Hypnosis and Placebos: Response Expectancy as a Mediator

- 99 -
anales de psicología
1999, vol . 15, nº 1, 99-110
© Copyright 1999: Servicio de Publicaciones de la Universidad de Murcia
Murcia (España). ISSN: 0212-9728
Hypnosis and Placebos: Response Expectancy as a Mediator
of Suggestion Effects
Irving Kirsch
University of Connecticut
Titulo: Hipnosis y placebos: la expectativa de respuesta
como mediador de los efectos de la sugestión
Resumen: En este artículo se revisan datos que sugieren
que los efectos de la hipnosis y de los placebos comparten
un mismo mecanismo: la expectativa de respuesta. Los
placebos y los procedimientos hipnóticos tienen efectos
terapéuticos porque modifican las expectativas de los
clientes. Se conciben ambos como ejemplos del fenómeno
más amplio de la sugestión. Otras variables pueden jugar un
papel importante en la respuesta a las sugestiones, pero aún
falta por saber si afectan a dichas respuestas a través de una
modificación de las expectativas o tienen un impacto
directo sobre las respuestas. Se discuten también los efectos
negativos de la utilización de la idea del trance en la terapia.
Palabras clave: Hipnosis, placebo, sugestión, expectativa de
respuesta
Abstract: In this paper, data are reviewed indicating that
hypnotic and placebo effects share a common mechanism:
response expectancy. Placebos and hypnotic procedures
may have therapeutic effects because they change the
client's expectations. Both placebo and hypnosis are
viewed as instances of the broader phenomenon of
suggestion. While other variables may play important roles
in the response to suggestions, it remains to be known
whether they affect responses by changing expectations or
have a direct impact on them. The potential disdvantages of
using the idea of trance in therapy are also discussed.
Key words: Hypnosis, placebo, suggestion, response expectancy
Hypnosis is a procedure in which a person
designated as hypnotist suggests changes in experience
to a person designated subject (Kirsch,
1994; Kirsch & Lynn, 1995). A representative
range of hypnotic suggestions are sampled in
standardized scales of suggestibility, most of
which are highly reliable and substantially correlated
with each other (Council, in press).
Suggestions on these scales are generally
thought to be of three basic types. Ideomotor
suggestions are suggestions that a particular action,
such as an arm rising in the air (arm levitation),
will occur automatically, without
awareness of volitional effort. Challenge suggestions
are suggestions that the person cannot
perform an act that is normally under voluntary
control, such as bending an arm (arm rigidity).
Cognitive suggestions are suggestions for
various cognitive or perceptual distortions,
such as selective amnesia, pain reduction, and
hallucinations. Note that some cognitive sug-
* Dirección para correspondencia / Address for correspondence:
Irving Kirsch. University of Connecticut. Storrs,
CT 06269-1020. United States of America.
E-Mail: irving@uconnum.uconn.edu
gestions contain challenges. Amnesia suggestions,
for example, require that the person not
recall particular information. These hypnotic
phenomena have been characterized as the
"domain of hypnosis" (Hilgard, 1973). An
adequate theory of hypnosis should be able to
account for the full range of behavior contained
in this domain.
There are many books and journal articles
focusing on the topic of hypnosis and hypnotizability.
In contrast, despite notable exceptions,
such as the work by Amigó (in press)
and Capafons (in press) at the University of
Valencia, the broader subject of suggestion has
been relatively neglected. There are historical
reasons for this state of affairs. The effects of
mesmerism and hypnosis have seemed so unusual
that many observers dismissed them as
fraudulent, and others attributed them to some
special condition or state. It seemed impossible
for suggestion to have such extraordinary
effects, without there being some kind of special
state creating them.
The results of research on hypnosis in the
twentieth century have clearly invalidated that
conclusion. One of the first things that was
100 I. Kirsch
anales de psicología, 1999, 15(1)
learned through experimental research on
hypnosis was that all behaviors seen in hypnosis
can also be obtained without hypnosis.
Clark Hull (1933, p. 391) wrote:
The only thing which characterizes hypnosis as such
and which gives any justification for calling it a "state"
is its generalized hypersuggestibility. That is, an increase
in suggestibility takes place upon entering the
hypnotic trance. The difference between the hypnotic
and normal state is therefore quantitative rather than
qualitative. No phenomenon whatever can be produced
in hypnosis that cannot be produced to lesser
degrees by suggestions given in the normal waking
condition. the essence of hypnosis lies in the fact of
change in suggestibility [emphasis in the original].
These data suggest that suggestion, rather than
hypnosis, is the fundamental phenomenon on
which we should focus.
This conclusion is reinforced by two more
recent sources of data. One is the work of
nonstate theorists, and in particular, the experimental
work of T. X. Barber and his followers.
It was Barber (1969) who most clearly
demonstrated that all of the observed effects
of hypnosis, including the increase in suggestibility
that was observed following a hypnotic
induction, could be duplicated by nonhypnotic
procedures. Thus, hypersuggestibility, the single
remaining justification that Clark Hull
could find for retaining the trance concept,
can be produced by nonhypnotic methods.
Importantly, the Barber Suggestibility Scale
differed from previous scales in that it included
a measure of subjective or experiential
response to suggestion, as well as a measure of
behavioral response. Thus, Barber was able to
demonstrate that thesubjective responses to
suggestion did not depend on the induction of
a hypnotic trance. This, of course, cast doubt
on the concept of hypnotic trance, doubt
which has been reinforced by the failure to
find any physiological or even self-report
markers of the hypothesized state. It turns out
that the most prominent experts cannot distinguish
the self-reports of hypnotized subjects
from those of subjects who have not been
hypnotized (Kirsch, Mobayed, Council, &
Kenny, 1992).
Paradoxically, it is the work of E. R. Hilgard
(1965) on suggestibility as a trait that
most clearly indicates the need for a change in
focus from the topic of hypnosis to the topic
of suggestion. Hilgard's data indicated that the
effect of hypnotic inductions on suggestibility
were generally quite small. A person who responds
to six of the twelve suggestions contained
in a typical hypnotic suggestibility scale
is likely to respond to five of them even without
the induction of hypnosis (Kirsch, 1997).
