Evans, F. J., & Orne, M. T. The disappearing hypnotist: The use of simulating subjects to evaluate how subjects perceive experimental procedures. International Journal of Clinical and Experimental Hypnosis, 1971, 19, 277-296.



Institute of the Pennsylvania Hospital and University of Pennsylvania

Abstract: This study examines the effects of a temporary power failure while S was hypnotized during a tape-recorded session. It was necessary that the power failure be perceived by S as an accident and not as an experimental deception. In a previous study (Orne & Evans, 1966), Ss simulating hypnosis with a "blind" E continued faking throughout the "power failure," apparently suspecting they were being observed. Therefore, no conclusions could be drawn about the behavior of the hypnotized Ss. After the procedure was modified, simulating Ss ceased faking as soon as the hypnotist left the room, thereby demonstrating that they believed that a power failure had actually occurred and that they were not being watched. Hypnotized Ss spontaneously, although slowly and with subjective difficulty, terminated hypnosis by themselves. The spontaneous behavior of the hypnotized Ss in the absence of a hypnotist seems inconsistent with predictions based on motivational or role-taking theories of hypnosis.

The current major theories of hypnosis agree that individual difference, motivational, and interpersonal factors are important components of the phenomenon. Some theorists (e.g., Barber, 1969; Sarbin, 1950) have argued that these aspects are wholly sufficient to explain hypnosis, while others (e.g., Hilgard, 1965; Kubie & Margolin, 1944; Orne, 1959) insist that, in addition, hypnosis involves an altered state of the individual. Because there is considerable overlap in these views, it has been difficult to operationalize the differences. One situation in which these

Manuscript submitted January 17, 1971.

1 This study was supported in part by contract Nonr-4731 (00) from the Office of Naval Research and by grant MH 19156-01 from the National Institute of Mental Health, Public Health Service.

2 Several colleagues were involved in the successful completion of this investigation. We wish to thank Barbara Marcelo, St. Christopher's Hospital for Children, Philadelphia, for her role as the blind E and Mary Jo Bryan for her help as technician. The comments of Harvey D. Cohen, Mary R. Cook, A. Gordon Hammer, John F. Kihlstrom, William A Mitchell, David A. Paskewitz, and especially Charles Graham and Emily Carota Orne were particularly valuable. We also wish to thank Howard A. Kaiser, Betty H. Marx, Toby L. Parcel, and Susan Jo Russell for their help in various ways.




differing theoretical views should lead to different predictions is the rather unlikely event that the hypnotist is prevented from terminating hypnosis. Occasionally an anxious patient may ask, "What would happen if the hypnotist had a heart attack while I was hypnotized ?" Dynamic implications aside, this question is concerned with whether hypnosis must be terminated by the hypnotist or whether it is contingent upon a continuing interpersonal relationship and would therefore become terminated automatically if the hypnotist became unavailable. The purpose of this paper is to present an empirical test of this question.

State and Motivational Theories of the Termination of Hypnosis

"Motivational" and "role playing" theorists regard hypnosis as a special kind of interpersonal relationship between S and the hypnotist which motivates S to behave like a hypnotized person or to enact the role of a hypnotized individual. The induction of hypnosis consists of establishing this special role relationship. If hypnosis does not exist beyond this specifically defined relationship, one would predict that the disappearance of the hypnotist would automatically and immediately terminate hypnosis. If the special relationship does not continue, then hypnosis cannot exist. 3 The major difficulty with operationalizing a test of this view involves defining the hypnotic relationship as ended in a way that permits a test of the continuation of the hypnotic effects without reinstating the special relationship.

Adherents of a "state" approach argue that hypnosis involves a qualitative change in the organism. Hypnosis is a special condition within the hypnotized individual -- in particular, an altered state that allows Ss to experience distortions of perception and memory as subjectively real. Although individual difference, motivational, and interpersonal factors are important in bringing about the induction and termination of hypnosis, hypnosis cannot be accounted for by these factors alone. Regardless of how hypnosis is initiated, it involves changes within the individual and, consequently, if the hypnotist disappeared, hypnosis would not automatically cease. On the contrary, time and effort would have to be expended by the individual to reverse the

3 In our attempt to derive a testable hypothesis differentiating the two types of theories of hypnosis, we do not claim that adherents of either theory have even implicitly made these predictions. Thus, Sarbin (1950) has argued that role involvement may not be at a conscious level. It seems that such a view cannot spontaneously be distinguished from a state theory. Nonetheless, if the defining relationship is stressed as the means by which role taking is established (Barber, 1969; Sarbin, 1950; Sarbin & Andersen, 1967), then the termination of the relationship ought to terminate the need to continue the role.



altered state of consciousness and attention which these theorists believe constitutes a significant aspect of hypnosis.

What Happens if the Hypnotist Does Disappear?

