Journal of Personality and Social Psychology Copyright 1985 by the American Psychological Association, Inc.

1985, Vol. 48, No. 1, 47-53 0022-3514/85/S00.75

Conditioned Placebo Responses

Nicholas J. Voudouris, Connie L. Peck, and Grahame Coleman
La Trobe University
Bundoora, Victoria, Australia

Following earlier animal research, we attempt to condition placebo effects in

human subjects. Four groups of 8 voluntary subjects were told that the experi-
menters would test a powerful new analgesic cream over three sessions by assessing
its ability to reduce experimentally induced pain. The analgesic cream was, in
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

fact, a placebo. In the first session all subjects were tested with and without the
This document is copyrighted by the American Psychological Association or one of its allied publishers.

cream to assess their placebo response. In the second session, to condition two
groups (with differing stimulation levels) to experience pain relief in response to
the placebo, we repeatedly paired a reduction in nocioceptive stimulation with
placebo administration. (Subjects were unaware that stimulation levels were
manipulated). To condition the other two groups (with different stimulation levels)
to experience an exacerbation of the pain, we paired an increase in nocioceptive
stimulation with placebo administration. In the third session, all subjects were
again tested for placebo response. Results suggested that placebo responses are
conditionable in the laboratory in both a positive and negative direction. The
clinical implications of a learning theory of placebo behavior are discussed.

Although the placebo effect has been used in the therapeutic or experimental milieu.
for centuries in the practice of medicine and Surprisingly, little work has been done to test
is a significant component of all therapies, it this theory. Herrnstein (1962) attempted to
remains a poorly understood phenomenon. demonstrate placebo effects in the rat. Pavlov
The placebo literature is marked by inconsis- (1927) had noted that the effects of morphine
tencies and lack of replication, and the term on dogs were present in experimentally ex-
placebo has often been ill-defined. Although perienced animals before injection. Herrnstein
it has been established that variables such as (1962) administered scopolamine hydrobro-
the therapeutic milieu, subject-therapist in- mide (a drug that disrupts the learned behav-
teraction, and treatment methodology influ- ior of rats in a predictable manner) to labo-
ence placebo response, their relative status ratory rats. Once the rats were conditioned
and mechanism of operation remain obscure. to a number of such injections, they showed
The search for a consistent "placebo re- a similar reaction when given injections of
sponder" has involved examination of many saline. Herrnstein reasoned that the rats'
traits and circumstances, yet no such individ- behavior toward the saline (placebo) injection
ual has been found (Parkhouse, 1963). This was an example of "simple Pavlovian condi-
situation is due to the complex, multideter- tioning" (p. 677). The conditioned stimulus
mined nature of placebo behavior and to the (saline injection) eventually elicited those re-
absence of an adequate theoretical framework sponses normally associated with the uncon-
within which to test empirical propositions. ditioned stimulus (scopolamine injection) after
Ullman and Krasner (1969) have suggested the scopolamine response was paired with
one such theory: that the bases for the occur- the injection stimulus. Herrnstein concluded
rence of placebo manifestations are condi- that "There appears to be no reason to
tioning effects. A conditioning theory of the suppose that the placebo effect in human
placebo would explain placebo effects as being patients differs in any way from that dem-
conditioned responses to the stimuli present onstrated here, other than in degree of com-
plexity" (p. 678).
Requests for reprints should be sent to Connie L. More recent animal studies have in fact
Peck, Department of Psychology, La Trobe University, emphasized just that, the complexity of such
Bundoora, Victoria, Australia, 3083. phenomena. Pihl and Altaian (1971) paired

47
 48 N. VOUDOURIS, C. PECK, AND G. COLEMAN

amphetamine with saline injections in rats. is paired with a therapeutic effect (reduced
As with the amphetamine, saline injections pain) or an antitherapeutic effect (increased
elicited increased activity. However, in their pain). In addition, we studied two levels of
second experiment, saline injections, after pain because this may be a situational factor
being paired with the tranquilizer chloro- that interacts with the strength of the placebo
promazine, failed to depress activity in rats. response (Bonica, 1979).
The authors concluded that the placebo effect
in rats is in some way, as yet unclear, depen-
Method
dent on the specific drug in question and
aspects of the stimuli present in the experi-
Subjects
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

mental chamber. Although the placebo effect

This document is copyrighted by the American Psychological Association or one of its allied publishers.

