1. ARTICLE IN PRESS
Journal of Psychosomatic Research xx (2008) xxx – xxx
Consistency of the placebo effect☆
Ben Whalley a , Michael E. Hyland a,⁎, Irving Kirschb
a
School of Psychology, University of Plymouth, Plymouth, UK
b
Department of Psychology, University of Hull, Hull, UK
Received 9 May 2007; received in revised form 4 October 2007; accepted 15 November 2007
Abstract
Objective: The existence of reliable personality predictors of (acquiescence and absorption) and response expectancy were
the placebo effect is controversial. For prediction to be possible, assessed as potential predictors of the placebo effect. Results:
the response to placebo must be reliable. We tested the consistency Placebo effects across trials were highly correlated (r=.60 and .77)
of the placebo effect by assessing the response to four trials of when placebos bore the same name but were not significantly
placebo analgesic treatment. Methods: Two identical experimental correlated when placebos had different names. Placebo effects
pain stimuli were administered simultaneously to matching fingers were significantly associated with response expectancy but not
on both hands. Pain sensation was compared between one finger, with acquiescence or absorption. Conclusions: Context-specific
which was treated with a placebo cream and the other which was predictions of placebo response (e.g., expectancy) are possible, but
not treated. Two placebo creams were used, each with a different personality predictors will not be consistent across contexts.
label. The procedure was repeated between 1 and 8 days later using © 2008 Published by Elsevier Inc.
the same creams and order of presentation. Two personality traits
Keywords: Placebo effect; Placebo responders; Pain; Test–retest reliability
Introduction one suggests that this method does not increase drug–
placebo differences [5,6].
When placebo effects are large, it can be difficult to An alternative method is to identify self-report measures
demonstrate drug effects in clinical trials [1,2], and this has of individual differences that predict placebo responding.
stimulated interest in identifying responders in order that that Although significant correlations between placebo response
drug–placebo differences can be more easily detected [3]. and the personality trait of acquiescence have been reported
One method for doing so is to use placebo run-in or [7–11], the general consensus is that reliable correlates of the
“washout” periods in which all participants in a trial placebo response have not been found [12].
(including those assigned to verum) are given a placebo, Both placebo washout periods and the search for
usually for a period of 7 to 10 days. Following this period, personality correlates of placebo responding implicitly
participants who have responded to the placebo are excluded assume that the placebo response is stable over time—
from the trial [4]. However, meta-analysis of antidepressant without stability, reliable predictions would not be possible.
trials comparing those with a run-in period to those without Early reports suggested that individual differences in
placebo responding are inconsistent [13–15], but these
studies were underpowered and did not use reliable measures
☆
of placebo response. Despite continuing efforts to identify
This study received ethical approval from the Faculty of Science
placebo responders, no further studies of the consistency of
Research Ethics Committee, University of Plymouth.
⁎ Corresponding author. School of Psychology, University of Plymouth, placebo responses have been reported.
PL48AA Plymouth, UK. In this study we addressed three important and unresolved
E-mail address: m.hyland@plymouth.ac.uk (M. E. Hyland). questions. First, we tested the hypothesis that consistency in
0022-3999/07/$ – see front matter © 2008 Published by Elsevier Inc.
doi:10.1016/j.jpsychores.2007.11.007

2. ARTICLE IN PRESS
2 B. Whalley et al. / Journal of Psychosomatic Research xx (2008) xxx–xxx
placebo responding can be found under optimal conditions. gloves. Expectancy ratings were taken after the cream was
To this end, responses to identical placebo treatments were applied but before the pain stimulus.