Historically, a variety of opinions have been expressed about the fate of a hypnotized S suddenly and unexpectedly left alone (see Orne & Evans, 1966). Although it has been generally held that under these circumstances a deeply hypnotized S would eventually lapse into normal sleep from which he would ultimately arouse himself, little empirical evidence is available. The only directly relevant study was conducted by Dorcus, Brintnall and Case (1941). The secretary interrupted a session while S was deeply hypnotized and reminded the hypnotist -- sufficiently loudly for S to overhear -- of an important appointment downtown. The hypnotist left immediately without any word to S. Twenty Ss remained hypnotized, defined in terms of the length of time they stayed in the experimental room, for a mean time of approximately 28 minutes. The control group of 25 "relaxing" Ss remained for a shorter period of approximately 20 minutes. However, no systematic attempt was made to determine whether Ss perceived E’s departure as genuinely due to external demands or whether they recognized the situation as an experimental manipulation. Therefore the interpretation of these results remains equivocal. 4

It is difficult to envisage a situation in which a hypnotist can legitimately disappear without creating suspicion or, worse yet, disturbing the rapport between himself and S. Few appointments or emergencies are sufficiently urgent that a hypnotist who is genuinely concerned for the welfare of his S can simply disappear without having some time available to terminate hypnosis, apologize, and explain his predicament to S. It is implicitly part of the rules on conducting experimentation that E concern himself about S’s welfare. Therefore, unless the hypno-

4 Due to a printer's error, it was reported in our earlier study (Orne & Evans, 1966) that the control group mean was 23 minutes instead of 20. The difference between the two groups is still, however, insignificant (t = 1.6). As Dorcus et al. (1941) reported their data in rather broad categories, this t test might be misleading. The data were grouped in 10-minute intervals up to 50 minutes; this interval was scored as 55, which leads to a very conservative t-test. Dorcus et al. (1941) reported that the two control Ss, who remained with their eyes closed for "50 minutes and over," did so for a shorter time than the seven hypnotized Ss who fell into this category. This difference is significant (p < .05, one-tailed Fisher's exact test). The control Ss performed in a fashion similar to the simulating Ss who were not deceived, i.e., the unhypnotized Ss continued with the experiment in a way that is consonant with the expectation that hypnotized Ss would remain hypnotized.



tist's disappearance is caused through no fault of his own, the situation seems implausible. It is likely that S’s credulity was strained by the inappropriately abrupt departure of the hypnotist in the study by Dorcus et al. (1941).

A more satisfactory method of conducting the relevant experiment is suggested by the use of a tape recording to induce hypnosis. Provided S is capable of experiencing deep hypnosis when it is induced by the hypnotist's tape-recorded voice, the special hypnotic relationship or rapport, to the extent that it can be said to exist, is between the voice on the tape recorder and S. A sudden, mechanical failure of the tape recorder provides a plausible and dramatic method of interrupting hypnosis. By a fortuitous accident, hypnosis can be interrupted precipitously in a fashion which will not interfere with S’s relationship with the laboratory. The hypnotist "disappears," as it were, through no fault of his own. The crucial issue in this design then becomes: Can a sudden mechanical failure be created artificially, yet still be able to bring about a situation in which S perceives the accident as genuine? Or, does the artificiality of the situation override the more subtle aspects involved in the hypnotist's abrupt departure? 5

Deception, Plausibility, and Empirical Validation

To obtain meaningful evidence about the specific outcomes predicted by the two different theoretical points of view, it is necessary that the experimental situation be perceived by S as E intends it to be seen: the hypnotist's disappearance must be perceived as truly accidental, not as a planned part of the experiment nor as unprofessional or inappropriate. It is not sufficient to end an experiment by an act of fiat on the part of E. Such a situation undoubtedly would be highly anxiety-provoking for some Ss, and would fail to deceive others.

Evaluating the success of justifiable deception in psychological experiments is difficult. Even when the deception seems highly plausible at face value, the active role played by S in a psychological experiment (Orne, 1959; Orne, 1962; Orne & Evans, 1965) makes it mandatory that his perceptions in the situation be evaluated. Typically, a pact of ignorance is established in which S does not wish to indicate that he has seen through E's deception (for that would invalidate his own experi-

5 The relationship as such is not interfered with in the long-term view of the interaction; however, at least during the interval while the tape is not playing and the observer is not present, no object of interaction exists with whom to play the defined role or to define what a good S is supposed to do. In this reciprocal sense, if these aspects are the features which define the occurrence of hypnosis, S would have no reason for continuing to be hypnotized.



mental results), nor does E wish to discover that Ss were not as naive as he had hoped (for that would also invalidate his experiment!). 6 The psychopharmacologist has developed a partial solution with the double-blind methodology. For the social psychologist who employs deception as a tool, the problem is no less acute (Stricker, 1967).

The Simulation Design and Evaluating Deception

Orne (1959, 1962, 1970, 1971, in press) has proposed several strategies (not strictly control groups in the classical sense) for evaluating S’s perceptions about the nature of the experimental procedure. One such strategy particularly suited to hypnosis research is that of Ss' simulating with a "blind" E. A S who is insusceptible to hypnosis is asked to simulate the performance of deeply hypnotized Ss. Provided E is blind to the identity of both real and simulating Ss, he cannot reliably detect which S is which by observing the classical hypnotic phenomena. 7 This technique is particularly well adapted to exploring S’s potential awareness of deception.

There are two special features of the simulating technique: (a) Both hypnotized and simulating Ss are treated alike. As E is "blind," he cannot provide cues for the hypnotized Ss which are not also provided for simulators. (b) The behavior of simulators provides a means of evaluating the contribution of Ss' expectations and perceptions in accounting for the obtained results.

Specifically, the behavior of simulating Ss in the present study will provide evidence about two important aspects. First, their behavior during the "accidentally" produced power failure will provide clues about the plausibility of the deception. Depending on their reaction to the cessation of the tape recording, it will be possible to evaluate whether already suspicious Ss perceive the power failure as a deliberate manipulation by E or whether they accept it as a plausible accident. Only when these highly suspicious Ss accept the procedure as a genuine power failure can meaningful conclusions be drawn about what

6 For an example of a highly plausible deception and subsequent evidence suggesting that the existence of such a pact of ignorance in fact raises questions about the generality of the obtained results, see Milgram's (1965) studies on obedience and a critique of this work by Orne and Holland (1968).

