all the necessary caveats apply about the quality of reporting on random studies, but i can't help linking to the bbc's fascinating report on what's going on when you sense ghosts or otherwise get a good case of the howling fantods . ignoring the very real possibility of methodological funniness, it appears that regular people can reliably detect "spookiness" in places that are claimed to be "spooky", even if they don't know that they are supposed to be in a "spooky" place:

"The results were striking: participants did record a higher number of unusual experiences in the most classically haunted places of Hampton Court, areas such as the Georgian rooms and the Haunted Gallery."

""Hauntings exist, in the sense that places exist where people reliably have unusual experiences," Dr Richard Wiseman told BBC News Online. "The existence of ghosts is a way of explaining these experiences.""

the scientists posit that there aren't really any ghosts, but rather that there some subtle perceptual cues that get consistantly interpreted as spooky. i'll give them the benefit of the doubt and assume that these cues are less obvious than standing in a dimly lit, cobwebbed crypt with organ music wafting in. it'd be interesting to see if they could reduce the perceptual cues to their constituent parts, so that you could more reliably manipulate the general level of "spookiness". until then, as one commenter aptly states, " To the believer no proof is required, to the sceptic, no proof is sufficient."