assessed on two occasions. If consistency cannot be found The study was conducted over two identical experimental
with all other factors held constant, it is unlikely to be found sessions, between 1 and 8 days apart (mean 2.5 days, S.D.=
at all. Second, we tested the hypothesis that an individual's 2.0). Each session consisted of two trials. On each trial,
response to placebo is reasonably consistent across participants simultaneously received two pain stimuli, each
contexts. To do this, we measured responses to two to a finger of a different hand. One finger was treated with a
placebos that differed only in name. If consistency in the placebo analgesic cream; the other finger was untreated. For
response to placebo can be disrupted by a simple change in one trial, the placebo was labeled “Trivaricaine;” for the
name, then the search for a generic placebo responder other, it was labeled “Ibuprofen”, and order of presentation
would be impossible. Third, we tested the extent to which was balanced across the sample. Different fingers were used
acquiescence, absorption, and response expectancy pre- for each trial within a session. For half the trials, placebo was
dicted placebo responding. administered to a finger on the right hand and for half to a
Acquiescence and absorption are personality traits. finger on the left hand. After the pain stimulus was applied,
Acquiescence has been reported to be associated with pain intensity was recorded. At the end of their final session,
placebo responding [7–11], and absorption has been participants completed the two personality measures and
reported to be a correlate of imaginative suggestibility were thanked and debriefed.
[16]. If there is stability in placebo responding they seem
likely candidates in the search for predictors. In contrast to Measures and pain stimulus
acquiescence and absorption, response expectancy, which
has been identified as a central factor in placebo responding Two identical Forgione-Barber Strain Gauge Pain
[17,18], is a context-specific variable. Unlike personality Stimulators were used. The device consists of a doughnut-
traits, which are presumed to be relatively stable across shaped weight (900 g) at the end of a movable bar (231 g)
situation and time, context-specific variables vary greatly as that pivots vertically from a stand at one end. The subject's
a function of the context in which they are assessed. Because finger is placed in a grooved notch on the top of a 50-mm
they are context-specific, response expectancies might stand so that the bar can be lowered gently onto the finger.
predict placebo responding even where it is inconsistent The bar tapered at the point of contact and is approximately
over time and situation. 2 mm wide. The pain stimulator delivers approximately
2041 g of force to a finger. Two sets of the apparatus were
adjusted for the right or left hand, so that the remaining three
Method fingers could rest on the platform between the stand and the
bar attachment.
Participants and procedure Measurements of pain and pain expectancy were taken
on 11-point scales, with separate ratings taken for each
Students and nonfaculty staff at the University of hand. End points were marked from no pain at all (0) to
Plymouth were recruited via poster for the study; most of the worst pain you can imagine (10). For pain ratings,
the students received partial course credit in return for their participants were asked “How intensely did the lever hurt
participation, but the remainder of the participants received [with/without] the cream?” and for expectancy measure-
no compensation. All participants had to be aged over ments participants were asked “How intensely do you think
18 years, and exclusion criteria were any preexisting skin the lever will hurt [with/without] the cream?” For each drug
condition, any previous trauma to the fingers to be tested, or (Trivaricaine and Ibuprofen) and at each time point
any medical condition affecting bones or joints (e.g., (Session 1 and Session 2), we calculated the following
osteoporosis, arthritis). A total of 81 participants were variables for analysis: placebo scores, subtracting the score
recruited, of whom 71 (88%) completed both test sessions for untreated finger from the score for the treated finger,
and are included in the analysis below (six participants did and expectancy scores, subtracting expected pain intensity
not give consent to begin the first trial and four failed to for the untreated finger from expected intensity for the
return for the second test session). Our sample was 67% treated finger.
female, and the mean age was 20.6 years (S.D.=5.4). Acquiescence was measured with the 56-item Bass
During the consent procedure, participants were told that Acquiescence scale [19]; valid responses to each item are
the experiment was to compare the reliability of two “yes,” “no,” and “uncertain.” Bass scale scores are calculated
analgesic creams: a well-known over-the-counter remedy by summing “yes” responses. Absorption was measured with
(“Ibuprofen”) and a medication used in hospitals for the the 34-item Tellegen absorption scale [20]; valid responses
relief of pain (“Trivaricaine”). Both creams were placebos are “yes” (1) or “no” (0). Absorption scores in this sample
but were presented in realistic nonbranded chemists' (mean=19.1, S.D.=5.9) were near identical to previous
packaging and consisted of white aqueous cream. To college-based samples [21]. Internal reliability for both
administer the creams, the experimenter donned surgical scales was acceptable: Tellegen α=.80, Bass α=.87.

