Ghosts are the mind's way of interpreting how the body reacts to certain surroundings, say UK psychologists.
A chill in the air, low-light conditions and even magnetic fields may trigger feelings that "a presence" is in a room - but that is all they are, feelings.
Dr Wiseman's team used hundreds of volunteers
This explanation of ghosts is the result of a large study in which researchers led hundreds of volunteers around two of the UK's supposedly most haunted locations - Hampton Court Palace, England, and the South Bridge Vaults in Edinburgh, Scotland.
Dr Richard Wiseman, of the University of Hertfordshire, and his colleagues say their work has thrown up some interesting data to suggest why so many people can be spooked in the same building but provides no evidence that ghosts are real.
In Hampton Court - alleged to contain the ghost of the executed Catherine Howard, 5th wife of Henry VIII - the volunteers were asked to face their fear.
They had to record any unusual experiences, such as hearing footsteps, feeling cold or a presence in the room, as well as marking the location and intensity of the experience on a floor plan.
The top image was taken in South Bridge Vault 9
The bottom image was taken just five minutes later
How does one explain the white fuzz? An artefact of processing?
Before this, candidates were also asked to reveal any prior knowledge of hauntings at the site.
The researchers then examined the distribution of unusual experiences.
In a "normal" setting, you would expect the ghostly encounters to be evenly spaced, but in classic haunting, they would be clustered around certain places.
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.
And in the Edinburgh vaults, the result was the same - the vaults considered most haunted were the locations where the most unusual encounters occurred during the study.
The researchers interpret this as evidence that hauntings are a real phenomenon because they are concentrated in specific places over time.
Indeed, it is known for people from different cultures to consistently report similar experiences over perhaps hundreds of years.
THE ROLE OF MAGNETISM
When we measure houses where pervasive haunts occur, the place where the occupants find they can sleep, by trial and error, has the most consistent and normal field strengths
"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."
But are the ghosts real? Dr Wiseman and his colleagues are not so sure.
They claim, somewhat paradoxically, that the hauntings exist but the ghosts do not.
"People do have consistent experiences in consistent places, but I think that this is driven by visual factors mainly, and perhaps some other environmental cues," he said.
Making detailed measurements at each place, such as temperature, light intensity and room space, Dr Wiseman thinks that people are responding unconsciously to environmental cues and the general "spookiness" of their surroundings.
He cites examples of mediums successfully indicating haunted areas of buildings with no prior knowledge of them.
Spiritualists interpret this as evidence that the ghosts are there, but another explanation is that the mediums are simply more sensitive to the environmental cues that result in haunted feelings - not sensitivity to the ghosts themselves.
Sceptics have long maintained that ghostly encounters are influenced by a person's knowledge of the place and its history, the "prior knowledge hypothesis".
But this study refutes that explanation, as the statistics showed that prior knowledge did not affect the areas in which strange experiences were recorded.
"We found little if no evidence that people's prior knowledge mattered," said Dr Wiseman. "If anything, it made them veer away from having experiences in the known haunted sites."
Dr Wiseman and colleagues report their data in the British Journal of Psychology.