You are on page 1of 12

Journal oj Personality and Social Psychology

1965, Vol. 1, No. 3, 189-200


Pennsylvania Hospital, The Institute, Philadelphia, and University of Pennsylvania

Rowland and Young found that hypnotized Ss were willing to carry out such
apparently antisocial actions as grasping a dangerous reptile, plunging their
hand into concentrated acid, and throwing the acid at an assistant. However,
only informal attempts were made to show that the Ss perceived the acts as
dangerous, or to demonstrate that the requested behavior exceeded the broad
limits of social control implicit in the context of a psychological experiment.
The experiment conducted by Young was replicated exactly, and his results
confirmed. However, nonhypnotizable Ss simulating hypnosis to a "blind" E,
and even normal, waking control Ss, complied with the same requests. Ss in-
variably reported they were convinced the activities were safe because they
were participating in research conducted by competent, responsible scientists.
It is concluded that the tasks were perceived by Ss as being within the limits of
legitimate requests made in an experimental context. The present study does
not answer the question whether antisocial behavior can be elicited under
hypnosis. However, it demonstrates the nature of control groups essential for
a valid test of the antisocial hypothesis, and illustrates the broad range of
behavior legitimized in an experimental context.

SOCIAL CONTROL IN THE PSYCHOLOGICAL experimental investigation is not a passive

EXPERIMENT entity. Experimental evidence (Orne, 1959;
There has been increasing awareness by Orne & Scheibe, 1964) has demonstrated that
behavioral scientists that the subject in an the subject takes an active role in interpreting
1 the nature of the investigation and makes
This study was conducted at the University of
Sydney, Australia, during a visit by the senior implicit assumptions about the hypotheses
author, June-August 1960. It was supported in part being investigated which influence his per-
by Research Grant AF-AFOSR-88-63 from the Air formance in the experimental situation. Nor
Force Office of Scientific Research and by a grant is an experimenter free from the influence of
from the Human Ecology Fund. his own investment in the hypotheses he is
We wish to thank A. G. Hammer, then acting
Head of the Department of Psychology, for his investigating. In a series of studies Rosenthal
cooperation with the use of departmental facilities. has shown that experimenters who have dif-
We wish to express our appreciation to the Depart- ferent hypotheses about the outcome of a
ments of Chemistry and Biochemistry, University of particular experiment may obtain results
Sydney, and the Radiology Department of the
Royal Prince Alfred Hospital, Sydney, for donating which are congruent with their hypotheses
equipment. We are indebted to the Department of (for example, Rosenthal, 1964; Rosenthal &
Zoology, University of Sydney, for loaning the Fode, 1963). Such studies imply that it is
harmless reptiles. We are particularly indebted to
Fred Matthews, herpetologist, for his willingness to necessary to consider the particular nature
loan, tend, and handle the venomous reptile through- of the special interpersonal interaction which
out the study. Without his generous help this exists between the subject and the experi-
study would not have been completed.
We are appreciative of the valuable comments menter in psychological experiments.
made during the preparation of this report by Peter Orne (1960, 1962) has emphasized that
B. Field, Lawrence A. Gustafson, Ulric Neisser, the experimental context legitimizes a very
Donald N. O'Connell, Emily C. Orne, and Ronald
E. Shor. We also wish to thank Eleanor DeRubeis broad range of behavioral requests. Subjects
for her valuable editorial assistance. have implicit faith that experimenters are
When the experiment was conducted the junior responsible people, that they will not be asked
author was supported by a University of Sydney
Research Studentship in the Department of Psy- to carry out tasks which are devoid of mean-
chology. ing, and that regardless of appearances they

