Skip to content
AL | Apatheia Labs
B

Bias Detection

Statistical Imbalance Analysis

Performs statistical analysis of directional bias in reporting. Calculates omission ratios, measures framing imbalance, and quantifies selective presentation.

Directional ScoringOmission AnalysisFraming MetricsSelection Bias DetectionPattern Identification

The Problem

Selective citation looks thorough. A report that cites 20 sources appears well-researched — until you check what was left out. When every omission favours the same side, that’s not oversight. It’s a pattern. But proving it requires systematic comparison of what was included against what was available, and statistical testing to determine whether the imbalance could have occurred by chance.

How It Works

  1. 1Compare cited passages against full source documents
  2. 2Calculate pro/con omission ratios
  3. 3Measure space/time allocation between perspectives
  4. 4Identify systematic omission patterns
  5. 5Compute directional bias score (-1.0 to +1.0)

Inputs

  • • Document corpus
  • • Source materials
  • • Citation mapping

Outputs

  • • Bias scores
  • • Omission inventory
  • • Framing analysis report

What You Get

DOCUMENT: Editorial Investigation Report
SOURCE MATERIALS: 12 documents analysed

OMISSION INVENTORY:
  Total omissions detected: 8
  Pro-prosecution: 8 | Pro-defence: 0

DIRECTIONAL BIAS SCORE: +1.0 (All omissions favour prosecution narrative)

STATISTICAL SIGNIFICANCE:
  Binomial test (H0: random omission direction)
  p = 0.004 (significant at p < 0.01)
  Probability of this pattern by chance: 0.4%

FRAMING RATIO:
  Subject-as-suspect: 132 minutes | Subject-as-cleared: 10 minutes
  Ratio: 13.2:1

PATTERN: 100% prosecution-favoring omission pattern across all source materials.

Works With

KContradiction

Feeds SELECTIVE_CITATION type findings into Bias Detection, which then tests whether selective citation follows a systematic directional pattern.

AArgumentation

Provides argument strength data — the engine checks whether weak arguments consistently support one side.

LAccountability

Uses bias scores to establish breaches of impartiality duties, packaging statistical evidence for regulatory complaints.

EEntity Resolution

Ensures omissions are correctly attributed to the right source documents and authors.

Use Cases

Broadcast documentary analysis

Comparing a programme’s content against its source materials to quantify what was included, what was omitted, and whether the omission pattern is statistically significant.

Expert report evaluation

Assessing whether a professional’s report cited evidence selectively, checking whether omitted studies consistently contradict the report’s conclusions.

Regulatory investigation review

Analysing whether an investigating body’s final report reflects a balanced assessment or exhibits systematic directional bias in its source selection.

Technical Approach

  • Source-to-report comparison using document alignment to map every claim back to available source materials, identifying coverage gaps
  • Omission extraction classifies each gap by type (exculpatory, contextual, procedural, temporal, contradicting) and direction
  • Binomial significance testing calculates the probability that the observed directional pattern could arise by chance, providing p-values
  • Framing ratio calculation measures allocation of space, time, or emphasis between perspectives using word count, segment duration, and prominence weighting