7 Although special experiments can be designed to differentiate real from simulating Ss, the experimental model uses simulators as a quasi-control group to evaluate the experimental procedure. Under these circumstances, the intent is not to study whether the two groups of Ss can be differentiated. Fortunately, the cues available to the hypnotist in conducting typical hypnotic experiments are not sufficient to discriminate between these two groups. Therefore, the E remains blind as is required by the model.



hypnotized Ss would do in such an emergency. Secondly, as a further check on the validity of the experiment, the behavior of the hypnotized Ss can be evaluated against the performance of the simulating Ss in terms of how they would expect a hypnotized person to react in such an emergency.

An Unsuccessful Study -- As Indicated by Simulating Ss

A study (Orne & Evans, 1966) was designed in which a tape recorder served as the hypnotist. While S was carrying out several suggestions (his arm stuck to the chair, his eyes glued together, his foot keeping time with hallucinated music), a switch was thrown, shutting off a bright desk light and bringing the tape recorder to a slurring halt in mid-sentence --presumably a power failure. The E unsuccessfully tried to restart the tape recorder, muttered "the damn fuse," and rushed from the room. The S was then observed for 30 minutes through a one-way screen. The lights and tape were then turned on again. The details of the procedure are the same (with some important modifications) as those reported below for the present study.

Six Ss were deeply hypnotized and six other Ss had been given standard simulating instructions (Orne, 1971). The E, of course, was unaware of S’s status and could not guess better than chance which Ss were simulating and which were hypnotized.

The objective behavior of the simulating Ss in the first study is presented in Table 1. Five of the six simulators continued to simulate hypnosis, keeping their eyes closed, tapping their feet to "hallucinated" music, and keeping their arms on the chair. The reason for the continuation of the simulating behavior throughout this period was most likely that they expected that E would try to trick them in order to detect simulation. They later reported to the nonblind E (who had originally given the simulation instructions) that they actually did suspect deception and that they did not feel the power failure was genuine.

Once it became clear that the manipulation was not successful, the procedure was examined with more care. Three distinct factors in this situation seemed to make the deception less than adequate: (a) In order to facilitate observation, all experimental rooms had conspicuous one-way screens. Since all Ss had completed several hypnosis experiments in the same room, it was hoped that Ss would have become sufficiently adapted not to notice them. (b) When the "power failure" occurred, a bright desk lamp, which was plugged into the same wall socket as the tape recorder, went off simultaneously with the tape recorder. The overhead ceiling light stayed on, however, partly because it facilitated visual observation through the one-way screen. (c) In addition to the tape recorder, a small two-channel Offner polygraph



was situated in the experimental room. It was plugged into a wall socket at the other side of the room and did not stop recording when the power failure occurred. A high hum is always associated with the polygraph; its continued functioning would have been obvious to Ss even with their eyes closed. Since it is quite common for a fuse to blow on one circuit while another circuit in the same room continues to function (and since physiological data for Ss during the breakdown period were also desired), we had hoped that the situation would be plausible despite the continued functioning of the polygraph and the overhead lights.

It should be emphasized that these flaws in the experimental procedure, so blatant and obvious in retrospect, represent the kind of procedural detail which is hardly ever reported in print, nor did they appear particularly meaningful prior to the completion of the first study. Only when the situation failed to deceive the simulating Ss did we consider these design problems. Inadequacies of the procedure identified by the simulators' behavior were also corroborated by the comments made later by the simulating Ss during the post-experimental discussion.

The behavioral evidence convinced us that the design of the study had been inadequate. Highly motivated and suspicious simulating Ss had not been convinced that the power failure was an accident. Consequently, it was not possible to draw clear inferences about the behavior of the (presumably less suspicious) hypnotized Ss. It was most parsimonious to assume that they, too, could have suspected the power failure was not an accident, but rather a ruse of E. Given this possibility, the experiment would not have been testing the effects of a sudden termination of hypnosis, but rather only Ss' ability to second guess E’s ruses -- a trivial finding in this context. The only feasible conclusion



was that a valid operational procedure for testing the experimental hypothesis had not been successfully executed.

The present study was an attempt to modify the procedure in important ways so that a viable experimental procedure could be established. The role of simulating Ss was to provide objective evidence about whether the experimental situation convinced motivated, suspicious Ss that the power failure which occurred was indeed accidental.



The Ss were selected from a large population of college undergraduates who had volunteered initially to participate in hypnosis research. They had completed at least the Harvard Group Scale of Hypnotic Susceptibility, Form A (HGSHS:A) of Shor and E. Orne (1962) and the Stanford Hypnotic Susceptibility Scale, Form C (SHSS:C) of Weitzenhoffer and Hilgard (1962). After qualifying with extreme scores (upper and lower five percent of the score distributions) on these scales, Ss were clinically evaluated at least twice on the diagnostic rating scale measuring hypnotic depth described by Orne and O'Connell (1967). The Ss who still qualified at the end of this procedure were either deeply hypnotizable or essentially insusceptible Ss.

In addition to the assessment of hypnotic susceptibility, Ss had participated in at least two experimental hypnosis sessions in the room in which the reported study took place. During these sessions they became acclimated to tape-recorded hypnotic induction procedures. The Ss were told that they were participating in an extensive investigation measuring depth of hypnosis and interrelating several different methods of assessment: standardized scales, objective behavioral tests, subjective impressions of depth of hypnosis (recorded continuously by S on a circular hypnotic depth indicator, HDI), and physiological indices (including galvanic skin potential, GSP). The Ss selected for the present study were scheduled for an additional two sessions as part of this ongoing research, and insusceptible Ss received the simulation instructions reported below. The physical setting and general procedures were the same in both sessions. The accident occurred late in the last session.

