Skip to main content

Educational tool only. All groups exist on a spectrum of control. Individual experiences vary. Based on publicly available reports, ex-member accounts, court records, and expert analyses — not medical or legal advice.

CLCI Hub

Source Hierarchy

How CLCI Hub ranks the evidence behind every claim — what counts as primary, secondary, and tertiary; how court records, government reports, academic work, journalism, books, and survivor testimony are weighed; and how we handle contested evidence and the distinction between allegation and finding.

This page is a longer, more detailed companion to the Source Policy. The Source Policy is the canonical short-form document covering tiers and confidence weighting; this page adds explicit guidance on contested evidence, allegations versus findings, and the boundaries between “reported” and “documented.”

The three reliability tiers

Primary sources

Primary sources are first-order public-record documents produced through a process with adversarial review or institutional accountability. They are the strongest evidence available and a single substantive primary source can move a profile's confidence to High on its own.

Secondary sources

Secondary sources are independently produced material reviewing or analysing primary material. They carry substantial evidentiary weight and a profile supported by multiple corroborating secondary sources can also reach High confidence.

Tertiary sources

Tertiary sources are aggregators or commentary built on top of primary and secondary material. They are useful pointers but rarely sufficient alone.

Ex-member testimony

Ex-member testimony is the most contested category in cult-research sourcing. Critics of the field have argued that ex-member accounts are inherently unreliable; defenders have argued that they are often the only window into organisations that operate under strict internal secrecy. The honest position is that ex-member testimony varies enormously in evidentiary weight, and treating it as a single category is itself misleading.

We apply a consistent four-level framework:

Allegation versus finding

Distinguishing allegation from finding is one of the most important editorial moves on CLCI Hub. The language we use signals which we are reporting.

We never present an allegation as a finding. Where a court has explicitly declined to find on a question (acquittal, dismissal for lack of evidence, statute-of-limitations bar), we say so. Where reporting on an allegation has been retracted, we treat that retraction as a primary source about the unreliability of the original report.

How conflicting evidence is handled

When primary and secondary sources disagree, we report the disagreement and explain how we weighted the evidence. We do not silently pick a side. Standard patterns we apply:

Why anonymous reports are not treated as conclusive

Anonymous reports come from a wide range of motivations: genuine survivors protecting themselves from retaliation; current members defending the group without identifying themselves; bad-faith actors pursuing personal grievances; researchers protecting their own identity from a litigious organisation. We cannot, from outside, distinguish these cases reliably enough to base a published claim on a single anonymous source. We can and do use anonymous accounts as a pointer: if a pattern emerges across many anonymous reports we will look for primary or secondary sources that document the same pattern.

This is a deliberately conservative posture. We accept that it sometimes means a real pattern of abuse remains uncaptured on a profile until a primary source emerges. We believe the alternative — publishing serious claims about named individuals on anonymous evidence — would do more harm to the field's credibility, to survivors' long-term ability to be believed, and to innocent third parties than the conservative posture costs.

Sources we will not use

How sources flow into confidence ratings

The source mix on a profile determines its overall confidence rating. Briefly: High confidence requires multiple primary sources plus secondary corroboration. Medium confidence is credible secondary work with limited primary support. Low confidence is fragmentary anecdotal reports, single-source claims, or active controversy. Low-confidence ratings are explicitly provisional and marked as such on the profile.

See also Source Policy · Confidence Levels · Scoring Appeals · Living Persons