Some of the effects of hypnosis may not
be suggested by the hypnotist. The best known
of these is so called spontaneous amnesia,
which in past centuries was a hallmark of hypnosis,
but currently is virtually nonexistent. Although
not directly suggested by the hypnotist,
these phenomena are far from spontaneous.
Instead, they too are products of suggestion--
in this case, suggestions that are transmitted by
the culture. Young and Cooper (1972) demonstrated
this quite nicely. They told one group
of subjects that hypnotized people experience
spontaneous amnesia, and they told a second
group that hypnotized subjects do not experience
spontaneous amnesia. The subjects were
later tested for their beliefs about hypnosis.
Forty-eight percent of those in the first group
agreed, with the statement "If I were to be
hypnotized I would not remember what had
happened after I woke up (Young & Cooper,
1972). In contrast, only 15% of the second
group expected to experience spontaneous
amnesia. When later hypnotized and tested for
so called "spontaneous" amnesia, 37% of the
subjects in the first group displayed it, compared
to only 10% of the in the second group.
Thus, the occurrence of amnesia was hardly
"spontaneous." This is further confirmed by
another interesting finding in these data.
Across both groups, 75% of the subjects who
expected amnesia experienced it, whereas
none of those who did not expect amnesia experienced
it. This illustrates a central theme of
this paper: the mediating role of expectancy in
the ability of suggestion to elicit a response.
What makes hypnosis interesting is people's
responses to such suggestions as analgesia,
amnesia, age regression, nonvolitional
movements, and positive and negative halluciHypnosis
and Placebos: Response Expectancy as a Mediator of Suggestion Effects 101
anales de psicología, 1999, 15(1)
nations. Without phenomena such as these,
there would be little to interest scholars or lay
persons in hypnosis. So if these responses do
not require hypnosis for their production, and
if the effect of hypnosis is merely to enhance
them to a slight degree, then clearly, the focus
of investigation should be on the broader topic
of suggestion and its effects.
Broadening the focus of attention from
hypnosis to suggestion also has the advantage
of adding the very important topic of placebo
effects. Placebos are typically viewed as artifacts
to be controlled in treatment outcome research.
However, the documented effects of
placebos on pain, asthma, tension, anxiety, depression,
blood pressure, heart rate, sexual
arousal, skin conditions, nausea, vomiting, gastric
motility, and angina (reviewed in Kirsch,
1990) reveals it to be a very important phenomenon,
well worthy of study in its own
right.
There are, of course, both similarities and
differences between hypnotic and placebo
phenomena. Whereas responses to the kind of
suggestions typically used in hypnosis are notoriously
trait-like, individual differences in response
to placebos appear to be very unreliable.
There does not appear to be a "placebo
reactor," comparable to the "hypnotizable subject."
Still, the stimulus for hypnotic and placebo
responses are suggestions, and hypnotic
inductions are so nonspecific that they might
well be regarded as placebos. In fact, placebo
substances have been used successfully as
hypnotic inductions (Baker & Kirsch, 1993;
Glass & Barber, 1961). Most important, as will
be shown in this paper, the effects of hypnotic
and placebo suggestions are mediated by response
expectancy (Kirsch, 1985, 1990). But
first, it may be worthwhile to define what is
meant by the term suggestion.
What is a Suggestion?
A suggestion is a communication indicating
that an individual will experience a particular
response. It differs from an instruction or
command in that the response is to occur
nonvolitionally, rather than emitted intentionally.
Often this distinction is clear in the phrasing
of the suggestion. "Raise your hand" is an
instruction, whereas "Your hand is getting
lighter and beginning to rise" is a suggestion.
The first calls for an intentional act; the second
for a nonvolitional response. Similarly,
"Take this medication" is an instruction. In
contrast, "It will help you sleep" is a suggestion,
because in suggests to the person that
taking the pill will automatically induce sleep.
Sometimes, the suggestive nature of a
communication is not apparent from its linguistic
character. In fact, suggestions need not
be linguistic utterances at all. Suggestive information
may be conveyed by the size, shape,
and color of a pill, for example, or by the behavior
of a model. Furthermore, the suggestive
nature of words depend on more than the
words themselves. The command "Sleep!" for
example, can be interpreted as either an instruction
or a suggestion, depending on the
context in which it is delivered. When given to
a child who is still awake at 11 p.m., it is
clearly a command; when given to a volunteer
at a stage hypnosis performance, it is a suggestion.
What is important, then, is the meaning
of the communication, as it is understood by
the person to whom it has been directed.
Of course, most hypnotic subjects want to
experience suggested phenomena, and they intentionally
do their best to bring them about.
However, the experience of automaticity or
nonvolition is part of the communicated suggestion,
so that simple behavioral compliance
would not be experienced as a successful response.
Statements are often interpreted as suggestions,
rather than commands, because the response
is not experienced as being under direct
volitional control. Most people cannot intentionally
reduce pain, see things that are not
there, sleep, or forget, in the same direct way
that they can raise or lower their arms. If they
can accomplish these responses at all, it is by
doing other things that are under volitional
control. For example, they may think about
other things to distract themselves from the
102 I. Kirsch
anales de psicología, 1999, 15(1)
pain, or they might try to imagine the suggested
hallucination. Their behavior is much
like that of an actor who thinks of sad experiences
in order to produce tears. The thoughts
of the sad experiences are voluntary acts
aimed at producing the nonvolitional response
of tears. Still, it is the contextually determined
meaning of the communication that differentiates
a suggestion from an instruction or command.
The stage director who tells an actor to
cry is giving an instruction; whereas the physician
who says "Crying is a side effect of this
medication" and the hypnotist who says "You
are beginning to cry" are giving suggestions.
Finally, it is important to note that whether
or not some verbal or nonverbal stimulus is a
suggestion does not depend on the response of
the subject. Suggestions are suggestions, even
when subjects do not respond to them.
Suggestions, Expectancies, and Nonvolitional
Responses
A suggestion is a particular type of stimulus. It
is a stimulus that conveys information that a
nonvolitional response will occur. Sometimes
the response occurs, sometimes it does not. So
the question is, what is it that determines the
response to a suggestion? In my work on hypnosis
and placebo effects, I have focused on
the role of expectancy as a mediating variable.