Simulating subjects. The Ss chosen to simulate hypnosis had been completely unresponsive to hypnotic effects in the previous sessions and were themselves convinced that they were insusceptible to hypnosis. They had scored 4 or less on at least two standard scales and no higher than 2 in at least two diagnostic sessions. Simulating Ss were given an accurate (and truthful) description of their task. In summary, each simulator was told:



Your task during the next two sessions is to try to fool another hypnotist by pretending to be deeply hypnotized. Although you are not susceptible to hypnosis, your task is to pretend to be a deeply hypnotized S, to fool another E who is a trained, expert hypnotist. The new E knows that some Ss will really be hypnotized and that others will only be acting, but he does not know which Ss are hypnotized and which ones are faking. In previous research it has been found that a "blind" E cannot with certainty pick out which Ss are simulating. The task is not an easy one for you, but it can be done successfully. If E becomes sure that you are actually simulating, however, he will stop the experiment and your services will be no longer required. You should make use of whatever clues you glean from the instructions, the experimental procedure, your past experiences with hypnosis, and your expectations and conceptions about what deep hypnosis is like. Your task is even more difficult in this study, because you have to simulate with the hypnotic depth indicator and the physiological record at all times; one of the important aims of the experiment is to determine whether you can fool the machine, as well as the hypnotist. Good luck!

Hypnotizable subjects. The six deeply hypnotizable Ss in each study could experience all major phenomena of deep hypnosis, with marked subjective conviction about the reality of their experiences. All Ss could experience challenge suggestions, positive and negative hallucinations, complete posthypnotic amnesia, and would carry out posthypnotic suggestions compulsively. They had scored 8 or more on HGSHS:A and SHSS:C and in the 5-range of the diagnostic scale. In accordance with the simulating model (Orne, 1971), hypnotized Ss did not know that simulating Ss were participating in the study.

Experimental Session Procedure

As in the first of the two sessions, the experimental session commenced after the application of the recording electrodes. Minimal interaction took place between the "blind" E-observer and S. As before, a short passive hypnosis was induced to measure psychophysiological baselines. Hypnosis was induced a second time, and S completed a memory experiment lasting about 30 minutes. During these periods the "blind" E occasionally monitored the GSP recording in the room. The "blind" E then made a final judgment about whether she believed S was hypnotized or simulating.

Hypnosis procedure before tape breakdown. The tape was started again and hypnosis was reinduced. Arm levitation and arm rigidity suggestions were administered. Emphatic suggestions were given that S’s right hand and arm were stuck tightly to the table and arm of the chair and that they would remain stuck until E said, posthypnotically: "Your arm is comfortable now." After S tested that his arm was stuck, by trying to lift it, suggestions were given that he was at a concert



listening to his favorite music, tapping his foot in time with the music. "... You're enjoying it [the concert] so very much that you hum the tune to yourself, that you tap in rhythm to the music. You go on tapping* your foot. Keep on tapping your foot in rhythm with the music." At the predetermined point during this suggestion (marked by the asterisk above), the power failure was made to occur.

The power failure. The tape recorder was plugged into a hidden extension cord which led behind a large Ampex tape console into the adjacent observation room. A silent mercury-switch was thrown in the observation room, stopping the tape recording in mid-sentence with a marked slur. Simultaneously, the desk light turned off and the overhead lights went out. The larger "computer" recorder and table polygraph both stopped. Immediately the "blind" E sprang up, switched the light twice, manipulated the control dials of the tape recorder console, and after muttering softly "the damn fuse," she hurriedly left the room, slamming the office door. She then quietly entered the adjacent observation room.

The S was observed for 40 minutes from the time the light went off. After 40 minutes the electrical power was switched on again. The lights flashed on instantaneously. The tape recorder started turning, but took a few seconds to warm up before the sound recommenced. The tape restarted at exactly the same point as it had stopped. Thus, the suggestion about the concert continued where it had left off 40 minutes earlier. Two and one-quarter minutes after restarting the tape, the "blind" E quickly re-entered the room, appearing slightly disheveled. If S still had his eyes closed, E said nothing. If S was not apparently hypnotized at this point, E re-induced hypnosis by counting from 1 to 20, the standard cue previously used.

The tape continued to the end of the suggestion about the concert and several additional suggestions were administered. Hypnosis was terminated, and, after amnesia was tested, a brief inquiry was conducted by the "blind" E, during which S was asked what he thought had been the purpose of the experiment. If S mentioned the power failure, E said that a fuse had blown, but further discussion about the tape breakdown was avoided as much as possible. The "blind" E removed the posthypnotic suggestion that S's arm was stuck to the chair and then left to find the "non-blind" E. A more exhaustive inquiry about S's perception of the meaning of the experiment and the nature of his subjective experiences was conducted.

Procedural Differences from the Original Unsuccessful Study

Several important changes were made in the replication study procedure which, it was hoped, would eliminate some of the cues that indi-



cated to the simulators in the earlier study that the power failure was, in fact, part of the experiment. With these changes, it was hoped that the power failure was perceived as a genuine accident.

