Riverhead DooksPublished by The Berkley Publishing Group200 Madison AvenueNew York, New York 10016

Copyright 1995 by Rupert SheldrakeBook design by Brian MulliganCover design 1995 by Richard Rossiter

I

All rights resbrved. This book, or parts thereof,may not be reproduced in any form without permission.

first American ledition: 1995Riverhead trade paperback edition: October 1996Riverhead trade paperback ISBN: 1-57322-564~9

TI1e Putnam erkley World Wide Web site address ishttp://www.berkley.com

The Library of Congress has catalogued the Riverhead hardcoveredition as follows:

Sheldr,ke, RUP9rr. .Seven experiments that could change the world : a do-it-yourself

guide to revolutionary science / Ruper-t Sheldrake.p. cm.ISBN 1-57322-014-0Includes bibliographical references ;md index.1. Science------Experiments-Popular works. I. Tide.Q1B2.3.S4B 1995 95-12047 CIP507'.24--

I\CHAPTER SEVEN

THE EFFECTS OF

EXPE~IMENTERS

EXPECTA TIONS

SELf-FULFILLING r~OrHECIES

Frequently things turn out just as expected Of prophecied, notbecause of a mysterious knowledge- of the future but because peo-ple's behavior tends to make the prophecy came true. Fr exam-pIe, a teacher who predicts that a student will fail may rreat thestudent in ways that make failure more likely, thus fulfJling theoriginal prophecy. Thc tendency for prophecies to be self-fulfdlingis weil known in the realms of econoncs, politics, and religion. Iris also a matter of practiC3.1 psychology. Various ways of using thesepowers are the bases of coumless self-help boaks, showing howavoiding negative attitlldcs and adopting positive ones help to bringabout remarkable successes in politics, business, and love. Likewiseconfidence and optimism play an important part in the practice ofmedicine and hnling-and in sports, fighting, and many atheractivities.

However Olle chooses to interpret it, positive and nCf,r.Hive ex-pectations often intluence what actually happens. Self-fulfillingprophecies arc commonplace. So how does this apply (0 seicnee?

209

SCIENTIFIC lLLI 'JNS-------

Many scientists cJrry out experiments with strong expectationsabout the outrome, and with 'deep~rootedassumptiom about whatis and what is not possible. Can their expectations influence theirresults? The answer is yes.

First, expecta1tions affect the kinds of questions that are asked inexperiments. And these questions in turn share what kinds ofanswers will be found. This is explicitly acknowledged in quantumphysics, where the design of the experiment determines what kindof o~tcome is possible; for example, whether the amwer will be inwave or particle form. But this principle is perfectly general. "The'structure of the examination is like a stenciL It determines howmuch of the total truth will appear and what pattern it will sug-gest." 1

Second, experimenters' expectations affect what they observe,giving a tendency to see what they want to see and to ignore whatthey da no~ want to see. This tendency can lead to unconsciousbiases in observation and in the recording and analysis of data, tothe dismissal of unfavorable results as errors, and to a very selectivepublication of results, as I discussed in the Introduetion to Part 3.

Third, and more mysteriously, experimenters' expectations mayafTect what actually happens. Just how mysterious this process mightbe is the quelstion this chapt~r explores. !

rXPElUMlNTEI< tFFlCTS

A pioneeriIig piece of industrial research, carried out at the Haw-thorne plant of the Western Eleetric Company in Chieago in1927-9, has become fam.iliar to generations of students of socialpsychology. It revealed what is now generally known as the "Haw-thorne etTect. "2 This study was designed to find out the effects onproductivity of various dunges in rest periods and refreshments.But, to the surprise of the investigatofs, output increased by about30 percent irrespective of the particular experimental treatments.The attention they were being given had a greater effect on the

210

TllE EFFECTS OF EXPERIMENTERS' EXPECTATlONS

workers than the particular physical conditions they were workingunder.

The Hawthornc effect may play a part in many kinds of re-search, at least in psychology, medicine, and animal bchavior. In-vestigators affect the subjects of their investigation merely bypaying attention to them. Moreover, they may not only luve J.general influence, owing to their attention and interest, but also 3specific influence on the way their subjects behave. In general,subjects tend to behave in accordance wirh experimemers' expec-tatiollS.

