21(4) 489 –493
© The Author(s) 2010
Reprints and permission:
What determines sensory preferences? Although this research
question has been intensively addressed by various disciplines,
which mechanisms underlie the shaping of preferences, and in
particular how sensory preferences are modulated by various
cognitive processes, is still a critically debated research ques-
tion (e.g., Lazarus, 1984; Lichtenstein & Slovic, 2006; Zajonc,
1984). Remarkably, it has been demonstrated that explicit
choices, traditionally considered a reflection of preferences,
can in fact shape preferences. Brehm (1956) showed that after
having made an explicit (i.e., overt) choice between two
objects evaluated as similarly desirable, participants rated the
chosen option as more desirable and the nonchosen option
(i.e., rejected) as less desirable than during the first evaluation.
Since this pioneering experiment, numerous studies have rep-
licated this choice-induced preference modulation (see Harmon-
Jones & Mills, 1999).
This effect has classically been interpreted on the basis of
the cognitive-dissonance-reduction assumption (Festinger,
1957). In this framework, awareness that the choice conflicts
with the desirable aspects of the rejected option and with the
undesirable aspects of the chosen one elicits discomfort (i.e.,
cognitive dissonance). Such discomfort could be reduced by
devaluing the rejected option and overvaluing the chosen one
(see Festinger, 1957; Sharot, De Martino, & Dolan, 2009).
However, this classical interpretation is still a matter of debate
(see Gawronski, Bodenhausen, & Becker, 2007; Harmon-
Jones, Amodio, & Harmon-Jones, 2009; Lieberman, Ochsner,
Gilbert, & Schacter, 2001). In particular, the level of process-
ing at which such choice-induced preference modulation could
take place is controversial. On the one hand, cognitive disso-
nance reduction is considered to be based on conscious strate-
gies and mediated by accessibility of dissonant cognitions to
awareness (see Allen, 1965; McGregor, Newby-Clark, &
Zanna, 1999; Rosenberg & Hovland, 1960; Wicklund & Brehm,
1976). Such interpretation would require participants to
explicitly remember the choices they made (see Gawronski
et al., 2007; Lieberman et al., 2001). On the other hand, this
preference modulation could rely on implicit processes, as
previous choices can modulate preferences without requiring
conscious or intentional recollection of them. Indeed,
Géraldine Coppin, Laboratory for the Study of Emotion Elicitation and
Expression (E3 Lab), Department of Psychology, Faculty of Psychology and
Educational Science, University of Geneva, 40 Bld. du Pont d’Arve, 1205
I’m No Longer Torn After Choice:
How Explicit Choices Implicitly Shape
Preferences of Odors
Géraldine Coppin1,2, Sylvain Delplanque1,2, Isabelle Cayeux3,
Christelle Porcherot3, and David Sander1,2
1Laboratory for the Study of Emotion Elicitation and Expression, Department of Psychology, University of Geneva;
2Swiss Center for Affective Sciences, University of Geneva; and 3Firmenich, Geneva, Switzerland
Several studies have shown that preferences can be strongly modulated by cognitive processes such as decision making and
choices. However, it is still unclear whether choices can influence preferences of sensory stimuli implicitly. This question was
addressed here by asking participants to evaluate odors, to choose their preferred odors within pairs, to reevaluate the odors,
and to perform an unexpected memory test. Results revealed, for the first time in the study of olfaction, the existence of
postchoice preference changes, in the sense of an overvaluation of chosen odors and a devaluation of rejected ones, even when
choices were forgotten. These results suggest that chemosensory preferences can be modulated by explicit choices and that such
modulation might rely on implicit mechanisms. This finding rules out any explanation of postchoice preference changes in terms
of experimental demand and strongly challenges the classical cognitive-dissonance-reduction account of such preference changes.
preferences, decision making, implicit processing, olfaction
Received 6/16/09; Revision accepted 8/26/09
Coppin et al.
postchoice evaluation changes have been demonstrated not
only in young children and capuchin monkeys (Egan, Santos,
& Bloom, 2007) but also in patients with anterograde amnesia
(Lieberman et al., 2001).
The aim of the present experiment was thus to test directly,
in a normal population, the hypothesis that sensory prefer-
ences can be implicitly shaped by explicit choices. Thus, we
adapted the free-choice paradigm (Brehm, 1956) and also had
our participants perform an unexpected memory test concern-
ing their choices. We hypothesized that explicit memory of a
choice is not necessary to observe a postchoice preference
change, and we therefore predicted that choice-induced pref-
erence modulation would be observed for both remembered
and forgotten choices.
