Classroom sound can be used to classify teaching practices in college science courses

Area of Science:

• Educational Technology
• STEM Education Research

Background:

• Active learning pedagogies yield superior student learning gains compared to traditional lectures.
• Understanding the adoption of active learning by STEM faculty is crucial for student retention and effective education.
• Current methods for assessing teaching practices are often resource-intensive and lack scalability.

Purpose of the Study:

• To develop and validate a machine learning algorithm, Decibel Analysis for Research in Teaching (DART), for analyzing classroom audio recordings.
• To quantify the extent of active learning strategies employed in higher education STEM courses.
• To compare teaching method utilization across different course types.

Main Methods:

• Development of a machine learning algorithm (DART) to analyze audio recordings based on volume and variance.
• DART predicts time spent on single voice (lecture), multiple voice (discussion), and no voice (thinking) activities.
• Application of DART to 1,486 audio recordings (1,720 hours) from 67 STEM courses.

Main Results:

• Varied patterns of lecture and non-lecture (active learning) activities were observed across analyzed courses.
• Courses for STEM majors showed significantly more use of multiple and no voice strategies than courses for non-STEM majors.
• DART achieved approximately 90% accuracy in identifying active learning presence.

Conclusions:

• DART offers a scalable, cost-effective, and accurate method for inventorying active learning across numerous educational settings.
• The tool can systematically assess and compare teaching strategies, facilitating broader adoption of evidence-based pedagogies.
• Findings highlight differences in active learning implementation between major-specific and general education STEM courses.