The tendency for experiments to yield the expeeted remIts isknown as the "experimcnter eifect," or more precisely the "exper-imenter expectancy effeet." Most researchers in the behavioral andmedical sciences are well aware of this tendency and try to guardagainst it by the use of "blind" methodologies. In "single-blind"experiments, (he subjeets do not know what treatment they arebeing subjected to. In "double-blind" experiments, the experi-menters do not know either. The treatments are coded by a thirdparty, and the experimenter does not know the code until after thedata have been collected.

Important though experimenter effects are in research on hu-man beings and aninuls, no one knows how widespread they are inother fields of seienee. The conventional assumptioll is thJt experi-mentereffects are widely enough recognized already, and are con-fined to animal behavior, psychology, and medicine. They arelargdy ignored in other areas of science, as can easily be seen byvisiting a scientifie library and looking through the journals indifferent fie1ds. In research in biology, chernistry, physics, and en-gineering, double-blind mcthods are rarely, if ever, employed. Sci-entists in these fields are generally innocent of the possibiliry thatexperimenters rnight ullconseiously affect the systems they arestudying.

Lurking in the background is the alarming thought that much ofestablished scicnce 1113y reftect the influence of the experimenters'expectations, even through psychokinetic or other paranormal in-

2/ /

SCIENTlflC ILLt )NS

fluences. These expectations may not on1y include those of indi-vidual investigators but also the consensus among [heir peers.Scientific paradigms, models of reality shared by professionals, havea gteat influence on [he general pattern of expectation and couldinfluence the 6utcome of countless experitilents.,

Ir is sometimes suggested, in a joking way, that nudear physicistsdo no~ so rnuch discover new subatomic particles as invent them.Ta start with, the particles are predicied on theoretical grounds. Ifenough professionals believe they are likely to be found, costlyaccelerators alnd coUiders are built to look for thern. Then, sureenough, the expected particles are detected, as traces in bubbleehambers or on photographie fums. The more often they are de-tected, the easier they become to fmd again. A new consensus isestablished: they exist. The success of this investment of hundredso[ millions of dollars then jl1stifies yet further expense on eVCI1bigger atom smashers to find yet more predicted particles, and soon. The only limit seerns to be set not by nature herselfbut by thewillingness of the U.5. Congress to go on spending billions ofdollars on this pursuit.

In (he physical sciences, although there has been very little em-pirical research on experimenter effects, there luve been manysophisticated discussions of the role of the observer in quantumtheory. Such observers, discussed philosophically, sound like thedetached minds of idealized objective scientists. But if the activeinfluencc ofthe experimenter's mind is taken seriously, then manypossibilities open up-even the possibility that the observer's mindmay have psychokinetic powers. Perhaps "mind over matter" phe-nomena take place in the microscopic realm of quantum physics.Perl1aps the mind can influence the probabilities of happeningswhich are "probabilistic," not rigidly determined in advance. Thisidea is (he basis o~much speculation among parapsychologists,3 aodis one way of trying to explain the interaction of mental and physi-cal processes in the br:lin. 4

In the realm of animal behavior, as I disCllSS below, there isac tual exper~mental evidence for (he effects bf experimenters' ex-pectations on the behavior of animals. But in most areas ofbiology,

I

212

TI-lE EFFECTS OF EXPERIMENTERS' EXl'ECTAT10NS

the possibiliry of such effects is usually ignored. An embryologist,for examplc, may well recognize the need to guard against biascdobservation aod to use appropriate statistical procedures but is Ull-likely to take seriously the idea that his expectations ean, in samemysteriolls way, influence the development of embryonic tissuesthemselves.

In psychology and medicine, experimenter effects are generallyexplained in terms of influences transmitted by "subtle cues." Bmjust how subde these cues may be is another question. It is gener-ally assumed that (hey depend on1y on recognized fOrIllS of sensoryconullllnicaon, in turn dependent on1y on the well-known prin-ciples of physics. The possibiliry that they inc1ude "paranormal"influences such 35 telepathy and psychokinesis is not discl1ssed inpolite scientific society. I bel.ieve that it is better to face this possi-bility than to ignore it, and propose an investigation of experi-rnetHer efieets tlut takes imo account the possibility of "mind overmatter" effects. But first it is importanr to consider what has a1-ready been established.

How rEOPLl BlHAVl AI EXI'[CHD

People generally behavc as expected. If we expect people to be[riendly they are more like1y to be so than if we expect thcm (0 behostile, and treat thern accordingly. The patients of Freudian ana-lysts tend to luve Freudian dreams, while patients of Jungian ;lna-lysts luve Jungian dreams. There are countless examples from altrealn1S of human experience (hat illustrate this principk.