Accepting a suggestion means coming to expect
that the suggested event will occur.
That expectancies are determinants of behavior
is a central tenant of cognitivebehavioral
learning theory (Tolman, 1955) social
learning theory (Bandura, 1977; Rotter,
1954; Mischel, 1973), behavioral decision theory
(Edwards, 1954), achievement motivation
theory (Atkinson, 1957; Heckhausan, 1977),
and the theory of reasoned action (Ajzen &
Fishbein, 1980). In all of these theories, behavior
is postulated to be a multiplicative function
of expected outcomes and their values.
In these theories, expected outcomes are
generally thought of as external events, such as
food, water, money, school grades, recognition,
and affection. In developing response expectancy
theory, I thought it important to distinguish
between these stimulus expectancies and
response expectancies (Kirsch, 1985). Response
expectancies are expectancies for the occurrence
of nonvolitional responses, such as pain,
alertness, fear, sadness, and joy. Like expected
stimuli, expected responses are valued outcomes,
and as such they are determinants of
voluntary behavior. We may drink coffee in
the morning to help wake us up, and we may
avoid it in the evening, if we think it will keep
us from getting to sleep. However, response
expectancies seem to have a property that
stimulus expectancies do not have: They are
self-confirming. When people expect to feel
alert, they often do feel alert; and when people
expect to stay awake, they may find themselves
unable to sleep.
Like stimulus expectancies, response expectancies
are derived from direct and vicarious
experience. Experience with active drugs
and medications, for example, leads us to expect
those drugs and medications to have particular
effects. Similarly, being told that a drug
has a particular effect or observing its effect
on others can produce an expectation of that
effect when the drug is ingested. However,
data indicate that direct experience is more
powerful than vicarious experience in shaping
response expectancies (Wickless & Kirsch,
1989). Response expectancies are somewhat
similar to self-efficacy expectations, and in
some circumstances the two constructs overlap
considerably (Kirsch, 1985). For example,
expected anxiety (a response expectancy) is the
primary determinant of self-efficacy for approaching
a feared object or situation
(Schoenberger, in press),and a hypnotic response
expectancy is closely related to the belief
that one is capable of experiencing a suggested
effect. Unlike self-efficacy expectations,
however, response expectancies are anticipations
of outcomes that will occur and are only
partly based on evaluations of one’s ability.
The response expectancy that coffee will make
one alert, for example, is likely to be independent
of self-efficacy judgments. Also, because
the experience of automaticity is central
Hypnosis and Placebos: Response Expectancy as a Mediator of Suggestion Effects 103
anales de psicología, 1999, 15(1)
to hypnosis, expectancies for responding are
more properly termed response expectancies.
The self-confirming nature of response expectancies
indicate that they might mediate the
effects of suggestion. The logic behind this
conclusion is as follows. First, the suggestive
nature of a communication is defined by the
recipient's interpretation that a nonvolitional
response is expected to occur. Second, response
expectancy is defined as an expectancy
of a nonvolitional response (Kirsch, 1985), in
contrast to intention, which has been defined as
the expectancy of a voluntary behavior (Ajzen
& Fishbein, 1980). Thus, a connection between
response expectancy and suggestion is
implied by the definitions of these terms. Perhaps
we are hard-wired in such a way that expecting
a subjective experience produces that
experience, in the same way that deciding
(termed intending in reasoned action theory) to
emit a voluntary act (e.g., lifting one's arm)
produces that act. The two clearest examples
of this phenomenon are placebo effects and
hypnosis.
Placebo Effects
Although the mechanisms by which placebos
produce their effects have not yet been established,
the data exclude some hypotheses and
indicate some parameters within which any
successful theory must fit. Most importantly,
explanations of placebo effects must account
for their specificity. Data indicating the highly
specific nature of placebo effects indicate that
they cannot be explained fully by such global
factors as rapport, trust, faith, hope, anxiety
reduction, or endorphin release.
First, it is clear that placebo effects are not
entirely due to the quality of the doctor/
patient relationship. Although such relationship
factors as touch have been demonstrated
to affect some therapeutic outcomes
(Whitcher & Fisher, 1979), there are ample
data indicating that the effects of placebos depend
on their information value. With relationship
factors held constant, different placebos
produce different results. For example,
placebo injections are more effective that placebo
pills (Traut & Passarelli, 1957). Also, placebo
effects are readily obtained in relatively
sterile, nonclinical, experimental settings.
Second, placebos produce both positive
and negative effects, and they do so in the
same people. Furthermore, the specific nature
of the effect (i.e., whether the placebo affects
gastric motility, sexual arousal, pain perception,
etc.) depends on the information available
to the recipient. Placebo analgesics, for
example, have very different effects than placebo
tranquilizers. These characteristics of
placebo effects cannot be explained by faith,
trust, hope, or any other hypothesis based on
the valence of placebo-induced expectancies.
Third, recent data from my laboratory indicate
that placebo pain reduction cannot be
explained by mechanisms like anxiety reduction
or endorphin release (Montgomery &
Kirsch, 1996). These are global mechanisms
that would affect the entire body. We obtained
a placebo effect by administering the placebo
in the guise of a local anesthetic and applying a
pain stimulus to treated and untreated parts of
the body. Because the pain stimulus was applied
simultaneously to both the treated (by
placebo) and untreated locations, the differences
in reported pain could not have been
due to any global changes in sensitivity, perception,
or affect. After all, one cannot be anxious
in one hand and calm in the other.
Taken together, these data demonstrate
that placebo effects are specific to the information
with which the placebo is administered.
This specificity makes theories based on global
mechanisms implausible. An adequate theory
of placebo responding must be able to account
for the specific nature of the effects.
Classical Conditioning
Classical conditioning is one of the most
frequently proffered theoretical explanations
of placebo effects (Turkkan, 1989; Wickramasekera,
1980). According to conditioning
accounts of placebo effects, effective medical
treatments are conditioning trials, during
104 I. Kirsch
anales de psicología, 1999, 15(1)
which the vehicles (pills, capsules, etc.) in
which active medication (the US) is delivered
become conditional stimuli (CSs), thereby acquiring
the capability of evoking the effects of
medication as conditional responses (CRs).
