remains unchanged.
A third issue related to publication and dissemination brought up in the interviews is related to the rise of pre-prints. In general, all interviewees who referred to pre-prints view them by and large favourably, and none of them argued that their negative effects pre-prints outweigh their benefits. Nonetheless, two major challenges related to pre-prints were discussed in the interviews: Firstly, pre-print servers, by facilitating access to research, inadvertently also decrease the barriers to publishing bad research on visible platforms. At least under certain circumstances this becomes a problem because, secondly, not all readers of papers published on pre-print platforms are aware that pre-prints have not been subjected to formal quality control, such as peer review, and thus should be read carefully, especially by non-experts.
Interestingly, an issue generally considered a major challenge in the open science transition was only briefly touched upon in the interviews, but not expounded in greater length by any interviewee, namely high open access fees. While largely omitting discussions of high open access fees might reflect the view shared by several interviewees that Research funding Organizations (RFOs) rather than journals should be regarded as the engines of the open science transition.
Reviewing, evaluating and editing was a major issue in many interviews inasmuch the necessity to develop incentive compatible performance assessment systems that reward open science was discussed. Beyond that, two interviewees with a background in research management mentioned that in their view guidelines and trainings also for reviewers (one interviewee referred to ethics reviewers, the other to grant reviewers) would be desirable to ensure they have the necessary understanding and awareness of open science issues.