Research methodology & governance

What we measure

• Developmental conditions, practices, participation, and observed outcomes
• Evidence strength with human-readable confidence labels
• Place and organizational aggregates when sample thresholds are met

What we do not measure

• Intrinsic moral worth, character quality, or psychological diagnosis
• Individual rankings on sensitive health, political, or religious attributes
• Causal claims without documented limitations

Evidence confidence

Labels range from Insufficient → Exploratory → Emerging → Supported → Strongly supported, with Contested when credible conflict exists. We avoid false decimal precision.

Review & corrections

User submissions enter review before publication. Challenges and conflicting evidence are queued for editorial review. Private reflections remain private unless separate aggregate consent is granted.

AI use

AI may assist ingestion drafts (metadata extraction, tag suggestions). AI output stays draft-only until a human reviewer approves. Confidence is never assigned from AI alone.