How the
Scoring Works
The framework, the dual-metric system, the scoring rules, and why each design decision was made.
Every criterion applied in this project is derived verbatim from Daniella Mestyanek Young and Amy Reed's The Culting of America: What Makes a Cult and Why We Love Them (Otterpine, 2026). Young and Reed define a cult as a group that meets these ten conditions. The composite scoring system was developed independently to extend analytical range for dataset-scale application — but the criteria themselves are Young and Reed's.
A cult is a group that:
Charismatic Leadership
A defined authority figure — or a central idea functioning in place of a person — whose directives are taken as truth and whose challengers are discredited. Does not require a living human leader; posthumous authority that is institutionally ratified and structurally unrevisable scores as high as or higher than a living leader.
Sacred Assumptions
Certain beliefs maintained against counter-evidence, with mantras repeated and alternatives dismissed. The key test is not whether beliefs exist but whether they are enforced against documented contradicting evidence. The most extreme form is architectural prevention of counter-evidence from being generated at all.
Transcendent Mission
A mission so large it justifies sacrifice, treats doubts as betrayal, and provides meaning in ways that override individual judgment. Organizations that structurally encourage internal dissent as an improvement mechanism score N/A — the structural opposite of this criterion.
Sublimation of Individuality
Identity demands, appearance and lifestyle conformity, and rest-as-weakness culture. Includes institutional statements that inflate membership value, creating a reputational double-bind where departure signals either rejection or poor judgment.
Isolation
Information environment narrows, outside perspectives are dismissed, the world shrinks. Isolation through institutional ecosystem completeness — parallel schools, hospitals, media, employment — scores equivalently to geographic compound isolation. The developmental outcome is functionally identical.
Private Vernacular
Specialized vocabulary that marks membership identity, encodes a way of understanding reality difficult to access from outside, and terminates inquiry rather than enabling it. Standard professional field vocabulary does not check this criterion — the test is whether vocabulary operates as epistemological closure, not merely institutional naming.
Us-Versus-Them
More-enlightened-than-outsiders framing, defectors characterized as broken or corrupt, disagreement framed as bigotry or betrayal. Appropriate labor-management framing in unions and symmetrical partisan framing in political parties are distinguished from pathological enemy-construction.
Exploitation of Labor
Sacrifice extracted as virtue, labor monetized through institutional control. Financial extraction coerced through doctrinal framing with salvific or mission stakes is labor extraction. The delivery mechanism — financial, physical, or psychological — does not moderate intensity scoring. Compensation engineered to create exit barriers rather than fairly reward labor also checks this criterion.
High Exit Costs
Departure produces social, economic, or identity costs; exit is framed as betrayal. Spiritual absolutism — where departure means eternal damnation, complete family rupture, and total social network dissolution — scores at the same level as physical confinement. The absence of physical restraint does not moderate the score.
Ends Justify the Means
Institutional harm tolerated in pursuit of mission, cover-ups occur, perpetrators are protected. Multi-generational non-correcting harm patterns score at the ceiling regardless of mechanism. The existence of internal dissenters who made the courageous choice within the same constraints establishes that compliant choices were genuine institutional choices, not forced outcomes.
Every organization receives two scores that are derived independently and never converted between each other. Their divergence is analytically meaningful, not a problem to resolve.
Young's Original Score
0–10 binary checklist
Each of the ten criteria either checks or does not check. Produces three bands: Not Culty (0–2), Kinda Culty (3–5), Super Culty (6–10). Must be derived from direct application of Young's checklist — never mechanically converted from composite intensity. Mechanical conversion produces systematically inflated results.
Composite Cultiness Score
Formula-based 0–100%
Formula: (Breadth ÷ 10) × (Mean Intensity ÷ 10) × 100. Breadth = criteria with non-N/A scores. Mean Intensity = average of those scores. Adds two dimensions the binary instrument cannot capture: intensity variance and breadth-intensity interaction. Produces six tiers from Healthy Group through Cult.
Each criterion receives either a score of 1–10 or N/A. N/A is not a floor score. N/A designates structural absence — the criterion describes a dynamic that is inapplicable because the organization lacks the architecture for it, or because the organization's documented behavior is the structural opposite of what the criterion describes.
Never assign N/A to soften a low score. Never assign a floor number when the evidence says the dynamic is absent. When body text explicitly states a dynamic is not operative, the score must be N/A — not 1 or 2. The distinction matters because N/A criteria are excluded from both the breadth count and the mean intensity calculation. Phantom scoring — floor numbers on structurally absent criteria — artificially inflates composite scores and produces misleading results.
Criterion structurally inapplicable — dynamic not operative and organization lacks the architecture for it
Essentially no evidence of this dynamic
Mild, occasional, or incidental presence
Moderate, recurring presence with documented examples
Strong, systematic presence across multiple documented behaviors
Extreme, defining feature with documented evidence of harm
Scores are anchored to publicly documented, verifiable behaviors — not reputation, not impression, not general public perception. Acceptable sources include court records and regulatory findings, investigative journalism from publications meeting T1–T2 credibility standards, peer-reviewed academic scholarship, government reports, and institutional self-documentation. Each criterion score includes at least one specific, verifiable example with citation.
Where evidence is limited, contested, or primarily indirect, confidence is rated Low and noted explicitly. The methodology does not paper over uncertainty.
AI-assisted scoring is used to generate proposed assessments at scale. Every proposed score passes through human review before entering the dataset. The reviewer verifies that each score is consistent with the body text, that N/A designations have structural rationale, that cited sources support the claims made, and that the assessment reflects consistent application of the methodology across the ideological and cultural spectrum.
The AI proposes. The human decides. No score enters the published dataset without that review.