itself has not been studied extensively within Thirty-two subjects, 12 male and 20 female university
a conditioning paradigm, animal studies in students, voluntarily participated in the study and were
paid $6 for participating. The subjects' ages ranged from
which the effect is used to examine other 19 to 34 years. Before participating, all subjects completed
phenomena are common (Sherman, Proctor, a consent form explaining the nature of the noxious
& Strub, 1982; Wilder, Pescor, Miller, & stimulation and warning people with heart conditions,
Norrell, 1971). Moreover, there have been no with hypertension, or currently on medication or other
drugs from taking part.
conditioning studies of the placebo effect in
humans.
In the current study we explore the possi- Apparatus
bility that placebo effects can be conditioned Noxious stimulations of one second each were admin-
in human subjects. Support for a learning istered by means of an iontophoretic pain generator.
theory explanation of the placebo could pro- Iontophoresis, by the repulsion of positive potassium ions
from the positive pole of an electric current, drives these
vide a valuable theoretical framework and ions into the skin, causing a prickling sensation at lower
facilitate a more complete understanding of levels of stimulation and a cramping sensation at higher
the genesis of placebo behavior. Because even levels. The degree of noxious stimulation is dependent
active treatments may contain placebo com- on the amount of current and the duration of adminis-
tration, and is independent of skin resistance (Benjamin
ponents and results from these "are apt to
& Helvey, 1963).
be due to a combination of both placebo and The apparatus consisted of a plastic clamp, which is
nonplacebo effects" (Shapiro & Morris, 1971), attached to the flexor surface of the forearm and which
a learning theory of placebo would imply incorporated a small bowl on the upper surface. The
bowl uses the upper surface of the arm as its base and is
that a history of therapy in which the patient
filled with a 3% solution of potassium chloride (the
had experienced relief from aversive symp- contact medium between the electrode and skin). Enclosed
toms at the hands of the medical profession within the bowl is a metal anode plate. A gauze pad
would make him or her more likely to respond saturated with a 9% solution of sodium chloride wrapped
positively to future treatment whether it is around a silver-silver chloride cathode plate is placed on
the surface of the arm. This gauze base acts as an
active or placebo. Similarly, a history of
insulator. The source of the DC current is a series of
disappointed expectations of relief in which batteries enclosed within the console, and one can increase
the doctor's treatment prescription failed or decrease the current by manipulating a simple poten-
would lead to a lowered probability that tiometer dial on the console face. The stimulation increases
or decreases accordingly. An electronic display gives a
future therapies would be efficacious because
constant readout on the level of stimulation, measured
the placebo component of treatment would in milliamps (mA). The duration of stimulation is preset
contain negative expectations of relief that with a built-in timer (set at one second for this study).
could mitigate against active components. The placebo analgesic was in the form of a cream.
Although previous studies have concentrated Simple cold cream was mixed with a linalol (CioHujO),
which contained rub-down cream in the ratio 8:1, so
almost exclusively on the positive response,
that the cream would have a distinct smell. The cream
both positive placebo effects (results congruent was removed by means of cleansing the skin with a 70%
with expectations of relief) and negative pla- alcohol solution. In order to control for possible changes
cebo effects (results that violate expectations in skin conductance that would be due to the application
of the cream at areas surrounding the iontophoretic cup,
of relief) need to be studied in order to gain
a constant current source was used in the construction
a complete understanding of the phenomenon. of the iontophoretic pain generator. This ensured that
In the present study we attempt to assess the differences in skin conductance across both trials and
effect of conditioning trials in which a placebo subjects would not influence pain levels.
 CONDITIONED PLACEBO RESPONSES 49

Table 1
Experimental Design

Stimulation intensity

Session 1: Session 2: Session 3:

Treatment condition Preconditioning test Conditioning Postconditioning test

Group 1"
With-placebo rating 5 2 5
No-placebo rating 5 5 5
Group 2b
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

With-placebo rating 8 5 8
This document is copyrighted by the American Psychological Association or one of its allied publishers.