The classical conditioning model has the advantage
of being able to account for the specificityof
placebo reactions. It predicts side effects
as well as therapeutic effects, and it is
consistent with the observation that placebos
evoke the same responses as the active drugs
they are replacing. However, there are two
problems with conditioning explanations of
placebo phenomena. First, they are based on
an outdated account of classical conditioning.
Second, there are abundant disconfirming data.
Traditional accounts of classical conditioning
suggest that parings of the CS with the US
lead to the automatic evocation of URs following
presentation of the US. In contrast, contemporary
theorists (e.g., Rescorla, 1988) view
classical conditioning as a means by which information
is acquired. Conditioning trials endow
the CS with information value, so that it
becomes a cue for the occurrence of the US.
As a result, the CR may not be the same response
as the unconditional response (UR). Instead,
the CR is an anticipatory response that
prepares the organism for the onset of the US.
Also, pairings of conditional and unconditional
stimuli result in conditioning only under those
circumstances in which they impart information
value to the CS. One way of interpreting
this is that the effect of conditioning trials on
behavior is mediated by expectancies, so that if
expectancy change is blocked, so too is the effect
on behavior.
This contemporary understanding of conditioning
phenomena can explain how placebos
(the CS) come to generate internal representations
of active medications (the US), and
since responses have stimulus value, it is not
much of a stretch to invoke this model to account
for the acquisition of response expectancies
as a function of conditioning trials
(Kirsch, 1985). However, this leaves the occurrence
of the placebo response (the presumedCR)
unexplained. It is not difficult to understand
the occurrence of salivation following a
stimulus that signals food. The conditional
stimulus elicits the expectation of food, the
thought of food causes the organism to salivate.
Similarly, because of their previous association
with active medication, pills elicit expectations
of particular changes. But why
should the expectation produce the expected
effect?
Unlike contemporary accounts of classical
conditioning, traditional stimulus substitution
models seem to explain the occurrence of the
placebo responses. However, leaving aside the
data that led to the replacement of that model
by contemporary cognitive models, data from
many placebo studies are inconsistent with
traditional conditioning models: First, with
tranquilizers as UCs, conditioning trials
weaken the placebo response instead of
strengthening it (Meath, Feldberg, Rosenthal,
& Frank, 1956; Pihl & Altman, 1971; Rickels,
Lipman, & Raab, 1966; Segal & Shapiro, 1959;
Zukin, Arnold, & Kessler, 1959). Also, there is
an inverse relation between the strength of the
US (i.e., the tranquilizer) and the magnitude of
the placebo effect that is presumed to be the
CR (Rickels et al., 1966). Second, placebo effects
can resist extinction over periods as great
as two years or more (Boissel, et al., 1986;
Coryell & Noyes, 1988; Traut & Passarelli,
1957). Third, when people expect effects that
are different from those produced chemically
by the drug, the placebo effect is consistent
with the expectation rather than with the UR
(Fillmore & Vogel-Sprott, 1992; Hull & Bond,
1986; Kirsch & Weixel, 1988). Finally, under
some circumstances, placebos can produce effects
that are stronger than those of the active
drugs that are presumed to be the USs
(Frankenhaeuser, Post, Hagdahl, & Wrangsjö,
1964; Ikemi & Nakagawa, 1962; Lyerly, Ross,
Krugman, & Clyde, 1964; Ross, Krugman,
Lyerly, & Clyde, 1962; Wolf, 1950). All of
these data are inconsistent with explanations of
placebo effects based on simple Pavlovian
models.
Hypnosis and Placebos: Response Expectancy as a Mediator of Suggestion Effects 105
anales de psicología, 1999, 15(1)
Psychophysiological Explanation and
the Hypothesis of Unmediated Expectancy
Effects
Faith, hope, rapport, and anxiety reduction
are psychological intervening variables that
have been hypothesized to mediate the relation
between expectancy and expected response.
The data I reviewed earlier in this paper
indicate that none of them are necessary
for evocation of placebo effects. It is possible
that there are no intervening psychological
variables between a response expectancy and
an expected response (Kirsch 1985). In this
sense, the relation between expectancy and response
may be the same as that between intention
and intended response. The relation between
an intention and a voluntary response
has been characterized as “immediate� (Ajzen
& Fishbein, 1980), meaning that it is not mediated
by any psychological intervening variables
between the intention and the response. Indeed,
recent work in cognitive and social psychology
indicate that previously formed intentions
may activate behavior automatically, with
no need to focus attention on the intended response
at the time that it is initiated (Bargh, J.
A., & Gollwitzer, 1994; Kirsch & Lynn, 1997,
1998). Similarly, the relationship between sadness
the perception of loss is presumed to be
immediate, in this sense of the term. One may
ask, for example, "Why are you sad?" and the
answer may be "Because my mother just died."
There seems no need to ask, "Why did that
make you sad?" I have hypothesized that the
relation between response expectancies and
nonvolitional responses may have this same
immediate quality. If this is the case, further
explanation requires consideration of physiological
rather than psychological variables.
One of the pitfalls to avoid in psychophysiological
theorizing is that of treating a
physiological variable as a mediating variable
between a psychological variable and a dependent
variable. For example, a finding that
endorphin release may be required for certain
kinds of pain reduction tempts one to consider
it an explanation of placebo pain reduction.
Notwithstanding the importance of such data,
they do not constitute an explanation. Instead,
they are phenomena in need of explanation.
How is it that placebos enhance endorphin release
(if and when that does in fact occur)?
The problem that is illustrated by the endorphin
release example is that which occurs
when any psychological variable (such as suggestion)
is hypothesized to produce a physiological
effect. Following the conventions proposed
by Michael Hyland (1985), discovering
an apparent cause of a physical effect ought to
trigger a search for the physiological substrates
of the psychological event. Hyland proposed a
principle of mind/brain complementarity,
based on Bohr's principle of wave/particle
complementarity in physics. Briefly stated, it
suggests that mind states and their corresponding
brain states are complementary descriptions
of the same underlying event, a position
that Hyland and I have shown to be a logically
necessary consequence of virtually all monist
mind-body philosophies (Kirsch & Hyland,
1987). If mind states and body states are descriptions
of the same underlying event, it is
technically improper to state that one causes
the other. Instead, psychophysiological explanation
requires establishing three sorts of relations
between variables. These are: (1) causal
relations between mind states and other mind
states, (2) causal relations between physiological
states and other physiological states, and (3)
identity relations between mind states and
their corresponding physiological states.
