More Options

Eyewitness Testimony and the Paranormal


Richard Wiseman, Matthew Smith, and Jeff Wisman

Volume 19.6, November / December 1995

Much of the evidence relating to paranormal phenomena consists of eyewitness testimony. However, a large body of experimental research has shown that such testimony can be extremely unreliable.

For example, in 1887 Richard Hodgson and S. John Davey held seances in Britain (in which phenomena were faked by trickery) for unsuspecting sitters and requested each sitter to write a description of the seance after it had ended. Hodgson and Davey reported that sitters omitted many important events and recalled others in incorrect order. Indeed, some of the accounts were so unreliable that Hodgson later remarked: The account of a trick by a person ignorant of the method used in its production will involve a misdescription of its fundamental conditions . . . so marked that no clue is afforded the student for the actual explanation (Hodgson and Davey 1887, p. 9).

In a partial replication of this work, Theodore Besterman (1932) in Britain had sitters attend a fake seance and then answer questions relating to various phenomena that had occurred. Besterman reported that sitters had a tendency to underestimate the number of persons present in the seance room, to fail to report major disturbances that took place (e.g., the movement of the experimenter from the seance room), to fail to recall the conditions under which given phenomena took place, and to experience the illusory movements of objects.

More recently, Singer and Benassi in the United States (1980) had a stage magician perform fake psychic phenomena before two groups of university students. Students in one group were told that they were about to see a magician; the other group, that they were about to witness a demonstration of genuine psychic ability. Afterward, all of the students were asked to note whether they believed the performer was a genuine psychic or a magician. Approximately two-thirds of both groups stated they believed the performer to be a genuine psychic. In a follow-up experiment the researchers added a third condition, wherein the experimenter stressed that the performer was definitely a magician. Fifty-eight percent of the people in this group still stated they believed the performer to be a genuine psychic!

These studies admirably demonstrate that eyewitness testimony of supposedly paranormal events can be unreliable. Additional studies have now started to examine some of the factors that might cause such inaccuracy.

Clearly, many supposedly paranormal events are difficult to observe simply because of their duration, frequency, and the conditions under which they occur. For example, ostensible poltergeist activity, seance phenomena, and UFO sightings often occur without warning, are over within a few moments, take place under poor lighting or weather conditions, or happen at a considerable distance from observers. In addition, some people have sight/hearing deficiencies, while others have observed these phenomena under the influence of alcohol, drugs, or when they are tired (especially if they have had to wait a relatively long time for the phenomena to occur).

It is also possible that observers’ beliefs and expectations play an important role in the production of inaccurate testimony. Different people clearly have different beliefs and expectations prior to observing a supposed psychic — skeptics might expect to see some kind of trickery; believers may expect a display of genuine psi. Some seventy years ago Eric Dingwall in Britain (1921) speculated that such expectations may distort eyewitness testimony:

The frame of mind in which a person goes to see magic and to a medium cannot be compared. In one case he goes either purely for amusement or possibly with the idea of discovering ‘how it was done,’ whilst in the other he usually goes with the thought that it is possible that he will come into direct contact with the other world (p. 211).

Recent experimental evidence suggests that Dingwall’s speculations are correct.

Wiseman and Morris (1995a) in Britain carried out two studies investigating the effect that belief in the paranormal has on the observation of conjuring tricks. Individuals taking part in the experiment were first asked several questions concerning their belief in the paranormal. On the basis of their answers they were classified as either believers (labeled “sheep”) or skeptics (labeled goats”). [Gertrude Schmeidler, City College, New York City, coined the terms sheep and goats.]

In both experiments individuals were first shown a film containing fake psychic demonstrations. In the first demonstration the “psychic” apparently bent a key by concentrating on it; in the second demonstration he supposedly bent a spoon simply by rubbing it.

After they watched the film, witnesses were asked to rate the “paranormal” content of the demonstrations and complete a set of recall questions. Wiseman and Morris wanted to discover if, as Hodgson and Dingwall had suggested, sheep really did tend to misremember those parts of the demonstrations that were central to solving the tricks. For this reason, half of the questions concerned the methods used to fake the phenomena. For example, the psychic faked the key-bending demonstration by secretly switching the straight key for a pre-bent duplicate by passing the straight key from one hand to the other. During the switch the straight key could not be seen. This was clearly central to the trick’s method; and one of the “important” questions asked was whether the straight key had always remained in sight. A second set of “unimportant” questions asked about parts of the demonstration that were not related to the tricks’ methods.

Overall, the results suggested that sheep rated the demonstrations as more “paranormal” than goats did, and that goats did indeed recall significantly more “important” information than sheep. There was no such difference for the recall of the “unimportant” information.

This is not the only study to investigate sheep/goat differences in observation and recall of “paranormal” phenomena. Jones and Russell in the United States (1980) asked individuals to observe a staged demonstration of extrasensory perception (ESP). In one condition the demonstration was successful (i.e., ESP appeared to occur) while in the other it was not. All individuals were then asked to recall the demonstration. Sheep who saw the unsuccessful demonstration distorted their memories of it and often stated that ESP had occurred. Goats tended to correctly recall the demonstration, even if it appeared to support the existence of ESP.

In addition, Matthew Smith in Britain (1993) investigated the effect that instructions (given prior to watching a film containing a demonstration of apparent psychic ability) had on the recall of the film. Individuals were split into two groups. One group was told that the film contained trickery; the other group was told that it contained genuine paranormal phenomena. The former group recalled significantly more information about the film than the latter group.

All of the above experiments were carried out in controlled laboratory settings. However, another recent study suggests that the same inaccuracies may exist in a more natural setting, namely, the seance room.