No-placebo rating 8 8 8
Group y
With-placebo rating 2 5 2
No-placebo rating 2 2 2
Group 4"
With-placebo rating 5 8 5
No-placebo rating 5 5 5

Note. Because of individual differences in the pain ratings of subjects associated with a particular amount of stimulation,
the amount of stimulation administered in each case is given by the subject's rating figure. Thus, for example, a rating
of 5 indicates that the intensity of the stimulation was that average mA value corresponding to the subject's previous
ratings of 5, during ascending/descending and random trials.
• Positive placebo, stimulation low, n = 8. " Positive placebo, stimulation high, n = 8. c Negative placebo, stimulation
low, n = 8 . d Negative placebo, stimulation high, n = 8.

Experimental Design at the same time of day. For four of the subjects, there
was a 48-hr gap between the second and third sessions.
Before predetermined levels of pain intensity could be
Session 1: Preconditioning test. Session 1 was ap-
reliably induced in subjects, it was necessary to determine
proximately 50 min long. The subject was seated at a
the relation between stimulus intensity (the independent
small table in a bare room with the iontophoresis equip-
variable, measured in mA) and pain rating (the dependent
ment attached to his or her arm as previously described.
variable measured on an 11-point scale) and also to
The experimenter administered the stimuli and recorded
identify any session effects, should they exist. A pilot
the subject's responses through a one-way mirror in a
study was undertaken for this purpose with six subjects.
small observation chamber adjacent to the experimental
Results of this study showed that iontophoretic pain was
room. The subject was given a set of written instructions
linear in nature and no session effects were established
to be read in conjunction with a tape recording of the
over a three-day period (Voudouris, 1981).
same. He or she was instructed to respond to each
In Table I we show the experimental design for the
stimulus verbally, rating the intensity of the pain on an
main study. Eight subjects were randomly assigned to
11-point scale ranging from 0 (no pain) to 10 (intolerable
each of four groups. Sessions I and 3 involved pretests
pain).
and posttests to determine the effects of conditioning
Because of the large individual differences in pain
carried out during Session 2. In these sessions, the levels
responsivity to the same stimuli, we decided to use
of stimulation remained constant across trials in which
equivalent pain-rating levels for each subject rather than
the placebo cream was used and those in which it was
equivalent intensity of stimulation. Thus after the in-
not used. In Session 2, the intensity of the stimulation
structions a series of ascending and descending trials
during the conditioning sessions was manipulated. In
were administered and an approximate range of stimu-
Groups 1 and 2, it was always lowered by three rating
lation values corresponding to the ten pain levels was
points when the placebo was administered and returned
established. This was followed by a brief interval during
to the Session 1 level when the placebo was removed. In
which the experimenter readministered 15 randomly
Groups 3 and 4, the intensity of stimulation was increased
chosen values from the range previously denned.
by three rating points when the placebo was administered,
The subject then rested for five minutes while the
but was held constant in trials in which the placebo was
experimenter calculated average values (mA levels) for
absent. Levels of initial stimulation (Session 1) differen-
each of the ratings on the 10-point pain scale. Subjects
tiated among the four groups. Differences between Sessions
in Groups 1 and 3 subsequently received average ionto-
1 and 3 were analyzed for the effects of the conditioning
phoretic stimulation values corresponding to their prior
session (Session 2).
pain ratings of "Level 5" and "Level 2," whereas subjects
in Groups 2 and 4 received average iontophoretic stim-
ulation values corresponding to their own pain ratings of
Procedure 8 and 5. The intensity of stimulation received by Groups
All subjects attended three sessions. For all but four 1 and 3 (low-intensity groups) was less than that received
subjects, the sessions were held on consecutive days and by Groups 2 and 4 (high-intensity groups) because using
 50 N. VOUDOUR1S, C. PECK, AND G. COLEMAN

