The Cultiness Spectrum — The Dataset

What Was
Assessed

370 active American organizations. 38 calibration anchors. Publicly available, openly documented, and ongoing.

Each organization in the dataset receives a complete assessment across all ten criteria. Every entry includes:

Per-criterion intensity scores (1–10 or N/A) with documented rationale

Evidence-based body text for each criterion with source citations

Confidence rating (High / Medium / Low) per criterion

Young's Original Score (0–10, binary checklist, independently derived)

Composite Cultiness Score (formula-based, 0–100%)

Composite tier classification

Trajectory assessment (Stable / Escalating / Declining / Defunct)

Political compass position (Economic axis and Authority axis, independent of cultiness scores)

One-paragraph summary assessment


Religious denominations

Evangelical, Catholic, mainline Protestant, high-control NRMs, and historical formations

~60

Political movements & parties

Across the full ideological spectrum — assessed by identical criteria

~35

Military formations

All US service branches, special operations formations, and historical military institutions

~20

Federal agencies

Intelligence community, law enforcement, regulatory agencies, and cabinet departments

~30

Corporate employers

Tech, finance, pharma, defense contractors, and retail

~25

Media institutions

Broadcast, print, digital, and ideological media formations

~20

Educational institutions

Universities, K-12 systems, and homeschool formations

~15

Advocacy & civil society

Civil rights, labor, environmental, and ideological advocacy organizations

~30

Historical calibration anchors

Domestic and international historical formations used to bracket the scoring spectrum

38

The dataset includes 38 calibration anchors — historical and international organizations used to bracket the scoring spectrum and ensure consistency across assessment sessions. Anchors span from organizations scoring at the Cult ceiling (100% composite, 10/10 Young's) to the Healthy Group floor (5% composite, 0/10 Young's). They are audited against current methodology after each versioned methodology update.

Calibration anchors include both extreme cases (Aum Shinrikyo, Peoples Temple, the Khmer Rouge) and reference floor cases (Costco, at 5% composite, serves as the Healthy Group anchor — demonstrating what high institutional loyalty and strong employee satisfaction look like in the complete absence of formation-system architecture).


The dataset is under active development. Every score change is recorded in an immutable audit log with timestamp, rationale, and the methodology version under which it was made. The analytical principles governing scoring are documented in a versioned methodology reference — currently at V4.0, reflecting refinements from systematic divergence review conducted in 2026.

Scores are analytical assessments anchored to publicly verifiable documented behaviors, not definitive determinations. Where evidence is limited or contested, this is noted explicitly in the entry.


Public Repository

The dataset, methodology documentation, scoring engine, and full audit trail are publicly available on GitHub.

View on GitHub

The dataset is not complete. Approximately 25% of entries currently have body texts that are thin, template-generated, or pending evidence-based revision. These are flagged in the repository and represent active work rather than finished assessments. Scores on entries with low-confidence body texts should be treated accordingly.

Evidence availability varies substantially across organizational types. High-profile religious movements, political organizations, and publicly traded corporations have extensive public documentation. Some smaller or more opaque organizations have limited public records, and their scores reflect that limitation.

The methodology has evolved over time. Earlier assessments were scored under less refined standards than the current V4.0 methodology. Systematic divergence review has corrected the most significant inconsistencies, but cross-batch variation remains a known limitation. A full rescore against current methodology is in progress.

Read the MethodologySee the Findings