It is here that the data on the specificity of
placebo effects becomes particularly useful. It
suggests that we need to look for physiological
substrates of very specific psychological states.
Even the psychological construct of expectancy
is likely to be too broad. Instead, we
need to establish the physiological correlatives
of specific expectancies, such as expectations
of alterations in arousal, pain sensitivity, nausea,
and so on.
106 I. Kirsch
anales de psicología, 1999, 15(1)
Hypnosis
Hypnotic Inductions as Nondeceptive
Placebos
What is a hypnotic induction? Charcot induced
hypnosis by clanging gongs, flashing lights, applying
pressure to subjects' heads. Braid
thought that eye-fixation was necessary, but
the Spiegels have subjects roll their eyes, and
many hypnotists merely ask subjects to close
them. Most contemporary inductions include
suggestions for relaxation, but increased alertness
can be suggested instead, and relaxation
can be prevented by having subjects pedal a
stationary bicycle. The only common ingredient
to these inductions is the label hypnosis. As
Sheehan and Perry (1976, p. 72) noted, "it is
not the procedural conditions per se that are
important but whether or not the subject perceives
them as part of a context that is 'appropriate'
for displaying hypnotic behavior."
When the effect of administering a drug is
found to be independent of the its specific ingredients,
the drug is deemed to be a placebo.
Similarly, hypnotic inductions must be expectancy
manipulations, akin to placebos, because
their effects on suggestibility are independent
of any specific component or ingredient. In
fact, it is possible to produce all of the suggestive
effects of hypnosis by giving subjects placebos
and telling them that the medication
produces a hypnotic state (Baker & Kirsch,
1993; Glass & Barber, 1961). If hypnosis is an
altered state or condition, then it is a state that
is produced by placebos. This establishes a
clear association between hypnosis and placebo
effects.
Recall that placebo effects are highly specific,
in that the nature of the effect depends
on the information that is presented to subjects.
A similar specificity is found in hypnotic
responses. Subjects' responses during hypnosis
depend on their expectancies of how a hypnotic
subject should respond. These expectancies
have been shown to affect the qualitative
experience of trance (Henry, 1985), spontaneous
arm catalepsy (Orne, 1959), spontaneous
amnesia (Young & Cooper, 1972), the ability
to resist suggestions (Lynn, Nash, Rhue,
Frauman, & Sweeney, 1984; Spanos,Cobb, &
Gorasssini, 1985), the ability to breach suggested
amnesia (Silva & Kirsch, 1987), and the
nature of "hidden observer" reports (Spanos &
Hewitt, 1980). In sum, hypnotic inductions are
as nonspecific as placebos, but hypnotic and
placebo-induced experiences and behaviors are
as specific as the expectancies that mediate
their occurrence.
The seminal study of McGlashan, Evans,
and Orne (1969) has been interpreted as indicating
that hypnosis and placebo effects are
not related. That study purported to demonstrate
that hypnotic analgesia was more effective
than a placebo among highly responsive
subjects, thus indicating that there was more to
hypnosis than expectancy effects. As important
as that study was, however, there was a fatal
flaw in its design. One of the essential
characteristics of placebo controlled investigations
is that the placebo match the treatment
for which it is serving as a control. The importance
of this requirement derives from the fact
that different placebos have different effects.
In a review of double-blind drug studies, for
example, Evans (1974) concluded that placebo
morphine was considerably more effective
than placebo Darvon, which in turn was more
effective than placebo aspirin. In the
McGlashan et al. study, the placebo was presented
as an analgesic and administered in
Darvon capsules. It stands to reason that subjects
who have experienced hypnoticallyinduced
hallucinations and amnesia during the
selection procedure would expect greater pain
relief from hypnosis than from Darvon.
In a more recent study, a colleague and I
compared hypnotic analgesia to to the effects
of two different placebos. One of these was
described to subjects as a "pain-reducing analgesic."
The other was described as a "hypnotic
drug" that "increases suggestibility." We replicated
the superiority of hypnosis to placebo
when the placebo was presented as a painrelieving
drug. However, placebo and hypnosis
were equally effective when the placebo was
Hypnosis and Placebos: Response Expectancy as a Mediator of Suggestion Effects 107
anales de psicología, 1999, 15(1)
presented as a drug that induces hypnosis
(Baker & Kirsch, 1993). Furthermore, in both
the placebo condition and the hypnosis condition,
expectancy was significantly correlated
with pain reduction, whereas the association
between hypnotizability and pain reduction
was only marginal.
Individual Differences in Responsiveness
Expectancy determines the circumstances
under which a good hypnotic subject experiences
and displays hypnotic phenomena. It
also determines the kind of phenomena that
good subjects experience and display. But
what determines the degree to which a subject
responds? Does expectancy produce hypnotic
experiences only in susceptible subjects, or is it
also one of the determinants of hypnotic susceptibility?
Expectancy is one of the few stable correlates
of hypnotizability (Kirsch & Council,
1992). Although early studies indicated that
these correlations were only moderate, much
higher correlations, some as high as .71 and
.84, have been reported in more recent studies
(Council et al., 1983; Council et al., 1986;
Johnson et al., 1989; Kirsch, 1991). Still, correlation
does not establish causality. It is possible
that expectancy is an epiphenomenon rather
than a cause of responsiveness. More convincing
evidence of causality is provided by studies
in which manipulated expectancies produced
changes in responsiveness. Kirsch, Council,
and Mobayed (1987) demonstrated that altered
expectancies can account for more variance
than trait hypnotizability (i.e., pre-manipulation
responsiveness) in subsequent hypnotic suggestibility.
In another study (Wickless &
Kirsch, 1989), the effect of an expectancy manipulation
was so strong that 73% of the subjects
scored in the high range of responsiveness
(9-12) on form C of the Stanford Scale
and the remaining 27% scored in the moderate
range (5-8). Not one subject scored in the low
range (0-4).