comparable intensities would, during the conditioning Group I t—1 •

session, have resulted in unacceptably high pain levels Group II •—• *+
for the negative-effect groups, especially Group 4.
Group III 4—4 -
After calculation of appropriate stimulation intensities,
the pretest was administered. This consisted of four Group IV o—-o —
blocks of five trials each, two blocks with and two without = 28
E
the application of the placebo cream. These blocks were .E 26
alternated, the first always being a no-placebo block. D

Subjects were told | 24

This test is designed in part to test a new, fast-acting | 2-2

Q.
analgesic preparation for its ability to rapidly reduce i 20
skin sensitivity and to act as a local pain killer. Recent
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

data have suggested that it may prove to be a harmless, | 18

o
This document is copyrighted by the American Psychological Association or one of its allied publishers.

yet extremely effective and convenient local anaesthetic.

They were also told
I '' 6
0 14
a.
In order to assess the effectiveness of the analgesic
preparation as a pain-killer, the level of stimulation 1 ''*
during this session will remain constant. I l-0
.a

As previously, subjects were asked to rate pain verbally s °8

and the experimenter recorded each response. c O6
Beginning with a block of five no-placebo trials, the S
S 04
stimuli were separated by 20-s gaps and a break of
approximately 2 min for the removal of the placebo
cream occurred between blocks, as well as approximately I °'2
2 min for the cream to "take effect" or "wear off." The I ao 1
pre- conditioning post-conditioning
experimenter applied the placebo cream preparation
liberally to the forearm, concentrating on those areas Figure 1. Mean differences between placebo and no-pla-
surrounding the apparatus where the subject reported cebo ratings.
experiencing the pain. The apparatus remained in place
during application of the cream but was removed while
the skin was cleansed. Cream applied at the beginning
of a placebo block remained active during those five Session 3: Postconditioning test. The posttest was
trials until its removal in preparation for the next block. identical to the pretest administered in Session I. Subjects
The placebo cream was removed with cotton soaked in were again given four blocks of five trials, two blocks
an alcohol solution. During the 20 trials the stimulation without and two blocks with placebo. Subjects were once
level remained constant. Subjects in Groups 1 and 4 again told that stimulation would remain constant. In
received "Rating 5" pain; Group 2, subjects received this session, stimulation did remain constant at levels
"Rating 8" pain; and Group 3, subjects received "Rating identical to those administered during the pretest.
2" pain.
Session 2: Conditioning. Conditioning sessions were
approximately 70 min long. As in the preconditioning Results
session, the subject was asked to rate each stimulus,
For each subject, four scores were calcu-
administered by the experimenter from the other room.
Although subjects were again instructed that the stimu- lated: placebo and no-placebo mean ratings,
lation level would remain constant, it in fact varied for the preconditioning session (Session 1)
during the session between placebo and no-placebo trials. and placebo and no-placebo mean ratings,
Each subject received fourteen blocks of five trials each; for the postconditioning session (Session 3).
seven of these blocks were with the placebo cream and
seven were without. As in the preconditioning session, Conditioning session data were not included
stimuli were separated by 20-s intervals and approximately in the analyses, because any effect that was
4 min occurred between blocks. The first block was due to the conditioning would be manifested
always a no-placebo block. as a change in mean pain ratings from pre-
In Groups I and 2, subjects were administered less to postconditioning tests. Figure 1 shows the
pain in with-placebo trials; in Group 1, with-placebo
trials were given at "Level 2" and no-placebo trials at mean pain ratings. It is of interest to note
"Level 5." In Group 2, with-placebo trials were admin- that a mild placebo effect is evident for all
istered at "Level 5" and no-placebo trials at "Level 8." conditions in Session 1, because all four
In Groups 3 and 4, the stimulation was increased when means for placebo trials are lower than those
the placebo was applied; in Group 3, with-placebo trials
were administered at "Level 5" and no-placebo trials at
for no-placebo trials.
"Level 2"; in Group 4, with-placebo trials were given at We conducted a three-way ( 4 X 2 X 2 )
"Level 8" and no-placebo trials at "Level 5." analysis of variance with two repeated mea-
 CONDITIONED PLACEBO RESPONSES 51

