Citability score · President
How citable is Theodore Roosevelt?
Strong citability (82.2/100). This profile is deep, well-structured, and freshly sourced — the cleanest signal in the retrieval set. Citability measures how likely an LLM is to cite this profile when a voter asks about the office holder.
- Measured:
- Dataset version:
- 1.20260727-1
- Empirical component:
- measured
Why this score
The composite is a weighted blend of the three deterministic heuristics: content depth (43%), structured data (36%), and source freshness (21%). The measured citation rate is published as evidence but is not currently weighted into the composite — recent measurement cycles found a 0% corpus-wide cite-rate, so weighting it only penalized the sampled profiles. The bench cadence continues unchanged; the weight re-arms for operator review on the first nonzero cycle.
Content depth (81%)
81%Deep, sourced pages earn citations; thin pages do not. narrative 1500/800w · sources 4/3 · 1/4 claim sections.
Structured data (73%)
73%Machine-readable JSON-LD lets an LLM extract verified facts. 8/11 JSON-LD-enriching fields populated.
Source freshness (100%)
100%Recently re-verified provenance reads as trustworthy. median source age 3d (fresh ≤ 90d, stale ≥ 730d).
Measured citation rate (0%)
0%0% citation rate across 2 panel trial(s), measured 2026-07-29. The measured citation rate is published as evidence but is not currently weighted into the composite — recent measurement cycles found a 0% corpus-wide cite-rate, so weighting it only penalized the sampled profiles. The bench cadence continues unchanged; the weight re-arms for operator review on the first nonzero cycle.
This trial protocol asks a fixed panel of models profile questions and records whether each answer cites this page’s canonical URL; the 0% result stays visible as evidence, while its composite weight is retired unless a future cycle returns a nonzero rate — see our methodology.
Profile health
Which signals are at the floor and which are gaps, measured against our content-depth and structured-data floors.
- OKContent depth vs floors — narrative 1500/800w · sources 4/3 · 1/4 claim sections
- GapStructured data (JSON-LD population) — 8/11 JSON-LD-enriching fields populated
- OKSource freshness — median source age 3d (fresh ≤ 90d, stale ≥ 730d)
- OKMeasured citation rate — 0% citation rate across 2 panel trial(s), measured 2026-07-29
How to raise this score
- Add sourced key policy positions.
Citability over time
Across 6 bench runs (2026-06-07 → 2026-07-29), this profile's composite held flat 0 points (57.6 → 57.6).
Its measured citation rate held flat 0 points (0% → 0%) across the model panel.
| Run date | Composite | Cite rate | Trials |
|---|---|---|---|
| 57.6 | 0% | 3 | |
| 57.6 | 0% | 3 | |
| 57.6 | 0% | 3 | |
| 57.6 | 0% | 3 | |
| 57.6 | 0% | 3 | |
| 57.6 | 0% | 2 |
See the corpus-wide citability trend for how every scored profile is trending together.
About this score
Citability measures how likely an LLM is to cite this profile when a voter asks about the office holder. It is computed offline from the published profile — no live model call is made when this page loads. Read the full Theodore Roosevelt profile or our sourcing methodology.
Related on The Candidate
- Theodore Roosevelt — full profileThe canonical sourced profile this citability score measures.Open
- All citability scoresBrowse every scored federal office-holder profile.Open
- Our methodologyHow we source facts and structure every profile for machine readability.Open
- Public read APIThe citability score is also exposed as a documented JSON endpoint.Open