The Emotional Engine of Audience Activation: Emotive Rhetoric and Audience Reactions in Presidential Campaign Rallies

Emotional appeals are central to campaign communication, shaping political behavior and voter preferences and often contributing to the affective polarization that increasingly characterizes contemporary partisan conflict. But how do leaders leverage emotions when appealing to the public, and how do these appeals resonate and mobilize support? Research on emotional rhetoric in politics mostly relies on controlled settings and self-reported measures, missing how audiences organically respond to emotional appeals in real-world contexts. To address this gap, we explore the resonance of emotive rhetoric in a naturalistic setting by examining organic audience reactions during presidential campaign rallies, where activating the audience is central to successful events. Our analysis draws on an original dataset of 260 rally speeches from the 2016 and 2024 U.S. presidential campaigns, comprising over 165,000 sentence-level units. Working with both text and audio, we quantify rhetorical appeals by valence and apply machine learning to classify audible audience responses (applause, cheers, chants, booing). We treat these responses as behavioral expressions of collective emotions and fit regression models to identify which emotional appeals are associated with stronger responses, and how strategic intertwining of positive and negative appeals predicts distinct engagement patterns. This study will contribute a novel analytical perspective to emotional mobilization in political campaign discourse, provide empirical insights into affective crowd dynamics, and advance the literature on the experience and expression of emotions in a collective.