sures. Factor A represented the four placebo- forms of therapy, an understanding of its
treatment conditions; Factor B represented etiology is of paramount importance. If the
the placebo/no-placebo repeated measure; etiology of placebo effects lies in a condition-
and Factor C represented the preconditioning/ ing (learning theory) model, every therapeutic
postconditioning repeated measure. Because encounter experienced by a patient becomes
the error terms for B, C, and the B X C part of a placebo conditioning (or reinforce-
interaction were nearly identical, we con- ment) history, thereby becoming a determi-
ducted Bartlett's test for homogeneity of vari- nant of how an individual will respond to
ance (Winer, 1962). Because the three error future placebo components of therapy. Fur-
terms resulting from that test were not sig- thermore, these data suggest that some of
nificantly different from each other, the error these conditioned placebo effects (experiences)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

terms were pooled. will be negative ones. If an individual has a

There was a significant effect due to the predominantly positive history of experience
treatment, F(3, 28) = 50.98, p < .05, which with a particular therapy, this model is pre-
reflects the manipulation of stimulation in- dictive of a predisposition for positive response
tensity and is of little interest. A clear overall to future placebo components. Moreover, if
placebo effect was observed, f\l, 84) = 12.4, the individual has a predominantly negative
p < .05, reflecting the predominance of pos- outcome history (in which expectancies have
itive placebo effects in all four of the precon- been violated), then he or she would be
ditioning treatments and two of the four predisposed to a negative placebo response.
postconditioning treatments. Of primary in- Given, then, that there appeared to be a
terest is a significant A X B X C interaction, placebo response from the subjects tested in
F(3, 84) = 3.76, p < .05, which may be used this study, must one conclude that most or
as a basis for investigation of conditioning all of these subjects have had a positive
effects. Post hoc tests in which we used the outcome history with regard to analgesics?
Scheffe technique (ScheSe, 1959) showed that What of the subject's conditioning histories
the conditioning effect for the high positive for medication in general? It seems possible
placebo group was significantly different from to conjecture that here there are two separate
the effect for the high negative placebo group, phenomena. At one conceptual level, the
F(3, 84) = 3.92, p < .05. Figure 1 indicates subjects are learning about and responding
that this conditioning effect was in the ex- directly to the immediate perceived efficacy
pected direction. The corresponding effect for of a particular medication. At a second level,
the low positive group compared with the many such experiences may have built up a
low negative group was not significant, al- much larger accumulated learning history
though it is evident from Figure 1 that the about analgesics in general, or indeed medi-
effect was similar to that for the high intensity cation in general, both of which will also
groups. No significant effects were found for determine the nature and magnitude of an
Factor B (no-placebo condition). individual's response on any one occasion.
The complexity of this behavior becomes
Discussion even more evident when one considers that
this argument can be applied not just to the
This study suggests that one can condition medication itself, but to the therapeutic agent
both positive and negative responses in the dispensing it, the milieu in which it is ad-
laboratory by pairing increased or decreased ministered, and the specific relation between
stimulation with a neutral therapeutic agent. this agent and the individual receiving the
These results, although not perfectly analo- treatment.
gous, are similar to those found in animal When evidence that there may be an inter-
research (Herrnstein, 1962; Pihl & Altman, action between the placebo and active com-
1971) and they support the notion of a ponents of therapies is considered (Dinner-
conditioning model as a basis of the placebo stein & Halm, 1970; Dinnerstein, Lowenthal,
response. Given that the placebo constitutes & Blintz, 1966), this conditioning theory has
a substantial and vital component in the profound implications for medicine and psy-
majority of medical treatments and in all chology. Chronically ill individuals or intrac-
 52 N. VOUDOURIS, C. PECK, AND G. COLEMAN

