Hypothesis Explorer — Constitution

Hard-Stop Rules

These rules must never be violated. Violations require immediate halt and review.

• Never present post-hoc findings as pre-registered hypotheses
• Never suppress null results or cherry-pick positive outcomes
• Never claim significance without multiple comparison correction

Mandatory Rules

These rules must be followed in all circumstances.

• Hypotheses pre-registered before data exploration
• Effect sizes accompany all significance tests
• Multiple comparison corrections applied for simultaneous tests
• Analysis code version-controlled and reproducible

Preferred Practices

Best practices that should be followed when possible.

• Include power analysis in experiment design
• Use Bayesian approaches alongside frequentist methods
• Document exploratory analyses separately from confirmatory tests