A variety of psychophysiological measures were recorded on an Offner Type R polygraph. Standard recording techniques were used. These included occipital and parietal monopolar EEG, horizontal eye movements, GSR, GSP, respiration, and EMG leg activity. The recording electrodes were led into the observation room but by a route not visible to S. The Ss were told that their responses would be played directly into a large Ampex tape recorder for subsequent computer analyses. The tape deck (which blocked the corner of the room where the leads entered the observation room) was started, but actually recorded nothing. In addition, a small two-channel Stoelting polygraph was in the experimental room. In the previous study, a similar small machine was used. One channel recorded respiration, the other was used to provide necessary coded signals between the experimental and observation rooms for the standard on-line monitoring of the physiological data. 8 As in the previous study, the Stoelting recording was visible to the S, and E periodically examined it. Following the power failure, the cessation of these recording devices presumably added to the plausibility of the deception, because nothing was apparently being recorded.

Instead of the one-way screen, an old but relatively unknown method of observation was adopted. Hanging on the walls of the experimental room were three framed silk screen paintings that fitted into the natural decor of the room. One of the paintings covered a hole in the wall. Silk screen acts like a one-way screen when looking from the dark to the lighter side. 9

When the "power failure" occurred, not only did the tape recorder slur to a halt, but the Stoelting and Ampex equipment also stopped, and all lights in the room went out. These occurrences were intended to help convince S that not even his physiological responses were being recorded. Thus, the power failure affected all the equipment in the room rather than just selected parts of it, as in the earlier study. When E left the room, even the hallway lights were off. During daytime sessions, a small amount of daylight passed through the drapes. At night, only dim

8 As there are no consistent differences which indicate the presence of hypnosis on standard psychophysiological measures, these data will not be reported. They were added primarily to provide emphasis for S that continuous recordings were being taken and to provide additional physical restraint during the session. Hopefully, the electrodes added to the plausibility of the study.

9 In fact, this was the system employed in the original one-way observation room at the Harvard Psychological Clinic prior to the development of one-way glass.



light entered from a distant parking lot light that had been specially and strategically installed for observational purposes. Further, a female medical student was chosen as the blind E in the replication study. It was hoped her ingenuousness and confused reaction to the fuse problem would be credible, making it less likely that Ss would feel they were being deceived when a power failure apparently occurred.


Simulating subjects. In the earlier study simulating Ss continued to pretend they were hypnotized throughout the 30-minute observation period. These performances provided objective evidence that they perceived the accidental power failure as part of the experimental procedure, thereby invalidating the experiment. This time, simulating Ss performed quite differently compared to simulators in the earlier study. Here, five of the six simulators immediately ceased simulating when the observer left the room. They opened their eyes within seconds after she closed the door. They remained active and alert for the full 40 minutes of observation. When the tape was recommenced, simulators immediately resumed simulating as the observer's footsteps approached the door. On each of the three objective criteria -- time to open eyes, cease tapping, and move the stuck arm (see Table 2) -- mean performance time for the simulators was significantly shorter (p < .01, p < .02, p < .05, respectively) in the present study than in the earlier one.

Two conclusions are suggested by the behavior of the simulators, both of which were confirmed in post-experimental inquiries. (a) The plausibility of the accident was confirmed. Simulators believed a fuse had truly blown and did not consider as a possibility that the accident was part of the experiment. They perceived no need to continue faking


during this period because it did not occur to them that they were being watched. This conclusion was reinforced by the varieties of private behavior which simulating Ss would not have indulged in had they believed they were being observed. (b) Simulators apparently believed that truly hypnotized Ss would not notice the power failure, and, as indicated by their return to simulation when the power was restored, believed hypnotized Ss would remain hypnotized for more than 40 minutes.

Having successfully demonstrated that the simulators behaved as if they believed they were not being observed --establishing by their behavior the plausibility of the accident -- we felt in a position to evaluate the performance of the deeply hypnotized Ss. If highly motivated simulating Ss did not doubt the reality of the accidental power failure, it seems unlikely that the hypnotized Ss with less reason to be suspicious would seriously entertain any doubts about the veridicality of the procedure.

Hypnotized subjects. Deeply hypnotized Ss showed no immediate behavioral signs of being aware of the power failure nor of E departing (see Table 2). Five of the six Ss remained passive for several minutes; behaviorally, they looked exactly as they had before the power failure. Then, after a mean of 10.7 minutes, they ceased foot-tapping; they opened their eyes after 16.5 minutes; and, while looking around the room, first moved their "stuck" arms after an average of 19.2 minutes. Their behavior was consistent with the hypothesis that they would arouse themselves from hypnosis after a period of time (ranging from 1 to 29 minutes) following the departure of E. This sequence of events was confirmed by their subsequent subjective reports. It appears that hypnosis is not limited in its duration by the immediate presence of the hypnotist (observer) or of an immediate hypnotic surrogate, the tape-recorded procedure, nor does hypnosis persist as long as other unhypnotized Ss believed it would.

Some Comparisons between the Two Studies

The dramatic difference in the performance of the simulating Ss in the two studies has been described. Changes in the procedure that appeared relatively minor, but which actually involved a great deal of planning and testing, led to a situation which was perceived entirely differently by the two simulating groups. In the first study the behavior of the simulators indicated that the deception was transparent; in the second, the behavior of the simulators indicated that no deception was suspected. No inference about the behavior of the hypnotized S was possible in the first situation, because the procedure could have been,



for them, equally transparent; there is no experimental way of evaluating their real perceptions. In the latter study, however, inferences can be drawn because the evidence indicated that the deception was convincing enough to be accepted. Under these circumstances, it is all the more remarkable that the performance of the hypnotized Ss in both studies was consistent and could not be differentiated statistically. In both instances the hypnotized Ss behaved significantly differently from the simulating Ss -- they did not remain hypnotized as long as the non-deceived simulators in the unsuccessful study, and they remained hypnotized longer than the deceived simulating Ss in the second study. 10 This evidence suggests that hypnotized Ss were not responding maximally to the cues inherent in the situation, but rather were responding to those cues derived from their ongoing subjective experience.

