Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

How News Feels: Understanding Affective Bias in Multilingual Headlines for Human-Centered Media Design

Mohd Ruhul Ameen, Akif Islam, Abu Saleh Musa Miah, Ayesha Siddiqua, Jungpil Shin

TL;DR

This work tackles affective bias in multilingual news headlines by performing large‑scale, fine‑grained emotion analysis on Bengali headlines. It employs zero‑shot inference with the Gemma‑3 4B model to classify 27 emotions (GoEmotions‑28 taxonomy) across 300,000 headlines, revealing a robust dominance of negative emotions and notable outlet‑level variations. The study advances beyond coarse sentiment by providing a detailed, emotion‑centric view of framing and proposes a Bias‑Sensitive News Interface that visualizes emotional cues to promote mindful consumption and transparency in journalism. The framework demonstrates practical offline reasoning on consumer hardware, highlights implications for reducing affective overload, and charts a path for integrating emotion‑aware analytics into real‑world news ecosystems.

Abstract

News media often shape the public mood not only by what they report but by how they frame it. The same event can appear calm in one outlet and alarming in another, reflecting subtle emotional bias in reporting. Negative or emotionally charged headlines tend to attract more attention and spread faster, which in turn encourages outlets to frame stories in ways that provoke stronger reactions. This research explores that tendency through large-scale emotion analysis of Bengali news. Using zero-shot inference with Gemma-3 4B, we analyzed 300000 Bengali news headlines and their content to identify the dominant emotion and overall tone of each. The findings reveal a clear dominance of negative emotions, particularly anger, fear, and disappointment, and significant variation in how similar stories are emotionally portrayed across outlets. Based on these insights, we propose design ideas for a human-centered news aggregator that visualizes emotional cues and helps readers recognize hidden affective framing in daily news.

How News Feels: Understanding Affective Bias in Multilingual Headlines for Human-Centered Media Design

TL;DR

This work tackles affective bias in multilingual news headlines by performing large‑scale, fine‑grained emotion analysis on Bengali headlines. It employs zero‑shot inference with the Gemma‑3 4B model to classify 27 emotions (GoEmotions‑28 taxonomy) across 300,000 headlines, revealing a robust dominance of negative emotions and notable outlet‑level variations. The study advances beyond coarse sentiment by providing a detailed, emotion‑centric view of framing and proposes a Bias‑Sensitive News Interface that visualizes emotional cues to promote mindful consumption and transparency in journalism. The framework demonstrates practical offline reasoning on consumer hardware, highlights implications for reducing affective overload, and charts a path for integrating emotion‑aware analytics into real‑world news ecosystems.

Abstract

News media often shape the public mood not only by what they report but by how they frame it. The same event can appear calm in one outlet and alarming in another, reflecting subtle emotional bias in reporting. Negative or emotionally charged headlines tend to attract more attention and spread faster, which in turn encourages outlets to frame stories in ways that provoke stronger reactions. This research explores that tendency through large-scale emotion analysis of Bengali news. Using zero-shot inference with Gemma-3 4B, we analyzed 300000 Bengali news headlines and their content to identify the dominant emotion and overall tone of each. The findings reveal a clear dominance of negative emotions, particularly anger, fear, and disappointment, and significant variation in how similar stories are emotionally portrayed across outlets. Based on these insights, we propose design ideas for a human-centered news aggregator that visualizes emotional cues and helps readers recognize hidden affective framing in daily news.
Paper Structure (15 sections, 9 figures, 4 tables)

This paper contains 15 sections, 9 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Data Collection and Preprocessing
  5. Emotion Classification Framework
  6. Model and Inference Setup
  7. Local Inference Pipeline and Optimization
  8. Results and Discussion
  9. Proposed Design
  10. System Overview
  11. News Feed Analysis View
  12. Bias-Sensitive News Interface
  13. Detailed Emotion Analysis Panel
  14. Design Significance
  15. Conclusion

Figures (9)

  • Figure 1: Sample Dataset of News Headlines and Content (Bangla)
  • Figure 1: Distribution of Dominant Emotions. Negative categories (anger, sadness, disappointment) appear more frequently than positive ones, indicating a prevalence of negative sentiment in multilingual headlines.
  • Figure 2: Top 10 Topics by Frequency in the Dataset
  • Figure 2: Mean Valence Score by Dominant Emotion. Positive emotions such as joy, pride, and gratitude exhibit high valence, while negative emotions such as anger, fear, and sadness correspond to low valence scores, indicating strong affective polarity.
  • Figure 3: Rolling Mean of Arousal Scores Over Index. Arousal levels remain relatively stable with moderate fluctuations, implying consistent emotional intensity across headlines.
  • ...and 4 more figures