An Empirical Trace Verification Framework for Government Statement Accountability
Year: 2026
Type: preprint
Abstract: Public officials routinely make factual, statistical, predictive, and causal claims that influence policy, public opinion, and democratic decision-making. Existing fact-checking systems often rely on ad hoc methodologies, expert judgment, and narrative assessments that can be difficult to reproduce or audit independently. This paper proposes the Empirical Trace Verification Framework (ETVF), a formal architecture for transforming public statements into empirically testable propositions linked to transparent evidential chains. The framework combines natural language processing, claim formalization, evidence mapping, statistical verification, causal inference, and reproducibility auditing to evaluate government statements systematically. Rather than producing binary true/false judgments, ETVF generates evidence-weighted assessments accompanied by confidence scores, uncertainty analyses, and complete evidential trace reconstructions. Empirical Validation & Reproducibility Package (ETVF-AUDIT-001) To demonstrate operational feasibility, this deposit includes the complete empirical validation audit (ETVF_Audit_001_Complete_Package.zip) examining UK government statements regarding long-term reductions in violent and neighbourhood crime since 2010 using cryptographically verified primary datasets from the Office for National Statistics (ONS) and regulatory correspondence from the UK Statistics Authority (UKSA). This will be part of the Duty Primacy audit tool kit #dutyprimacy
Source: HAL
Author: Davidson, Craig Kyrle Strachan
Language: English
Subfield: Sociology and Political Science
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7236318198
