Scientists' confidence might be shaken if they were forced to acknowledge that they are unwitting actors in their own experiments. But, as A.J. ELLISON shows, some have already had to face up to this possibility

NEW SCIENTIFIC THEORIES that do not seem to agree with what is thought to be known are frequently rejected without a careful examination. Innovators who have been able to overcome their own educational conditioning to create original ideas are obliged to put up with this sort of response from the scientific 'establishment' until the day when their ideas become generally accepted - and, probably, become a new orthodoxy. The idea that scientific investigators and their collaborators are an integral part of their own experiments is such an idea.

I had an experience of the inability of scientists to face this possibility in my own laboratory. I was studying a well-known British psychic. He was apparently creating paranormal physical effects in a complex apparatus in which an infra-red beam was produced and its intensity measured by electronic circuitry. The psychic was apparently able to cause a sudden drop in the reading indicating the beam's intensity. (Whether it was the beam's actual intensity or the measuring instrument that was affected is uncertain.)The effect was repeated a number of times and seemed so clear and definite that I fetched three colleagues who were not part of the team, as independent witnesses. All fully understood the 'normal' electrical engineering and physics. They watched the effect being produced several times to order. Two of them were fascinated, declaring that they did not understand how it could possibly occur. They readily agreed to their names being quoted as witnesses. The third stated that 'there must be an explanation', even if he had not yet found it, and practically ran from the laboratory. He reduced the stress of this clash between what he was seeing with his own eyes and the received ideas of orthodoxy by refusing to admit the facts at all.

Brian Inglis, writer on the paranormal and consultant to The Unexplained, has come across examples of biological experiments in which success or failure apparently depended on psychic influences from the researcher

Every researcher knows of experiments that did not give the expected results; the phenomenon is very common. Usually it is assumed that something went wrong and the experiment is repeated until it does give the 'correct' result. One wonders how many research students who do not clearly understand what the result of an experiment is supposed to be produce anomalous results -sometimes by their own unrecognised psychic ability - and are told to repeat the procedure until they get a more acceptable result. Perhaps the spoon-bending children who appeared following Un Geller's television shows were able to produce paranormal results because they did not know that they were impossible - at least, according to their physics teachers.

The psychical researcher Rhea White has made the point forcefully: the experimenter has been a neglected variable in parapsychological research. . . . There could hardly be a more significant area of investigation than the role of the experimenter, because not only may the achievement of extrachance results depend on the experimenter, but the experimenter may also affect the nature of the results obtained.

Perhaps the most important factor in successful experiments in psychical research is intense and sustained enthusiasm, and a desire on the part of the experimenter to get the best out of the subjects. This level of enthusiasm scemed to be present during the experiment with the British psychic mentioned above. It appears when thc right people are present, and it must be carefully nurtured. Such an occasion is like a comet: its imminent arrival can be recognised, but it is not repeatable to order. The impressive paranormal events that can be brought about spring up and die down, and the skilful researcher must be ready for them -equipped to record them on audio- and videotape, and by other means, and to have witnesses present.

This phenomenon is by no means confined to psychical research. Brian Inglis describes the experience of the noted biologist Neil Miller. Miller wanted to find out whether rats could learn to control certain bodily functions - an ability that would upset conventional ideas about the workings of their nervous systems. With difficulty he found an assistant willing to collaborate on the experiments. They discovered that rats could indeed learn to alter blood pressure, heart rate, the temperature of one ear independently of the other, and other functions. Even though this sounded wildly improbable, other researchers were able to repeat the experiments and published their results during 1959.

The favoured few Psychologists are learning a great deal about the surprising ways in which human beings can influence the results obtained insupposedly objective tests. A classic demonstration was made by Robert Rosenthal and flenore Jacobson in an American elementary school. They asked teachers to administer a nonverbal intelligence test to children in all six grades (ages 5 to 11). They then misled the teachers: they told them that the results indicated that certain children, who were named to the teachers, would show marked gains in their scores during the coming year. In reality these children had been selected randomly. They represented one fifth of all the pupils. Eight months later the teachers again administered the tests. The graph (left) shows the results. In all grades there was an average increase in test scores; but in grades I and 2 the children who had been marked out as promising made much greater gains than the rest. Some of the classes were tested by an educator from another school. She obtamed the same results. In some unknown way the teachers' expectations about their pupils had called forth improved results. Such effects can appear in very unexpected places. Before this work with schoolchildren, Rosenthal had studied behavioural scientists working with rats. Some researchers were told that their rats had been bred for intelligence; others were told that theirs had been bred for dullness. The results they obtamed in such apparently straightforward tests as maze-running confirmed the supposed brightness or stupidity of the rats - which were, of course, a randomly selected sample.

