“Receiving this fellowship is an honor and a wonderful opportunity to expand my research interests with resources and a three-year plan of action while also working to strengthen research data management (RDM) services and better advocate for researchers at UH,” Boehm said. “My hope is to expand this to the greater RDM community in scholarship, leading to some gradual shifts in service practices.”

Boehm’s research addresses gaps between funder data management requirements for research grant projects and the resources available to academic researchers. Often funders and RDM practitioners approach requirements from the scholarly defined ideals presented in the data science and library and information science disciplines. While this is the ultimate aim, Boehm’s focus is on what researchers are experiencing in reality. The goal is to learn more about these gaps pertaining to how the University and other public Research 1 academic institutions work with researchers. With attention to context in service and training, by learning from research partners instead of simply presenting best practices, there is greater potential to increase advocacy and communicate more clearly to funders about these realities.

“The ECLF provided me with time and space to develop and execute a research study,” Vinson said. “While this was valuable in its own right, it also had a very positive unanticipated outcome – I was able to draw on what I had learned from the study and tap into the Fellowship funding to launch a project to retain student employees in a remote work project during the COVID-19 campus closure. Along with colleagues in Metadata and Digitization Services, we were able to engage over 25 student employees from across the UH Libraries in the creation of hundreds of hours of high-quality video captions.”

Vinson’s research involved the comparison of transcription method accuracy, specifically, how to make audiovisual (AV) archival collections more accessible to deaf and hard of hearing users through closed captions. Positing that online machine-generated transcripts would be the path forward for making collections accessible, Vinson developed a small research study to compare four methods of transcription, including professional transcriptionist services, artificial intelligence machine-generated services, free “hacks,” and untrained study participants drawn from library student and staff employees. In order to create a study that reflected the types of materials that would be encountered in an archival AV collection, Vinson selected video samples from the UH collection with a diverse range of recording issues that could affect transcription quality.