Methods you can trust and inspect
Reference scoring and research scoring
We will keep published reference procedures separate from experimental variants. Researchers can explore new scoring methods without silently changing established instruments or past results.
| Reference scoring | Experimental research scoring |
|---|---|
| Preserves documented source keys | Permits investigator-defined weights and formulas |
| Uses a specified version of an established instrument | Supports newly authored or shortened instruments |
| Reports published evidence and limitations | Labels changes and validation status prominently |
| Reproducible reference scoring procedures | Versioned experiment configurations and comparisons |
| Uses reference norms only where justified | Allows appropriate study-specific reference groups |
Every instrument version and scoring configuration should be immutable. Science must be reproducible—even when the scoring improves.
A new algorithm can be computed precisely without being scientifically validated. Results must say which is which.