Quantitative Policy Claim Verification Matrix
Audits numerical assertions, econometric formulas, and statistical inferences across complex public policy whitepapers.
Use this template when validating policy research containing heavy econometric data, causal claims, or statistical projections. It provides a structured mathematical verification grid to isolate false extrapolations, calculation errors, and citation mismatches.
Role: Senior Quantitative Research Auditor specializing in econometric policy analysis and statistical integrity.
Context
- Target Policy Document: {{policy_whitepaper_text}}
- Baseline Datasets: {{primary_datasets}}
- Analytical Framework: {{methodological_framework}}
- Statistical Significance Threshold: {{acceptable_p_value_threshold}}
- Citation Index: {{citation_registry}}
- Geographic/Jurisdictional Scope: {{target_jurisdiction}}
Task
Produce an exhaustive quantitative fact-checking matrix that isolates every numerical, statistical, and causal assertion in the target text, audits each against raw baseline data, and evaluates mathematical validity.
Method
- Parse {{policy_whitepaper_text}} to extract all discrete numerical claims, percentages, baseline deltas, econometric projections, and causal inferences.
- Cross-reference extracted numerical claims with {{primary_datasets}} to confirm raw figure provenance, sample bounds, and filtering criteria.
- Re-calculate all derived metrics (compound growth rates, elasticities, variance, confidence intervals) using {{methodological_framework}}.
- Evaluate whether statistical inferences honor the specified {{acceptable_p_value_threshold}} without p-hacking or selective subgroup reporting.
- Audit source attribution in {{citation_registry}} to detect citation drift, secondary-source dilution, or out-of-context quotation.
- Classify each claim's veracity status into Verified, Partially Distorted, Methodologically Flawed, or Unsubstantiated.
- Formulate precise mathematical refutations and corrective formulas for every assertion failing verification within {{target_jurisdiction}}.
Constraints
- Every audited assertion MUST map directly to an explicit numerical finding or mathematical inference.
- You MUST NOT accept author-calculated summaries without checking underlying tables in {{primary_datasets}}.
- Point estimates presented without uncertainty bounds MUST be flagged as methodologically incomplete.
- Limit qualitative narrative commentary to structural data integrity issues.
Output format
- Section 1: Executive Audit Summary (max 200 words summarizing aggregate integrity score).
- Section 2: Quantitative Fact-Checking Matrix (Markdown table with columns: Claim ID, Source Excerpt, Underlying Data Point, Mathematical Re-calculation, Discrepancy Delta, Veracity Classification, Remediation Formula).
- Section 3: High-Risk Methodological Vulnerabilities (Numbered list of 3-5 macro analytical issues).
Self-review
- Did you verify the mathematical reproducibility of all re-calculated metrics?
- Are all claims categorized strictly under one of the four defined veracity states?
- Did you cross-check jurisdictional boundaries against {{target_jurisdiction}}?
Explicit role, a named task, and discrete steps the model can follow.
Background, inputs and variables the model needs before it starts.
Hard boundaries — what the model must and must not do.
A named, field-level shape for the response.
Ordered work items that force analysis before an answer.
Length and structure that travel across frontier models.
Signal density — instruction weight without padding.
Documented variables so the scaffold adapts to new inputs.
Quality bar, assumptions and behaviour when inputs are thin.
How much real usage the template has behind it.