01
Blinded session management
Sessions are created, scheduled and conducted against opaque codes. Mediums never see target identity; coordinators see only what their study's blinding configuration permits.
For researchers
Study anomalous identity claims with reproducible infrastructure.
The Continuity Project maintains the operational machinery that blinded, longitudinal identity studies normally lack: coded participants, randomized assignment, independent judging panels, immutable audit trails and analysis-ready exports. The infrastructure is available to external investigators under explicit, study-scoped roles.
Available infrastructure
02
Randomized target assignment
Targets are allocated to sessions by seeded randomization, recorded before the session opens and reproducible from the stored seed.
03
Longitudinal continuity testing
Repeated independent sessions against the same coded target, with continuity features extracted per session and compared across the chain.
04
Cross-medium testing
The same coded target can be assigned to multiple unconnected mediums, allowing agreement between independent encounters to be measured directly.
05
Independent judging
Four judging designs — target ranking, reading ranking, statement scoring and continuity judging — with per-judge randomized ordering and locked submissions before any result is visible.
06
Preregistration support
Hypotheses, primary and secondary outcomes, exclusion rules and analysis plans are filed and version-locked before data collection begins.
07
Audit logging
Every sensitive operation writes an append-only entry with user, acting role, action, resource, study and request metadata. Target identity access is logged field by field.
08
Anonymized export
Analysis-ready CSV and JSON datasets containing coded identifiers only, generated from the same records the judging interface used.
09
Public protocol publishing
Protocol versions, amendments and deviations are published so an unaffiliated team can run the same design and compare outcomes.
Work with the project
Access conditions
Collaborators receive explicit, study-scoped roles. No role grants access to private target information by default, and every identity field read is logged individually. Coded datasets are released under a data use agreement once a study closes; identifying material is never shared.
Protocols are written so an unaffiliated team can run them. Instruments, coding manuals and randomization procedures are published with each protocol version — see protocols and preregistrations.
Criticism is welcome
The advisory board includes independent reviewers appointed specifically to attack the design. Objections that change a design are recorded in the protocol history with attribution, and dissenting commentary may be published alongside any study report.
Conceptual background is available on the optional conceptual framework page. No study depends on it.