will not be permitted to suffer any harm be- chological experiment has received few ex-
cause of obvious social sanctions. plicit tests.
In a series of informal experiments in our
laboratory, it has been impossible to devise a
task which the subject perceives as completely It is generally considered that a hypnotized
"meaningless" within the context of an ex- subject relinquishes considerable social and
periment. For example, subjects were con- behavioral control to the hypnotist. A subject
fronted with a stack of paper, each page frequently reports that he felt compelled to
containing rows of random digits. The experi- carry out the commands of the hypnotist;
menter instructed subjects to continue adding that he could not resist the suggestions made
the rows of numbers successively, and after by the hypnotist. This apparent increase in
accurately completing each page, to tear it the amount of social control relinquished by
into a minimum of 32 pieces. Although sub- the hypnotized subject to the hypnotist has
jects were given no reason to justify the task, raised the unresolved question whether a
they continued this apparently meaningless subject can be compelled, under hypnosis, to
endeavor beyond the tolerance limits of the perform apparently antisocial acts or be-
experimenters (Orne, 1962). havior which is perceived as injurious and
Frank (1944) has reported some informal dangerous to self or others. Estabrooks
experiments in which subjects continued (1943), Rowland (1939), Weitzenhoffer
meaningless and impossible tasks, including (1949), Wells (1941), Wolberg (1945), and
trying to balance a marble on a small steel Young (1952) have stated that, provided
ball and transferring spilled mercury to a adequate techniques are used, hypnotized
small bottle with a wooden paddle, even when subjects may be compelled to carry out ap-
an assistant tried to prevent them from trying parently antisocial actions, while Erickson
to complete the tasks. No justification was (1939), Meares (I960), and Schilder and
given for performing the tasks other than Kauders (1927) have disagreed with this
that it was an experiment. viewpoint. The extensive literature present-
In the same study, Frank also reported ing the conflicting viewpoints has been re-
that subjects continued eating several un- viewed elsewhere (Barber, 1961; Orne,
savory, unsalted soda crackers for time 1960; Weitzenhoffer, 1953). The present
periods in excess of what would seem reason- investigation is concerned with two studies
able for such an unpleasant task, and longer frequently cited as evidence that hypnosis
than subjects who were told they could stop can be used to induce antisocial behavior.
eating them as soon as they wished. Shor Rowland (1939). Two deeply hypnotized
(1962) has reported that subjects were willing subjects were asked to reach through the
to accept extremely high levels of electric window of a box and pick up a large, active,
shock when requested to select a level of diamondback rattlesnake. This request was
intensity which was as high as they could rationalized to them by suggesting that the
tolerate for experimental purposes. Milgram snake was a coil of rope. One hypnotized
(1963) has shown that subjects continue to subject immediately complied, but was pre-
administer what they believe are extremely vented from handling the snake by a pane of
high levels of electric shock, exceeding ap- invisible glass. The other subject came out of
parently dangerous levels, to another "sub- the hypnotic state and refused to continue
ject" in the context of a learning experiment. with the experiment. Another two hypnotized
The limits of boredom, tolerance, pain, and subjects attempted to grasp the snake, even
when no attempt was made to delude them
fatigue which are accepted as reasonable re-
about what it was. Similarly, two hyp-
quests within an experimental situation seem notized subjects were requested to throw con-
extremely broad. However, the actual range centrated sulphuric acid, which they had
of social and behavioral control legitimized been told was very dangerous, at the face of
by the special contract implicit in the sub- an assistant. The two subjects complied with
ject's agreement to participate in a psy- this request.
By way of control, forty-two persons, of every out without the intervention of hypnosis it
age and degree of sophistication, were asked to must be shown that:
come to the laboratory and pick up the snake. . . .
With one exception all the persons were not only 1. Subjects in hypnosis carry out actions
badly frightened at the appearance of the snake, which are not performed by nonhypnotized
but would not come close to the box . . . [Row- control subjects.
land, 1939, p. 116]. 2. Control subjects must be treated in an
identical fashion to hypnotized subjects;
Young (1952). In a replication of Row- both in regard to explicit instructions as well
land's study, Young asked eight deeply hyp- as implicit cues.
notized subjects to carry out similar tasks. If these actions are to be designated as
Young reported, antisocial or self-destructive it must be shown
The results show that seven of the eight subjects that they are perceived as such by the sub-
would enter into a situation which unhypnotized jects, i.e., truly dangerous or harmful to
observers shrank from, the subjects carrying out themselves or others. The implicit cues are
suggestions to handle snakes and throw nitric acid of crucial importance. A subject is aware of
under conditions from which they themselves re-
coiled in the waking state [p. 4051. certain realities imposed by the experimental
situation. It is as clear to a subject as it is to
One major aim of the present study was any scientist that no reputable investigator
to confirm these results by replicating in can risk injuring a subject during the course
'exact detail the procedures outlined by of an experiment. A subject knows that an
Young (1952). It is apparent from the ac- experimenter will outline in advance any
counts of both Rowland and Young that possible specific and deliberate danger which
strong pressure was placed upon the hyp- could be associated with his actual par-
notized subjects to comply with the requested ticipation in a study. Consequently, any re-
antisocial acts. However, Rowland did not quested behavior which appears to a subject
exert similar pressure to comply when testing to be dangerous at face value may be re-
the independent, informal waking control interpreted in the context of a laboratory
group, nor did Young confront his hyp- situation. In spite of the apparent objective
notized subjects with the same type of pres- danger of a task it may nonetheless be per-
sure in the subsequent waking condition. ceived to be harmless because the subject
Consequently, it is doubtful if the informal realizes that necessary precautions will be
controls used by these investigators provide taken to avoid possible injury to him. If an
any answer to the fundamental question apparently dangerous task is requested of a
being considered: does the degree of social subject during an experiment the subject's
control which occurs under hypnosis actually compliance, or refusal, may depend on
exceed the amount of social and behavioral whether he perceives that he is expected, or
is not expected, to carry out the task. It is
control already existing in the experimental particularly vital in an experiment which
situation? depends on a contrived situation to determine
what the subject, in different groups, per-
SOCIAL CONTROL, ANTISOCIAL BEHAVIOR, ceives about the experimental situation and
AND HYPNOSIS what is implicitly communicated to the sub-
The basic question may be formally stated: ject within different groups.
does the degree of social and behavioral con- Even though the experimenter is extremely
trol under hypnosis exceed that which is careful to treat all groups alike, subtle and
unintentional cues may be differentially com-
legitimized by the special social and be- municated by him to subjects in different
havioral control implicit in the experimental experimental groups, particularly when the
situation? experimenter knows to which experimental
In order to determine that subjects have group a specific subject belongs. These
indeed been compelled to carry out any experimenter influences on results have been
actions which they would not have carried demonstrated in both animal (Rosenthai &

