What can we do to create and support an environment where research integrity is at the core of the research endeavor? The Duke Office of Scientific Integrity (DOSI) invited institutional and departmental leaders as well as researchers to answer this question at a research town hall held on September 17th, 2021.
“While it's important to speak up when there are problems or issues, it's also important to think about how to be part of the solution”
The National Academics of Sciences Engineering Medicine (NASEM) Report “Integrity in Scientific Research: Creating an Environment that Promotes Responsible Conduct of Research” concluded that the concept of research integrity embodies “individual’s commitment to intellectual honesty and personal responsibility” while for an institution “is a commitment to creating an environment that promotes responsible conduct by embracing standards of excellence, trustworthiness, and lawfulness and then assessing whether researchers and administrators perceive that an environment with high levels of integrity has been created”. The moderator of our town hall, Dr. Geeta Swamy, Associate VP for Research and Vice Dean for Scientific Integrity, framed the conversation in the light of the NASEM Report. She mentioned that while in the past research integrity and research misconduct were perceived as synonyms, this was an inaccurate definition. “Research integrity is about making sure that we are striving for excellence, can trust our findings and create a system for quality assurance and accountability. And this involves commitments beyond just the highest level of institutional leadership. We need to make sure that we're doing the best we can, follow rules and policies, ethical guidelines, embrace best practices, participate in continuing learning and be accountable on all counts. While it's important to speak up when there are problems or issues, it's also important to think about how to be part of the solution.” Swamy concluded.
Duke is asking for feedback on the revised Data Policy
Assistant VP for Research Dr. John Dolbow, also Professor in the Department of Mechanical Engineering and Materials Science, opened the institutional panel with a presentation on the process of revising the Duke Data Policy, an effort led by the Data Policy Initiative at Duke. “While [Principal Investigators] (PIs) can certainly adopt rigorous procedures for conducting science during the course of a project, an issue that we're increasingly facing regards the challenges that come after the project concludes. Unfortunately, researchers aren't necessarily preparing for these challenges at the beginning of a project because they can't imagine being confronted with them until it’s too late. We are addressing these kinds of challenges here at Duke when it comes to research data and policies related to things like ownership, access retention, which were last revised at Duke in 2007. Obviously quite a bit has changed over the last decade”, Dolbow said. Revising the old policy was “a timing necessity” – he added – “especially now when we're seeing federal funding agencies that begin to roll out new requirements around data management and sharing - the most recent being NIH’s announcement of their new policy that will take effect in January of 2023”.
The research Data initiative was launched by the Office of Research and Innovation in the Spring 2021 with two main components: 1) to engage the research community in determining what Duke policies regarding research data should be; and 2) to understand what processes and resources should be in place in order to support the policy over the long term. The Office of Research and Innovation is currently sharing the draft of the revised data policy to the entire research community which was asked to provide feedback by October 1st, 2021 via a REDCAP survey. The draft will be then revised based on the input from the research community and the final policy is expected to take effect in January 2023.
Duke Office of Scientific Integrity implemented a wealth of programs and resources to enable the environment for conducting quality research
Research Integrity Officer Dr. Donna Kessler presented the main initiatives of DOSI to support a strong research culture. For clarity, we outlined them below:
- The Faculty & Staff Responsible Conduct of Research Program was established with the goal of strengthening our foundation in research integrity and ethics. The program requires all faculty and staff engaged in research to participate in ongoing Responsible Conduct of Research (RCR) courses. The ongoing RCR education requirement for faculty and staff is to complete one online, self-directed course (RCR 100) and one collaborative (RCR 200) course every three years.
- The Research Quality Management Program and Clinical Quality Management Program were launched in the School of Medicine which hosts about 80% of the Duke research enterprise. The Research Quality Management Program will be soon expanded to the entire Duke Campus.
- The LabArchives electronic research notebook is Duke-supported, multi-disciplinary, and designed to help improve research documentation. It is offered at no cost to all Duke researchers.
- To support researchers in their publication efforts, Duke provides free access to all faculty and staff to a writing tool that scans for matching text, iThenticate. The application enables quick upload and scanning of manuscripts and applications against all published documents, and reports matching texts. It is a secure system which does not share user’s scanned document content.
- Duke’s Data Management Plan (DMP) Guidance Document is designed to help researchers develop a data management plan for their research group and outlines specific considerations, best practices, and links to helpful resources.
- The Survey of Organizational Research Climate (SOURCE), developed, validated and conducted by the National Center for Professional and Research Ethics, University of Illinois was distributed at Duke in February/March 2021. Survey results will be disseminated to all departments and units to provide an opportunity for improving research climate.
