Lecture Capture Transcript Accuracy

This project evaluated the implementation and perceived impact of automated lecture transcripts and closed-captioning within the institution's Synergistic Guided Learning (SGL) curriculum. Rowan-Virtua records all live lectures using the Echo360 Lecture Capture system. Beginning in the 2022-2023 academic year, the team incorporated automated speech recognition (ASR) to generate transcripts and closed captions for those lecture recordings, with the explicit aims of improving accessibility, usability, and usefulness of recorded teaching materials and addressing ADA and Universal Design for Learning (UDL) considerations.

To ensure transcript quality, work-study medical students and volunteers reviewed and edited the ASR-generated transcripts, then released updated versions for student access. The project therefore combined automated technology with human quality control to produce usable, searchable text versions of lecture content that could be consumed alongside or in place of audio/video recordings. Use and perceived impact were assessed through a survey administered via Qualtrics to the student body. The survey results provided a preliminary but encouraging picture: 88% of respondents indicated that adding transcripts and closed captions to lecture recordings is valuable and should continue. Student feedback highlighted practical uses for the transcripts, including clarifying notes taken during live lectures and searching for key terms and concepts within recorded material.

These findings suggest that the availability of transcripts and closed-captioning for medical school lecture recordings can be a valuable resource for facilitating student learning. In particular, transcripts appear to support students' ability to review, verify, and locate specific content within long lecture recordings, enhancing study efficiency and comprehension. The combination of ASR-generated text with human editing helped address accuracy concerns common to automated speech recognition, while making timely transcript releases possible for student use.

As a preliminary study presented at Rowan-Virtua Research Day, the work offers an evidence-based example of integrating educational technology into medical education in ways that align with accessibility legislation and pedagogical frameworks such as ADA and UDL. While the data reported are initial and based on self-reported student perceptions, the strong positive response rate underscores the potential value of extending or refining transcript and captioning services. The project demonstrates a practical implementation pathway, leveraging existing lecture capture infrastructure (Echo360), adding ASR for scalability, and using student editors for quality control, to produce accessible learning resources that students find useful for note clarification and targeted review.

Future directions implied by this work include ongoing monitoring of transcript accuracy and usage patterns, evaluation of impacts on objective learning outcomes, and consideration of sustainable workflows for captioning and editing as use scales. For now, Garwood and colleagues present a replicable model showing that thoughtfully deployed transcript and closed-captioning services can support accessibility and learning in a contemporary medical school curriculum.