Researchers report an algorithm that uses sound to identify teaching practices in college classrooms. Previous studies have found that classes with active learning, in which students learn through talking and problem-solving, result in higher learning gains and student retention than lecture-only classes, prompting efforts to shift science, technology, engineering, and math (STEM) college courses from lecture-based teaching to active learning. Kimberly Tanner and colleagues designed Decibel Analysis for Research in Teaching (DART), a machine-learning algorithm that rapidly analyzes classroom audio recordings, to quantify the frequency of different teaching practices in a classroom. For 1,486 recordings from 67 college courses at community colleges and universities, DART distinguished the amount of classroom time spent with no voices or thinking/writing time, one voice or lecture/question-and-answer time, and multiple voices or discussion time. DART identified teaching styles with an approximately 90% accuracy rate and could be applied to both small and large classes. The amount of time spent on active learning activities was higher for courses for STEM majors than courses for non-STEM majors, and 88% of analyzed courses used active learning in at least half of the class sessions. DART could aid systematic analyses of the use of active learning in classrooms, according to the authors. - Read at PNAS
Article #16-18693: “Classroom sound can be used to classify teaching practices in college science courses,” by Melinda T. Owens et al.