These data provide strong evidence for a
causal relation between expectancy and hypnotizability,
but they still leave some variance in
responsiveness unexplained. It is possible that
expectancy is the sole proximal determinant of
hypnotizability and that the residual variance is
a result of measurement error. Conversely, the
unexplained variance may be due to a talent or
personality characteristic, the nature of which
is yet to be established.
Placebos, Hypnosis and "Sensory Suggestibility"
There is an important difference between placebos
and hypnosis that is worth noting for
practical reasons. The administration of placebos
entails deception. Hypnosis does not. Physicians
and psychotherapists are justifiably reluctant
to use placebos for this reason. Because
the use of hypnosis does not require deception,
it can be used as a Nondeceptive
means of exploiting the therapeutic potential
of suggestion. For example, practitioners who
use hypnosis to enhance cognitive-behavior
psychotherapy inform their clients that hypnosis
does not involve going into a trance, but is
instead a method of helping them become
deeply involved in creating new experiences
(Kirsch, 1993).
The deceptive nature of placebos makes
that phenomena similar to the tests from
which the concept of "secondary suggestibility"
(Eysenck & Furneaux, 1945) was derived.
In fact, the labels deceptive and nondeceptive suggestion
might be preferable to direct and indirect,
so as to avoid confusion with the very different
concepts of direct and indirect suggestions
used by the Ericksonians. Gheorghiu's (1989)
"indirect-direct" approach to measuring "sensory
suggestibility" occupies an intermediate
ground between the nondeceptive suggestions
used in hypnosis and the deceptive procedures
used in earlier attempts to measure sensory
suggestibility.
There is, however, a dimension on which
placebo and hypnotic suggestions differ from
those used to measure sensory suggestibility.
108 I. Kirsch
anales de psicología, 1999, 15(1)
Both placebo and hypnotic suggestions elicit
response expectancies. They are suggestions
that changes will occur within the individual.
In contrast, tests of secondary or sensory suggestibility
are more closely related to stimulus
expectancies. They are suggestions that the
stimulus will change. The placebo literature
demonstrates clearly that response expectancies
are self-confirming. Stimulus expectancies
may be less so, perhaps because internal states
are more ambiguous than external stimuli. In
any case, suggesting that a person has become
less sensitive to pain is not the same thing as
suggesting that the pain stimulus has been reduced
in intensity. The degree to which they
elicit comparable results and the correlations
between them are worthy topics for future investigation.
Conclusions
The data presented in this paper indicate that
hypnotic responses and placebo effects share a
common mechanism, that of response expectancy.
They share a common mechanism because
they are subsets of a broader phenomenon:
the phenomenon of suggestion. Like placebos,
hypnosis produces therapeutic effects
by changing client's expectancies. But unlike
placebos, hypnosis does not require deception
in order to be effective. Whereas placebos are
presented deceptively as pharmacological
treatments, hypnosis is presented honestly as a
psychological procedure. Furthermore, honestly
informing clients about what has been
learned through research about the nature of
hypnosis may reduce resistance and increase
responsiveness to hypnotic interventions.
Expectancy is not the only variable mediating
the effects of suggestion on behavior, but
it is certainly a critical variable. Among the
questions that remain to be answered are the
following: What other variables can be shown
to influence response to suggestion, and what
are the relations between those variables and
expectancy? Specifically, is expectancy the final
link in the causal chain between suggestion
and response, so that other mediating variables
affect response by means of their impact on
expectancy, or are there other mediating variables
that have direct (immediate) effects on
response?
Hypnotic procedures have evolved as if
their creators were aware of the importance of
shaping subjects' response. Skilled practitioners
are permissive. They present and respect
choices, often in the form of therapeutic double-
binds, so that either choice promotes improvement.
They prevent failure by beginning
with easy tasks that the client is almost certain
to accomplish, and they define tasks so that
failure is impossible. They evaluate performances
at any level as indications of success,
and they structure expectations so that even
small improvements are seen as significant beginnings.
They are alert to random fluxuations
and capitilize on those that occur in a desired
direction. They also prepare clients for setbacks
by labeling them in advance as inevitable,
temporary, and useful learning opportunities.
These practices, which have evolved from
clinical and experimental hypnosis, can be
used to maximize expectancy effects in nonhypnotic
psychotherapy as well.
There is one mistake, however, that most
clinical hypnotists continue to make, and that
is the continued use of the altered state concept
and terminology. The idea of going into a
trance scares many clients and inhibits them
from experiencing hypnotic effects. In contrast,
debunking the altered state myth and
presenting hypnosis from a cognitivebehavioral
perspective enhances subsequent
responsiveness.
Not only does trance terminology scare
clients away from the potential benefits of
hypnosis, it also scares many professionals
needlessly. There are hospital administrators
who prohibit the use of hypnosis in their institutions
and therapists who are reluctant to put
their clients into an altered state, fearing that
the clients might get stuck in that state. Despite
the involvement of such prominent figures
as Binet, Freud, Hull, and Hilgard, hypnosis
remains stigmatized as a mysterious,
quasi-mystical procedure. The trance concept
Hypnosis and Placebos: Response Expectancy as a Mediator of Suggestion Effects 109
anales de psicología, 1999, 15(1)
surely deserves at least some of the blame for this unfortunate state of affairs.
References
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and
predicting social behavior. Englewood Cliffs, NJ: Prentice-
Hall.
Amigó, S. (in press). Self-Regulation Therapy: Suggestion
Without Hypnosis. In I. Kirsch, A. Capafons, E.
Cardeña, & S. Amigó (Eds.), Clinical hypnosis and selfregulation:
Cognitive-behavioral perspectives. Washington,
DC: American Psychological Association.
Atkinson, J. W. (1957). Motivational determinants of risktaking
behavior. Psychological Review, 46, 359-372.
Baker, S. L., & Kirsch, I. (1993). Hypnotic and Placebo Analgesia:
Order Effects and the Placebo Label. Contemporary
Hypnosis, 10, 117-126.
Bandura, A. (1977). Social learning theory. Englewood Cliffs,
NJ: Prentice Hall.