Compared with the richness of personal experience and anec-dotal accounts, experiments on the cffects of cxpectation on peo-pIe's behavior seem contrived and trivial. Nevertheless, they areimportant in that they enable this eifect to be investigated empiri-cally and brought within the realm of scientific discourse. Andhundreds of experimellts havc indeed shown dut expcrimenterscan affect the outcome of psychologieal investigations, biasingthem in the direction of their expectations.5

213

SCIENT!FIC ILLU

. ,

Here 1S one exafl1ple. A group of fourteen psychology graduatestudents was given "special training" in a "new method oflearningthe Rorschach procedure," in which they would be asking peoplewhat patterns they saw in ink blats. Seven of thern were led tobelieve that experienced psychologists obtained more human thananimal images from their subjects. The other seven were given thesame ink blats but told that they had been found by experiencedpsychologists to give rise to a higher proportion of animal images.Surb enough, the second group obtained significantly more animalimages than th~ first. I I

Less trivial is the empirical demonstration that the etlects ofsuch e~pectations are not confined to short-term laboratory exper-iments. In schools, for instance, the way teachers treat pupils andhence the way the children learn is strongly influenced by expecta-

,

tions. The textbook example is called the "Pygl11alion experi-ment," carried out in an elementary school in San Francisco by theHarv:ud psychologist Rohert Rosenthal and his colleagues. Theseprestigious scientists created expectations in the teachers that cer-tain children in their classes were about to bloom intellectually andwould show remarkable gains in the current school year. The psy-chologists created this beliefby adrninistering a test to all the chil-dren in the school, describing je as a new technique for predictingintellectual "blooming," calling it the "Harvard Test of InflectedAcguisition." Within each dass, the teacher was then given thenames of the 20 percent of the children who had scored highest. Infact, it was an ordinary non-verbal intelligence test, and the namesof those most likely co "bloom" were chosen at randorn.

At the end of the school year, when all the children were testedagain with the same intelligence cest, in the first grade, the "prom-isillg" children scored an average of 15.4 IQ points more than thecontral children; in the second grad'e 9.5 points more. Not onlydid these "promising" children tend to SCOre better, but there wasalso a tendency for teachers to rate thern as more appealing, ad-juseed, affeceionate, curious, and happy. This dIeer showed upmuch less from the third grade upwards, probably because theteachers had their 4wn expectations about the children; the expec-

214

lE EFFECTS F EXl'ER1MENTERS' EXPECTATlNS

tations created by RosenthaI and his colleagues had rnuch less er-feet when they had to compete wirh established reputations. 6

Many subsequent studies have conflTllled and amplifled these con-

clusions.7

A criticism levded against RosenthaI and his colleagues was thattheir own strang commitment to fmding experimenter efTects hadbiased their own results. Rosentl1al replied that if this were so Itwould rnerely prove his point in another way:

We could perform a study in which we randornly dividedexpectancy investigators into two groups: in the first, expec-tancy experiments would be conducted as usual, while in thesecond, special safeguards would be employed so that theexpectancy of the principal investigator could not be COlD-muoicated to the experimenters. Suppose that the averageexpectancy effect for the first group waS seven, and fr thesecond, zero. We would still view this as evidence for thephenomenon of expectancy effects!8

Although in the medical and behavioral sciences double-blind pro-cedures are routinely ernployed to guard againsl: experimenter ef-fects, these methods are only partially effective. Some effects ofexpectancy still persist, and are most clearly seen in the placeboeffect in medical research.

THE PLACEBO tFFECT

Placebos 3re tre:ltments wich no speciflc therapeutic value whichneverthe1ess help to make rnany people better. Medic31 researchershave found that placebo effects are all-perv3sive in rnedicine, Ifphcebo effects are not controlled in therapeutic studies, ehe find-ings are generally considered unreliable. ~lacebo effeccs have beenfound in many conditions, including cough, mood changes, anginapeccoris, headache, seasickness, anxiety, hypertension, statusasthmaticus, depression, COlTlIl10n cold, lymphos3rcon13, gastric se-

215

SCIENTIFIC ILLI NS \'HE EHECTS OF EXPERIMENTER$' EXPECTATIONS

cretion and motiliry, dermatitis, rheumatoid arthritis, fever, warts,insomnia, and pain symptoms from a variety of sources.9

