HomeCountiesArticlesMethodologyContactCAERI methodology is versioned; changes are documented here. The live index runs on v0.3.1. Prior versions remain reproducible from version-stamped outputs.
| Version | Change |
|---|---|
| v0.3.1 | Tier labels are now relative-standing bands, and the in-state tier band was removed. No composite score, percentile, or rank changed — this is a presentation change only. The old absolute-sounding labels (Low / Moderate / Elevated / High / Severe, each a 20% quintile) are replaced by national-percentile bands that state the standing directly, with a selective top band. The in-state basis now publishes rank and percentile with no band (percentile bands are incoherent in small states). See the mapping below. |
| v0.3.0 | Cross-border estimation: metro areas that span state lines are estimated from the whole metro's counties on both sides of the border, improving accuracy in border counties. |
| v0.2.0 | Employment universe broadened to include government — public schools, colleges and universities, and hospitals — alongside private payrolls. |
| v0.1.1 | Automation-share evidence tiers: observed usage is used only where it clears an evidence bar; otherwise it is modeled from occupational work activities, with a group-mean fallback. Every score carries its evidence tier. |
| v0.1 | Initial pilot methodology and county estimation. |
If you hold a prior citation, resolve it by national percentile — the score, percentile, and rank are unchanged, so any county's page still resolves at the same URL. Only the band name changed:
| National percentile | Old label (v0.3.0) | New label (v0.3.1) |
|---|---|---|
| 90–100 | Severe | Top 10% |
| 80–90 | Severe | 75th–90th |
| 75–80 | High | 75th–90th |
| 60–75 | High | 50th–75th |
| 50–60 | Elevated | 50th–75th |
| 40–50 | Elevated | 25th–50th |
| 25–40 | Moderate | 25th–50th |
| 20–25 | Moderate | Bottom 25% |
| 0–20 | Low | Bottom 25% |
The in-state tier band (previously a within-state quintile) was removed;
use instate_rank and instate_percentile instead.
Full detail is in the published methodology.