Testing an Anomalous Photograph Against Ordinary Image-Forming Mechanisms: A Null Hypothesis Framework and a Single-Case Demonstration from Battle Mountain, Virginia, USA
Year: 2026
Type: preprint
Abstract: Photographs offered as evidence of anomalous phenomena are usually evaluated qualitatively, with little attempt to state which ordinary explanations are excluded or how improbable the image is under each. This study developed a reproducible null-hypothesis testing framework for a single anomalous photograph and demonstrated it upon an unedited 2015 smartphone image showing two pale, blue-white forms within the shaded tree line at Battle Mountain, Rappahannock County, Virginia, USA, site of an 1863 Confederate infantry assault. The global null hypothesis, that the forms are ordinary lighting artifacts, was decomposed into component tests for sensor noise, lens-flare ghosting and scene lighting, the last employing a scan statistic with a generalized Pareto tail correcting for the look-elsewhere effect after veiling glare had been removed. Sensor noise (p ≤ 0.00001) and lens flare (p = 4.2 × 10⁻⁶) were rejected against an alpha of 0.001, while scene lighting was rejected against patches at the forms' own illumination level (p < 2.2 × 10⁻¹⁶) but not against the full glare-corrected tree line (p = 0.95 and 1.0). The author concluded that the outcome depends upon which comparison set is physically appropriate, and that although the analysis excludes the tested mechanisms, it cannot establish the cause.
Source: Research Square
Author: Thomas Alan Woolman
Language: English
Subfield: Social Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7226018941