Much of the success of therapy through the ages can be attrib-tlted to the placebo effect, irrespective of the kind of therapy, or ofthe theories supporting it. Arid there can be no doubt that it plays alarge role in modern medicine as weil. A survey of a wide range ofdrug trials has revealed that placebos are, on average, about a thirdto a half as effective as specific medication-a big eifect for blankpills that cast almost nothing. But placebos are not just blank pills.They can also be forms of blank counseling or psychotherapy, oreven blank surgery. For exampIe, one surgical procedure for thetreatment of angina pain involved the binding of the mammaryarteries. When the effectiveness of this procedure was tested, theappropriate incision was made in control patients, but no arterywas bound. "Relief from the angina pain was the same among thereal and sham surgery groups. In addition, both groups showedphysiological changes, including reduction in the inverted T -waveof the EKG recording."1 0

So what are p~cebos? The history of the word itself is revealing.It is the first word of a chant in medieval funeral rites, "placebodomino"-I shall please the Lord. II The word was used to refer toprofessional mourners who were paid to "sing placebos" at the bierofthe decehsed in place ofthe family, whos~ role it was originally.Over the course of scveral centuries the connotations of the termgradually became derisive; it was used to describe flatterers, syco-phants, and soeia1 parasites. It first appeared in a medical dictionaryin 1785, in a pejorative sense, defmed as "a eommonplace methodor medicine."I2

The professional placebo singers in the Middle Ages no doubttended to lack any speciflc devotion to the deceased. Neverthelesstheir c11anting was generally believed to be of value as part of anacknowledged ritual. Modern placebos are given in a therapeuticsetting, and also depend for their power on current beliefs andexpectations, both of the doctor and the patient. Any method oftreatment in apy culture, traditional'or modern, occurs in a context

216

in which the particular techniques are viewed by the patient asplausible anel the therapists as potentially effective.

Doetors are often quick to ascribe the efficacy of traditional or"unscientific" medicaJ systems to the placebo phenomenon, andalso to impute the lIse of placebos to other kinds of physician. Dutthey tend to exempt their own kind of mcdicine. In one survey ofattitudes to placebo effects, surgeons excluded surgery, internistsexcluded medication, psychotherapists excluded psychotherapy,and psychoanalysts exc1uded psychoanalysis. 13 Moreover, in medi-cal research, placebo effecrs are generally regarded as a nuisance.But perhaps the negative attitudes of physicians to placebos is justas weil, since it is the other siele of the coin of their faith in thespecial efficacy of their own techniques, which therefore tend towork better-because of the placebo effect!

Tbe largest placebo effects oeeur in double-blind trials in whichboth patients and physicians believe a powerful new treatment isbeing used. If the treatment is believed by the doctors to be lesseffective, a smaller placebo effeet is obtained. In single-blind trials,in which thc doctors know which patients luve been given eheplacebo but the patients do not, placebos are still less effcctive. Inopen conditions where the patients know they are receiving place-bos, the efTects are smallest of all. In other words, treatments workbest if they are thought to havc powerful beneflCial efTects by bothdoctors and patients. Conversely, in trials where the aetive medica-tions are labelcd as placebos, the drugs give poorer clinical rc-sults. 14

Thus lowcred expectations lead to a lowered placebo efTecLThis is the case wich new "wonder drugs" that arouse high hopesto start with, but fail to live up to expectations. This pattern wasrecognized by the nineteenth-century French physician ArmandTrousseau, who advised his colleagues to "treat as maHy patients aspossible with the new drugs while they stiU have the power toheal. "15 There are many modern examples. For instance, at onetime the drug chlorpromazine was hailed for its efficacy in treatingschizophrenia, but then [aith in its powers gradually wJned. In

217

SCIENTIFIC ILL '1'15 TUE EFFECTS GI-' EXPERIMENTERS' EXPECTATlNS

successive trials it was found to be less and less effective. The effects,

of placebos declined in paralleL "Perhaps as the investigators beganto realize that the new 'wonder drug' was not as powerful as theylud hoped, their expectations, and possibly their interest in thepatiems declined. "16 Here is another particularly striking example,from the 1950s:

A man with advanced cancer was no longer responding toradiation treatment. He was given a single injection of anexperimental drug, Krebiozen, considered by same at thetime to be a "miracle eure" (it has since been discredited).The results were shocking to the patient's physician, whostated that his tumors "melted like snowballs on a hot stove."Later the man read smdies suggesting the drug was ineffec-tive, and his cancer began to spread once more. At this poinehis doctor, acting on a hunch, administered a placebo intra-venously. The man was cold the plain water was a "new,improved" form ofKrebiozen. Again, his cancer shrank awaydramatically. Then he read in the newspapers the AmericanMedical Association's official pronouncemcnt: Krebiozen wasa worthless medication. Thc man's faith vanished, and he wasdeJ.d within days.17