Many individuals have reported experiencing extraordinary phenomena during dark-room seances. Eyewitnesses claim that objects have mysteriously moved, strange sounds have been produced, or ghostly forms have appeared, and that these phenomena have occurred under conditions that render normal explanations practically impossible.

Believers argue that conditions commonly associated with a seance (such as darkness, anticipation, and fear) may act as a catalyst to produce these phenomena (Batcheldor 1966). Skeptics suggest that reports of seances are unreliable and that eyewitnesses are either fooling themselves or being fooled by fraudulent mediums.

The authors carried out an experiment in the United Kingdom to assess both the reliability of testimony relating to seance phenomena, and whether paranormal events could be produced in a modern seance. We carried out our experiment, titled “Manifestations,” three times. Twenty-five people attended on each occasion. They were first asked to complete a short questionnaire, noting their age, gender, and whether they believed that genuine paranormal phenomena might sometimes take place during seances.

A seance room had been prepared. All of the windows and doors in the room had been sealed and blacked out, and twenty-five chairs had been arranged in a large circle. Three objects — a book, a slate, and a bell — had been treated with luminous paint and placed onto three of the chairs. A small table, the edges of which were also luminous, was situated in the middle of the circle. Two luminous maracas rested on the table.

Following a brief talk on the aims of the project, the participants were led into the darkened seance room. Richard Wiseman played the part of the medium. With the help of a torch, he showed each person to a chair, and, where appropriate, asked them to pick up the book, slate, or bell.

Next, he drew participants’ attention to the table and maracas. Those participants who had picked up the other luminous objects were asked to make themselves known, and the “medium” collected the objects one by one and placed them on the table.

He then pointed out the presence of a small luminous ball, approximately 5 centimeters in diameter, suspended on a piece of rope from the ceiling. Finally, he took his place in the circle, extinguished the torch, and asked everybody to join hands.

The medium first asked the participants to concentrate on trying to move the luminous ball and then to try the same with the objects on the table. Finally, the participants were asked to concentrate on moving the table itself. The seance lasted approximately ten minutes.

Clearly, it was important that some phenomena occurred to assess the reliability of eyewitness testimony. The maracas were therefore “gimmicked” to ensure their movement during the seance. In the third seance the table was also similarly moved by trickery. Finally, we also used trickery to create a few strange noises at the end of each seance.

All of the ungimmicked objects were carefully placed on markers so that any movement would have been detectable. After leaving the seance room, the participants completed a short questionnaire that asked them about their experience of the seance.

No genuine paranormal phenomena took place during any of the seances. However, our questionnaire allowed us to assess the reliability of participants’ eyewitness testimony.

Would participants remember which objects had been handled before the start of the seance? As the maracas were gimmicked, we had to ensure that they were not examined or handled by anyone. Nevertheless, one in five participants stated that they had been. This was an important inaccuracy, as observers are likely to judge the movement of an object more impressive if they think that the item has been scrutinized beforehand.

This type of misconception was not confined to the maracas. In the first two seances, the slate, bell, book, and table remained stationary. Despite this, 27 percent of participants reported movement of at least one of these. In the third seance the table was gimmicked so that it shifted four inches toward the medium, but participants’ testimony was again unreliable, with one in four people reporting no movement at all.

An interesting pattern develops if the results are analyzed by separating the participants by belief. The ball, suspended from the ceiling, did not move at any time. Seventy-six percent of disbelievers were certain that it hadn’t moved. In contrast, the same certainty among believers was only 54 percent. In addition, 40 percent of believers thought that at least one other object had moved, compared to only 14 percent of disbelievers. The answers to the question “Do you believe you have witnessed any genuine paranormal phenomena?” perhaps provide the most conclusive result for the believer/ disbeliever divide. One in five believers stated that he or she had seen genuine phenomena. None of the disbelievers thought so. This would suggest that while we are all vulnerable to trickery, a belief or expectation of paranormal phenomena during seances may add to that vulnerability.

The results clearly show that it is difficult to obtain reliable testimony about the seance. Indeed, our study probably underestimated the extent of this unreliability as the seance lasted only ten minutes and participants were asked to remember what had happened immediately afterward.

Although a minority of participants believed that they had observed genuine paranormal phenomena, it does not seem unreasonable to assume that these individuals might be the most likely to tell others about their experience. Our results suggest that many of their reports would be fraught with inaccuracies and it might only take a few of the more distorted accounts to circulate before news that “genuine” paranormal phenomena had occurred became widespread.

In short, there is now considerable evidence to suggest that individuals’ beliefs and expectations can, on occasion, lead them to be unreliable witnesses of supposedly paranormal phenomena. It is vital that investigators of the paranormal take this factor into account when faced with individuals claiming to have seen extraordinary events. It should be remembered, however, that such factors may hinder accurate testimony regardless of whether that testimony is for or against the existence of paranormal phenomena; the observations and memory of individuals with a strong need to disbelieve in the paranormal may be as biased as extreme believers. In short, the central message is that investigators need to be able to carefully assess testimony, regardless of whether it reinforces or opposes their own beliefs concerning the paranormal.

Accurate assessment of the reliability of testimony requires a thorough understanding of the main factors that cause unreliable observation and remembering. Research is starting to reveal more about these factors and the situations under which they do, and do not, occur. Indeed, this represents part of a general movement to increase the quality of the methods used to investigate psychic phenomena (Wiseman and Morris 1995b). Given the important role that eyewitness testimony plays in parapsychology, understanding observation is clearly a priority for future research.