3. ARTICLE IN PRESS
B. Whalley et al. / Journal of Psychosomatic Research xx (2008) xxx–xxx 3
Table 1 Table 3
Means and (S.D.'s) for pain intensity scores (N=71) Correlations of placebo and expectancy scores
Time 1 Time 2 Placebo scores
Placebo
name Untreated Treated Untreated Treated Time 1 Time 2
Trivaricaine 6.56 (1.80) 5.32 (1.89) 6.83 (1.78) 5.46 (1.95) Expectancy scores Trivaricaine Ibuprofen Trivaricaine Ibuprofen
Ibuprofen 6.61 (1.75) 5.69 (1.86) 6.68 (1.73) 5.78 (2.03)
Trivaricaine Time 1 .25 ⁎ .20 .14 .12
Ibuprofen Time 1 .47 ⁎⁎ .18 .36 ⁎⁎ .15
Trivaricaine Time 2 .59 ⁎⁎ .22 .47 ⁎⁎ .33 ⁎⁎
Data analysis Ibuprofen Time 2 .27 ⁎ .61 ⁎⁎ .68 ⁎⁎ .58 ⁎⁎
⁎ P≤.05.
Statistical analyses were performed with SPSS 14. A ⁎⁎ P≤.01.
within-subject analysis of variance (ANOVA) was used to
test for the presence of a placebo effect and differences in its
magnitude as a function placebo name and experimental indicated that the delay between test sessions did not
session. Because the intertrial interval varied widely, we significantly affect the reliability of placebo responses.
tested whether the length of interval had effect by entering
interval is a covariate in the above ANOVA. Pearson Were responses temporally consistent?
correlations were calculated to assess consistency of
individual differences in the placebo effect and the prediction First, we tested whether a person who responded to a
of the placebo effect by personality variables initial placebo at Time 1 was more likely to respond to the same
response-expectancy ratings. The mediating effect of placebo at Time 2. Large correlations between placebo
expectancy ratings was analyzed with multiple regression scores for the same placebo at Times 1 and 2 indicate a
analyses in which prior placebo effect ratings were strong temporal relationship for responses to the same
controlled statistically. placebo at different times (Trivaricaine r =.60, P≤.001;
Ibuprofen r =.77, P≤.001).
Results
Were responses consistent across contexts?
Did the creams generate a placebo effect?
We tested whether a person who responded to one
placebo would be more likely to respond to a second,
Mean pain ratings with and without placebo on each trial
differently labeled placebo. Given the high correlations
are presented in Table 1. We carried out repeated-measures
between responses to the same placebo across sessions, we
ANOVA (drug label, session, and treatment) to evaluate the
combined scores for each placebo, creating a mean placebo
effects of the creams as placebos. A significant main effect
Trivaricaine score and a mean placebo Ibuprofen score.
was found for placebo treatment [F(1,70)=43.61, P≤.001,
There was no significant relationship between mean scores
η2=.38, S.M.D.=0.60], indicating that the placebo creams
for the two placebos: r =.10, P N.41. An analysis using
had a moderate beneficial effect when compared with no
individual scores from each trial (e.g., correlating response to
cream. No other significant main effects or interactions were
Trivaricaine at Time 1 or Time 2 with response to Ibuprofen
found, suggesting that the two labels were equally effective
at Time 1 or 2) produced similar results.
and they were equally effective at Sessions 1 and 2. Re-
peating this analysis with intersession interval as a covariate
Did our personality measures correlate with the
placebo effect?
Table 2
Correlations of personality measures with placebo and expectancy scores Correlations between the personality measures used and
Tellegen absorption a
Bass acquiescence a placebo and expectancy responses are shown in Table 2. No
correlations were found with placebo responses, although
Placebo responses
Trivaricaine Time 1 .06 −.05 one correlation was found between Time 1 Trivaricaine
Trivaricaine Time 2 .02 .02 expectancy and absorption. No further correlations were
Ibuprofen Time 1 −.06 −.02 found with placebo or expectancy scores when collapsing
Ibuprofen Time 2 −.06 −.04 across drug, session, looking at trials in the order of
Expectancy scores
administration, or when correlating the maximum placebo
Trivaricaine Time 1 .28 ⁎ .13
Trivaricaine Time 2 .10 .10 response exhibited in any of the trials. A follow-up
Ibuprofen Time 1 .03 .07 analysis, testing for a moderating effect of personality on
Ibuprofen Time 2 .01 .01 responses across context, indicated that neither acquies-
⁎ Pb.05. cence nor absorption affected consistency of response
a
Intercorrelation between the Tellegen and Bass scales was .44. across contexts.