table patients who have a history of negative or injections appear to operate internally, in
outcome from a variety of treatments will some unseen way, and so may not be as
render future treatment at a disadvantage. effective as a more readily understood and
The effects of active components in treatment observable application. Although there are
may be negated by conditioned negative re- no known direct physiological analgesic prop-
sponse to the treatment's placebo component. erties of a cold cream preparation such as
If future placebo response can be predicted the one used in this study, it is conceivable,
from conditioning history, prescription of albeit highly unlikely, that the cream itself
therapy could take this into account to max- reduced sensations of pain. Given the posttest
imize the effectiveness of the selected remedial results obtained, however, it appears than any
measure. A person's treatment history might
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

such effect would be of an insignificant mag-

This document is copyrighted by the American Psychological Association or one of its allied publishers.

become a useful piece of data when treatment nitude.

modalities are being chosen. Questions such The number of actual placebo/no-placebo
as the following would become more impor- events that the subjects experienced in this
tant: What are the patient's past experiences study may have contributed to their belief in
(personal or vicarious) with drug treatment, the stated nature of the experiment, a con-
hospitalization, surgery, psychotherapy, ra- dition not always achieved in placebo exper-
diation treatment, and so on? What is the iments with wary university students. In the
patient's estimation of the potential efficacy majority of studies, the subjects are presented
of the procedure? with only one placebo; it would seem more
The notion of conditioning has implications logical to repeat comparisons of placebo and
for the use of pure placebo treatment as well. no-placebo experiences to enhance the effect.
If such treatment is to be used, it should only This study, following from the animal re-
be attempted with patients possessing a pos- search, is the first attempt known to us to
itive conditioning history in that particular condition a placebo response in human sub-
therapeutic modality. The most effective jects. Clearly it needs replication and exten-
strategy may be to combine the use of a sion. Future research should address the issue
placebo with an active component (e.g., a of whether the placebo effect can be condi-
drug), then remove the active ingredient at a tioned in clinical populations, the number of
later stage (Fordyce, 1976). Unfortunately, trials needed to establish an effect, and the
pure placebos are most often attempted with permanency of the conditioning effect and its
those therapeutically intractable patients for resistance to extinction. Finally, the applica-
whom nothing has worked. If this theory is bility and utility of this effect will depend on
correct, then placebos would hold out little the range of therapeutic modes and experi-
hope for these patients because they have no mental paradigms in which it can be dem-
conditioned positive effects, and may serve onstrated. What has been labeled the most
merely to further their history of ineffective neglected asset in the care and treatment of
treatment. patients may, with future research, become
The high rate of the placebo response an important therapeutic strategy.
attained even before conditioning trials is a
remarkable one, especially because an exper-
References
imental pain paradigm is used. Beecher's
(1959) review, which showed that approxi- Beecher, H. K. (1959). Measurement of subjective respon-
mately one-third of clinical patients react to ses: Quantitative effects of drugs. New York: Oxford
placebo administration, is seldom replicated University Press.
Benjamin, F. B., & Helvey, W. M. (1963). Iontophoresis
in the laboratory, in which an average of of potassium for experimental determination of pain
approximately 15% respond (Evans, 1974). endurance in man. Proceedings of the Society for
The use of an "analgesic cream" in this study Experimental Biology and Medicine, 113, 566-568.
(rather than an ingested pill) may have been Bonica, J. J. (1979). Letter. Pain, 7, 203-207.
Dinnerstein, A. J., & Halm, J. (1970). Modification of
responsible for the high rate of placebo re-
placebo effects by means of drugs: Effects of aspirin
sponse. Application of the placebo to the and placebos on self-rated mood. Journal of Abnormal
actual site of stimulation may also have in- Psychology, 75, 308-314.
creased the placebo response. Placebo pills Dinneretein, A. J., Lowenthal, M., & Blintz, B. (1966).
 CONDITIONED PLACEBO RESPONSES 53