In summary, the behavior of the hypnotized S is not consistent with the view that the cessation of the ongoing relationship will instantaneously terminate hypnosis (as it did with the simulators who accepted the accident in the second study). Nor is it consistent with the behavior of the simulators who, in the first study, believed that the "accident" was faked and that they were being observed throughout. The hypnotized Ss appeared to be initially unaffected by the accident, but, in accord with a feeling of free-floating anxiety that "something had happened" (expressed later in post-experimental interviews), were able to terminate hypnosis by themselves. This task was typically described by them as a difficult one.

The Return of E

Of considerable interest is the behavior of both groups of Ss when the power was turned back on. About 2 1/2 minutes after the tape recorder came back on, E could be heard walking down a corridor and then entered the room. As soon as the lights were turned on, the simulating Ss (who had temporarily stopped role playing) quickly resumed a position similar to the one occupied when E left, typically taking great care to put the arm, which was supposed to be stuck to the arm of the chair, back again, and adjusting the HDI (which several of them had randomly moved and played with in the interim). As they heard E returning, and as the doorknob turned, they resumed tapping and closed their eyes. They clearly wished to convey the impression that they were still hypnotized, providing additional evidence that they did not seriously contemplate that they might have been observed during the preceding 40 minutes. In addition, this behavior provides strong evi-

10 Using simulating Ss does not provide a control group in the classical sense, and traditional statistical tests may not be appropriate. The results reported were significant using Fisher's exact test (p < .05).



dence about a popular notion of what would happen if, in an actual situation, hypnosis was inadvertently terminated by the absence of the hypnotist; clearly the simulating Ss believed a hypnotized person would remain in trance for at least this length of time.

In contrast to the expectations communicated by the simulating S, the behavior of the hypnotizable Ss in both studies was counterexpectational. Excluding the one S who did not terminate hypnosis, 10 of the 11 susceptible Ss in the two studies made no attempt to re-enter hypnosis when E returned. Rather, they turned towards E with a puzzled look, sometimes with a nervous laugh, apparently seeking an explanation of what had happened.


The experimental findings of the second study considered in conjunction with the previously reported results (Orne & Evans, 1966) have important methodological and theoretical implications for hypnosis research.

The Use of Simulating Ss to Test the Adequacy of an Experimental Deception

In this study the purpose of the simulating Ss was to provide a rigorous test of how S perceived what was intended to appear to be a spontaneous power failure. Since simulators are instructed to mimic the behavior of deeply hypnotized individuals without the benefit of specific task instructions, they are forced to rely upon their previous knowledge about hypnosis and their analysis of the cues in the situation. These individuals, therefore, would be particularly sensitive to subtle nuances in the behavior of the hypnotist and even more so to subtle aspects of the experimental situation which might serve to communicate how they ought to behave.

The two experiments taken together illustrate how the simulating Ss provide a conservative estimate of the kind of cues made available by the total experimental procedure. In the first experiment the simulating group was not deceived because sufficient care was not taken to solve the technical problems that are involved in properly staging the power failure. Therefore, it was not appropriate to draw inferences about the behavior of the deeply hypnotized group in that study. Although significant differences were found between the behavior of the two groups, the behavior of the simulators showed that the deception was not convincing. Because of the inadequacy of the manipulated deception, it could not be asserted, with any degree of certainty, that



the power failure was perceived as genuine by the hypnotized group. In the second experiment, great care was taken to assure the plausibility of the deception and, indeed, the behavior of the simulating Ss was evidence that the power failure was seen as genuine. Because the simulating group now perceived the experimental situation as it was intended -- the deception was successful -- we felt justified in assuming that the behavior of the deeply hypnotized group permitted ecologically valid inferences to be drawn about what would happen in a nonexperimental situation, were the hypnotist to actually disappear.

One very interesting observation deserves special comment. The hypnotized Ss in the first and second studies behaved similarly. Not only does this observation challenge the view that hypnotized individuals are particularly sensitive to subtle cues in a situation -- not surprisingly, a group of individuals asked to deceive E are more alert to deception than the deeply hypnotized group -- but it also illustrates the conservative nature of the simulating control group. In retrospect, the data suggest that the behavior of the deeply hypnotized group may have been ecologically valid even in the first experiment. On the other hand, we were not in a position to evaluate their behavior until the data from the second experiment became available.

The purpose of the simulating control is to test the adequacy of an experimental procedure, and it will sometimes provide a more rigorous test than is actually necessary. On the other hand, a control procedure which forces the investigator to perfect his experimental procedure is highly desirable since it will minimize the likelihood of generalizing inappropriately from laboratory findings. While further data will be needed, the observations in these two studies, as well as the logical analysis of the situation, suggest that, if a deception is not recognized as such by simulators, it is extremely unlikely to be recognized as such by other Ss not given simulating instructions.

The use of simulators in this context is in some ways analogous to carrying out a test of significance. A test of significance may turn out to be insignificant, leading the investigator to accept the null hypothesis when it may not, in fact, be true. This is the price one must pay in order to minimize the probability of accepting random fluctuations as significant differences. Similarly, when simulators see through a deception, it becomes logically necessary to view the experimental procedure as inadequate for the "real" group even though, as in this instance, this group did not necessarily see through the deception. The virtue of the simulation group in a study such as this is to minimize the probability that the investigator will delude himself about the effectiveness of a deception manipulation.



