How Scores Change Over Time

CLCI scores are not fixed at publication. New evidence emerges; leadership changes; reforms happen; groups dissolve, splinter, or rebrand. This page documents the reasons we change a score, the public changelog that records every change, and what readers should infer about score volatility.

Why scores need to change

A score frozen at publication would soon be inaccurate. High-control groups are dynamic institutions: founders die, leadership transitions hand the organisation to figures with different temperaments, internal reform movements succeed or fail, court findings emerge that document previously denied practices, and entire eras can end. A static score would either misreport the present pattern (if we set it at the height of the organisation's control practices and never updated) or systematically understate older patterns (if we set it at a reformed present and lost the historical record). The score is therefore a current-best snapshot, with the historical record preserved separately on the timeline.

Categories of update

New evidence

The most common reason for a score change is that the public record has grown. A new court ruling, a substantive investigative piece, a new academic study, the release of previously sealed material, or a substantial on-the-record account from a named former member can move the rating in either direction. New evidence updates go through the standard scoring-appeals evaluation: we cross-check against existing sources, weight by tier, and either move the score, move the confidence rating, or both. Every evidence-driven change is logged in the public changelog with the source cited.

Reform

Documented structural reform earns score credit per the Scoring Appeals page's reform criteria: leadership change, formal governance change, documented disciplinary-practice change, financial-transparency mechanisms, sustained over time, independently verifiable. Reform credit typically manifests as a one- to three-point reduction on the relevant BITE axis and sometimes as a Modifier shift from a positive value toward zero. The historical pattern is preserved in the timeline.

We are deliberately slow on reform credit because the risk of premature credit — granting it on the basis of leadership statements that have not yet been operationalised — is more harmful to readers than the risk of belated credit. A genuine reform programme will be visible across multiple independent sources over multiple years; that is the standard we apply.

Leadership change

Leadership change does not automatically move the score in either direction. A new leader who maintains or extends the previous control practices does not cause the score to move; a new leader who materially loosens or tightens control patterns does, on the same evidence standards that apply elsewhere. Where leadership succession produces a meaningfully different operational entity — different doctrinal emphasis, different disciplinary practice, different scope — we sometimes split the profile into a leader-specific entry separate from the broader-movement entry, per the entity-handling rules on the Scoring Appeals page.

Splinter and breakaway groups

Splinter groups receive their own profiles. The parent group's profile is not automatically rescored by the existence of a splinter; the splinter is not automatically rescored to match the parent. Each entity is documented against its own public-record evidence. Where splinters and parents are cross-linked, we explain the relationship and how the operational patterns differ.

Dormant or defunct groups

Groups that have clearly ceased operation receive a defunct or historical status marker on the profile and the score reflects the operational pattern at the time of dissolution. We do not raise or lower the score of a defunct group on subsequent retrospective reporting unless the retrospective work substantively revises understanding of the operational pattern (e.g. a long-after-the-fact court case, an archive release, a newly published memoir from a senior figure). Where a previously defunct group has been revived under new leadership, we typically create a new profile rather than reactivating the historical one.

Splinter merger or reabsorption

Where a splinter has rejoined its parent organisation or two splinter groups have merged, we maintain the historical profiles and create or update the current-entity profile to reflect the merged organisation. The timeline records the merger date.

Duplicate merge or canonical realignment

Where the dataset contains two profiles documenting the same organisation under different names (alias, regional naming, leadership-era confusion), we merge them, set the canonical slug, and 301-redirect the alias slugs to the canonical URL. The merge itself is logged in the changelog; the score on the merged profile reflects the consolidated evidence base.

Editorial correction

Where a profile was scored in a way that the published methodology does not authorise (e.g. an axis raised on anonymous evidence alone, a Modifier applied without written explanation), the change is an editorial correction rather than an evidence-driven update. We log corrections explicitly as such so readers can distinguish methodology-driven updates from internal-error corrections.

Versioning

Every score change carries a version increment internal to the profile. The current public surface of versioning includes:

lastReviewed — ISO date of the most recent factual review. Shown in the profile header and used by the sitemap.
lastUpdated — ISO date of the most recent edit of any kind (factual or editorial). Shown in the profile header.
changeLog — Append-only list of substantive edits. Each entry includes the date, a short description of the change, and where applicable the source IDs that drove the change. Surfaced inline on the profile.

The plan's longer-term direction is to add a per-versionscoreHistory record alongside the changelog so readers can see the trajectory of a profile's rating over time (previous score, current score, dated). This is not yet implemented across all profiles; it will be in a forthcoming rollout.

Public changelog routes

Substantive score and content updates surface on dedicated routes (planned for a forthcoming stage and listed here for transparency):

/updates — Aggregate index of recent changes;
/updates/scores — Score changes specifically;
/updates/corrections — Corrections resolved;
/updates/new-groups — New profiles published;
/updates/merged-entries — Profiles consolidated under a canonical slug.

These routes are not yet live; the changelog data underlying them is captured in profile-level fields today and will be aggregated into the update routes in the forthcoming stage. The site RSS feed will also extend to include score updates.

What readers should expect about score volatility

Most profiles are stable. Once a profile has High confidence it rarely moves by more than one or two points across the full BITE total, and the movements that do happen are well-supported by new primary evidence.
Low-confidence profiles are more volatile. A Low-confidence rating can move several points in a single review when a substantive new source emerges.
Reform credit is incremental. A reforming group typically sees its score drop by a small amount over multiple reviews rather than a large amount in one update.
New scandals can move scores quickly. A court ruling, government inquiry, or major investigative piece documenting previously undocumented patterns can move a profile's score and confidence in a single review.