Barber, T. X. (1969). Hypnosis: A scientific approach. New
York: Van Nostrand Reinhold.
Bargh, J. A., & Gollwitzer, P. M. (1994). Environmental
control of goal-directed action: Automatic and strategic
contingencies between situations and behavior. Nebraska
symposium on motivation, 41, 71-124.
Boissel, J. P., Philippon, A. M., Gauthier, E., Schbath, J.,
Destors, J. M., & the B.I.S. Research Group. (1986).
Time course of long-term placebo therapy effects in
angina pectoris. European Heart journal, 7, 1030-1036.
Capafons, A. (in press). Applications of Emotional
Self-regulation Therapy. In I. Kirsch, A. Capafons, E.
Cardeña, & S. Amigó (Eds.), Clinical hypnosis and selfregulation:
Cognitive-behavioral perspectives. Washington,
DC: American Psychological Association.
Coryell, W., & Noyes, R. (1988). Placebo response in panic
disorder. American Journal of Psychiatry, 145, 1138-1140.
Council, J. R. (in press). Measures of hypnotic responding.
In I. Kirsch, A. Capafons, A. Cardeña, S., & Amigó
(Eds.), Clinical hypnosis and self-regulation: Cognitivebehavioral
perspectives. Washington, DC: American Psychological
Association.
Council, J. R., Kirsch, I., & Hafner, L. P. (1986). Expectancy
versus absorption in the prediction of hypnotic
responding. Journal of Personality and Social Psychology, 50,
182-189.
Council, J., Kirsch, I., Vickery, A. R., and Carlson, D.
(1983). "Trance" vs. "skill" hypnotic inductions: The
effects of credibility, expectancy, and experimenter
modeling. Journal of Consulting and Clinical Psychology, 51,
432-440.
Edwards, W. (1954). The theory of decision making. Psychological
Bulletin, 51, 380-417.
Evans, F. J. (1989). The independence of suggetibility, placebo
response, and hypnotizability. In V. A. Gheorghiu,
P. Netter, H. J. Eysenck, & R. Rosenthal (Eds.),
Suggestion and suggestibility: Theory and Research (pp. 145-
154). Berlin: Springer-Verlag.
Evans, F.J. (1974). The placebo response in pain reduction.
Advances in Neurology, 4, 289-296.
Eysenck, H. J., & Furneaux, W. D. (1945). Primary and secondary
suggestibility: An experimental and statistical
study. Journal of Experimental Psychology, 35, 485-503.
Fillmore, M., & Vogel-Sprott, M. (1992). Expected effect of
caffeine on motor performance predicts the type of response
to placebo. Psychopharmacology, 106, 209-214.
Frankenhaeuser, M., Post, B., Hagdahl, R., & Wrangsjö, B.
(1964). Effects of a depressant drug as modified experimentally-
induced expectation. Perceptual and motor
skills, 18, 513-522.
Gheorghiu, V. A. (1989). The development of research on
suggestibility: Critical considerations. In V. A. Gheorghiu,
P. Netter, H. J. Eysenck, & R. Rosenthal (Eds.),
Suggestion and suggestibility: Theory and Research (pp. 3-55).
Berlin: Springer-Verlag.
Glass, L. B., and Barber, T. X. (1961). A note on hypnotic
behavior, the definition of the situation, and the placebo
effect. Journal of Nervous and Mental Diseases, 132,
539-541.
Heckhausen, H. (1977). Achievement motivation and its
constructs: A cognitive model. Motivation and Emotion,
1, 283-329.
Henry, D. (1985). Subjects' expectancies and subjective experience
of hypnosis. Unpublished doctoral dissertation,
University of Connecticut.
Hilgard, E. R. (1965). Hypnotic susceptibility. New York: Harcourt,
Brace & World.
Hilgard, E. R. (1973). The domain of hypnosis: With some
comments on alternate paradigms. American Psychologist,
28, 972-982.
Hull, C. L. (1933). Hypnosis and suggestibility: An experimental
approach. New York: Appleton-Century Crofts.
Hull, J. G., & Bond, C. F. (1986). Social and behavioral consequences
of alcohol consumption and expectancy: A
meta-analysis. Psychological Bulletin, 99, 347-360.
Hyland, M. E. (1985). Do person variables exist in different
ways? American Psychologist, 40, 1003-1010.
Ikemi, Y., & Nakagawa, S. (1962). A psychosomatic study of
contagious dermatitis. Kyoshu Journal of Medical Science,
13, 335-350.
Johnston, J. C., Chajkowaski, J., DuBreuil, S. C., & Spanos,
N. P. (1989). The effects of manipulated expectancies
on behavioural and subjective indices of hypnotisability.
Australian Journal of Clinical and Experimental Hypnosis,
17, 121-130.
Kirsch, I. (1985). Response expectancy as a determinant of
experience and behavior. American Psychologist, 40, 1189-
1202.
Kirsch, I. (1990). Changing expectations: A key to effective psychotherapy.
Pacific Grove, CA: Brooks/ Cole.
Kirsch, I. (1991). The social learning theory of hypnosis. In
S. J. Lynn & J. Rhue (Eds.), Theories of hypnosis: Current
models and perspectives (pp. 439-466). New York: Guilford
Press.
Kirsch, I. (1993). Cognitive-Behavioral Hypnotherapy. In J.
W. Rhue, S. J. Lynn, & I. Kirsch (Eds.), Handbook of
110 I. Kirsch
anales de psicología, 1999, 15(1)
clinical hypnosis (pp. 151-171). Washington, DC: American
Psychological Association.
Kirsch, I. (1994). APA definition and description of hypnosis:
Defining hypnosis for the public. Contemporary Hypnosis,
11, 142-143.
Kirsch, I. (1997). Suggestibility or Hypnosis: What Do Our
Scales Really Measure? International Journal of Clinical and
Experimental Hypnosis, 45, 212-225.
Kirsch, I., & Council, J. R. (1992). Situational and personality
correlates of suggestibility. In E. Fromm and M.
Nash (Eds.), Contemporary hypnosis research (pp. 267-292).
New York: Guilford.