Implicatiom of Behavior of Hypnotized Ss

Since it appears clear that the situation in the second experiment was perceived as an actual power failure which happened to occur during an experiment, it is possible to infer how hypnotized Ss would respond if the hypnotist had actually disappeared. Hypnotized Ss tended to arouse themselves from hypnosis after some time had elapsed but in all cases well before the 40-minute period expired. For example, they stopped tapping their feet after a mean of 10.7 minutes, and first opened their eyes, which were supposedly tightly shut until after hypnosis had terminated, after a mean of 16.5 minutes. No S terminated hypnosis as soon as E left the room, nor did any remain hypnotized for as long as the 40 minutes. Neither the stopping of the tape recorder, the abrupt darkening of the room, nor the departure of the research assistant, leaving S alone in the room, was sufficient to terminate hypnosis immediately.

Typically, the first sign that S was about to arouse himself was a movement of the hypnotic depth indicator which often preceded other behavioral signs by several minutes. The Ss had received instructions to move the depth indicator in accordance with how deeply hypnotized they were, and it is interesting that the power failure did not deter them from continuing to follow the initial instructions they had received. As Ss actually roused themselves, they appeared confused when they first opened their eyes, looking aimlessly about the room. They subsequently reported during the inquiry with the original non-blind E that they experienced a vague feeling that something had changed or was wrong, which gradually became more pronounced and made them somewhat uncomfortable, bringing them out of trance, until eventually they no longer felt hypnotized. These subjective reports seemed highly concordant with their behavior, which suggested that hypnotized Ss aroused themselves slowly with some difficulty.

In evaluating these findings it is important to keep in mind that hypnotized Ss were selected from a large population for their ability to enter deep hypnosis. Their behavior, therefore, can only be generalized to other Ss who manifest the traditional criteria of deep hypnosis. (All Ss were in the 5 range, according to the specifications of Orne and O'Connell, 1967.) Certainly, phenomena of this kind demand Ss who have the ability to respond to hypnosis and who are motivated to do so, factors which are at least as important as the hypnotic induction procedure. There is little doubt that Ss who are not capable of entering very deep hypnosis, but are "hypnotized" by being exposed to a trance induction procedure, would arouse themselves almost immediately



when the power failure occurred. The question to which we have addressed ourselves in this research is the extent to which deep hypnosis, the phenomenon classically described as artificial somnambulism, can or cannot be explained exclusively as a consequence of S's motivation and expectations about how to enact the role of a hypnotized individual. While the importance of these factors is recognized by the authors (see Orne, 1959, 1970; Evans, 1968), this experiment was designed to see whether these factors are sufficient to account for deep hypnosis.

The Counterexpectational Nature of These Findings

The behavior of the simulating Ss permitted inferences to be drawn concerning Ss' expectations about what deeply hypnotized individuals would do if the hypnotist actually disappeared. In the first study, the simulators, who recognized that the power failure was a deception, continued to act as if they were hypnotized for the entire 30-minute period. In the second study, in which the power failure was more realistically contrived, simulators stopped simulating as soon as they believed themselves to be alone. When they heard E coming down the hall, however, they quickly assumed the appropriate simulated hypnotic behavior, making certain that they appeared to be simulating hypnosis as E re-entered the room.

It is clear from the behavior of the simulators in both the first and second studies that Ss believe the appropriate behavior of hypnotized individuals would be to remain in trance for a considerable period of time after the disappearance of the hypnotist -- certainly more than 30 to 40 minutes. The deeply hypnotized Ss behaved contrary to these expectations and beliefs. The hypnotized Ss terminated trance during the observation period and, of course, did not make any pretense of being hypnotized when E returned. If deeply hypnotized Ss are merely enacting a role, they certainly do not see their role as ending immediately with the hypnotist's departure; in other words, if hypnosis is only role enactment, the role is not audience-dependent in the usual sense of the word "role."

Usually it is difficult to determine whether the hypnotized S's behavior can be ascribed to hypnosis or whether it might equally well and more parsimoniously be understood as S's responding according to his expectations of how deeply hypnotized individuals ought to behave. In this experiment, however, the behavior of the deeply hypnotized S cannot be explained as a function of either his prior expectations or of the cues inherent in the experimental situation. Therefore it is difficult to see how the behavior of the hypnotized Ss in this study can be



explained solely as playing the role of a hypnotized individual. Rather, these data are more congruent with the view that hypnosis involves some as yet unspecified alterations in the S's state of consciousness.


BARBER, T. X. Hypnosis: A scientific approach. New York: Van Nostrand, 1969.

DORCUS, R. M., BRINTNALL, A. D., & CASE, H. W. Control experiments and their relation to theories of hypnotism. J. gen. Psychol., 1941, 24, 217-221.

EVANS, F. J. The case of the disappearing hypnotist. Paper presented at the meeting of the American Psychological Association, New York, September 1966.

EVANS, F. J. Recent trends in experimental hypnosis. Behav. Sci., 1968, 13, 477-487.

HILGARD, E. R. Hypnotic susceptibility. New York: Harcourt, Brace & World, 1965.

KUBIE, L. S., & MARGOLIN, S. The process of hypnotism and the nature of the hypnotic state. Amer. J. Psychiat., 1944, 100, 611-622.

MILGRAM, S. Some conditions of obedience and disobedience to authority. Human Relat., 1965, 18, 57-76.

ORNE, M. T. The nature of hypnosis: Artifact and essence. J. abnorm. soc. Psychol., 1959, 58, 277-299.

