Gender and Tone in Recorded Economics Presentations: Audio Analysis with Machine Learning

Abstract

This paper studies gender and tone in economics presentations using a replicable,
scalable, machine-learning approach. We train a deep convolutional neural network to
impute labels for gender, age, and tone-of-voice. We apply this to recorded presentations
from the 2022 NBER Summer Institute to measure tone at a high-frequency
level. We find that female economists are more likely to speak in a positive tone and
less likely to be spoken to in a positive tone, even by other women. We find that male
economists are significantly more likely to sound angry or stern compared to female
economists.

Keywords: machine learning, audio analysis, gender, seminar dynamics, tone

JEL Classification: A1, C8, C45