Congratulations to the research team that just completed its report on Supporting Big Data Research at Temple University. Freely available on TU Scholarshare, I think the report will be widely read and have real impact in contributing to the conversation on support for data research at Temple.
Rather than use this space to regurgitate the report’s important findings, I asked the team about the research experience itself. Here is an edited version of that “interview” conducted via email. Contributing comments are from: Will Dean, Fred Rowland, Adam Shambaugh and Gretchen Sneff, as well as Bernadette Mulvey (executive director, Information Technology Services) who assisted in the work.
From your perspective, what was the value of using this semi-structure interview method for the questions you had about support for big data research at Temple?
GS: We asked big data researchers and researchers who used data science methods about their research methods, their training, their data sharing and publishing practices, etc., These questions are relevant to us as a university unit that supports academics and research. The specific questions we asked had been tested and modified, as needed, by the project coordinating team at Ithaka S+R.
Four of us on the study team conducted interviews so using a semi-structured interview method helped ensure that we covered all the topics of interest in each interview. Because it was semi-structured, we could also go into more depth or breadth, as needed.
The transcripts of our conversations with Temple researchers were shared with a central coordinating research team at Ithaka S+R. Because the study teams at all 21 participating institutions used the same semi-structured interview questions, Ithaka S+R was more easily able to work with responses across institutions.
AS: The semi-structured approach allowed me and the other interviewers an opportunity to understand the complexities of data science research. We were able to ask follow-up questions when necessary, and the researchers had the opportunity to reflect on the specific circumstances surrounding their research.
FR: Since I was unfamiliar with this very complex topic, it helped to be able to have a free-ranging conversation to tease out the many, many things I did not understand. It also allowed me to pursue some issues that I found interesting. It allowed me (and us) to get a sense of how the researchers thought about and processed their work.
What did you find challenging about using a qualitative approach in understanding how big data research is supported at Temple?
WD: It was hard to know how much weight to give a participant’s comments, suggestions, and explanations, since we were only interviewing a small number of people and only those who agreed to speak with us. Though we tried to recruit a variety of participants from across disciplines, it was hard to know whose valuable insights we did not hear.
BM: The qualitative approach, to me, left a lot of room for interpretation. It was interesting for me as a technology veteran and how I interpreted the interview transcripts as opposed to the other team members. It was difficult to understand the science that was being called out on top of understanding their use of technology and the resources that they have available in the science community.
In terms of the process of participant recruitment, data collection and management, the analysis and report writing – what aspects of the process were most challenging?
AS: The biggest challenge came from synthesizing the report into a cohesive document that did justice to the varied landscape of data science research at Temple.
GS: Surprisingly, collaborative writing and document sharing was challenging. We were all used to using Google Drive to collaborate on writing documents. We switched to using Microsoft OneDrive at the start of the project. Because document sharing operated differently than expected, a couple of us lost work that had taken us hours to do.
Our study was done during the pandemic and interviews were conducted over Zoom. Connectivity and audio quality problems were sometimes an issue. Recruiting participants who met study criteria—big data researchers or researchers using data science methods—required more time and effort than expected.
What was the most interesting, or surprising thing, that you learned from the project?
FR: Thinking back on that first pass through the 14 interviews, when we understood so little, it is interesting – though not altogether surprising – that we accomplished so much with this report. It shows what a long persistent slog through dense material – step by step sense-making – can reveal.
AS: I’m not sure I found anything particularly surprising, but all of it was interesting! As a liaison librarian, I found this work invaluable in understanding the evolving research needs of faculty on Temple’s campus.
WD: The most interesting thing, to me, was the window into the work of the researchers that our interviews provided. Research products are often compact and cohesive (a paper delving into a single intervention, a clean dataset of a clear variables), and it was enlightening to hear researchers talk about the complexities of their work and the essential labor of many collaborators that make it happen.
GS: The Temple researchers we spoke with, for the most part, expressed a commitment to sharing their research data. Their interest in sharing data went beyond meeting the mandates and requirements of funders and publishers, as many expressed it as an ethical principal or a value they held, a practice of importance not only to them but to the scientific research community.
BM: For me, it was interesting to not only hear about the research itself and the resources that the researchers have and don’t have but I also learned a lot about resources that the Temple University Library offers.
Again, a big thank you to the team and their willingness to share some reflections on the process. Nice work!