IThe same principles apply to medical research itself. Believers andIlon-believers in new forms of treatment tend to obtain very differ-ent results: "Quantitatively, the pattern is consistent. The initial 70to 90 perdent effectiveness in the enthusias~s' reports [decreased] to30 to 40 percent 'baseline' placebo effectiveness in the skeptics'

repor~."18A remarkable feature of placebos is that patients not only benefit

from them, but also exhibit coxic responses or side effects. In onesurvey of sixty-seven double-blind drug trials involving 3,549 pa-tients, 29 percent of the patients showed various side effects whilethey were being treated with the placebo, including anorexia, nau-sea, headache, dizziness, tremor, and skin rash. 19 The side effectswere sometimes so severe that they required additional medical

71R

intervention. Moreover, they showed a relationship to the doctors'or patients' expectatiollS about the active drug being used in thetria1.20 For example, in a large-scale double-blind trial of oral con-traceptives, 30 percent of the women who were administered theplacebo reported decreased sex drive, 17 percent increased head-ache, 14 percent increased menstrual pain, and 8 percent increasednervousness and irritability.21

lust as thc power ofblessings is mirrored by the power of curses,so the beneficial effects of placebos are mirrored by the negativeeflects of procedures expected to bring abollt hann, technicaUyknown as "negative placebos" or "nocebos." Spectacular examplesin Africa, Latin America, and elsewhere are known to anthropolo-gists as "voodoo deaths," brought about by belief in the power ofbevvitchment. Less spectacubr nocebo effects have also been dem-onstrated in laboratory experiments, as in a study in which subjectswere told that a mild eleetric current was being passed throughtheir head by means of applied electrodes, and warned that thismight give rise to a headache. Although there was in fact no cur-rent, two-thirds ofthe subjects developed headaches. 22 Both place-bos and nocebos depend on prevailing cultural beliefs, includingthe belief in scientific medicine. "Simply put, belief sickens; beliefkills; belief heals. "23

THl INFLLllNCl OF EXrlCTANCV ON ANIMALS

Animals respond to different people differently, as every pet ownerand animal trainer knows. They recognize people they are used (0,and tend to be on their guard with strangers~ They seem to sensewh ether people are friendly, gauge their fear or confIdence, andrespond to their expectations. From a commonsense point ofview,based on everyday experience, it 1S hardly surprising that scientistswho do experiments on animals have a personal influence on theanimals. The experimenters' attitudes and expectations affen theanimals they work with.

The classical experiments on the etfects of experimenters' ex-

219

SCIENTIFIC ILLUS_

pectations on animals were carried out in the 1960s by RohertRosenthai and his colleagues. They used students as experimentersand rats as subjects. The rats came from a standard laboratary strain,but were divided at tandam iuta two groups, labeled "Maze-Bright" and "Maze~Dull." Thc students were told that these ani-mals were the ptducts of generations of selective breeding atBerkeley for good and poor performance in standard mazes. Thestudents naturally expected the "bright" rats to learn quicker thanthe "dulI" ones. Sure enough, this is what they found. Overall the"bright" ra~ made 51 percent more correct I1esponses and learned29 percem faster than the "duli" rats. 24

These I fllldiogs have been conflrmed in other laboratories andwirh other kinds of learning.2s Comparable experimenter effectsluve even been obscrved with flatworms, lowly creatures that livein mud at the bortom of ponds and in similar aquatic environ~ments. In Olle such study, a sample of essentially identical PlanariawornlS was divided into two groups, one of which was described asastrain showing few head turns and body contractions ("low-response-producing worms"), and the other as a frequent turnerami contracter ("high-response-producing wonns"). With theseexpectations in mind, the experimenters found on average fivetimes more head turns and tweIllty times more contractions in the"high-response..l.producing" worms. 26

These expectancy efTects, like those in Rosenthal's rat experi-ment, were shown by undergraduate students, who may be espe-cially prone to see: or even to pretend to see, what they are told toexpeet. Seasoned observers might generally show smalier expec-taney etTects. This was the case, for example. when more experi-enced I researchers were working wich Pfanaria. The number ofcontractions in "high-response-prodllcing" Planaria was found tobe two to sevcn times greater than in "low-response~producing"worms, compared with the average of twcnty times greater foundby uHdergraduates.