4. ARTICLE IN PRESS
4 B. Whalley et al. / Journal of Psychosomatic Research xx (2008) xxx–xxx
Did response-expectancy predict the placebo effect? and response expectancies independently predicting variance
in outcome [26].
To evaluate initial response expectancies prior to any Another possibility is that response to placebo is due to
experience with either placebo cream, we calculated each implicit, as well as explicit, response expectancies. Work in
participant's initial placebo response and initial expectancy other fields has alerted researchers to the extent to which
response, ignoring the name of the label. Initial placebo nonconscious processing can influence behavior without
response and expectancy response were significantly provoking verbal reports [29,30], and this has also been
correlated (r=.33, P≤.01). Table 3 shows the relationship found in the effect of conditioning on placebo response [18].
between placebo response and expectancy for Trivaricaine It follows that verbal reports of expectancies may not
and Ibuprofen in both sessions. Subsequent expectancies exhaustively measure the extent to which people process and
were significantly associated with prior placebo effects, respond to features of a given placebo. Future research
suggesting that people's response expectancies were altered should explore this possibility.
by previous experience. However, expectancies for Triva- We found a strong association between experienced
caine [t(70)=1.10, ns] and Ibuprofen [t(70)=−.07, ns] did not placebo effects and subsequent expectancies, suggesting that
change between Sessions 1 and 2. the experience of placebo effects altered participants'
expectations for future responding. This may account for
larger observed correlations between expectancies and
Discussion placebo responses in later trials.
A practical implication of these data is that initial
We found that with contextual factors held constant, response to a placebo, as might be obtained during the run-
responses to placebo were highly reliable indicating that, in phase of a clinical trial, is likely to be the best way of
under these circumstances, there is temporal stability in identifying placebo responders. Placebo washouts are
placebo responding. However, participants' responses to two controversial, and they have been criticized both because
differently labeled placebos were not consistent, indicating they are ineffective and because they may have unexpected
that even very small changes in the context of presentation consequences on trial samples [31,32], although as Lee et al.
can affect individual differences in the placebo response and [33] have recently noted, they continue to be widely used.
confirming Gelfand's early prediction that placebo correlates Our data suggest that a general rejection of placebo washout
may vary according to the “entire placebo situation” [22]. periods may be premature. We note that much of the data on
These data suggest that the search for a generic placebo wash-out periods has come from trials of selective serotonin
responder—one who responds consistently to placebo across reuptake inhibitors (SSRIs) antidepressants for which a large
different situations—can never be successful, and our failure component of the “active” treatment is in fact placebo-
to find general associations between placebo responses and mediated [34], and under these conditions, washing out
personality measures is therefore unsurprising. placebo responders would reduce effect sizes in both arms of
There may not be one placebo response but several. It the trial. Ironically, placebo washout periods may be
would seem that there are multiple mechanisms involved and effective only in trials of medications that do not generate
they differ as a function of the context in which the placebo large placebo effects.
is presented [23–26]. For example, spirituality has been Expectancy predicts placebo responding to some degree
found to be a consistent predictor of a placebo when the and, although it is correlated with prior experience, it is
placebo is contextualized as a spiritual therapy [26], and partially independent. However, some caveats are in order.