Several years later Inglis visited Miller and discovered that he had been unable to repeat the results at this later date. The progressive decline of the results was inexplicable to Miller. He might have been completely discredited as a scientist if it had not been found in the 1960s that human beings could learn to control bodily functions, such as heart rate, previously thought to be completely automatic and independent of conscious control. Inglis suggests that psychic abilities of Miller or his assistant,or someone else involved in the experiment, might have played a role in his early success -and for some reason this ability declined subsequently.

A similar decline in experimental success occurred with Albert B. Sabin, the famous discoverer of a polio vaccine. He thought he had obtained reliable evidence that the virus that causes herpes also caused many human cancers. Later he was quite unable to obtain the same results.

There are many examples of carefully controlled scientific trials that have produced such strongly conflicting results that the effects of the experimenters' beliefs and attitudes must be seriously considered. The best known data on the effects of the experimenters' expectations about the outcomes of the experiment were provided by Robert Rosenthal in 1966. They are widely quoted, but also widely criticised. However, he makes it clear that careful thought must go into the design of experiments in human behaviour, and experimental psychologists must be suspected of the same kinds oferrors as parapsychologists.

Albert B. Sabin shows how his polio vaccine is administered. At one time Sabin thought he had demonstrated that certain human cancers are carried by viruses; but after a while he found himself unable to obtain the same experimental results. Was the change due to random disturbances - or to some change in his own attitude to his work?

An interesting study was carried out by the American parapsychologist J.C. Crumbaugh in 1958. He had been unsuccessful in his attempts to demonstrate paranormal effects and had decided that the personality of the experimenter must be an important factor. With the help of the staff at the Parapsychology Laboratory at Duke University, North Carolina, he designed a series of experiments. He worked with 16 subjects and 16 experimenters, working in long and short sessions. He studied the personalities and attitudes of both subjects and experimenters, and rated them according to their degree of self-confidence or insecurity, and their degree of belief or disbelief in ESP. He obtained results in the expected direction -that is, with the self-confident participants who believed in the possibility of success. He then refined the experimental design and repeated it, using only long sessions. No significant ESP results showed up.

Nevertheless, though he used 16 experimenters, Crumbaugh remained the principal experimenter: it is still possible that his own conviction that he could not succeed in ESP experiments inhibited his results. For there is no doubt that different experimenters obtain different results in similar psi tests.

Brian Inglis delivered the J.B. Rhine lecture to the Parapsychological Association in 1980. Its theme was of great importance in this area. His title was: 'Power corrupts; scepticism corrodes'. Many parapsychologists tend to be sceptical of the existence of the phenomena they study, though often they are quite unconscious of this resistance. Consequently they refuse to accept almost any effect, no matter how unequivocal the evidence for it, for fear of being 'conned'.

The orthodox scientist will protest at this point. 'This is manifest nonsense!' he will say. 'You are suggesting that whatever nonsensical pseudo-scientific idea anyone comes up with is verifiable!'

Not at all: this does not necessarily follow. Take the morphogenetic field as an example (see page 1461). The field was postulated to explain the course of past evolution, and the development of living individuals throughout the course of their lives. These processes existed long before there was any Rupert Sheldrake to propose the existence of the field. But if its existence is confirmed at some future date, it will be by the demonstration of further effects and properties of the field manifesting themselves in new experimental arrangements. And it may be that, initially, some experimenters are more successful in demonstrating it than others. It is possible that these discoveries will not be made until scientific thinking is substantially in accord with them, when there will be a great weight of detailed thinking behind the notion of the morphogenetic field.

The opinions prevailing among the nonscientific public may also be highly important in determining what experimental results can be obtained. And these can be swayed by the opinions of noted scientists. However, it is probably the whole formed by the subjects and the experimenter that is of primary importance.

