domains / families / research / result-reliability-checker

result-reliability-checker

separate

Assesses whether study results are trustworthy by auditing design integrity, sample structure, statistical handling, bias control, validation chain, and claim discipline. It identifies where results are robust, fragile, overfit, under-validated, or overclaimed. Always separate reported findings from reliability judgment. Never fabricate references, PMIDs, DOIs, trial identifiers, study features, or validation claims.

When to use

• assessing whether results are reliable enough to treat as usable evidence
• identifying weak points in study design and statistics
• evaluating if validation is convincing or superficial
• checking if conclusions exceed what methods support
• auditing for small-sample risk, overfitting, and weak validation
• assessing trustworthiness of claimed treatment-effect findings
• checking if omics paper findings are stable
• auditing mechanism papers for overclaiming

When not to use

• generic paper summary
• result restatement
• replacement for full systematic risk-of-bias appraisal
• patient-specific clinical decision support
• guaranteeing truth from partial snippets with no methods/results basis
• inventing missing methods, statistics, validation details, or references
• certifying a paper as definitive evidence without uncertainty disclosure

Skill page

Source