- More reporting resources were put in place to support a strong research culture while research concerns can be reported as follows:
- Anonymous reporting of concerns: Duke Integrity Line: 1-800-826-8109 and https://values.duke.edu/
- Faculty, postdoc, and student ombuds (confidential advice) - https://oie.duke.edu/ombuds
- Conflicts of interest – email DOSI COI office; firstname.lastname@example.org
- Research Integrity Officer: (919) 668-5115; email@example.com
Data management is at the core of research integrity. Sophia Lafferty-Hess, a data management consultant at the Duke Libraries’ Center for Data and Visualization Services presented about the many services available to all Duke researchers to improve their data management practices, services that complement the workshop series on Data Management that the Libraries are hosting. “We have a team of data experts that provide consulting and instruction on a variety of topics that support data driven research, including data science, mapping data visualization and data management. We provide our services to the entire community with no charge at all stages of the research process and we're available, right now, both for in person and virtual consultations.”
Lafferty-Hess followed up with some considerations about data sharing: “Funders requiring DMPs expect to see that project teams communicate how they plan to share their data to comply with these policies, build public trust in research and expand the data value chain.”
Duke Libraries provide assistance to researchers who want to find out the best options for data sharing for their particular needs. For example, the Duke Research Data Repository was created to provide long term public access to data. “Sharing data via repositories is also a key way to make the data FAIR, the gold standard for data stewardship: Findable, Accessible Interoperable and Reusable”, Lafferty-Hess added, emphasizing that the overarching goal of all these resources is to normalize the culture around data sharing, make the process as easy as possible, and increase the visibility of data that's being shared through Scholars @Duke.
Overcoming research integrity challenges in humanities and social science research
In the increasingly complex research environment, humanists may face ethical challenges that are quite different from order disciplines. From history to literature and art, research and scholarly work are becoming more and more interdisciplinary. “Scholars need to understand the principles from other disciplines and be aware of the potential for bias that may limit their analysis” – we learned from Dr. Victoria Szabo, Research Professor of Visual and Media Studies and Director of the Digital Humanities Initiative at the John Hope Franklin Humanities Institute. “One of the issues we encounter is the black box white coat problem, which is that for some of us who don't have a lot of quantitative background we might be overly credulous of things that we receive out of some new system or software package that we've adopted”.
And there are many other challenges in the humanities world, as Szabo described. One could be how to find the right balance regarding privacy and confidentiality when analyzing the past. Another challenge regards managing risks when using virtual reality and data visualizations. It has become a common difficulty to encourage collaborations and inter-disciplinary communication when most scholars see their work as solitary and independent. Setting up explicit work agreements, criteria for rewarding interdisciplinary work, and standards of assessment of the scholarly work could help overcoming these challenges – Szabo concluded.
Dr. Kate Bundorf, Professor of Health Policy and Management, Sanford School of Public Policy and Margolis Center for Health Policy, talked about the importance that researchers think about research integrity before they actually start their research. And she emphasized the importance of establishing mechanisms that help and motivate them to do the right thing. For Dr. Adriane Lentz-Smith, Associate Professor of History and Associate Chair of History at the Trinity College of Arts & Sciences, a key word for research integrity is transparency which allows “someone to determine if you've used the right kind of data to make your interpretations trustworthy.” A major ethical challenge for historians and for her in particular is writing about “people who are either still living - as with my current book which takes place in the 1980s - or about people who have lived in in the memory of others; you get into a tricky terrain when you're dealing with history and memory. The question is to whom you have obligations and what to do when your obligation to knowledge and your obligation to your subject interact”. While recognizing the multiple individual commitments to research integrity, Lentz-Smith said that “so much comes back to an institution like Duke to make research possible through a well-funded, deeply supported library, and through time to write, think and work. An institution that stresses the value of the deliberative process is crucial to setting the tone that allows individual scholars to figure out where they're going to place themselves.“
Dr. James Moody, Professor in the Department of Sociology, Trinity College of Arts & Sciences and Faculty Director of the Duke Network Analysis Center (DNAC), Social Science Research Institute analyzed the multiple dimensions to be considered when analyzing replicability and reproducibility in research. “Reproducibility is about accuracy, about the ability to get the same result from the same data, using the same procedures”, Moody said. He was interested in finding a solution to enhance reproducibility, but a solution which is not reactive, which solves the problem at heart and is seen as helpful by the researchers. In other words, he looked for a solution which allow all incentives to be aligned, he said. Moody’s solution to reproducibility crisis is called “PROVE - Pre-publication Replication of Validated Evidence”. He drafted a research proposal for testing the effectiveness of this mechanism. PROVE’s essence is checking papers before they go public in order to identify errors, correct them on time and prevent embarrassment. “Most authors are not trying to cheat. They want the work to be right as much as anyone else. Finding problems is therefore a good thing because it saves embarrassment. ”, Moody explained. Therefore, if PROVE would be implemented, someone’s work would be published only after it is checked and the author works with the institution to make sure it is accurate. Of course, there are implementation challenges to be overcome, if this new process is found effective – he admitted. It takes individual and institutional will and courage to overcome barriers and build a culture where research integrity is at its heart and goes beyond compliance.
Enabling the Environment for Conducting Research with Integrity at Duke
September 17, 2021 | 10:00 am to 11:30 am