About
It started as a methodological problem, not a belief
The Continuity Project exists because the standard way of testing this question cannot answer it.
Accurate information about a deceased individual is not logically identical to continuation of that individual.
The project began from a methodological problem rather than a belief. Decades of accuracy studies asked whether a medium can produce correct information about a deceased person. Even a striking positive answer leaves the central question untouched: information about someone can, in principle, arrive by many routes, and none of those routes requires that the person still exists as an integrated unit.
Accuracy is therefore the wrong primary measure. If something persists, it should behave like a persisting thing: it should be recognisably the same across encounters that cannot influence one another, it should carry idiosyncrasies forward, and it should fail in structured rather than random ways.
So we test persistence of identity across independent, blinded encounters. Multiple mediums, no shared information, repeated sessions, blinded judges. The comparison of interest is not one reading against a biography, but one encounter against another.
What follows from that
- The unit of analysis is the chain of sessions, not the individual reading.
- High-base-rate accuracy contributes little; reproducible idiosyncrasy contributes a great deal.
- A result can be negative in an informative way — and negative results are published in full.
How we work
Roles are explicit and separated: coordinators run studies, mediums see only codes, judges see only anonymized transcripts, analysts see only coded data, and administrators manage permissions under audit. Protocols are reviewed before use, consent is documented and revocable, and every sensitive operation writes an immutable audit entry.
What we do not claim
No participant is required to believe in survival after death. The project investigates measurable claims, not personal beliefs, and no study outcome is described as proof about any individual. The falsification conditions are published in advance — what would change our minds.
Researchers who want the conceptual vocabulary behind the design can read the optional conceptual framework. No study depends on accepting it.