Journal of Environmental Psychology 57 (2018) 83e86
Contents lists available at ScienceDirect
Journal of Environmental Psychology
journal homepage: www.elsevier.com/locate/jep
The impact of coffee-like scent on expectations and performance
Adriana Madzharov a, *, Ning Ye b, Maureen Morrin b, Lauren Block c
a
School of Business, Stevens Institute of Technology, 1 Castle Point on Hudson, Hoboken, NJ 07030, USA
Fox School of Business, Temple University, 1801 Liacouras Walk, Philadelphia, PA 19122-6083, USA
c
Baruch College, The City University of New York, One Bernard Baruch Way, New York, NY 10010, USA
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 August 2017
Received in revised form
14 April 2018
Accepted 21 April 2018
Available online 23 April 2018
The present research explores the effect of an ambient coffee-like scent (versus no scent) on expectations
regarding performance on an analytical reasoning task as well as on actual performance. We show that
people in a coffee-scented (versus unscented) environment perform better on an analytical reasoning
task due to heightened performance expectations (Study 1). We further show that people expect that
being in a coffee-scented environment will increase their performance because they expect it will increase their physiological arousal level (Study 2). Our results thus demonstrate that a coffee-like scent
(which actually contains no caffeine) can elicit a placebo effect.
© 2018 Elsevier Ltd. All rights reserved.
Handling Editor: Florian Kaiser
Keywords:
Olfaction
Performance
Coffee-like scent
Placebo effects
1. Introduction
Many people experience the scent of coffee on a daily basis: in
the workplace, at coffee shops, and in retail and service environments. Coffee houses such as Starbucks are known as the “third
place” where people spend hours of their day working, socializing,
or relaxing while immersed in a coffee-scented atmosphere
(Oldenburg, 1989, pp. 39; Walton, 2012, pp. 1). Beyond stand-alone
coffee shops, numerous retailers have built coffee counters into
many of their outlets (e.g., Uniqlo, Club Monaco; Cheng, 2014).
Given the prevalence and popularity of coffee-scented environments in people’s daily lives, it is important to understand how
coffee scent might influence behavior. Can merely smelling a
coffee-like scent without actually ingesting coffee have an effect on
people’s behavior?
The vast majority (80%) of the world’s population consumes
caffeinated products such as coffee or tea every day, making
caffeine “the single most widely consumed psychoactive ingre€ther & Giesbrecht, 2013, pp. 251). Studies suggest that
dient” (Eino
ingesting caffeine can enhance self-reported levels of alertness and
* Corresponding author.
E-mail addresses: adriana.madzharov@stevens.edu (A. Madzharov), ning.ye@
temple.edu (N. Ye), maureen.morrin@temple.edu (M. Morrin), lauren.block@
baruch.cuny.edu (L. Block).
https://doi.org/10.1016/j.jenvp.2018.04.001
0272-4944/© 2018 Elsevier Ltd. All rights reserved.
feelings of “energy,” and decrease perceived mental fatigue (Glade,
2010). In addition, caffeine has been shown to boost performance
on tasks requiring sustained vigilance, and improve accuracy in
€ther
solving reasoning problems and making correct decisions (Eino
& Giesbrecht, 2013; Glade, 2010; Smith, 2002).
We posit that because of these well-known effects of coffee and
caffeine, people will associate not only the ingestion of coffee but
merely the smell of coffee with the same types of effects. Prior
research has shown that ambient scent can enhance people’s ability
to recognize and recall brand stimuli (Morrin & Ratneshwar, 2003)
and promote approach behavior online (Vinitzky & Mazursky,
2011). The scent literature has also established that scents can
bias perceptions and prime certain behaviors because of the se, Poels, Janssens, & De
mantic associations they evoke (Douce
guen, 2012; Krishna, Elder, & Caldara, 2010;
Backer, 2013; Gue
Madzharov, Block, & Morrin, 2015). In a similar way, we propose
that coffee scent will be associated with the positive effects that
caffeine is known to have on alertness and behavioral performance.