Lawson, 1963) and human research (Rosen- PROCEDURE s

thai & Fode, 1963). Selection of Subjects4
It is essential, therefore, that the non-
Hypnosis group. Several volunteer undergraduate
hypnotized control groups and the hypnotized students were individually tested for susceptibility
subjects are treated in an identical manner. to hypnosis. After at least two hour-long screening
The real-simulating hypnosis model de- sessions, six subjects were selected to participate
scribed by Orne (1959, 1960, 1962) pro- in the study as the real hypnosis group. These
vides one method of making reasonably sure subjects could readily achieve deep somnambulistic
hypnosis manifesting such typical phenomena as
that experimental and control subjects are catalepsy and rigidities, positive and negative hal-
treated alike. With this procedure both lucinations, posthypnotic suggestions, and complete
hypnotized and control group subjects are posthypnotic amnesia.
run "blind" by an experimenter, because he Simulating hypnosis group. Another six subjects
who showed no hypnotic responses during at least
is not informed, and cannot readily detect, two extensive, individual hypnotic induction sessions
which subjects are hypnotized, and which participated in the study as simulators. Each simu-
subjects are awake but simulating hypnosis. lating subject was told that he was participating
The procedure has similar aims to those of in research investigating the nature of hypnosis, but
he was not told that the study involved allegedly
the "blind" designs used in the evaluation antisocial behavior. He was told that his task was to
of new drugs in psychopharmacology. try to fool an experienced hypnotist, in an experi-
The real-simulator design has been adopted mental setting, by pretending to be deeply hypno-
as the main control procedure in the present tized. Even though he would probably not experi-
ence or produce any real hypnotic response, he was
study. If both real and simulating subjects to act as if things were happening just as the
carry out behavior which is apparently anti- hypnotist said they would. Although the hypnotist
social or self-destructive, it would seem in- was described as an authority on the subject, it
appropriate to conclude that the hypnotized was stressed that it was possible, though very
group was compelled to carry out these difficult, for the subject to fool the hypnotist.
No special training or instruction about how to
actions because of the use of hypnosis: simulate was given to the subjects in this group,
rather, the behavior of both groups could be nor were they told what phenomena would be
interpreted more parsimoniously as an ap- involved in the experimental session. However, each
propriate response to the existing cues in the simulator had individually watched a good hypnotic
experimental situation. 3
Several procedural details have been deposited
In addition to the use of the crucial simu- with the American Documentation Institute. This
lating control group, the content of social and material contains additional detail about selection of
behavioral control existing within the experi- subjects, choice of a venomous reptile, the effects
of the nitric acid on the coin and safety procedures
mental situation was explored by including protecting the subject, description of the procedure
a normal waking control group. This inde- for substituting the colored solution for the acid,
pendent waking control group was treated, and a scale diagram of the apparatus. Order Docu-
as far as possible, in the same way as the ment No. 8203 from ADI Auxiliary Publications
hypnotized and simulating groups. An at- Project, Photoduplication Service, Library of Con-
gress, Washington, D. C. 20540. Remit in advance
tempt was also made to demonstrate that it $1.25 for microfilm or $1.25 for photocopies and
is possible to communicate to a subject that make checks payable to: Chief, Photoduplication
he should fail to carry out the apparently Service, Library of Congress.
dangerous acts. The procedure adopted was One experimenter (FJE), who selected sub-
jects, trained simulators, was assistant and observer
the same as our understanding of the informal during the study, subsequently confirmed that
controls used by Rowland (1939). In addi- hypnotized subjects were amnesic for the session,
tion, each hypnotized subject was retested that simulators experienced no real effects of the
in the waking condition, as his own control, hypnotic induction and suggestions, and conducted
postexperimental inquiries. The second experimenter
similar to the control procedure of Young (MTO) induced hypnosis for the main session, con-
(1952). A parallel procedure using subjects ducted the five groups in the experimental procedure,
with no previous experience with hypnosis and at the completion of each subject's experimental
participation, conducted a further independent in-
was used as an additional waking control quiry after being informed which were real and
group. simulating subjects.