1

Nevcrtheless, a two- to sevenfold increase isstill a brge effect, jnd obviously introduces a serious bias into theresults.

On the other hand, experienced observers may be stronglyI

220

'HE EFFECTS OF EXPERJMENTERS' EXPECTATiONS

committed to particular views of reality, directly or indirectly re-sulting in greater expeetancy efTects than those found among nov~iees wirh less personal comnlitment to particular theories. Thcymay create a climate of expectation among their colleagues andtechnicians, and this in turn may influence the way their animalsbehave.

Although expectaney effects were first systematically investi-gated in the 1960s and have now been demonstrated in hundredsof special studies.27 the general principle is by HO means new. Forexample, Bertrand Russell, writing with his customary wit andclarity, put it as foliows in 1927:

The manner in which animals learn has been much studied inrecent years, with a great deal of patient observation andexperiment. One may broadly say that all the animalsthat luve been carefully observed luve behaved so as to con-firm the philosophy in which the observer believed before hisobservations began. Nay, more, they have a11 displayed thenational characteristics of the observer. Animals studied byAmericans rush about frantically, and with an incredible dis-play ofhustle and pep, and at last achieve the desired result bychance. Animals observed by Germans sir still ami think, andat last evolve the solution out of their inner consciousness. 28

EXI'ERIMENTfR EFFECTS IN PARAPSYCHOLOCY

Experimenter effects are weil known to parapsychologists, for sev~eral reasons. First, it has long been known to experienced research-ers that subjects tend to show more psychic powers when they arefeeling relaxed, and in a positive and enthusiastic atlllosphere. Ifthey are anxious, uncomfortable, or treated in a formal and de-tached way by the scientiflc investigators, they do not perform soweil. In fact they may show 110 significant psychic powers at a11, no"psi effeets," in the jargon of parapsychology.

Second, it is a matter of conunon observation among research-

221

SCIENTtflC ILLUS s -HE EFFECTS OF EXPERIMENTERS' EXPECTATlNS

crs in this field that subjects who show considerable psychic abili-ties often tend to lose them when strangers COOle ioto the room asobservers. Tbe pioneering parapsychologist J. B. Rhine actuallyquantified this efTect in aseries of trials with a gifted subject, Hu-bert Pearce, having noticed that when someone called in to seePearce at work his scores at onee dropped down. "We began totake down evidence, sometimes inviting avisitor for that purpose,sometiInes availing ourselves of a casual caller. We reeorded thetime of entrance and exit on 7 visitors, one being present twice.They all produced a drop in Pearce's scoring. "29

Thc off-puttin~ effect of strangers is particularly strang whenthe strangers are sceptical, especially if they are hostile to the exper-iment itself or to the people involved. However, if strangers arefriendly, and especially ifthey help in some way in the experiment,ratber than ~ehaving as detached observers,lsubjects get used tothem and psi scores rise again.30 Sceptics usually take the failure ofparapsychologieal tests in the presence of scepties to mean thatpsychic powers eannot be detected under rigorous scientifie condi-tions, and therefore don't really exist. But the negative effeets ofscepties may weil be due to their off~puttingpresence and negativeexpectations, mediated by subtle aod not-so-subtle cues.

Third, it is weil known among parapsyehologists that some ex-perimenters consistently obtain positive results in their research,wbile others do not. This effeet was systematieally investigated inthe 1950s by two British researehers. One, C. W. Fisk, a retiredinventor, consistently obtained signifieant results in his experi-ments. Tbe other, D. J. West, later to become Professor of Crimi-nology at Camb1ridge, was usuaIly unsuceessful in detecting psyehiephenomena. In these c>"1'eriments, each investigator prepared halfof the test items, aod seored them at the end. The subjects did not

Iknow that rwo experimenters were involved, nor did they meettbem; they received the test items through the post and also re-turned them by mail. The results from Fisk's half of the experi~ment Ishowed highly significant effeets for clairvoyance andpsyehokinesis. West's data showed 110 deviation from chance. Theyconcluded that West was "a jinx. "31