We used olfactory stimuli, which have not, to our knowl-
edge, previously been used to study postchoice preference
changes. We therefore took advantage of the fact that olfaction
is particularly well suited to study implicit memory (e.g.,
Issanchou, Valentin, Sulmont, Degel, & Köster, 2002). More-
over, we considered that assessing the desirability or the pre-
dicted pleasantness of options, as has typically been done, may
not be the optimal way to investigate sensory preferences.
Indeed, discrepancies have been reported between experi-
enced utility—the subjective pleasantness experienced—and
predicted utility—beliefs about the subjective pleasantness
experience of outcomes (see Kahneman, Wakker, & Sarin,
1997). Therefore, we asked participants to perform their
choices based on the experienced pleasantness (i.e., experi-
enced utility) elicited by the actually presented odors rather
than on the predicted pleasantness (i.e., predicted utility) of
Thirty-seven University of Geneva students (25 females, 12
males; mean age: 23.6 ± 0.62 years) took part in this experi-
ment. Before starting, they completed a consent form. All par-
ticipants reported a normal sense of smell. They were
individually tested and were paid 10 Swiss francs for their par-
ticipation. During the day of testing, they were asked not to
wear any fragrance.
Eighteen odorants (provided by Firmenich, Geneva, Switzer-
land) were selected on the basis of their ratings of pleasant-
ness, familiarity, and intensity obtained from a previous study
(Delplanque et al., 2008). To hinder odor recognition, we
excluded very familiar odors (Rabin & Cain, 1984). We also
excluded odors that were extreme in valence or intensity. The
mean ratings of the selected odors are provided in Table 1.
Odorants were diluted in odorless dipropylene glycol to obtain
a roughly similar average intensity and were injected into
cylindrical felt-tip pens (provided by Burghart, Wedel, Ger-
many; see Delplanque et al., 2008, for further details). Each
odorant was coded by a random three-digit code.
Table 1. The 18 Odors Used and Their Mean Ratings
Odors included in all phases of the experiment
Aladinate (floral note)
Odors added during the fourth part of the experiment
Paracresol (animal note)
Vetyver (woody note)
Note: Pleasantness was rated on a scale ranging from 1, very unpleasant, to 10, very pleasant; intensity was rated on a
scale ranging from 1, not perceived, to 10, very strong. Standard deviations are given in parentheses.
Choices Implicitly Shape Odor Preferences
First, we assessed individual ratings of pleasantness for 12 of
the 18 odors. On the basis of these first ratings, pairs were cre-
ated for the choice phase. During this second phase, each par-
ticipant was presented with six odor pairs, four of them
corresponding to the conditions of interest: (a) two pairs of
odors that he or she had rated as similarly pleasant (i.e., difficult-
choice condition; mean rating differences = 0.3, SD = 0.08, on
the 10-point subjective scale described in the next section) and
(b) two pairs of odors that he or she had rated differently for
pleasantness (i.e., easy-choice condition; mean rating differ-
ences = 4.34, SD = 0.94, on the 10-point subjective scale
described in the next section). Participants were required to
choose the odor they preferred within each pair. In the third
phase, about 10 min later, participants again assessed the pleas-
antness of the 12 odors. Finally, participants rated the intensity
of these 12 odors, together with 6 new odors (see Table 1).
Critically, participants also indicated whether they had already
smelled each odor. If they answered “yes,” they were asked
whether they had chosen or rejected this odor during the choice
phase. Before this time, participants were not aware that they
would have to complete a memory task. During the entire
experiment, the order in which odors or pairs of odors were
presented was controlled. Participants were instructed to smell
each odor for no more than two inhalations.
After each odor was presented during the prechoice and the
postchoice phases, participants rated its pleasantness on a
computer screen. Participants used a mouse to move a vertical
marker across a horizontal line and click to indicate their rat-
ing. Participants rated the odor on a scale ranging from very
unpleasant, 1, to very pleasant, 10. During the last phase, par-
ticipants rated the subjective intensity of the odor on a scale
ranging from not perceived, 1, to very strong, 10.
