Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

The Face of Populism: Examining Differences in Facial Emotional Expressions of Political Leaders Using Machine Learning

Sara Major, Aleksandar Tomašević

TL;DR

This study investigates whether political leaders with populist rhetoric differ in facial emotional expressions from non-populist leaders by analyzing 203 processed YouTube videos across 15 countries using a deep-learning emotion detector. It leverages the Global Party Survey populism index to classify leaders and computes per-video mean scores for seven emotions (six core emotions plus neutral), focusing on negative emotions and neutrality. The findings show populist leaders express more negative emotions and display less neutral expressions on average, with moderate effect sizes and robust differences in subgroup comparisons. The work contributes an open-source computational workflow, a time-series emotion dataset, and a cautionary discussion of limits in sampling, labeling, and cross-cultural emotion interpretation, outlining directions for more nuanced future analyses of visual political communication.

Abstract

Populist rhetoric employed on online media is characterized as deeply impassioned and often imbued with strong emotions. The aim of this paper is to empirically investigate the differences in affective nonverbal communication of political leaders. We use a deep-learning approach to process a sample of 220 YouTube videos of political leaders from 15 different countries, analyze their facial expressions of emotion and then examine differences in average emotion scores representing the relative presence of 6 emotional states (anger, disgust, fear, happiness, sadness, and surprise) and a neutral expression for each frame of the YouTube video. Based on a sample of manually coded images, we find that this deep-learning approach has 53-60\% agreement with human labels. We observe statistically significant differences in the average score of negative emotions between groups of leaders with varying degrees of populist rhetoric.

The Face of Populism: Examining Differences in Facial Emotional Expressions of Political Leaders Using Machine Learning

TL;DR

This study investigates whether political leaders with populist rhetoric differ in facial emotional expressions from non-populist leaders by analyzing 203 processed YouTube videos across 15 countries using a deep-learning emotion detector. It leverages the Global Party Survey populism index to classify leaders and computes per-video mean scores for seven emotions (six core emotions plus neutral), focusing on negative emotions and neutrality. The findings show populist leaders express more negative emotions and display less neutral expressions on average, with moderate effect sizes and robust differences in subgroup comparisons. The work contributes an open-source computational workflow, a time-series emotion dataset, and a cautionary discussion of limits in sampling, labeling, and cross-cultural emotion interpretation, outlining directions for more nuanced future analyses of visual political communication.

Abstract

Populist rhetoric employed on online media is characterized as deeply impassioned and often imbued with strong emotions. The aim of this paper is to empirically investigate the differences in affective nonverbal communication of political leaders. We use a deep-learning approach to process a sample of 220 YouTube videos of political leaders from 15 different countries, analyze their facial expressions of emotion and then examine differences in average emotion scores representing the relative presence of 6 emotional states (anger, disgust, fear, happiness, sadness, and surprise) and a neutral expression for each frame of the YouTube video. Based on a sample of manually coded images, we find that this deep-learning approach has 53-60\% agreement with human labels. We observe statistically significant differences in the average score of negative emotions between groups of leaders with varying degrees of populist rhetoric.
Paper Structure (11 sections, 4 figures, 1 table)

This paper contains 11 sections, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Populism as performance
  3. Data & Method
  4. Party selection
  5. Video selection
  6. Video processing and analysis
  7. Data processing pipeline
  8. Results
  9. Discussion
  10. Data availability
  11. Acknowledgements

Figures (4)

  • Figure 1: Example frames showing a high detected score of negative emotions (left: angry 0.78, disgust 0.00, fear 0.09, happy 0.00, sad 0.07, surprise 0.00, neutral 0.04) and a high detected score of neutral expression (right: angry 0.07, disgust 0.00, fear 0.04, happy 0.00, sad 0.02, surprise 0.08, neutral 0.79).
  • Figure 2: Distribution of average scores of negative emotions (left) and neutral expression (right) of videos featuring populist and pluralist leaders.
  • Figure 3: Raincloud plot showing the distribution of average scores of negative emotions across four degrees of populism. Every data point is presented as a colored dot and above them boxplots and density plots are displayed. The dashed vertical bar indicates the score of 0.5.
  • Figure 4: Raincloud plot showing the distribution of average scores of neutral expression across four degrees of populism. Every data point is presented as a colored dot and above them boxplots and density plots are displayed. The dashed vertical bar indicates the score of 0.5.