222

Fourth, in resc3rch on psychokinesis it has repe3tedly beenfound that experimenters who find significant effects are them-selves good subjeets. For eX3Il1ple, Helmut Schmidt, the inventorof the Schmidt machine, a randorn number generator whose ou(-put can apparently be affeeted by willing certain patterns toemerge, has found that he is often his Qwn best subject. 32 Oneinvestigator, Charles Honorton, has even shown that psychokineticeffecrs on randorn number generators by the subjects in his experi-ments are more due to himse1f than to his subjectS. 33 The subjeetsshowed psycbokinetic powers when he was present; aod hc himselfshowed them when he was acting as the experimental subjecL Butthe psi dIeet was lost when he was not present and the subjeetswere tested by another experimenter. Honorton and his colleagueBarksdale concluded that such etTects showed that "traditiOIl3.1boundaries between subjects and experimenters cannot be easilymaintained." They interpreted their results as a "psi-medi3ted ex-perimenter effect."34

The implications of such experimenter effects are staggering. Ifparapsychologists can bring about psi-mediated experimenter cf-feets, whether intemionally or not, through their influenee overthcir subjeets, even at a distanee (as in the Fisk-West experiments),then thc conventional separation between experimenters and thesubjects of their investigation breaks down. Moreover, if peoplecan influcnee physical events such as radioactive decay, then theeonventional separation betv/een mind and matter breaks downtoo. But then why should psi-mediated experimenter effects beconfmed to parapsycholob'Y? Might they not occur in many otherfields of scienee?

Ho\' PAf\.ANOf\.MAL Is NOf\.MAL SCILNCE?

There is a good reason for the conventional taboo against parapsy-ehology, making it a kind of outcast from established science. Theexistenee of psyehie phenomena would seriously endanger the illu-sion ofobjectivity. Ir would raise the possibility that many cmpiri~

223

SCIENTlfIC ILL__ JNS

cal results in many flelds of seienee refleet the expeetations of theexperimenters through subtle uneonseious influences. Ironieally,the orthodox ideal of passive observation may weil provide excel-lent conditions for paranormal effects:

An experimemer preparing his apparatus, getting his animalsready, and then leaving them with some feeling of assurancethat the experiment will run and the animals will appropri-ately "da their thing" cannot but remind us of certain aspectsof magie, ritual, cr perhaps petitionary prayer. Something isdone with lthe confidence that it will produce a desired result,and the participant, once he has done this, psychologicailypUls a distanse between himself and the outcome. He is nottrying to make things happen, but just trusts that they will.

Such circumstances may provide an optimum opportu-nity for psychokinetic intervention.35

I

This possibility has indeed been raised in a paper in Nature entitled"Scientists confronting the paranormal, " by the physicist DavidBohm and others. They noted that the relaxed conditions neces-sary for the appe,rance of psychokinetic phenomena are also thosemost fruitful for scientific research in general. Conversely, tension,fear, and hostility tend not only to inhibit psi effects. but also toinfluence e~periments in the so-called hard-sciences too. "If any ofthose who participate in a physical experiment are tense and hos-tile, and do not really want the experiment to work, the chanees ofsuccess are greatly diminished. "36

The defenders of orthodoxy generativ reject or ignore the possi-biUty of paranormal influences under auy circumstances. The taskof keeping s2ienee psi-free iS

Iundertaken Iby organized groups of

Skeptics. These scientific vigilantes continually challenge any evi-dence for psi effects, rejecting it on one or more of the following

I ,groun,ds:

1. Incompetent experimentation.2. Selective observation, recording. and reporting of data.

224

THE EFFECTS OF EXPERIMENTERS' EXPECTATlONS

3. Unconscious or conscious deception.4. Experimenter efIects mediated by subtle cues.

Skeptics are right to point Out these possible sources of error inparapsychological research. But the same sourees ofbias are presentin orthodox research as weil. The very fact that parapsychologicalresearch is subject to such critical scrutiny makes researchers in thisfield unusually conscious of the effects of expectation. Ironicaily, itis in conventional, uncontroversial fjelds of research that the influ-ences of experimenters' expectations are most likely to pass unde-tected.

