Workshop Examines Opportunities, Challenges Using Multimodal Learning Technologies
Led by Northwestern’s Marcelo Worsley, “Multimodal Learning Analytics for Interaction Analysis Workshop” examined potential challenges for accessible, multimodal data collection and research
Marcelo Worsley listens as NYU's Xavier Ochoa discusses the need to give data collected back to the students.Marcelo Worsley listens as NYU's Xavier Ochoa discusses the need to give data collected back to the students.A workshop examining how to best develop opportunities to use computational techniques and multimodal data in learning sciences brought together students and researchers from Northwestern University and other institutions.
Held May 2-4 at the Seeley Mudd Building on Northwestern’s Evanston campus, participants in “Multimodal Learning Analytics for Interaction Analysis Workshop” discussed ways to increase awareness, develop best practices, and examine potential challenges to these quickly evolving tools.
“I believe that viewing human interaction on video is a sacred act,” said Shirin Vossoughi, Northwestern assistant professor of learning sciences. “We must use a sense of care with these interactions.”
Giving the data back to the students is important, said Xavier Ochoa, assistant professor of learning analytics at New York University. “I give the data back to you, the student, so you can do something different.”
Cynthia D’Angelo, assistant professor of educational psychology at University of Illinois Urbana-Champaign, is committed to collecting the right data, along with properly handling issues of privacy and consent. She discussed how she found that using headsets to collect data from middle-school students working in groups may be awkward, but that wearing the gear actually served as an important signal to the students that it is time to focus and that the researchers were monitoring them.
“We want the students to be clear about what we are doing,” she said.
Ananda Marin, assistant professor at University of California Los Angeles Graduate School of Education and Information Studies, agreed that researchers must consider privacy. “One of my worries is, when do our activities move into surveillance? We have to think about the ethics of the tools we are creating,” she said. “Who owns the data and where might it go?” She also emphasized the need to consider the subtext and cultural context of the learners.
“Our job is to make sense of the details,” Marin said. “As researchers, we must view ourselves as learners. We are learners of the condition of human possibility.”
The workshop was funded by NSF Award #1832234 - Catalyzing Research in Multimodal Learning Analytics.