← Back to Search
⚖️

Methodology & Definitions

Understanding how Comparipedia analyzes and scores information

🎭 Tone Analysis Metrics

Our tone analysis engine uses natural language processing (NLP) heuristics to evaluate the stylistic properties of the text. We analyze word choice, sentence structure, and grammatical patterns.

Objectivity

Measures how neutral and fact-based the writing is. High scores indicate the text avoids personal opinions, superlatives (e.g., "amazing", "genius"), and editorializing. It looks for a balance between direct reporting and the use of verifiable facts.

Logic: Penalizes opinion markers and "weasel words" (unattributed claims), while giving slight credit for neutral hedging.

Formality

Gauges academic and professional language use. High scores reflect formal, structured prose typical of peer-reviewed encyclopedias. It detects the use of active vs. passive voice and complex sentence structures.

Logic: Rewards formal conjunctions (e.g., "moreover", "consequently") and passive voice, while penalizing informal contractions and casual slang.

Confidence

Assesses how assertive and definitive the claims are. High scores mean the text makes strong declarative statements. Low scores indicate "hedging" where the text uses words like "may", "possibly", or "some argue" to soften claims.

Sentiment

Measures the overall emotional "vibe" of the content on a scale from -1.0 (Negative) to +1.0 (Positive). 0.0 represents a perfectly neutral tone.

Logic: Aggregates scores from a curated dictionary of over 100 positive and negative emotional descriptors.

Bias Level

Detects the presence of loaded language, one-sided framing, and emotional triggers. This metric highlights text that might be trying to influence the reader's opinion rather than just informing them.

📚 Citation Quality Scoring

Comparipedia evaluates every citation to determine its reliability and authority. The overall Citation Quality score (0-100) is a weighted composite of four factors.

Domain Authority

Domains are categorized into four tiers based on their historical reliability and editorial standards:

  • Tier 1 (95pts): Academic journals (.edu), Government agencies (.gov), and International organizations (UN, WHO).
  • Tier 2 (75pts): Major global news agencies (Reuters, AP) and high-prestige reference works.
  • Tier 3 (60pts): General news outlets and reputable niche publications.
  • Tier 4 (30pts): Blogs, opinion sites, and social media platforms.

Source Diversity

Measures the variety of source types used (e.g., combining news with academic and government data). A higher diversity score suggests a more comprehensively researched article that isn't relying on a single type of information source.

Overall Score Calculation

The final Citation Quality score is calculated using the following weights:

  • 40% Average Domain Score: The baseline credibility of the sources.
  • 20% Source Diversity: The breadth of source types.
  • 20% High-Quality Ratio: The percentage of sources from Tiers 1 and 2.
  • 20% Volume: The density of citations relative to article length.