Kirsch, I., & Hyland, M. E. (1987). How thoughts affect the
body: A metatheoretical framework. Journal of Mind and
Behavior, 8, 417-434.
Kirsch, I., & Lynn, S. J. (1995). The altered state of hypnosis:
Changes in the theoretical landscape. American Psychologist,
50, 846-858.
Kirsch, I., & Lynn, S. J. (1997). Hypnotic involuntariness
and the automaticity of everyday life. American Journal of
Clinical Hypnosis, 40, 329-348.
Kirsch, I., & Lynn, S.J. (1998). Social-cognitive alternatives
to dissociation theories of hypnosis. Review of General
Psychology, 2, 66-80.
Kirsch, I., & Weixel, L. J. (1988). Double-Blind Versus Deceptive
Administration of a Placebo. Behavioral Neuroscience,
102, 319-323.
Kirsch, I., Council, J. R., & Mobayed, C. (1987). Imagery
and response expectancy as determinants of hypnotic
behavior. British Journal of Experimental and Clinical Hypnosis,
4, 25-31.
Kirsch, I., Mobayed, C. P., Council, J. R., & Kenny, D. A.
(1992). Expert judgments of hypnosis from subjective
state reports. Journal of Abnormal Psychology, 101, 657-
662.
Lyerly, S. B., Ross, S., Krugman, A. D., & Clyde, D. J.
(1964). Drugs and placebos: The effects of instructions
upon performance and mood under amphetamine sulphate
and chloral hydrate. Journal of Abnormal and Social
Psychology, 68, 321-327.
Lynn, S. J., Nash, M. R., Rhue, J. W., Frauman, D. C., &
Sweeney, C. A. (1984). Nonvolition, Expectancies, and
Hypnotic Rapport. Journal of Abnormal Psychology, 93,
295-303.
McGlashan, T. H., Evans, F. J., & Orne, M. T. (1969). The
nature of hypnotic analgesia and placebo response to
experimental pain. Psychosomatic Medicine, 31, 227-246.
Meath, J. A., Feldberg, T. M., Rosenthal, D., & Frank, J. D.
(1956). Comparison of reserpine and placebo in treatment
of psychiatric outpatients. A. M. A. Archives of
Neurology and Psychiatry, 76, 207-214.
Mischel, W. (1973). Toward a cognitive social reconceptualization
of personality. Psychological Review, 80, 252-283.
Montgomery, G., & Kirsch, I. (1996). Mechanisms of Placebo
Pain Reduction: An Empirical Investigation. Psychological
Science, 7, 174-176.
Orne, M. T. (1959). The nature of hypnosis: Artifact and
essence. Journal of Abnormal Psychology, 58, 277-299.
Pihl, R. O., & Altman, J. (1971). An experimental analysis of
the placebo effect. Journal of Clinical Pharmacology, 11,
91-95.
Rescorla, R. A. (1988). Pavlovian conditioning: It's not what
you think it is. American Psychologist, 43, 151-160.
Rickels, K., Lipman, R., & Raab, E. (1966). Previous medication,
duration of illness, and placebo response. The
Journal of Nervous and Mental Disease, 142, 548-554.
Ross, S., Krugman, A. D., Lyerly, S. B., & Clyde, D. J.
(1962). Drugs and placebos: A model design. Psychological
Reports, 10, 383-392.
Rotter, J. B. (1954). Social learning and clinical psychology.
Englewood Cliffs, NJ: Prentice-Hall.
Schoenberger, N. E. (in press). Expectancy and fear. In I.
Kirsch (ed.), Expectancy, experience, and behavor. Washington,
DC: American Psychological Association.
Segal, M., & Shapiro, K. L. (1959). A clinical comparison
study of the effects of resperine and placebo on anxiety.
Archives of Neurology and Psychiatry, 81, 392-398.
Sheehan, P. W., & Perry, C. W. (1976). Methodologies of hypnosis.
Hillsdale, NJ: Lawrence Erlbaum Associates.
Silva, C. E., & Kirsch, I. (1987). Breaching amnesia by manipulating
expectancy. Journal of Abnormal Psychology, 96,
325-329.
Spanos, N. P., & Hewitt, E. C. (1980). The hidden observer
in hypnotic analgesia: Discovery or experimental creation?
Journal of Personality and Social Psychology, 39,
1201-1214.
Spanos, N.P., Cobb, P.C., & Gorassini, D. (1985). Failing to
resist hypnotic test suggestions: A strategy for selfpresenting
as deeply hypnotized. Psychiatry, 48, 282 -
292.
Tolman, E. C. (1932). Purposive behavior in animals and
men. New York: Appleton-Century-Crofts.
Traut, E. F., & Passarelli, E. W. (1957) Placebos in the
treatment of rheumatoid arthritis and other rheumatic
conditions. Annals of the Rheumatic Diseases, 16, 18-22.
Turkkan, J. S. (1989). Classical conditioning: The new hegemony.
Behavioral and Brain Sciences, 12, 121-179.
Whitcher, S. J., & Fisher, J. D. (1979). Multidimensional reaction
to interpersonal touch. Journal of Personality and
Social Psychology, 37, 87-96.
Wickless, C., & Kirsch, I. (1989). The effects of verbal and
experiential expectancy manipulations on hypnotic
susceptibility. Journal of Personality and Social Psychology,
57, 762-768.
Wickramasekera, I. (1980). A conditioned response model
of the placebo effect: Predictions from the model. Biofeedback
and Self-Regulation, 5, 5-18.
Wolf, S. (1950). Effects of suggestion and conditioning on
the action of chemical agents in human subjects--the
pharmacology of placebos. Journal of Clinical Investigation,
29, 100-109.
Young, J., & Cooper, L. M. (1972). Hypnotic recall amnesia
as a function of manipulated expectancy. Proceedings of
the 80th Annual Convention of the American Psychological Association,
7, 857-858.
Zukin, P., Arnold, D. G., & Kessler, C. (1959). Comparative
effects of phenaglycodol and meprobromate on anxiety
reactions. Journal of Nervous and Mental Disease, 129,
193-195.
(Artículo recibido: 6-5-98; aceptado 1-3-99)
(Article received: 6-5-98; acepted 1-3-99)