Specifically, we propose that an ambient coffee-like scent will
produce a placebo effect such that it will lead people to feel and
behave in ways similar to drinking coffee. Thus, we extend knowledge on ambient scents by identifying for the first time coffee-like
scent as an environmental stimulus that can influence behavior.
A placebo is “an inert substance or procedure that alters one’s
physiological or psychological response” (Geers, Kosbab, Helfer,
Weiland, & Wellman, 2007, pp. 563). A placebo effect is “a
84
A. Madzharov et al. / Journal of Environmental Psychology 57 (2018) 83e86
positive outcome resulting from the belief that a beneficial treatment has been received” (Beedie & Foad, 2009, pp. 314). Placebo
effects, from both the actual and perceived ingestion of caffeine,
have been demonstrated. In one study (Beedie, Stuart, Coleman, &
Foad, 2006), competitive cyclists were told they would receive in
random order a placebo, a low dose of caffeine, and a high dose of
caffeine (when in fact all were placebos), and their cycling power
would be measured after each treatment. The authors found a
dose-response relationship such that cycling improved with increases in the perceived doses of caffeine (Beedie et al., 2006, pp.
2160).
Expectancy theory describes how individuals learn relationships among events via repeated exposure, over time (e.g., when A
occurs, B occurs). Such learning allows individuals to predict outcomes in their environment (e.g., if A is observed to have occurred,
B is expected to occur). Expectancy theory thus helps to explain
placebo effects (Rescorla, 1988). Expectancy is the “experienced
likelihood of an outcome or an expected effect” (Price, Finniss, &
Benedetti, 2008, p. 571). What seems to be critical in placebo
studies is not whether or not an individual is in the placebo or true
treatment group, but whether or not the individual expects the
treatment they receive will work (Price et al., 2008). Thus, it is
largely the expectation or belief that the placebo will work that
elicits the expected outcome. In the present research, expectancy is
the extent to which a participant expects that smelling a coffee-like
scent will enhance their performance on an analytical reasoning
task.
We thus extend prior research by showing that merely smelling
a coffee-like scent leads to a placebo effect. We propose:
H1. A coffee-like ambient scent (vs. no scent) will lead to
improved performance on an analytical reasoning task. This effect
will be mediated by higher performance expectations.
2. Study 1
2.1. Method
2.1.1. Design and subjects
One hundred and fourteen (87% aged 18-21, 41% female) undergraduate business students participated for course credit in a
study that was run, in groups of about 30, as a single factor
(ambient scent: yes ¼ coffee-like scent, no ¼ unscented control)
between-subjects design with random assignment to condition.
2.1.2. Procedure
The experiment was conducted in a computer laboratory where
participants entered either a coffee-like-scented or unscented
environment and filled out a computer survey. We used a commercial electric diffuser to emit the coffee-like scent into the room
(Madzharov et al., 2015). The scent smelled like coffee but contained no actual caffeine or other stimulants. The scent diffuser
(kept hidden from view) was run for five minutes before the
beginning of each scented session, and then was used to emit the
odor every 15 s during the experimental sessions.
After entering the scented (or unscented) room, participants
were told that they would take part in an analytical reasoning task.
While in the room, participants indicated their performance expectations (“How confident are you that you will do well on the
task?”; “How well do you expect to do on the task?”; “How certain
do you feel you will score highly on the task?”; from 1 ¼ not at all to
7 ¼ very; a ¼ 0.97). Then they completed the analytical reasoning
task, which consisted of ten multiple-choice algebraic math problems from the GMAT (Graduate Management Admission Test),
which the subject population was expected to be familiar with, and
which has previously been used to measure performance (Ilyuk &
Block, 2015). Performance was measured by the number of correct responses (out of ten possible).
To minimize demand effects, participants were asked questions
about the ambient scent only after responding to the other measures. They were asked whether they noticed any scent when they
first came into the room (yes, no) and whether they noticed any
scent right now (yes, no). The majority of participants in the coffeelike scent condition reported they remembered noticing a smell
when they first came in (63%), and noticed a scent at the end of the
survey (70.4%). Finally, participants answered demographic questions and a hypothesis probe.
