METHODOLOGY
Editorial Divergence Rate (EDR)
JustSuss automatically collects and analyzes coverage of the same issue across multiple media outlets and quantifies the divergence in editorial direction using AI.
What is JustSuss
JustSuss is an AI-powered information reliability analysis platform that quantifies differences in factuality and editorial intent across news coverage.
Core principle: You judge. JustSuss provides the evidence.
What is EDR
The Editorial Divergence Rate (EDR) quantifies how far apart media outlets are in their coverage of the same issue, using a 3-dimensional independent vector model. The result is bounded between 0 and 1.7.
Three independent vectors
· Credibility deficit — combines reporting rationality and source quality
· Intentional directional bias — strength of editorial intent toward one side
· Emotional manipulation — degree of emotional language beyond factual reporting
These three metrics are normalized and combined into a single EDR value ranging 0–1.7. The exact formula is withheld to protect our proprietary algorithm.
EDR is further adjusted by a weight based on reporting style (factual through sensational, 6 tiers).
Reporting Style Categories
Reporting closer to purely factual receives a lower weight; reporting closer to sensationalism receives a higher weight.
EDR Grade Scale
| Grade | EDR Range | Meaning |
|---|---|---|
| Consensus | 0 ~ 0.3 | Outlets broadly agree in direction |
| Divergence | 0.3 ~ 0.6 | Noticeable differences in framing |
| Conflict | 0.6 ~ 1.0 | Clear opposition in editorial direction |
| Extreme Conflict | 1.0+ | Severe fragmentation in coverage |
Terms
Divergence Score
The Divergence Score measures how far each outlet's coverage deviates from the factual baseline built from all coverage, expressed as a relative distance from 0.0 (theoretical upper bound 1.7). It combines three dimensions: editorial bias, information transparency, and headline-body consistency. Scores closer to 0 indicate coverage faithful to the baseline; larger values indicate reporting skewed toward a particular perspective.
Bias signals
- Selective Emphasis — Highlighting facts favorable to one side
- Fact Omission — Omitting key counterarguments or contextual facts
- Subjective Judgment — Using evaluative or biased words unnecessary for factual reporting
- Labeling — Attaching positive or negative labels to specific subjects
- Framing Bias — Repeatedly presenting only one angle and restricting a multi-faceted story
- Emotional Tone — Rhetoric designed to provoke fear, anger, or sympathy
- Dogmatic Assertion — Presenting uncertain information or possibilities as confirmed facts
- False Causality — Imposing causal relationships not supported by objective evidence
- Overgeneralization — Extrapolating a single case to a broad, universal pattern
- False Dilemma — Reducing complex, nuanced issues to simple binary choices
- Fearmongering — Emphasizing threats and anxiety-inducing elements
- Understatement — Minimizing severity or presenting harmful actions as benign and trivial
Tone map and Stance Alignment
In the coverage tone map, redder means higher divergence and a larger area means more outlets took that tone. Stance Alignment measures whether outlets take the same direction (supportive or critical) toward a specific target such as a person, policy, or organization. It shows direction, while the Divergence Score shows intensity. 'Conflict' means outlets are split.
Data Coverage
Countries
🇰🇷 Korea · 🇺🇸 United States
🇬🇧 UK · 🇯🇵 Japan · 🇹🇼 Taiwan
Collection Cycle
Scout: every 30 min
Analysis: 5 min tick
Korean Outlets
30+ media outlets
Chosun, Hankyoreh, MBC, KBS, Yonhap, etc.
International
AP · Reuters · Guardian
+ Google Trends signals
Disclaimer
All analysis on this service is performed quantitatively by a proprietary AI pipeline designed to exclude subjective human editorial intervention. Results reflect statistical differences in reporting tendencies only and do not constitute a determination of factual accuracy for any specific report.
Source article links are provided on each card. Readers are encouraged to review primary sources before drawing conclusions.
The system undergoes continuous model refinement and data validation; nevertheless, as with any statistical analysis, misclassification or numerical errors may occur. Report an error: contact@justsuss.com