Flashcards: Enhancing Scientific Self-Correction — 18 cartões

Error-prone humans — role?

Cause of mistakes in science

Self-correction myth — belief?

Science automatically corrects flawed findings

Retraction — purpose?

Withdraws invalid or problematic publication

Honest errors — can lead to?

Retractions or corrections without misconduct

Careless errors — examples?

Typos, copy-paste mistakes, code errors

Negligent errors — include?

Poor design, weak analysis, low power

Error detection — tools?

Statcheck, GRIM, anomaly checks

Statcheck — function?

Recalculates p-values from reported stats

Retraction Watch — role?

Tracks retractions and reasons

Statcheck inconsistencies — signs?

Mismatched p-values, signs, or means

Optimistic rounding — effect?

Can falsely suggest significance

GRIM test — purpose?

Checks if means match data granularity

Randomization check — evaluates?

Baseline plausibility under random assignment

Blinding check — ensures?

Researchers unaware of group assignments

Measurement practices — issues?

Questionable scales, ad hoc items, low validity

Statistical power — meaning?

Ability to detect true effects

Low power — consequence?

Misses real effects or exaggerates findings

Data anomalies — detection?

Constraints violations, implausible patterns