Working Papers

Automated Annotation of Political Speech Recordings

Audio and video recordings are increasingly popular for studying the nonverbal dimensions of political speech. However, annotations such as timestamps and speaker tags that accompany these recordings are often inaccurate, disparate, or entirely lacking, which significantly limits the ability to systematically explore the political implications of nonverbal speech. The purpose of this paper is two-fold. First and foremost, I validate the empirical use of political speech recordings by investigating the accuracy of annotations across multiple countries and their sensitivity to the estimation of nonverbal features. Second, I develop a method that can automatically annotate speech recordings with timestamps, speaker tags, the text of speech, and nonverbal features with close to human accuracy, while still being invariant to language and speech setting (e.g. legislative vs. campaign speech). I demonstrate the potential of annotated speech recordings with two substantive applications based on a newly compiled dataset of more than 20,000 hours of aligned text-audio data of parliamentary speech from six European countries.

Partisan Conflict in Nonverbal Communication

With Frederik Hjorth

In multiparty systems parties signal conflict through communication, yet standard approaches to measuring partisan conflict in communication consider only the verbal dimension. We expand the study of partisan conflict to the nonverbal dimension by developing a measure of nonverbal conflict signaling based on variation in speakers’ vocal pitch. We demonstrate our approach using comprehensive audio data from parliamentary debates in Denmark spanning more than two decades. We find that nonverbal conflict signals reflect prevailing patterns of partisan polarization and predict subsequent legislative behavior. Moreover, consistent with a strategic model of behavior, we show that nonverbal conflict signals track the electoral and policy incentives faced by legislators. All results persist when accounting for the verbal content of speech. By documenting a novel dimension of elite communication of partisan conflict and providing evidence for the strategic use of nonverbal signals, our findings deepen our understanding of the nature of elite partisan communication.

Ongoing Projects

  • A Dynamic Theory of Descriptive Representation (w. Helene Helboe Pedersen)

  • Soy un Perdedor: What Happens When Politicians Lose? (w. Roman Senninger and Mathias Wessel Tromborg)