optimism has been found to predict for a positive but not a First, unlike the present study, placebo-controlled clinical trials
negative placebo [27]. Conceptualizing placebos as a contain double-blind instructions and participants are aware
“meaning response” [28] that acts through multiple mechan- that they might receive placebo. In contrast, the instructions in
isms—including motivation as well as expectancy and our study were similar to those in comparative drug trials
conditioning—may aid our understanding of these context- without placebo controls. Second, our population was
specific effects. primarily female and we cannot be sure that the results
Response expectancies are context-specific, and this generalize to males as well. Third, we used an experimental
makes them more reliable predictors of placebo response pain stimulus. Thus, we cannot be certain that our results will
than the generic trait measures included in this study. The generalize to clinical contexts. Future studies of consistency of
best measure of the influence of expectancy is the association response in clinical patients might employ a variant of the
between initial expectancies and the initial response to open-hidden paradigm [35], avoiding the ethical problems of
placebo. This reveals a modest association, suggesting that, withholding analgesics from patients in pain. Fourth, data on
at best, expectancy only partially explains the placebo effect. placebo analgesics might not generalize to other types of
These data suggest that there may be other context-specific placebos. For these reasons, tests of consistency of placebo
characteristics of the person that determine the response to a responding should also be conducted with other types of
placebo. Recent studies of placebos in therapeutic contexts placebos known to be effective, for example anxiolytics or
have found just such a pattern, with context-specific traits treatments for Parkinson's disease [35]. Additionally, the

5. ARTICLE IN PRESS
B. Whalley et al. / Journal of Psychosomatic Research xx (2008) xxx–xxx 5
experiment was not double blind because the experimenter had [12] Brody H. The placebo response. New York: HarperCollins, 2000.
access to previous trial data for each participant. To minimize [13] Wolf S, Doering CR, Clark ML, Hagans JA. Chance distribution and
the placebo “reactor”. J Lab Clin Med 1957;49:837–41.
the risk of unintentional bias, scores for each trial were [14] Lasagna L, Mosteller F, Von Felsinger J. A study of the placebo
recorded separately, and Trial 1 scores were not visible to response. Am J Med 1954;16:770–9.
experimenter or participant during the second trial. However, [15] Shapiro AK. The reliability and validity of a placebo test. J Psychiatr
we cannot rule out the possibility of experimenter effects. Res 1979;15:253–90.
We conclude that the search for predictors of placebo [16] Kirsch I, Braffman W. Imaginative suggestibility and hypnotizability.
Curr Dir Psychol Sci 2001;10:57–61.
responses in particular contexts is realistic because [17] Kirsch I. Response expectancy as a determinant of experience and
context-specific placebo responding is as reliable as behavior. Am Psychol 1985;40:1189–202.
many trait measures. However, the search for a generic [18] Stewart-Williams S, Podd J. The placebo effect: dissolving the
placebo-prone personality is unlikely to be successful. Our expectancy versus conditioning debate. Psychol Bull 2004;130:
324–40.
data suggest that focusing on context-specific predictions
[19] Bass BM. Development and evaluation of a scale for measuring social
of responses to particular types of placebos would be a acquiescence. J Abnorm Psychol 1956;53:296–9.
more productive enterprise. [20] Tellegen A, Atkinson G. Openness to absorbing and self-altering
experiences (“absorption”), a trait related to hypnotic susceptibility.
References J Abnorm Psychol 1974;83:268–77.
[21] Glisky ML, Tataryn DJ, Tobias BA, Kihlstrom JF, McConkey KM.
[1] Schatzberg AF, Kraemer HC. Use of placebo control groups in Absorption, openness to experience, and hypnotizability. J Pers Soc
evaluating efficacy of treatment of unipolar major depression. Biol Psychol 1991;60:263–72.
Psychiatry 2000;47:736–44. [22] Gelfand S, Ullmann L, Krasner L. The placebo response: an
[2] Walsh BT, Seidman SN, Sysko R, Gould M. Placebo response in experimental approach. J Nerv Ment Dis 1963;163:379–87.
studies of major depression: variable, substantial, and growing. JAMA [23] Hyland ME, Geraghty AWA, Joy OET, Turner SI. Spirituality predicts
2002;287:1840–7. outcome independently of expectancy following flower essence self-
[3] Charney DS, Nemeroff CB, Lewis L, Laden SK, Gorman JM, Laska treatment. J Psychosom Res 2006;60:53–8.
