---
name: evidence-evaluation
description: "Evaluate whether a claim deserves confidence before repetition, fluency, vivid examples or preference are mistaken for evidence. Use for research, proposals, fact checks and consequential AI-generated claims."
---

# Evidence Evaluation

Evaluate whether a claim deserves confidence before repetition, fluency, vivid examples or preference are mistaken for evidence. Use for research, proposals, fact checks and consequential AI-generated claims.

## When to use

- A team already prefers one explanation and is mainly gathering support for it.
- A claim appears in many places but may trace back to one source.
- A vivid example is dominating broader evidence or base rates.
- An AI answer sounds convincing but important claims remain unverified.

## Workflow

1. Write the exact claim and the decision it would change.
2. Separate direct observations, source claims, inferences and assumptions.
3. Trace repeated claims back to genuinely independent sources when source access is available.
4. Write what evidence would weaken the preferred conclusion before collecting more supporting material.
5. Compare vivid examples with a broader sample, base rate or reference class when relevant.
6. Classify the result conservatively: supported for this use, limited, disputed, unverified or unknown. Explain why.

## Required output

- Claim under review
- Evidence that supports it
- Evidence or tests that could weaken it
- Source independence / provenance notes
- Current confidence and limitations
- Decision implication

## Evidence and safety boundaries

- Do not treat source count as independent confirmation when sources repeat one origin.
- Do not use another generated answer as independent verification of an AI-generated claim.
- Do not upgrade confidence because wording is fluent, familiar or repeated.
- Keep unknowns visible rather than filling them with plausible prose.
## Evidence-linked lenses

Use these as candidate lenses, not diagnoses:

- [Confirmation Bias](https://cognitive-biases.github.io/biases/cognitive-bias-confirmation-bias/) — Confirmation bias is an umbrella label for several ways existing beliefs or hypotheses can influence information search and interpretation. It should not be reduced to one behaviour such as reading only agreeable news, and a preference for confirming tests is not irrational in every task or environment.
- [Illusory Truth Effect](https://cognitive-biases.github.io/biases/truth-judgment-illusory-truth-effect/) — Repeated claims receive higher truth ratings than comparable new claims on average, but the effect is heterogeneous and does not generalize equally to every kind of statement. A 2026 systematic review and meta-analysis covering 182 studies, 366 effect sizes and more than 31,000 participants found a reliable corrected average effect. A later 2026 paper with two preregistered experiments found little or no meaningful repetition effect for social-political opinion statements. The useful claim is therefore narrower than 'repeat anything and people will believe it.'
- [Availability Heuristic](https://cognitive-biases.github.io/biases/heuristic-bias-availability-bias/) — Availability is a judgment heuristic: people can use how easily examples or scenarios come to mind when estimating frequency or probability. That shortcut is not automatically an error because memorable or accessible examples can correlate with real frequency. Bias appears when accessibility is driven by factors that are not diagnostic of the quantity being judged.
- [Continued Influence Effect](https://cognitive-biases.github.io/biases/memory-bias-continued-influence-effect/) — The continued influence effect describes a specific pattern: false or outdated information can keep affecting later reasoning even after a clear correction. It does not mean the correction had no effect, that the person rejected the correction, or that belief in the false claim became stronger. Corrections often reduce misinformation reliance without removing it completely. That distinction matters because continued influence is much broader than the rarer backfire outcome.

## References

- [Decision skill](https://cognitive-biases.github.io/skills/evidence-evaluation/)
- [Evidence reviews](https://cognitive-biases.github.io/evidence/)

## Portability

This is an instruction-only Agent Skill. It requires no secrets, no executable scripts and no network access to run. If source-access tools are available, use the linked Cognitive Biases material to preserve review state and uncertainty.

Licence: CC BY-NC-SA 4.0. Commercial reuse requires prior written permission.