When Einstein's theory of relativity became the orthodoxy of the scientific world, conflicting theories and experiments had difficulty in gaining a hearing - just as relativity itself had at first been rejected as inconceivable. The most distinguished physicist to bring contrary evidence was an American, Dayton C. Miller. He repeated the experiment of A.A. Michelson and E.W. Morley, first performed in 1887, which had come to be regarded as the cornerstone of the experimental evidence for relativity. It involved comparing the time of travel of two light beams along different paths. When Michelson and Morley performed the experiment, they expected that the 'wind' of ether (the hypothetical medium in which light waves travelled) would affect a beam travelling parallel to the Earth's motion more severely than a beam travelling at right angles to it. But they could find no difference, and it was gradually accepted that the ether does not exist. (See page 803.) From 1921 to 1926 Miller repeatedly carried out the Michelson-Morley The favoured few experiment. His results varied - but he was convinced that they revealed an 'ether wind' with a speed of about 6 miles per second (10 kilometres per second) -one third of the Earth's speed around the Sun. This work was sufficiently respected to win Miller a prize of $1000 from the American Academy of Sciences. No one had demonstrated any flaw in it by the time of Miller's death in 1941. But the implication that ether drift was detectable was rejected. Other experiments supported relativity, and in due course the Michelson-Morley experiment was repeated with other types of radiation, such as radio and radar, with no sign of ether drift. In the 1950s Miller's results were closely analysed, and the investigators concluded that his results had been partly due to temperature fluctuations and partly due to random disturbances. But long before this reasoned criticism was made, the mass of scientists had assumed that something 'must' be wrong with such heretical results.

Patterns of guessing

	Left: S.G. Soal conducted a classic series of experiments apparently demonstrating the existence of telepathy. But he had altered certain figures in his results, thus giving his subjects a spuriously high success rate
Above: J.B. Rhine and his wife Louisa. The most impressive demonstrations of ESP by the Rhines occurred early in their career, when, as J.B. Rhine believed, their enthusiasm was at its height

There are, of course, ways in which the attitudes and beliefs of experimenters and subjects can influence the results without involving paranormal agencies. Subjects frequently tend to avoid calling the same card twice in a row during card-guessing experiments. This may be a quite unconscious habit, or it may be due to a belief that 'lightning doesn't strike twice in the same place'. In fact, if the drawing of the cards is truly random, all cards are equally likely to be drawn on any occasion, no matter what the previous card has been (provided each card is returned to the pack before the next draw). Someone else might have a preference for calling a particular type of card, and so on. If the card-drawing is not completely random, the biases of the subject may result in scores that are significantly above or below those that would be expected by pure chance - even if no paranormal factor is involved.

There is a tragic and highly controversial example of what may be an unusual case of the experimenter effect. The important experiments of S.G. Soal, a former president of the British Society for Psychical Research, have been widely regarded as outstanding in their class. They were card-guessing experiments that apparently demonstrated the existence of telepathy. A later computer analysis of his results indicated that the figure 1 had been altered to 4 or 5 in various places. Subsequently an independent observer at Soal's experiments claimed that she had seen him making such alterations. (See page 14.)

This was a shattering discovery to parapsychologists. Some of them have seriously suggested that an alternative to conscious fraud is possible. The changes may have been made quite unwittingly by Soal: his subconscious mind may have seized on this as a way of obtaining the results that he so passionately desired.

Cynics will scoff at this explanation, of course. They will classify Soal with those physical mediums who have been caught cheating when supposedly in trance during seances. But I am fairly sure that in at least some cases these mediums too were not cheating consciously and deliberately: they were in an 'altered state of consciousness', in which their normal mind was not in control.

Further reading Fritjof Capra, The tao of physics, Fontana 1976 R. Rosenthal, Experimenter effects in behavioral research, Appleton- Century-Crofts (New York) 1966 B.B. Wolman (ed.), Handbook of parapsychology, Van Nostrand Rheinhold (New York) 1977

Of course, it is deplorable that the experimental conditions during these seances and during Soal's experiments were so faulty that this 'cheating' could take place. We are learning rapidly, and the rate of learning is increasing. We can now record the activities of all participants in an experiment, and analyse the results completely automatically, thus circumventing most of these earlier difficulties. Thus the enthusiasm and optimism of experimenters and subjects are able to contribute to the success of the experiment, without distorting the results.

What the psychical researchers are bringing out - and this is perhaps the special importance of the topic - is the influence of human views and expectations in all experiments. This factor may well be the explanation of the anomalous results that do not find their way into the orthodox scientific literature. Gradually it is being appreciated by scientists - but not yet sufficiently.

On page 1494: some demonstrations of the experimenter effect are surveyed