subject being hypnotized, demonstrating motor and Behavior Requested of the Subject
challenge suggestions, simple positive hallucinations,
amnesia, and posthypnotic suggestion. Inanimate objects. The hypnotist suggested that
Additional control groups. Six subjects, who did the subject would have an irresistible urge to do
not meet the criteria for either good hypnotic sub- whatever the hypnotist told him. The subject was
jects or simulators, formed one of the additional then asked to open his eyes, to reach into the
control groups. The remaining two control groups compartment, to pick up and identify objects he
contained subjects recruited from the student popu- saw there, and to place them in a bag held by the
lation who did not know the experiment also hypnotist. A two-shilling coin, then a piece of
involved hypnosis. chalk, were placed in the compartment. All subjects
identified and removed these objects.
Harmless reptiles. The subject was then told that
Experimental Procedure a harmless, live reptile would be placed into the
The experiment follows the procedure described compartment. A thick-tailed gecko lizard (Gymno-
by Young (1952) as closely as possible.5 dactylus milii) was placed in the compartment.
Hypnosis was induced in a room adjacent to the When the lizard had been removed, a thin, 14-inch
experimental room. After the induction of hypnosis long, harmless, green tree snake (Ahaetulla punc-
the subject was led, with eyes closed, into the tulatus) was lowered into the compartment.
experimental room and seated in front of a large The remaining three activities described below
wooden box which stood on a table. objectively appeared to be dangerous or harmful,
The box was divided into two compartments by either to the subject himself, or to a third person
a removable partition. The subject could easily present in the room.
reach into the front compartment through a window. Venomous snake. An assistant wearing long, thick,
A sheet of plate glass could be raised or lowered X-ray gloves placed a 14-inch long venomous snake
silently across the window, sliding along felted into the compartment. The reptile, popularly known
grooves. The compartment was illuminated by two as the red-bellied black snake (Pseitdechis porphy-
concealed lights, so that no light was reflected riacus), is generally considered to be one of the
directly onto the surface of the glass. The two two most venomous snakes found in Australia,
lights were the only source of illumination in the capable of inflicting a mortal wound if the victim
room. From the position of the subject it was not is not treated immediately upon being bitten.
possible to detect the presence of the glass.6 The hypnotist asked the subject to pick up the
snake carefully about its middle, and place it in the
We are indebted to P. C. Young for his kindness bag. Unknown to the subject, the glass screen had
in providing an unpublished manuscript describing been lowered silently across the window. If an
apparatus and procedure in detail and for clarifying attempt was made to grasp the reptile, the subject
further details in private communications (1959). struck his hand firmly against the glass.
The invisibility of the glass was attested by the Coin in acid. The screen, and the snake, were
following: (a) Inability of the experimenters to removed. The subject then watched the hypnotist
detect if the glass was in position unless they had pour fuming concentrated nitric acid (15 molar)
prior knowledge of its position, (b) The witness in a small, shallow, Pyrex glass beaker. After watch-
and bruised knuckles of several members of the ing the acid actively dissolve a copper-alloy half-
Department of Psychology whose curiosity got the penny coin for approximately a minute, the subject
better of them, (c) The expressed concern of a was asked to remove the coin from the beaker as
graduate student when watching a subject reach quickly as possible with his bare hand, and then to
for the snake. Although he knew about the glass plunge both hand and coin into a nearby basin
screen, and although he had been watching closely, of soapy lukewarm water.
he did not see the screen being lowered, (d) Only Throwing acid. The experimenter stood between
one subject in a subsequent inquiry reported detect- the subject and the box as he carefully washed
ing the glass. This happened because of the forget- the subject's hand. While the subject was distracted,
fulness of the experimenters who failed to clean the acid was replaced with an identical beaker
the window before this particular subject was tested. containing a harmless, colored solution which was
Most subjects expressed spontaneous surprise if, the same color as the acid. The subject was simi-
and when, they discovered the glass in their attempts larly distracted if he did not remove the coin, so
to grab the snake, often making exploratory move- that changing the solution was not observed.
ments across the surface of the window. The central partition in the box was removed,
Elizabeth Erickson (1962) has argued that visual exposing the assistant who had been responsible
acuity may increase under hypnosis; hence, hyp- for recruiting and training the subject, sitting at
notized subjects would detect the presence of the the opposite end of the open-ended box. The
glass. This hypothesis would be relevant if clear- hypnotist suggested assertively that the subject was
cut differences were found between the hypnotized very angry with the assistant for placing him into
group and other crucial groups in the willingness to the unpleasant situation of the experiment. The
comply with requests to perform the antisocial subject was told he felt so angry that he would have
activities. an irresistible urge to throw the acid firmly into

the face of the assistant and authoritatively told RESULTS

to pick up the acid and throw it.
The number of subjects in each group who
attempted to comply with the requested tasks
Postexperimental Interview
is summarized in Table 1. A detailed analy-
The subject was taken back into the adjoining sis of the responses of each subject within
room, and hypnosis was terminated. Comprehensive the various groups is presented in Table 2,
interviews were conducted with each subject to
allow him to express his subjective reactions to the together with ratings of the degree of sub-
experiment, and his thoughts and comments about jective involvement in the tasks. The latter
the experiment. Appropriate questions were asked ratings were made by the assistant who was
to establish that the subject had remained amnesic aware of the group membership of hypno-
throughout the experimental session, or, in the case tized and simulating subjects, and must be
of simulators, to confirm that no trance effects
occurred during the experiment. interpreted accordingly.7
The data presented in Tables 1 and 2 are not
Treatment of Experimental Groups amenable to further statistical analysis. There is a
partial experimental dependence of the tasks be-
The procedure has been outlined as it was en- cause of the way in which they were, of necessity,
countered by the subjects in the hypnosis group. presented to the subject. Consequently, neither treat-
The hypnotized group of subjects was subsequently ing each antisocial action as a separate activity,
retested in the waking state, while still amnesic nor the alternative of tabulating the number of
for the hypnosis session, and an additional four acts actually attempted by each subject, is a valid
groups were tested. method of treating the data statistically. Even if
For clarity of exposition a detailed statement of either alternative were possible, the analysis of
the aims of testing each group will be presented frequency data, preferably by exact probability
below in the Results section. General modifications procedures, requires the determination of the ex-
in the experimental procedures for the respective pected frequencies of occurrence. Hypothetical ex-
waking control groups will be quite apparent. The pected frequencies of six subjects in any group
procedures outlined above were followed either ex- successfully attempting each act could be defended
actly, or as approximately as the nature of the as readily as hypothetical expected frequencies of
specific groups would allow. The major difference no subject in any particular group attempting each
between groups was the manner and attitude of the act. Even the statistical determination of expected
experimenter, and the subtle means whereby expecta- frequencies from marginal entries has limited merit
tions of compliance or noncompliance with the in groups of this size, for this would impose a level
apparently antisocial actions were conveyed. Dif- of precision on the data which is not warranted, as
ferences in the approach of the experimenter, in well as limiting the possibility of finding statistically
terms of the aims and purpose of each group, will significant differences. However, this does not re-
flect upon the size of the groups, for the experi-
be discussed as the results for each group are being
mental rationale is such that the mere occurrence
presented. of apparently antisocial behavior is significant if it
There was, of course, no difference in the treat- occurs in the crucial simulating control group. The
ment of the hypnotized subjects and simulating aim of the study is, in a real sense, to demonstrate
subjects because the hypnotist remained unaware the truth of the null hypothesis: to demonstrate that
of the subgroup membership of these subjects specially designed groups do not differ in their
throughout the experiment. propensity to respond in a defined manner.

