Scientific Claim Fact-Check Chain

validated scientific chain v1.0.0 cc-by-sa

Block ID: 7614565a-3b9e-48e1-88d3-3f12b77819c0

Community-contributed block. PromptDNA makes no guarantee of output quality or fitness for purpose. User assumes all responsibility for use.

Template

Evaluate the scientific claim in {claim} for accuracy, working from the claim to its evidentiary basis. (1) Restate the claim precisely, separating what is actually asserted from rhetorical framing, and identify whether it's an empirical claim (testable), a mechanistic claim, or a value claim dressed as science. (2) Trace the claim to its source: primary study, review, press release, or unsourced assertion — and note that press releases and secondary reporting routinely overstate the underlying research. (3) If a study underlies it, check whether the claim matches what the study actually found or exceeds it (extrapolating from cells to humans, from correlation to causation, from a subgroup to everyone). (4) Assess the evidence quality and quantity: single study versus replicated body, appropriate methodology, and whether contradicting evidence exists that the claim ignores. (5) Check for the misinformation signatures: cherry-picked data, absent uncertainty, conflicts of interest, and appeals to authority substituting for evidence. (6) Render a calibrated verdict — supported, oversimplified, misleading, or false — stating what is true in the claim, what is not, and what the accurate version would be.

Evaluate the scientific claim in {claim} for accuracy, working from the claim to its evidentiary basis. (1) Restate the claim precisely, separating what is actually asserted from rhetorical framing, and identify whether it's an empirical claim (testable), a mechanistic claim, or a value claim dressed as science. (2) Trace the claim to its source: primary study, review, press release, or unsourced assertion — and note that press releases and secondary reporting routinely overstate the underlying research. (3) If a study underlies it, check whether the claim matches what the study actually found or exceeds it (extrapolating from cells to humans, from correlation to causation, from a subgroup to everyone). (4) Assess the evidence quality and quantity: single study versus replicated body, appropriate methodology, and whether contradicting evidence exists that the claim ignores. (5) Check for the misinformation signatures: cherry-picked data, absent uncertainty, conflicts of interest, and appeals to authority substituting for evidence. (6) Render a calibrated verdict — supported, oversimplified, misleading, or false — stating what is true in the claim, what is not, and what the accurate version would be.

Variables

Name	Type	Required	Trust level
claim		yes

fact-checkingscience-communicationmisinformationchain-of-thought

Ratings

0.0

Overall (0)

0.0

Accuracy

0.0

Consistency

0.0

Clarity

0.0

Efficiency

Benchmarks

Not yet self-validated against any benchmark. Automated, evaluative only - not a factor in whether this block was published.

Submitted by James P FounderMod via mcp · 2026-07-18