The 1889-1894 Census of Hallucinations argued that apparitions coinciding with a real death occurred more often than chance, but the finding rests entirely on retrospective self report gathered by volunteers approaching people they knew, against a chance baseline its own authors had to construct. No prospective, controlled test of crisis apparitions has been conducted since.
What would resolve this A prospectively registered study defining in advance what counts as a qualifying apparition and a chance baseline independent of the reporting population, run forward in time rather than reconstructed from memory.
OpenStatistics; survey methodology; history of psychical researchHenry Sidgwick, Eleanor Sidgwick, Frederic W. H. Myers, Frank Podmore, Alice Johnson, Report on the Census of Hallucinations