2.2. Results
An ANOVA on expectations with ambient scent as a betweensubjects factor was significant (F(1, 112) ¼ 9.56, p < .01,
h2 ¼ 0.079). Participants in the coffee-like scent condition expected
to perform better than those in the unscented control (MCoffeeLikeScent ¼ 4.94, SD ¼ 1.45 vs. MNoScent ¼ 4.18, SD ¼ 1.73).
An ANOVA on performance with ambient scent as a betweensubjects factor was significant (F(1, 112) ¼ 11.83, p < .01,
h2 ¼ 0.096). Participants in the coffee-like scent condition scored
higher than those in the unscented control (MCoffeeLikeScent ¼ 5.44,
SD ¼ 1.95 vs. MNoScent ¼ 4.22, SD ¼ 1.86).
Using standardized variables, we performed a mediation test
using a bias-corrected bootstrap procedure (Hayes’ Model 4, Hayes
& Preacher, 2013), with scent as the independent factor (unscented
control ¼ 0, coffee-like scent ¼ 1), expectations as the mediator,
and the number of correct responses on the GMAT task as the
dependent variable. The results showed a significant indirect effect
of scent on performance via expectations (bIndirectEffect ¼ 0.13, 95%
CI [0.03, 0.29]), but the results indicated only partial mediation. The
analysis revealed that scent condition predicted expectations about
performance (bScentToExpectations ¼ .56, p < .01) and that expectations
predicted actual performance (bExpectationsToPerformance ¼ .24, p < .01;
see Fig. 1).
We conducted a follow-up study to better understand the underlying process. We wanted to see whether people expect that a
coffee scent would improve performance because they believe it
would lead to higher levels of physiological arousal (an effect
traditionally associated with the ingestion of caffeine). In Study 2,
we ask people what the effect would be of being in an environment
with a regular coffee scent (or a decaffeinated coffee scent, a floral
scent, or no scent at all). In this study we also seek to understand
whether the imagined presence of any scent would have a similar
effect on expectations regarding both performance and physiological arousal.
3. Study 2
3.1. Method
3.1.1. Design and subjects
Two hundred and eight participants from Amazon Mechanical
Turk took part for a small payment (Agemean ¼ 33, SD ¼ 9.70; 55%
female) in a single-factor design (imagined scent condition: regular
coffee, decaffeinated coffee, floral, no scent control), with random
assignment.
3.1.2. Procedure
We asked participants to imagine they were about to perform
on an analytic reasoning task in the presence of either: regular
coffee scent, decaffeinated coffee scent, floral scent, or no scent.
Participants were asked to indicate how alert and energetic they
A. Madzharov et al. / Journal of Environmental Psychology 57 (2018) 83e86
85
Effects of Coffee-Like Scent (vs. No Scent) on Performance
via Expectations (Study 1)
Expectations
about Performance
= .56, SE=.18
t =3.09, p < .01
Scent
(Coffee-Like Scent = 1,
No Scent = 0)
= .24, SE=.09
t = 2.62, p = .01
Total Effect:
= .62, SE=.18
t = 3.44, p <.01
GMAT Test
Performance
Direct Effect:
= .48, SE=.18
t = 2.65, p <.01
Note: analysis was conducted with standardized variables.
Fig. 1. Effects of coffee-like Scent(vs. No scent) on performance via expectations (study 1).
would feel (“I will feel energetic”; “ I will feel alert”; on 7-point
scales from 1 ¼ not at all to 7 ¼ very), as a measure of expected
physiological arousal. We averaged the answers to the two items
measuring the expected physiological arousal to create a single
index (r ¼ 0.96, p < .001). We also measured their expectations
about performance (same three items as in Study 1; a ¼ 0.87).