The evidence for experimenter effects in medicine and the be-havioral sciences is undeniable. And that is why "subtle cues" takeon such an important explanatory role. Almost everyone agreesthat subtle cues such as gestures, eye movements. body posture, andodors can influence people and animals. Skeptics are very keen onemphasizing the imp~rtance of such cues, and rightly so. A favor-ite example showing the importance of subtle commumcation isthe story of Clever Hans, a famous horse in Berlin at the turn ofthe century. This horse could apparently perform arithmctic in thepresence of its owner by tapping a hoof on the ground to COUlltout an answer. Fraud seemed unlikely, sinee the owner wouldallow other people (free of charge) to question the animal them-selves. The phenomenon was scientifically investigated in 1904 bythe psychologist Oskar Pfungst, who coneIuded that the horse \vasreceiving eIues from gestures made, probably unwittingly, by theowner and other questioners. pfungst found that he could gec thehorse to give the concct answcr sirnply by concentrating his J.ttCI1-tion o"n th-e number, though he was not aware of making any1110vement that would give the number away.37

No one denies that subtle cues from experirnenters, passingthrough normal sensory channe1s, can affect people and anim3.ls.Skeptics claim that such influences rnay explain many examples ofseemii:lgly telepathie communication. Dut granted a11 this, the pos-sibility remains that both subtle sensory cues and "paranormal" in-fluences playapart.

225

SCIENTIFIC IL. .ONS

The story of Pfllngst's investigation of Clever Hans has beentold again and again to generations of psychology students. What isless weB ~nown is that after Pfungst's inveJtigation, described in hisbook on Clever Hans pllblished in 1911, further studies on horseswith bmilar rnathernatical powers showed that more was involvedthan subtle sensory cues. For exarnple, whenMaurice Maeterlinckinvestigated the farnous calculating horses of Elberfeld, he COI1-cluded that they were somehow reading his mind, raeher than re-sponding co subtle sensory cues. After aseries of increasinglystringent tests, he fll1al1y thought of one which "by virtue of itsvcry simplicity, could not be exposed to any elaborate and far-fetched suspicioIlS." He took three cards with nurnbers on ehern,shuffied thern without looking at them, and placed them facedown on a board where the horse could see only their blank backs."There was therefore, at that mornent, not a human soul on earth

Iwho knew the figures. " Yet, without hesitation, the horse rappcdout the 11tlmber the three cards formed. This experiment suc-ceeded with thle other calculating horses too "as often as I cared totry it."38 These results go even beyond the possibility oftelepathy,since Maeterlinck himself did not know the answers when thehorses were tapping thern out. They imply either that the horseswete capable of clairvoyance, directly knowing what was on thecards, or precognition, knowing the nllmber that would be inMaeterlink's mind when he later turned the cards over.

For more than eighty years, the story of Clever Hans andPfungst has beet? told and retold as a triumph of scepticism. It hastaken on a mythic significance, enabling seemingly paranormal ef-fects to be explained in terms of subtle cues. But what if some ofthe subtle lfues are themselves paranormal? There is a taboo againsteven discussing this possibility, let alone investigating it. N everthe-less, the possible imponance of parapsychological influences wassuggested to Rosenthai by one ofhis colleagues at Harvard right atthe outset of his research on experimenter effects:

Had I the wit or comage to do so, I tould easily have con-I

duc ted a study in which experimenters with varying expecta-

??;;

THE EFFECTS OF EXPEIUMENTERS' EXPECTATrONS

tions for their subjects' responses were prevented from havingsensory contact wich those subjeets. My prediction, then andnow, was (and wauld be) that under these conditions noexpectaney effects eould occur. But I never did the study. 39

Maybe if sameone actuaIly did this study, Rosenthal's predictionwould turn out to be wrong. Maybe some of the effects of experi-menters' expectations are indeed paranormal. Such subtle influ-ences would not be opposed to subtle cues; they would usuallywork along with them, and operate just as unconsciously.

Although experimenter etTects are weil recognized in the medi-cal and behavioral sciences, the fact that they are explained--orexplained away-in terms of "subtle cues" prevents thern frombeing taken very seriously in other fields of investigation such asbiochemistry. Whereas a person Of a rat might pick up a scientist'sexpectations and respond accordingly, an enzyme in a test tubewould not be expected to respond to subtle body language, uncon-scious facial gestures, ete. Of course, there is a general recognitionof the possibility of biased observation, but this is not a result ofany actual inflllence on the experimental system itself. The scientistmay "see" a difference that fits his or her expectancy, but thedifference is supposed to be only in the eye of the observer, not inthe material studied.

Nevertheless, all this is merely an assumption. There has beeilpractically no research on the inftuence of experimenters' expectJ-tions in fields of science such as agrieulture, genetics, moleculJrbiology, chemistry, and physics. Since the material studied is as-sumed to be inuuune from such influences, precautions againstthem are assumed, to be unnecessary. Except in the behavioralsciences and in clinical research, double-blind procedures are rarelyemploycd.

I now suggest a variety of tests to explore the possibility thatexperimenter effects may be far more widespread than previouslythought.

227