ORNE, M. T. On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications. Amer. Psychologist, 1962, 17, 776-783.

ORNE, M. T. Hypnosis, motivation, and the ecological validity of the psychological experiment. In W. J. Arnold & M. M. Page (Eds.), 1970 Nebraska symposium on motivation. Lincoln, Neb.: Univer. of Nebraska Press, 1970. Pp. 187-265.

ORNE, M. T. The simulation of hypnosis: Why, how, and what it means. Int. J. clin. exp. Hypnosis, 1971, 19, 183-210.

ORNE, M. T. On the simulating subject as a quasi-control group in hypnotic research: What, why, and how. In E. Fromm & R. E. Shor (Eds.), Current trends in hypnosis research. Chicago: Aldine, 1972, in press.

ORNE, M. T., & EVANS, F. J. Social control in the psychological experiment: Anti-social behavior and hypnosis. J. Pers. soc. Psychol., 1965, 1, 189-200.

ORNE, M. T., & EVANS, F. J. Inadvertent termination of hypnosis with hypnotized and simulating subjects. Int. J. clin. exp. Hypnosis, 1966, 14, 61-78.

ORNE, M. T., & HOLLAND, C. H. On the ecological validity of laboratory deceptions. Int. J. Psychiat., 1968, 6, 282-293.

ORNE, M. T., & O'CONNELL, D. N. Diagnostic ratings of hypnotizability. Int. J. clin. exp. Hypnosis, 1967, 15, 125-133.

SARBIN, T. R. Contributions to role-taking theory: I. Hypnotic behavior. Psychol. Rev., 1950, 57, 255-270.

SARBIN, T. R., & ANDERSEN, M. L. Role-theoretical analysis of hypnotic behavior. In J. E. Gordon (Ed.), Handbook of clinical and experimental hypnosis. New York: Macmillan, 1967. Pp. 319-344.

SHOR, R. E. Hypnosis and the concept of the generalized reality-orientation. Amer. J. Psychother., 1959, 13, 582-602.

SHOR, R. E., & ORNE, E. C. Harvard Group Scale of Hypnotic Susceptibility, Form A. Palo Alto, Calif.: Consulting Psychologists Press, 1962.



STRICKER, L. J. The true deceiver.Psychol. Bull., 1967, 68, 13-20.

WEITZENHOFFER, A. M., & HILGARD, E. R. Stanford Hypnotic Susceptibility Scale, Form C. Palo Alto, Calif.: Consulting Psychologists Press, 1962.


El Hipnotista Desaparecido Sobre como los Sujetos Simuladores y Realmente Hipnotizados Perciben los Procedimientos Experimentales

Frederick J. Evans y Martin T. Orne

Resumen: Este trabajo examina los efectos producidos por una interrupcion de la corriente electrica en la conducta de sujetos sometidos a hipnosis por medio de cinta magnetofonica; para ello era necesario que esta interrupcion apareciera como casual y no como el resultados de un componente experimental como se desprende de un estudio anterior (Orne & Evans, 1966). En ese estudio los sujetos simuladores continuaron simulando el estar en hipnosis pues pensaron que la falla electrica era parte del experimento. En este estudio dejaron de seguir simulando apenas el experimentador abandono la pieza creyendo que la falla era real y no estaban siendo observados. En cambio los sujetos realmente hipnotizados salieron de su estado lenta y gradualmente. Esta conducta no parece estar en relacion con aquellas teorias que suponen que la hipnosis tiene que ver, o bien con motivacion, o bien con el desempeno de un rol.


Der verschwindende Hypnotiseur: Die Anwendung von simulierenden Vpn. zur Bestimmung der Wahrnehmungen Yon experimentellen Verfahren in Versuchspersonen

Frederick J. Evans und Martin T. Orne

Abstrakt: Dies Studium untersucht die Effekte eines zeitweiligen Versagens des elektrischen Stroms in einer Tonbandsitzung, wuhrend die Vp. hypnotisiert war. Es war notwendig, dass die Vp. das Versagen des Stroms als einen Zufall und nicht als ein Hintergehen im Rahmen des Experiments auffasse. In einem fruheren Studium (Orne und Evans, 1966) hatten Hypnose-simulierende Vpn. unter einem "blinden" Vl. das Imitieren wuhrend des Stromversagens fortgesetzt, wahrscheinlich in der Annahme, dass sie unter Beobachtung waren. Daher konnten keine Schlussfolgerungen uber das Benehmen der hypnotisierten Vpn. gezogen werden. Nachdem man das Verfahren modifiziert hatte, horten die simulierenden Vpn. mit dem Imitieren auf, sobald der Hypnotiseur den Raum verlassen hatte, womit sie ihren Glauben bewiesen, dass ein Versagen des Stroms tatsachlich eintrat und auch dass sie nicht beobachtet wurden. Hypnotisierte Vpn. beendeten selbst und spontan die Hypnose, obgleich langsam und mit subjektiver Schwierigkeit. Das spontane Benehmen von hypnotisierten Vpn. bei der Abwesenheit eines Hypnotiseurs scheint nicht mit den Voraussagungen in Einklang zu stehen, die auf motivierenden oder Rollespielenden Theorien der Hypnose basiert sind.

The preceding paper is a reproduction of the following article (Evans, F. J., & Orne, M. T. The disappearing hypnotist: The use of simulating subjects to evaluate how subjects perceive experimental procedures. International Journal of Clinical and Experimental Hypnosis, 1971, 19, 277-296.). It is reproduced here with the kind permission of the Editor-in-Chief of The International Journal of Clinical and Experimental Hypnosis.