---
name: research-claim-evaluation
description: Evaluate a specific research claim against relevant primary evidence, study design, uncertainty, and scope.
---

# Research claim evaluation

Use when asked whether a scientific or empirical claim follows from evidence. State the claim precisely, including population or system, comparison, outcome, magnitude, and causal language. Identify its original source if possible and search for relevant primary studies and credible syntheses, with dates and source limitations.

## Evaluate support

Read the relevant passages and methods rather than relying on a title or citation alone. Compare the measured outcome and conditions with the claim, and distinguish correlation, prediction, mechanism, and intervention effects. Examine study design, sample or data coverage, comparison, measurement, confounding or biases, precision, and competing explanations. For each material result identify effect direction, magnitude and uncertainty as reported; do not replace a missing effect size with a significance label.

Consider counterevidence and differences across studies, including design, populations and time points. Identify whether results generalize to the claim's target context. Separate a source's conclusions from your inference, and indicate whether a correction or retraction changes interpretation. Source access or search limits can prevent a firm assessment.

## Output and dependencies

Give a calibrated conclusion such as supported within stated conditions, mixed, unsupported by the available evidence, or insufficient evidence. Link each important claim to the specific source and explain the most consequential limitations. Report when a source was only available as an abstract or when searches were incomplete. No citation, quote, effect, or consensus should be invented; the evaluation is not a substitute for domain-specific professional judgment where consequences are significant.