On September 29, Interim VC/VP for Research Claudia Neuhauser and Santi Thompson of UH Libraries held an interactive webinar on data management best practices for UH researchers. They solicited engagement and feedback from participants concerning their current data management pain points.
What is Research Data?
The presenters kicked off the workshop by asking “what is research data?” Research data is the recorded factual material necessary to validate research findings— excluding preliminary analysis, communication with peers, laboratory samples and drafts of scholarly work.
From the Library of Alexandria to your portable hard drive, lost data equals trouble and expense. The presenters stumped participants with another question: “What percentage of hard drives fail within the first four years of storage?”
The answer: 20%.
Data Management Plans
There are many reasons to implement a documented data management plan. Many federal agencies now require them, they mitigate error and data loss, and they can help ensure compliance with institutional policies as well.
The seminar centered around the planning, execution and sunsetting of data management actions. Reproducibility, replicability and research misconduct were areas of focus. “Teach your students about this,” emphasized Neuhauser. The final chapter of every thesis should outline the details of how someone else can replicate the experiments from your data.
Conditions for Research Misconduct
“Environment” was listed as having one of the strongest effects on ethical decision making in research. Running labs under either an executive model or a competition model can create a toxic workplace and strain research integrity. Under the executive model, a PI demands that their students and researchers reach a specific outcome, while the competition model rewards whoever completes the research first. Both of these models make labs liable for research misconduct.
Almost 2% of researchers have confessed to some sort of research misconduct, and Neuhauser reminded the large audience of UH professionals that misconduct does in fact occur, it can cause data to be sequestered, and it has the potential to ruin careers. She also mentioned that with the advent of ChatGPT, more and more cases of plagiarism and fake references are being recorded.
Post-Project Considerations
Thompson discussed the post-project work of archiving and preservation in contrast to basic data storage. He delved into digital object identifiers and cite-ability as well as how to choose the best repository for your data.
Where are your funders expecting you to preserve your data? Where are your peers depositing and what are the parameters of the repository? All pertinent questions when deciding how to store your important work. He advertised the UH Dataverse Repository, a no cost open-access repository that is part of UH Libraries’ Cougar ROAR (Research Open Access Repositories) platform, which provides a digital object identifier and up to 10GB of memory per project.
Lastly, Thompson explained how metadata describes and outlines the way data is preserved in the same way the label on a jar of peanut butter describes the contents of its container.
Research Security
Other highlights included a discussion of the new thrust of research security across the United States. UH has made research security a central theme of its educational programming this year. New webpages have been published, and Neuhauser directed the participants to visit them soon and often for upcoming information regarding new security hires and the federal guidelines surrounding foreign collaboration, foreign travel and other important actions.