Divinity Atlas
Four Open Questions Point to Statistics, and All Four Are Missing the Same Thing
Five open questions in this atlas name Statistics as the discipline that could resolve them. One, on the population frequency of Human Design types, is not public and is excluded here. The remaining four span three centuries of claims, Highland second sight, a Victorian census of ghost sightings, a 2015 meta-analysis of parapsychology experiments, and a global network of random-number generators, and none of them share a subject. What they share is a gap: in every one of the four, no baseline was fixed before the fact, so no test now run can settle it.
Two Claims With No Baseline at All
The entire evidentiary record for Highland second sight is testimonial, collected by folklorists across three centuries of oral tradition, and no experimental or quasi-experimental protocol comparing a claimed seer's forewarnings to ordinary chance has ever been devised. The 1889 to 1894 Census of Hallucinations argued that apparitions coinciding with a real death occur more often than chance, but the finding rests entirely on retrospective self-report gathered by volunteers approaching people they already knew, tested against a chance baseline its own authors had to construct after the fact. Neither claim has ever been run forward. What would resolve the first is a prospective, blinded elicitation study, predictions logged in advance and scored by an independent judge; what would resolve the second is the same shape, a prospectively registered study fixing in advance what counts as a qualifying apparition and what counts as chance.
Two Disputes Over Data Already Collected
The other two questions are not about testimony at all. Bem, Tressoldi, Rabeyron and Duggan's 2015 meta-analysis reports a small positive effect across ninety retroactive-influence experiments and argues that bias-correction methods do not remove it; critics answer that the correction tools available cannot distinguish a genuine small effect from a literature filtered by what got written up, especially where the constituent studies were not pre-registered, a dispute live in the wider methodology literature and not confined to parapsychology. The Global Consciousness Project's own network and data archive are not in dispute, but three analytic choices, which events count as globally significant, where each analysis window begins and ends, and which statistic to compute, are made by people with access to the results, and critics argue that freedom alone is enough to produce the reported pattern from structureless data. What would resolve the first is a registered-report replication series with the analysis plan lodged in advance; what would resolve the second is an outside party fixing the event list and window boundaries before the project runs its analysis once against them.
The Same Gap, Asked From Four Directions
A folklore claim with no controlled test behind it, and a statistical dispute over how much freedom an analyst had after the data existed, look like different problems. Read together, all four questions are asking Statistics for the identical thing: a way to fix, before an outcome is known, what would count as success and what would count as chance, so that no result can be argued into shape after the fact. That is what none of the four currently has, and it is the entire content of the discipline's own open work here.