The continuity protocol

Hypotheses, outcome measures, scoring instruments and the analysis plan are filed before any data is collected. Nothing is added to the confirmatory set afterwards; later ideas are reported as exploratory.

Every target, sitter, medium and session receives an opaque randomized code, for example TCP-T-8F42, TCP-S-19AD or TCP-M-104. Identifying material is stored separately from operational records and is not reachable from a medium’s view of a study.

A medium sees “Session TCP-S-19AD” and nothing else. Identity is disclosed only when a specific approved protocol requires it, and that disclosure is logged.

Sessions are scheduled, recorded and transcribed under fixed conditions. Transcripts are segmented into discrete statements before scoring so that judging never depends on the narrative order in which material arrived.

Judges receive anonymized transcripts with no indication of medium, sitter or target. Where the judging design requires it, judges score sets of transcripts against multiple candidate profiles without knowing which pairing is the intended one.

The same claimed identity is revisited across sessions, time intervals and different mediums. Continuity measures compare later material against earlier material without exposing earlier material to the current medium.

Analysts work from coded data. Results are reported against the preregistered plan including effect sizes and confidence intervals, and inconclusive outcomes are published in the same form as positive or negative ones.

Studies are in preparation. No protocol has completed data collection yet.