QUANTITATIVE INVESTIGATION OF A HAUNTED CASTLE IN NEW-JERSEY

Twenty witnesses provided accounts of their own and others' experience s in an ostensibly haunted castle. The case was investigated with a va riety of methods. Quantitative tests were administered to experimental participants. Three sensitives and three skeptics toured the castle a nd attempted to detect locations of ghostly phenomena. The subjects th en evaluated a checklist that was composed of brief descriptions of th e witnesses' reports randomly interspersed with plausible but irreleva nt descriptions. One sensitive designated locations that were in signi ficant agreement with sites where phenomena had previously been report ed (p = .033); another gave responses to the descriptive items that si gnificantly resembled the witnesses' responses (p = .029). Sensitives' combined responses were also significant for locations (p = .014) and descriptions (p = .035) of the phenomena. Responses of skeptics were not significant. A random number generator (RNG) yielded no significan t differences between target and control sites. Infrared tests produce d no anomalies, but a Polaroid test and a videorecording test produced curious effects that were not identified as familiar artifacts. Three witnesses each chose a different likeness from a series of photograph s as the image that most closely resembled an apparition each had repo rted. Projective tests evaluated witnesses' psychological states, rece ptive psi potential, and poltergeist propensities.