First, the difference between prechoice and postchoice ratings
for each of the 12 odors was converted to standardized indi-
vidual z scores. Second, for each participant, we assessed
odor-recognition-memory performance by using parameters
based on signal detection theory (Corwin, 1989). If the odor
was presented during the experiment and declared so by a par-
ticipant, a hit was scored. If the odor was not presented during
the experiment but declared so, a false alarm was recorded.
From hit and false alarm scores, we then calculated four
parameters: hit rate (HR), false alarm rate (FR), discrimination
measurement (d′L), and response bias (CL). We also assessed
memory performance as a function of choice by using the
same procedure, with a hit being recorded if the odor was cho-
sen or rejected and the participant declared so accurately and a
false alarm being recorded if the odor was chosen or rejected
but the participant declared the opposite choice. For the analy-
ses performed on the subjective ratings (pleasantness and
intensity), we defined a trial as remembered if the participant
correctly recalled the choice he or she made. Otherwise, the
trial was considered forgotten.
We first assessed odor-recognition-memory performance. As
indicated by the participants’ mean hit rate, discrimination
index, and response bias (HR = 0.87, d′L = 3.33, CL = –0.47),
participants remembered the presented odors well and dis-
criminated them well from the distracting odors. In contrast,
memory performance for choice was globally poor (HR =
0.41). In addition, we investigated whether recall of choice
depended on the type of choice made (chosen vs. rejected) and
its difficulty. Thus, we performed a 2 × 2 repeated measures
multivariate analysis of variance (MANOVA), with choice
(chosen, rejected) and difficulty (difficult, easy) as the inde-
pendent variables and hit rate, discrimination, and response
bias as the dependent variables. Critically, neither main effect
(choice or difficulty) nor the Choice × Difficulty interaction
was statistically significant, all Fs(1, 36) < 1, ps > .29.
of preferences for odors
Choice-induced changes are typically reported when the choice
is difficult. In our case, this was when the pleasantness of the
two paired odors had been rated similarly before. A repeated
measures analysis of variance (ANOVA), with choice (chosen,
rejected) as the independent variable and the difference between
prechoice and postchoice ratings in the difficult condition as
the dependent variable, was significant, F(1, 36) = 15.15, p <
overvaluation of the chosen odors or the devaluation of the
rejected odors, we analyzed pleasantness scores in the difficult-
choice condition using a 2 × 2 repeated measures MANOVA
with phase (prechoice, postchoice) and choice (chosen,
rejected) as the independent variables. As displayed in
Figure 1, the interaction between these factors was statistically
significant, F(1, 36) = 17.10, p < .001, ηp
choices, pleasantness ratings were significantly decreased for
rejected odors, planned contrast, F(1, 36) = 8.75, p < .01, ηp
.20; the increase in pleasantness ratings for chosen odors was
marginally significant, planned contrast, F(1, 36) = 2.88, p =
antness scores in the easy-choice condition revealed only a sig-
nificant main effect of choice, F(1, 36) = 90.3, p < .001, ηp
.71, which simply reflected significantly higher pleasantness
ratings for chosen odors than for rejected ones.
2 = .29. To specify whether this effect was due to the
2 = .32. For difficult
2 = .07. The identical analysis performed on the pleas-
Coppin et al.
Role of explicit memory of the choice
To assess the role of memory for the choice in postchoice pref-
erence changes, we performed a 2 × 2 repeated measures
MANOVA with choice (chosen, rejected) and memory (remem-
bered, forgotten) as the independent variables and difference
score in the difficult-choice condition as the dependent vari-
able. This analysis revealed a main effect only of choice, F(1,
9) = 6.95, p < .03, ηp
the overvaluation of the chosen odors and the devaluation of
the rejected ones was significant for forgotten and remembered
choices combined (see Fig. 2). Two repeated measures
ANOVAs conducted on the difference scores for the forgotten
and remembered difficult choices separately confirmed signifi-
cant postchoice preference changes for both, F(1, 28) = 7.91
and F(1, 14) = 7.06, ps < .05, ηp
2 = .44, showing that the difference between
2s = .22 and .33.
Influence of pleasantness on choices
We further investigated whether the pleasantness of the odor
before the choice varied as a function of the participants’ choice.
A repeated measures ANOVA with pleasantness score before
the choice as the dependent variable and choice (chosen,
rejected) as the independent variable did not reach significance,
F(1, 36) = 0.23, p > .6.