EM, Borenstein M, Bowden CL, Caplan A, Emslie GJ, Evans DL, [24] Geers AL, Helfer SG, Kosbab K, Weiland PE, Landry SJ. Reconsider-
Geller B, Grabowski LE, Herson J, Kalin NH, Keck PE, Kirsch I, ing the role of personality in placebo effects: dispositional optimism,
Krishnan KRR, Kupfer DJ, Makuch RW, Miller FG, Pardes H, Post situational expectations, and the placebo response. J Psychosom Res
R, Reynolds MM, Roberts L, Rosenbaum JF, Rosenstein DL, 2005;58:121–7.
Rubinow DR, Rush AJ, Ryan ND, Sachs GS, Schatzberg AF, [25] Geers AL, Weiland PE, Kosbab K, Landry SJ, Helfer SG. Goal
Solomon S. National Depressive and Manic-Depressive Association activation, expectations, and the placebo effect. J Pers Soc Psychol
consensus statement on the use of placebo in clinical trials of mood 2005;89:143–59.
disorders. Arch Gen Psychiatry 2002;59:262–70. [26] Hyland ME, Whalley B, Geraghty AWA. Dispositional predictors of
[4] Quitkin F, McGrath P, Stewart J, Ocepek-Welikson K, Taylor B, Nunes placebo responding: a motivational interpretation of flower essence
E, Delivannides D, Agosti V, Donovan S, Ross D, Petkova E, Klein D. and gratitude therapy. J Psychosom Res 2007;62:331–40.
Placebo run-in period in studies of depressive disorders. Clinical, [27] Geers AL, Handley IM, McLarney AR. Discerning the role of
heuristic and research implications. Br J Psychiatry 1998:173. optimism in persuasion: the valence-enhancement hypothesis. J Pers
[5] Posternak MA, Zimmerman M, Keitner GI, Miller IW. A reevaluation Soc Psychol 2003;85:554–65.
of the exclusion criteria used in antidepressant efficacy trials. Am J [28] Moerman DE. Meaning, medicine, and the ‘placebo effect’. Cam-
Psychiatry 2002;159:191–200. bridge: Cambridge University Press, 2002.
[6] Trivedi MH, Rush J. Does a placebo run-in or a placebo treatment cell [29] Bargh JA, Chartrand TL. The unbearable automaticity of being. Am
affect the efficacy of antidepressant medications? Neuropsychophar- Psychol 1999;54:462–79.
macology 1994;11:33–43. [30] Cleeremans A, Destrebecqz A, Boyer M. Implicit learning: news from
[7] Fisher S, Greenberg RP. The curse of the placebo: fanciful pursuit of a the front. Trends Cogn Sci 1998;2:406–16.
pure biological therapy. In: Fisher S, Greenberg RP, editors. From [31] Reimherr FW, Ward MF, Byerley WF. The introductory placebo
placebo to panacea: putting psychiatric drugs to the test. New York: washout: a retrospective evaluation. Psychiatry Res 1989;30:191–9.
John Wiley, 1997. pp. 3–56. [32] Greenberg R, Fisher S, Riter J. Placebo washout is not a meaningful
[8] McNair D, Gardos G, Haskell D, Fisher S. Placebo response, placebo part of antidepressant drug trials. Percept Mot Skills 1995;81:
effect, and two attributes. Psychopharmacology (Berl) 1979;63: 688–90.
245–50. [33] Lee S, Walker JR, Jakul L, Sexton K. Does elimination of placebo
[9] McNair DM, Barrett JE. Two Bass Scale factors and response to responders in a placebo run-in increase the treatment effect in
placebo and anxiolytic drugs. Psychopharmacologia 1979;65:165–70. randomized clinical trials? A meta-analytic evaluation. Depress
[10] McNair DM, Fisher S, Sussman C, Droppleman LF, Kahn RJ. Anxiety 2004;19:10–9.
Persistence of a drug–personality interaction in psychiatric outpatients. [34] Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ
J Psychiatr Res 1970;7:299–305. 2005;331:155–7.
[11] McNair DM, Khan RJ, Droppleman LF, et al. Compatibility, [35] Colloca L, Lopiano L, Lanotte M, Benedetti F. Overt versus covert
acquiescence and drug effects. Excerpta Medica Internat Congress treatment for pain, anxiety, and Parkinson's disease. Lancet Neurol
1966;129:536. 2004;3:679–84.