Group* Grasp harmless Grasp venomous Take coin Throw acid

reptiles6 snake from acid at assistant

1. Real hypnosis S 5 S S
la. Retest real, as waking control X 2 3 2
2. Waking control : press for 6 3 1 1
fail (Rowland)
3. Simulating hypnosis 6 6 6 6
4. Waking compliance control S 3 5 5
S. Waking control : eliminate X 2 3 2
"shaping" procedures

subjects In each group; see text for description of groups.
Includes small lizard and harmless snake. Two subjects refused both. Eliminated from procedure in waking retest of Group la
and Group 5.

Group and subject Sex Grasp harmless Grasp venomous Take coin Throw acid
reptiles" snakeb from acldb'° at assistant6 <d

1. Real hypnosis
JB F Yes Yes (H) Yes (H) Yes (A, T)
YM F Yes Yes Yes (A) Yes (A T)
GP M Yes Yes Yes Yes (T)
CW F Yes Yes Yes Yes
EB° F No —
. .

SM F Yes Yes Yes (U) Yes

la. Retest real hypnosis subjects
in waking state
(Young control)1
JB F No No No
YM F — No Yes (H) Yes (A, T)
M — No Yes (H) No
CW F — Yes (H) Yes (H) Yes (A, T)
EB" F —
SM F —
— —
Yes —
No No
2. Waking control : press for
failure (Rowland control)
RF M Yes Yes Yes Yes (H, V)
JF F Yes Yes (A) No No
DF M Yes No No No
PH M Yes Yes (H) No No
HH F Yes No No No
EC F Yes No No No
3. Simulating hypnosis
KM F Yes Yes (H) Yes (A, U) Yes (V)
RW F Yes Yes Yes Yes (T)
ST F Yes Yes Yes (H, U) Yes (A, T)
JJ M Yes Yes Yes Yes (V)
MF M Yes Yes (H) Yes (H,U) Yes (H)
AT M Yes Yes Yes (H, U) Yes (V)
4. Waking compliance control
FP F No No No No
TM F Yes No Yes (H) Yes (H)
RS M Yes No Yes (A) Yes (H)
PP M Yes Yes (H) Yes Yes
AL M Yes Yes Yes Yes (V)
FJ M Yes Yes Yes (A) Yes (V)
5. Waking control: elimination
of "shaping" procedures
IR F Yes (H) Yes Yes (H)
NG M — No Yes (H) No
RM M — Yes Yes (H) Yes
RE M — No No No
MA F — No No No
F —
— Yes (?)« No No

Note,—Yea °= attempt to carry out requested activity; No • refused to carry out requested activity.
» Gecko lizard and harmless snake. Procedures omitted for Groups la and 5 (see text).
H •> hesitant in attempting task; A = agitation and emotional Involvement markedly apparent.
* U = unsuccessful in removing coin, but attempted to extent of placing fingers in acid.
T =• tentative throw, hardly splashing assistant; V — violent throw, for example, hurling solution and container.
• Subject almost hysterical at sight of lizard; remaining activities omitted. Counted failure (see Footnote 8).
' Hypnosis-group subjects retested.
« Subject reported detecting glass screen at end of experiment (see Footnote 6).

Replication of Results of Rowland (1939) easily and readily all suggestions given to
and 0} Young (1952) him. Each activity was preceded by a firm
Real Hypnosis: Group 1 statement—"I want you to ... ." The appar-
Treatment. The hypnotist told the subject ent dangers in the venomous reptile and acid
that he had an irresistible urge to carry out situations were stressed. The experimenter

