Game, Set, and Conflict: Evaluating Conflict and Game Frames in Indian Election News Coverage

Tejasvi Chebrolu, Rohan Chowdary, N. Harsha Vardhan, Ponnurangam Kumaraguru, Ashwin Rajadesingan

Abstract

This paper investigates the prevalence and drivers of "conflict framing" in Indian election news coverage. By analyzing nearly 70,000 headlines from seven major outlets during the 2014 and 2019 general elections, the study finds that media modality—specifically television versus print—is a primary driver of adversarial reporting, outweighing ideological bias. Furthermore, a comparison of campaign speeches to news reports reveals a significant "substantive filter," where political attacks are prioritized over policy issues like agriculture and infrastructure.

Introduction

In a healthy democracy, news media is tasked with informing citizens about policy and governance. However, journalists often adopt a "conflict frame," emphasizing disagreements, personal attacks, and the "horse race" aspect of elections. While this can increase audience engagement, it often leads to political cynicism and a lack of substantive knowledge among voters. This study addresses a gap in framing research, which has largely focused on Western contexts, by examining the unique media landscape of India, the world’s largest democracy.

Methodology

The researchers employed a multi-stage computational pipeline to analyze 69,400 headlines from seven prominent English-language outlets. The dataset included both print and television-based media, categorized by their perceived ideological leanings.

Table 1: Characteristics of the news sources analyzed in the study, including article counts, ideological bias, and modality.

The analysis involved:

  1. Conflict Detection: A DistilBERT classifier was fine-tuned to identify conflict frames in headlines (macro F1-score: 0.87).
  2. Role and Issue Extraction: GPT-4 Turbo and dictionary-based approaches were used to identify "Aggressors," "Targets," and 19 key electoral issues.
  3. Ground Truth Comparison: To assess media over-emphasis, 224 campaign speeches were matched to 607 news articles reporting on them, comparing what was said to what was reported.

Results: The Impact of Modality and Bias

A key finding of the study is that an outlet's ideological bias (Left-leaning vs. Right-leaning) does not significantly correlate with its use of conflict frames. Instead, modality is the primary driver: TV-based outlets are 2.03 times more likely to frame news as a conflict than print-based ones. This suggests that the structural demands of television—often requiring high-engagement, dramatic narratives—influence framing more than political alignment.

Additionally, the study found that national parties (BJP and INC) are featured in conflict-framed headlines at a rate over three times higher than regional parties. This indicates that the "national narrative" is perceived as inherently more adversarial by the media. Interestingly, most outlets showed "Aggressor Neutrality," not consistently portraying one specific coalition (NDA or UPA) as the aggressor more than the other.

Results: The Substantive Filter and Reporting Gap

The study highlights a stark disconnect between what politicians say and what the media reports. While campaign speeches frequently address substantive issues, the media acts as a "substantive filter." For instance, while "Opposition" attacks were mentioned in 88% of reports, substantive issues like farmer welfare appeared frequently in speeches but were reported in only about 30% of corresponding articles.

Figure 1: Scatter plot showing the disconnect between issues mentioned in speeches and their reporting in news articles.

Quantitatively, approximately 19.3% of all election headlines used a conflict frame. The reporting gap is particularly concerning: while 31.96% of speech-related articles used conflict frames, only 1.1% of those headlines actually mentioned a substantive policy issue.

Conclusion

The prevalence of conflict framing in Indian election news poses a challenge to informed democratic participation. The findings suggest that the medium of delivery and the scale of the political party (national vs. regional) are more significant factors in adversarial reporting than ideological bias. Future research should expand this analysis to regional language publications and incorporate multimedia elements to provide a more comprehensive view of the Indian media landscape.

Summary updated

BibTeX
@inproceedings{chebrolu2025game,
  author = {Tejasvi Chebrolu and Rohan Chowdary and N. Harsha Vardhan and Ponnurangam Kumaraguru and Ashwin Rajadesingan},
  title = {{Game, Set, and Conflict: Evaluating Conflict and Game Frames in Indian Election News Coverage}},
  booktitle = {WebSci},
  year = {2025},
  doi = {10.1145/3717867.3717900},
}
Figure