National Science Foundation Awards $600k Grant to Robert B. Davis Institute to Expand Community and Capacity of Video Data
Researchers at the Robert B. Davis Institute for Learning (RBDIL) at Rutgers Graduate School Education were awarded $600,000 by the National Science Foundation (NSF) to support the development of a multidisciplinary research community, which will address the challenges of increasing capacity to access video data. The goal of the project is to develop a tool and software for indexing language and non-language objects in video data sets. This study is expected to have significant and global ramifications for both the social science and educational research fields.
“There is currently no cost-effective method for accessing the content data locked within a video,” said Dr. Carolyn Maher, primary investigator of the grant and a distinguished professor of mathematics education at the GSE and director of the RBDIL. She indicated, “By building capacity through this project and research, we hope to increase access to an informative database of mathematics education research videos.”
At Rutgers, Co-PI Dr. Marjory Palius and Researcher Robert Sigley, plan to structure initial development and evaluation of effectiveness around accessing the collection of the RBDIL archives. This video collection comes from 25 years of longitudinal and cross-sectional studies of students learning mathematics. A portion of the RBDIL collection is accessible on the Video Mosaic Collaborative (VMC), an initiative funded by a prior NSF grant. The VMC is an interactive web portal that enables teachers and researchers to use videos of students’ learning to enhance research, teaching, and learning. In the current research, the project team will investigate which categories of videos from the archives will be amenable to the data analytics through processing with an automatic transcription tool.
The project team at Rutgers is working closely with the architect of the tool, Dr. Paul Wang, and his group at the University of Maryland. The team is also collaborating with a distinguished Advisory Board to expand community across disciplines by soliciting their input to stimulate cross- disciplinary dialogue. Advisory Board experts are providing discipline related search terms. The team will then build key vocabularies for searching through the video data, using a training set of videos to test the transcription tool’s accuracy in comparison to video files that have already been transcribed by other researchers.
“We are hoping to build a tool that will allow researchers and teachers to make new discoveries about mathematics instruction and learning styles,” said Dr. Palius, assistant professor of professional practice at the GSE, associate director of the RBDIL, and co-principal investigator on the grant. “This tool has the potential to aid in the coding of other video observations.”
By appyling this research with samples of RBDIL videos, the researchers will determine to what extent it is possible to apply automated transcription and indexing to a set of educational videos and what categories of videos are amenable to such processing for exploring large data sets. Their research has implications for best practice with future video creation, enabling a similar level of analysis on other data sets.
“To edit one hour of video and prepare the accompanying materials such as transcripts, the process ends up taking about 20 to 30 manual hours,” said Robert Sigley, a researcher on the grant. “If the tool achieves a certain accuracy rate, we can use it on the remaining videos that haven’t been analyzed and make thousands more hours of data available to the public.”
To learn more about the Robert B. Davis Institute for Learning and the Video Mosaic Collaborative, please visit their website.