conveyed by his verbal tone and nonverbal be persuaded to attempt to pick up the
manner and attitude that he expected compli- venomous snake and to throw the acid at the
ance. Further pressure and persuasion were assistant. Another subject removed the coin
adopted if the subject showed hesitancy about from the acid.
carrying out any of the activities. Amnesia for the previous day's hypnotic
After the experimental tasks were com- experience was again confirmed, and was then
pleted, hypnosis was terminated. A careful removed. Subjects reported that under hyp-
evaluation of the validity of the suggested nosis they felt more passive, were not par-
complete amnesia for the session was made, ticularly concerned with the consequences of
and depth of hypnosis was checked by testing their actions or what safeguards existed, and
a posthypnotic suggestion administered by generally were less disturbed by the situation
the hypnotist at the completion of the than they were in the waking state. The sub-
experimental tasks. jects who attempted any of the activities
Results. Five of the six hypnotized subjects claimed they were much more hesitant in the
carried out all of the suggested activities.8 waking state than they had been in the
The major result of the earlier studies of previous hypnotic state. These reported dif-
Rowland (1939) and Young (1952) was thus ferences were consistent with the observer's
replicated: hypnotized subjects will carry out ratings of such differences between the two
apparently antisocial and injurious activities performances, summarized in Table 2.
such as handling dangerous reptiles and
concentrated acid. Waking Control Used by Rowland (1939):
Group 2
Hypnosis Subjects Retested as Own Control Treatment. As far as could be ascertained,
(Young, 1952): Group 1
the six subjects in this group were treated
Treatment. The five subjects in the real similarly to the informal control group in
hypnosis group who had carried out the Rowland's (1939) study.
actions returned the following day. Each sub- The experimenter implicitly communicated
ject was asked, in the waking state, whether the expectation that the subject would refuse
he would be willing to carry out the suggested to carry out the requested activities. The
activities involving the snake and acid. This apparent danger was stressed at least as
was the control measure carried out by Young much as it was with subjects in other groups.
(1952). Although the verbal content of the The request to carry out the various actions
request implied the possibility of compliance, was phrased more in the form of a question
it was conveyed by the experimenter's general of whether he would be willing to comply
manner and nonverbal behavior that compli- with such obviously dangerous actions. How-
ance with the request was not necessarily ever, persuasion and pressure were applied
expected. if a subject refused to carry out any activity,
Results. As found by Young (1952), and although not as insistently as for subjects
as expected in the present study, subjects in the hypnotized and simulating groups.
generally refused to carry out the actions in Results. Three of the six subjects in this
the waking state, even though they had been group attempted to pick up the snake, but
quite willing to carry them out in the hyp- only one of them attempted to take the coin
notic condition. However, two subjects could from the acid, and throw the acid at the
The remaining subject became emotionally dis- assistant, both after considerable hesitation.
turbed at the sight of the harmless lizard and it It had been expected that subjects in this
was not possible to administer the remaining tasks. group would refuse to carry out the activi-
As the subject had passed the selection criteria, it ties.9 Some subjects could be persuaded to
was decided not to replace her. It is noted that she
was one of the two hypnotic-group subjects whose It was frequently noted in the records made
previous .experience with the experimenters was by the assistant that the experimenter exerted more
minimal, and it is possible that the interpersonal pressure with subjects in the three control groups
relationships existing were not as strongly developed in which noncompliance was predicted than our
as with other subjects. understanding of the parallel procedures of Row-

carry out the tasks. It appears that either havior is within the limits of what they per-
the experimenter could not refrain from being ceived as legitimate and reasonable within
more insistent than Rowland, or the psycho- the experimental context. Such a result would
logical experiment is more institutionalized demonstrate that the present tasks do not
for University of Sydney students in 1960 provide a test of whether hypnosis can be
than it was for University of Tulsa students used to compel the subject to carry out
in 1939. antisocial behavior.
Results. All six subjects who were simu-
Summary of Attempted Replication oj lating hypnosis attempted to comply with the
Studies of Rowland (1939) and apparently antisocial and injurious activities
Young (1952) requested of them. They objectively at-
The results obtained from the three groups tempted to carry out the various tasks at
reported in the present study are in substan- least as readily as the subjects who were
tial agreement with the findings reported by actually hypnotized.
Rowland and by Young. Their results, as far Qualitatively, it appeared that the simu-
as comparisons are relevant, have been lators were more hesitant than the hypnotized
confirmed, and their procedure has been subjects about grasping the snake, less suc-
replicated successfully. cessful in their attempts to remove the coin
from the acid, but more violent and less hesi-
Limits oj Social and Behavioral Control in tant about throwing the acid solution at the
the Experimental Situation assistant10 (see Table 2).
Rowland and Young concluded from their
Waking Compliance Control: Group 4
data that hypnosis was instrumental in pro-
ducing antisocial and dangerous behavior Treatment. There are special motivational
which could not be elicited otherwise. How- aspects involved when a subject simulates
ever, these results cannot answer the funda- hypnosis which are not present when a sub-
mental question of whether the apparently ject is actually hypnotized. A simulator is
antisocial behavior elicited under hypnosis motivated by two special features: he must
exceeds the limits of behavior which may be attempt to please one experimenter by
legitimized by the experimental situation per striving to deceive a second experimenter. It
se. To answer this question, it is necessary is possible that these special motivations may
to test whether the request to carry out the increase the range of social and behavioral
apparently antisocial and self-injurious acts control implicit within the unique type of
exceeds the limits of social and behavioral experimental situation which exists for the
control implicitly existing in the experimental simulating subject.
situation. In an attempt to evaluate this possibility,
an additional set of six subjects served as
Simulators: Group 3 the usual type of control group. They were
Treatment. Because the experimenter did told that they were a normal waking control
not know whether he was testing a hypno- group in a study employing hypnotized sub-
tized or a simulating subject at any time, jects. As far as possible, • they were treated
both the hypnotized and simulating groups identically to the hypnosis and the simulating
of subjects were treated alike, and were groups; the experimenter's manner conveyed
exposed to the same demand characteristics an expectation that subjects would comply
concerning compliance with the experiment- with his requests. This group differs from the
er's requests. Consequently, compliance with 10
For example, some simulators (and some mem-
the requests of the experimenter by the simu- bers of Group 4 below) hurled both solution and
lators would indicate that the requested be- container at the assistant. In comparison the hyp-
nosis group, though throwing the solution, some-
land and Young warranted. The excessive persuasion times failed to splash the assistant with it. These
was not intended: we do not know if this ac- subjective differences were observed by the as-
counted for the unpredicted successful attempts in sistant who was aware of the group membership
these groups. of the subjects.