The mean intensity ratings for all the odors are reported in
Table 1. There was no difference between the intensity ratings
of the chosen and the rejected odors, F(1, 36) = 2.17, p > .1.
Results showed, for the first time in the study of olfaction, the
existence of postchoice preference changes, as exhibited by an
overvaluation of chosen odors and a devaluation of rejected ones.
This finding indicates that preference shaping by decision-making
processes also applies to smell, suggesting that the underlying
mechanisms may not be modality specific. In this respect, further
studies might investigate the extent to which postchoice prefer-
ence changes could be transferred from one modality to another.
Moreover, we demonstrated the existence of postchoice
preference changes not only when choices were remembered,
but also, critically, when choices were forgotten. This last
point rules out any explanation of the choice-induced prefer-
ence changes in terms of experimental demand and, most
important, suggests that the mechanisms underlying this
choice-induced preference change may function at an implicit
level (see also Lieberman et al., 2001). Therefore, the classical
cognitive-dissonance model cannot accurately account for
these findings unless one assumes that dissonance elicitation
and reduction may be implicit.
The recent action-based model of cognitive dissonance
(Harmon-Jones et al., 2009) postulates that making a choice
between two similarly pleasant stimuli leads to conflicted
action tendencies, as they both elicit approach tendencies
automatically (Chen & Bargh, 1999; Custers & Aarts, 2005).
These conflicted action tendencies lead to an unpleasant emo-
tional reaction of dissonance that is reduced by postchoice
preference changes. In particular, a decrease in pleasantness of
a rejected stimulus will decrease the approach tendency toward
it (Veling, Holland, & van Knippenberg, 2008) and conse-
quently reduce the unpleasant emotional reaction elicited by
the initial conflicting action tendencies.
Testing alternative models of postchoice preference
changes could benefit from a better understanding of the brain
mechanisms underlying these psychological processes. For
instance, in the first functional magnetic resonance imaging
experiment on this topic, Sharot et al. (2009) have shown that
postchoice changes in stimulus evaluation can be predicted on
the basis of the activation in the caudate nucleus, a region
Chosen Odors Rejected Odors
Fig. 1. Mean pleasantness ratings of odors that were presented in the
difficult-choice condition, as a function of phase of the experiment (prechoice
or postchoice) and whether the odor was chosen or rejected during the
choice phase. Pleasantness ratings were converted to mean z scores. Error
bars represent standard errors of the mean.
Difference in Pleasantness Scores
Forgotten Choices Remembered Choices
Fig. 2. Mean difference in pleasantness scores in the difficult-choice
condition as a function of whether the odor was chosen or rejected and
whether this choice was forgotten or remembered correctly. Scores were
converted to z scores, and difference was calculated as the postchoice score
minus the prechoice score. Error bars represent standard errors of the mean.
Choices Implicitly Shape Odor Preferences Download full-text
modulated by the perceived value of a stimulus, as well as by
choices associated with rewards (Delgado, 2007).
To conclude, our results suggest not only that preference
acquisition can be determined by antecedent explicit choice,
but also that such changes might rely on implicit processes. An
important research question concerns the consolidation of
implicitly shaped preferences in long-term memory. In partic-
ular, future experiments could investigate how postchoice
preference changes evolve over time when choices are remem-
bered or forgotten.
The authors thank Maria-Inés Velazco, Christian Margot, and all the
members of the Human Perception and Bioresponses Department of
the Research and Development Division of Firmenich, SA, for their
precious advice and their theoretical and technical competence. The
authors also thank two anonymous reviewers and the editor for their
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with
respect to their authorship or the publication of this article.
This research was supported by the National Center of Competence
in Research for the Affective Sciences, financed by a grant from the
Swiss National Science Foundation (51NF40-104897), hosted by the
University of Geneva, and also funded by a research grant from
Firmenich, SA, to David Sander and Patrik Vuilleumier.
Allen, V.L. (1965). Effect of extraneous cognitive activity on disso-
nance reduction. Psychological Reports, 16, 1145–1151.
Brehm, J.W. (1956). Post-decision changes in desirability of choice
alternatives. Journal of Abnormal and Social Psychology, 52,
Chen, M., & Bargh, J.A. (1999). Consequences of automatic evalua-
tion: Immediate behavioral predispositions to approach or avoid the
stimulus. Personality and Social Psychology Bulletin, 25, 215–224.