The interaction of drugs with placebos in the control therapy and behavior change (pp. 369-410). New York:
of pain and anxiety. Perspectives in Biology and Med- Wiley.
icine. 10(1), 103-107. Sherman, J. E., Proctor, C., & Strub, H. (1982). Prior
Evans, F. J. (1974). The placebo response in pain reduc- hot plate exposure enhances morphine analgesia in
tion. In J. J. Bonica (Ed.), Advances in Neurology (Vol. tolerant and drug-naive rats. Pharmacology, Biochem-
4, pp. 289-296). New York: Raven Press. isiry and Behavior, 17, 229-232.
Fordyce, W. E. (1962). Behavioral methods for chronic Ullman, L. P., & Krasner, F. (1969). Cognitions and
pain and illness. St. Louis: C. V. Mosby Co. behaviour therapy. Behavior Therapy. 1, 202-204.
Herrnstein, R. J. (1962). Placebo effect in the rat. Science, Voudouris, N. J. (1981). The conditionahilily of positive
138, 677-678. and negative placebo effects in experimental pain.
Parkhouse, J. (1963). Placebo reactor. Nature, 199, 308. Unpublished honors thesis, La Trobe University, Bun-
Pavlov, 1. P. (1927). Conditioned reflexes (G. V. Anrep, doora, Victoria, Australia.
Trans.). London: Oxford Press. Wikler, A., Pescor, F. T., Miller. D., & Norrell, H. (1971).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Pihl, R. O., & Altman, M. A. (1971). An experimental Persistent potency of a secondary (conditioned) rein-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

analysis of the placebo effect. The Journal of Clinical forcer following withdrawal of morphine from physically
Pharmacology and New Drugs, 11, 91-95. dependent rats. Psycholopharmacologia, 20, 103-117.
Scheffe, H. (1959). The analysis of variance. New York: Winer, B. J. (1962). Statistical principles in experimental
Wiley. design. New York: McGraw-Hill.
Shapiro, A. K., & Morris, I. A. (1971). The placebo
effect in medical and psychological therapies. In S. L. Received December 16, 1983
Garfield & A. E. Bergin (Eds.), Handbook of psycho- Revision received May 15, 1984 •

Instructions to Authors

Authors should prepare manuscripts according to the Publication Manual of the American
Psychological Association (3rd ed.). Articles not prepared according to the guidelines of the
Manual will not be reviewed. All manuscripts must include an abstract of 100-150 words typed
on a separate sheet of paper. Typing instructions (all copy must be double-spaced) and instructions
on preparing tables, figures, references, metrics, and abstracts appear in the Manual. Also, all
manuscripts are subject to editing for sexist language.
APA policy prohibits an author from submitting the same manuscript for concurrent con-
sideration by two or more journals. Also, authors of manuscripts submitted to APA journals
are expected to have available their raw data throughout the editorial review process and for
at least 5 years after the date of publication. For further information on content, authors should
refer to the editorial in the March 1979 issue of this journal (Vol. 37, No. 3, pp. 468-469).
The reference citation for any article in any JPSP section follows APA's standard reference
style for journal articles; that is, authors, year of publication, article title, journal title, volume
number, and page numbers. The citation does not include the section title.
Blind reviews are optional, and authors who wish blind reviews must specifically request
them when submitting their manuscripts. Each copy of a manuscript to be blind reviewed
should include a separate title page with authors' names and affiliations, and these should not
appear anywhere else on the manuscript. Footnotes that identify the authors should be typed
on a separate page. Authors should make every effort to see that the manuscript itself contains
no clues to their identities.
Manuscripts should be submitted in quadruplicate (the original and three photocopies), and
all copies should be clear, readable, and on paper of good quality. Authors should keep a copy
of the manuscript to guard against loss. Mail manuscripts to the appropriate section editor.
Editors' addresses appear on the inside front cover of the journal.
Section editors reserve the right to redirect papers among themselves as appropriate unless
an author specifically requests otherwise. Rejection by one section editor is considered rejection
by all, therefore a manuscript rejected by one section editor should not be submitted to another.