informal controls employed by Rowland (see Results. Two subjects attempted to handle
Group 2 above) in that these subjects were the snake and to throw the acid, and a third
treated as formal experimental subjects and subject attempted to take the coin from the
compliance with the requests was taken for acid. The results are somewhat equivocal, but
granted, whereas such pressure was not ap- it appears that the "shaping" procedure em-
plied in Group 2, described above, or by ployed is not an essential determinant of
Rowland. subsequent behavior.
Results. Five subjects attempted to carry
out the two tasks involving the acid, but two Informal Controls
of these subjects refused to handle the ven- A number of colleagues were informally
omous snake. The performance of this group shown the experimental tasks. These faculty
was only slightly inferior to that of the hyp- members were treated in a fashion similar
notized and simulating groups. It would ap- to the treatment of the control group used
pear that consideration of the special motiva- by Rowland (1939).
tions which existed for subjects simulating They invariably refused to carry out even
hypnosis is not necessarily essential to the least objectionable of these three tasks,
account for the present results. the removal of the penny from the nitric acid.
The fellow faculty members who were asked
Elimination of "Shaping" Procedures: to perform these tasks could not be pressured
Group 5 or persuaded in any way to undertake them.
Treatment. It is possible that the pre- It is significant that their behavior was
liminary tasks, taking inanimate objects from identical to that of Rowland's informal
the compartment and handling harmless rep- control group.
tiles, could operate as a type of "shaping" Two aspects appear relevant in interpreting
or "conditioning" procedure, gradually lead- these informal observations:
ing the subject to tasks which in themselves 1. The faculty members were not in a
look frightening, but in the context of the formal experimental situation.
"shaping" procedure lose much of their 2. A different role relationship existed
fearsome potency. between faculty members and the experi-
A separate waking control group was con- menters compared to that existing between
ducted without the preliminary harmless subjects and experimenters. These differences
tasks. These subjects were asked to partici- emphasize the importance of the special con-
pate in a psychological experiment, but they texts provided by the combined subject-
were not told that they were controls in a experimenter relationship and the nature of
study involving hypnosis. Apart from the the experimental situation in terms of the
elimination of the innocuous tasks, the pro- willingness of the subjects to carry out the
cedure was essentially the same as for Group experimental tasks.
2. Unfortunately, the pressure exerted to gain
compliance was generally greater than it was Postexperimental Inquiry
for Group 2." With few exceptions, subjects who at-
There was some confusion between the two tempted to carry out the requested activities
experimenters when subjects in this group were reported in the postexperimental inquiry that
being tested. While the experimenter was testing they felt quite safe in the experimental situa-
some subjects in this group, the relevant question
was seen as to whether expectation of failure could even the simulating subjects, Group 3. Clearly two
be as readily communicated with and without the separate groups should have been tested. It is certain
preliminary "shaping" procedures, in which it was that some subjects within this group were treated
predicted the results would be similar to Group 2. quite differently from others. The confusion was
For other subjects the question was implicitly seen not recognized until after the completion of the
as whether subjects would comply as readily with study, and it is not known to what extent the
the tasks when the innocuous tasks were eliminated equivocal results are concordant with the differing
compared to the ease of gaining compliance when expectations that were undoubtedly induced. The
they were included. It was predicted from this view- results for the group are reported only for the
point that results would be similar to Group 4, and sake of completeness.

tion. Though they reported feeling rather by hypnotic techniques lies with the investi-
uncertain about the tasks, and reported gator, who must demonstrate that the so-
strong emotional reactions to the repugnant called antisocial behavior does exceed that
activities, the subjects reported that they which is legitimized by the experimental
were quite convinced that they would not situation, and that the behavior is perceived
be harmed because the context was an experi- by the subject as truly dangerous or anti-
mental one, presumably being conducted by social. It is our belief that it may not be
responsible experimenters. All subjects ap- possible to test the antisocial question in an
peared to assume that some form of safety experimental setting because of the prob-
precautions had been taken during the experi- lems of finding tasks which are not seen as
ment. Although no subject reported suspect- legitimized by the experimental context.
ing the presence of a glass screen, subjects 3. It may seem surprising that the waking
felt that the snake had either been milked control group subjects who were pressed to
of its venom or defanged. Some felt that the comply with the apparently dangerous be-
solution was not really acid, even when they havior were quite willing to follow the
plunged their hand into it; others suspected experimenter's commands. This confirms the
the assistant could duck in time to avoid anecdotal evidence reported above from our
the acid; or the glass screen would be in own laboratory: we still have not found an
place again (having previously discovered its aspect of behavior which is sufficiently safe
presence during the procedure with the to request of a subject, and which a subject
snake); or perhaps the assistant would even will refuse to carry out if the expectation
plunge himself into a nearby tub of water; of compliance is communicated to him. In the
or in one case, that the assistant was not present study the experimenter could virtu-
actually there—it was an illusion produced ally predetermine the nature of the resulting
by a complex arrangement of mirrors. behavior by deciding in advance whether he
would consciously, but subtly, communicate
DISCUSSION to a subject an expectation either of failure
1. No conclusions can be drawn from the or of compliance. Such a result is concordant
present investigation about the potential use with other findings (for example, Rosenthal,
of hypnosis to induce antisocial behavior. 1964) showing the relative ease with which
However, the study clarifies conclusions an experimenter may nonconsciously bias the
which may be drawn from two previous results of a study by subtly communicating
studies. After replicating the studies by his own expectations and hypotheses.
Rowland (1939) and Young (1952) and con- Failure to gain compliance with hypnotized
firming their results, it has been shown that subjects would indicate inadequate replica-
similar apparently antisocial behavior can tion of the exact details of Young's investiga-
be elicited in control groups treated in an tion. Because the experimenter is unaware of
identical fashion to a group of hypnotized the group membership of real and simulating
subjects. The apparently antisocial actions subjects outcome-bias should not differen-
were also carried out successfully by sub- tially affect performance of subjects in these
jects who were not hypnotized, indicating two crucial groups. It is partly because of the
that the tasks are within the broad range problem of outcome-bias that simulating
of activities which are perceived as legiti- subjects are included as the crucial control
mized by the nature of the situation: they group.
were requests made by experimenters, viewed 4. The purpose of the simulating group
by subjects as responsible scientists, in the was to examine whether the chosen behavior
context of a psychological experiment. exceeded the limits of behavior legitimized
2. The present study is essentially method- by the special nature of the experimental
ological, demonstrating experimental condi- context. The waking control subjects (Group
tions which are necessary to investigate the 4) also complied with the requested activities,
antisocial hypothesis. The burden of demon- and in this sense the simulating subjects were
strating the production of antisocial behavior not essential to demonstrate that the alleged