Corwin, J. (1989). Olfactory identification in hemodialysis: Acute
and chronic effects on discrimination and response bias. Neuro-
psychologia, 27, 513–522.
Custers, R., & Aarts, H. (2005). Positive affect as implicit motivator:
On the nonconscious operation of behavioral goals. Journal of
Personality and Social Psychology, 89, 129–142.
Delgado, M.R. (2007). Reward-related responses in the human stria-
tum. Annals of the New York Academy of Sciences, 1104, 70–88.
Delplanque, S., Grandjean, D., Chrea, C., Aymard, L., Cayeux, I.,
Le Calvé, B., et al. (2008). Emotional processing of odours: Evi-
dence for a non-linear relation between pleasantness and famil-
iarity evaluations. Chemical Senses, 33, 469–479.
Egan, L.C., Santos, L.R., & Bloom, P. (2007). The origins of cogni-
tive dissonance: Evidence from children and monkeys. Psycho-
logical Science, 11, 978–983.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA:
Stanford University Press.
Gawronski, B., Bodenhausen, G., & Becker, A.P. (2007). I like it,
because I like myself: Associative self-anchoring and post-
decisional change of implicit evaluations. Journal of Experimental
Social Psychology, 43, 221–232.
Harmon-Jones, E., Amodio, D.M., & Harmon-Jones, C. (2009).
Action-based model of dissonance: A review, integration, and
expansion of conceptions of cognitive conflict. Advances in
Experimental Social Psychology, 41, 119–166.
Harmon-Jones, E., & Mills, J. (Eds.). (1999). Cognitive dissonance:
Progress on a pivotal theory in social psychology. Washington,
DC: American Psychological Association.
Issanchou, S., Valentin, D., Sulmont, C., Degel, J., & Köster, E.P.
(2002). Testing odor memory: Incidental versus intentional learn-
ing, implicit versus explicit memory. In C. Rouby, B. Schaal, D.
Dubois, R. Gervais, & A. Holley (Eds.), Olfaction, taste, and
cognition (pp. 211–230). New York: Cambridge University Press.
Kahneman, D., Wakker, P.P., & Sarin, R. (1997). Back to Bentham?
Explorations of experienced utility. The Quarterly Journal of
Economics, 112, 375–406.
Lazarus, R.S. (1984). On the primacy of cognition. American Psy-
chologist, 39, 124–129.
Lichtenstein, S., & Slovic, P. (2006). The construction of preference.
New York: Cambridge University Press.
Lieberman, M.D., Ochsner, K.N., Gilbert, D.T., & Schacter, D.L.
(2001). Do amnesiacs exhibit cognitive dissonance reduction?
The role of explicit memory and attention in attitude change.
Psychological Science, 2, 135–140.
McGregor, I., Newby-Clark, I.R., & Zanna, M.P. (1999). “Remem-
bering” dissonance: Simultaneous accessibility of inconsistent
cognitive elements moderates epistemic discomfort. In E. Harmon-
Jones & J. Mills (Eds.), Cognitive dissonance: Progress on a piv-
otal theory in social psychology (pp. 25–42). Washington, DC:
American Psychological Association.
Rabin, M.D., & Cain, W.S. (1984). Odour recognition: Familiarity,
identifiability, and encoding consistency. Journal of Experimen-
tal Psychology: Learning, Memory, and Cognition, 10, 316–325.
Rosenberg, M.J., & Hovland, C.I. (1960). Cognitive, affective, and
behavioral components of attitudes. In C.I. Hovland & M.J.
Rosenberg (Eds.), Attitude organization and change: An analy-
sis of consistency among attitude components (pp. 1–14). New
Haven, CT: Yale University Press.
Sharot, T., De Martino, B., & Dolan, R.J. (2009). How choice reveals
and shapes expected hedonic outcome. Journal of Neuroscience,
Veling, H., Holland, R.W., & van Knippenberg, A. (2008). When
approach motivation and behavioral inhibition collide: Behavior
regulation through stimulus devaluation. Journal of Experimen-
tal Social Psychology, 44, 1013–1019.
Wicklund, R.A., & Brehm, J.W. (1976). Perspectives on cognitive
dissonance. Hillsdale, NJ: Erlbaum.
Zajonc, R.B. (1984). On the primacy of affect. American Psycholo-
gist, 39, 117–123.