antisocial actions were within the realm of ORNE, M. T. The nature of hypnosis: Artifact and
what subjects accept as reasonable requests essence. Journal of Abnormal and Social Psychol-
ogy, 1959, 58, 277-299.
in the experimental situation. However, the ORNE, M. T. Antisocial behavior and hypnosis:
simulating group of subjects cannot be elimi- Problems of control and validation in empirical
nated from future investigations of this type. studies. Paper presented at Colgate Symposium
It is not always possible to determine whether on Hypnosis, Colgate University, April 1960. (In
G. H. Estabrooks [Ed.], Hypnosis: Current prob-
waking control subjects have been treated lems. New York: Harper & Row, 1962. Pp. 137-
differently from the nonhypnotized subjects, 192)
even though an experimenter is not aware ORNE, M. T. On the social psychology of the psycho-
of such differential treatment. logical experiment: With particular reference to
5. The present investigation demonstrates demand characteristics and their implications.
American Psychologist, 1962, 17, 776-783.
the misleading conclusions that may be drawn ORNE, M. T., & SCHEIBE, K. E. The contribution of
when casual, informal "control" groups, of nondeprivation factors in the production of sen-
the type employed by Rowland (1939) and sory deprivation effects: The psychology of the
Young (1952), form the basis for evalu- "panic button." Journal of Abnormal and Social
Psychology, 1964, 68, 3-12.
ating experimental performance. It cautions ROSENTHAL, R. Experimenter outcome-orientation
against making untested assumptions about and the results of the psychological experiment.
the way in which subjects will behave in Psychological Bulletin, 1964, 61, 405-412.
an experimental situation, or making as- ROSENTHAL, R., & FODE, K. L. Psychology of the
sumptions about what aspects of behavior scientist: V. Three experiments in experimenter
bias. Psychological Reports, 1963, 12, 491-511.
are within the repertoire of the subjects, ROSENTHAL, R., & LAWSON, R. A longitudinal study
however reasonable these assumptions maj of the effects of experimenter bias on the operant
appear, when subjects are participating ii. learning of laboratory rats. Journal of Psychiatric
the social phenomenon known as "the Research, 1963, 2, 61-72.
ROWLAND, L. W. Will hypnotized persons try to
psychological experiment." harm themselves or others? Journal of Abnormal
REFERENCES and Social Psychology, 1939, 34, 114-117.
SCHILDER, P., & KAUDERS, 0. Hypnosis. (Trans, by
BARBER, T. X. Antisocial and criminal acts induced S. Rothenberg) Nervous and Mental Disease
by "hypnosis": A review of experimental and Monograph Series, 1927, No. 46.
clinical findings. Archives of General Psychiatry, SHOR, R. E. Physiological effects of painful stimula-
1961, S, 301-312. tion during hypnotic analgesia under conditions
ERICKSON, ELIZABETH M. Observations concerning designed to minimize anxiety. International Jour-
alterations in hypnosis of visual functions. Amer- nal of Clinical and Experimental Hypnosis, 1962,
ican Journal of Clinical Hypnosis, 1962, S, 131- 10, 183-202.
134. WEITZENHOEFER, A. M. The production of antisocial
ERICKSON, M. H. An experimental investigation of acts under hypnosis. Journal of Abnormal and
the possible anti-social use of hypnosis. Psychiatry, Social Psychology, 1949, 44, 420-422.
1939, 2, 391-414. WEITZENHOFFER, A. M. Hypnotism: An objective
ESTABROOKS, G. H. Hypnotism. New York: Button, study in suggestibility. New York: Wiley, 1953.
1943. WELLS, W. R. Experiments in the hypnotic produc-
FRANK, J. D. Experimental studies of personal pres- tion of crime. Journal of Psychology, 1941, 11,
sure and resistance: I. Experimental production 63-102.
of resistance. Journal of General Psychology, 1944, WOLBERG, L. R. Hypnoanalysis. New York: Grune
30, 23-41. & Stratton, 1945.
MEARES, A. A system of medical hypnosis, Phila- YOUNG, P. C. Antisocial uses of hypnosis. In L. M.
delphia: Saunders, 1960. LeCron (Ed.), Experimental hypnosis. New York:
MILGRAM, S. Behavioral study of obedience. Jour- Macmillan, 1952. Pp. 376-409.
nal of Abnormal and Social Psychology, 1963,
67, 371-378. (Early publication received June 9, 1964)