3.2. Results
An ANOVA on expected physiological arousal as a function of
imagined scent condition was significant (F(3, 204) ¼ 14.43, p < .01,
h2 ¼ 0.18). The LSD post-hoc tests showed that participants
believed they would feel more physiologically aroused in the regular coffee scent condition than in any of the other conditions
(MRegularCoffeeScent ¼ 5.25, SD ¼ 1.40 vs. MDecafCoffeeScent ¼ 4.56,
SD ¼ 1.23, p < .01; vs. MFloralScent ¼ 4.59, SD ¼ 1.79, p < .01; vs.
MNoScent ¼ 3.62, SD ¼ 1.39, p < .01). The decaffeinated and floral
scent conditions did not differ from each other (p ¼ .90) but were
both greater than the no scent control (ps < .01).
A similar ANOVA on expectations revealed a significant effect
(F(3, 204) ¼ 8.22, p < .01, h2 ¼ 0.11). The LSD post-hoc test showed
that participants in the regular coffee scent condition had higher
performance expectations than those in the other conditions (M
SD ¼ 1.24
vs.
MDecafCoffeeScent ¼ 4.43,
RegularCoffeeScent ¼ 4.98,
SD ¼ 1.21, p < .05; vs. MFloralScent ¼ 4.36, SD ¼ 1.05, p < .01; vs.
MNoScent ¼ 3.92, SD ¼ 0.96, p < .01). The decaffeinated and floral
scent conditions did not differ from each other (p ¼ .80) but were
both greater than the no scent control (ps < .05).
Using standardized variables, mediation analysis showed that
participants’ beliefs regarding their physiological arousal fully
mediated the effect of imagined regular coffee scent (no scent ¼ 0,
regular coffee scent ¼ 1) on performance expectations (bIndirectEffect ¼ 0.80, 95% CI ¼ [0.52, 1.09]; see Fig. 2 for details). Expected
physiological arousal also fully mediated the imagined effects of
regular (coded 1) versus decaffeinated (coded 0) coffee scent (bIndirectEffect ¼ 0.41, 95% CI ¼ [0.11, 0.75]; bScentToArousal ¼ 0.49, p < .01;
bArousalToPerformance ¼ .84, p < .01) and of regular coffee (coded 1)
versus floral (coded 0) scent (bIndirectEffect ¼ 0.38, 95% CI ¼ [0.11,
0.66]; bScentToArousal ¼ 0.47, p < .01; bArousalToPerformance ¼ .81, p < .01)
on performance expectations.
4. General discussion
We contribute to current understanding of how ambient scents
Beliefs about the Effect of Regular Coffee Scent (vs. No Scent) on
Expectations about Performance on Cognitive Tasks via Expected
Physiological Arousal Level (Study2)
= 1.15, SE=.19
t =6.14, p < .01
Imagined Scent
(Regular Coffee Scent=1,,
No Scent = 0)
Expected
Physiological
Arousal Level
Total Effect:
= .90, SE=.18
t = 4.96, p <.01
= .70, SE=.07
t = 10.62, p < .01
E
Expectations
about
Performance
P
Direct Effect:
= .10, SE=.15
t = .70, p =.49
Note: analysis was conducted with standardized variables.
Fig. 2. Beliefs about the effect of regular coffee scent (vs. No scent) on expectations about performance on cognitive tasks via expected physiological arousal level (study 2).
86
A. Madzharov et al. / Journal of Environmental Psychology 57 (2018) 83e86
work by identifying coffee-like scent as a specific sensory trigger
that can change behavior in a placebo-like manner. These findings
support prior literature on olfaction that has established the
importance of the sense of smell and the subtle but powerful influence that scents can have on perception and behavior (Krishna
et al., 2010). For the first time, we identify coffee-like scent as one
such influencer of analytical reasoning performance via
expectations.
Our findings are also informative for research on placebo effects
by suggesting that smelling certain scents can have an effect similar
to ingesting placebo substances. Placebo effects based on actual or
believed caffeine ingestion have been demonstrated for physical
performance tasks (e.g., Beedie & Foad, 2009). Here, we extend
these findings by showing that merely smelling a caffeineassociated scent (without ingesting any caffeine) can boost performance on an analytical reasoning task. Additional research is
needed to better understand people’s expectations regarding the
specific effect of scents presumed to contain caffeine on arousal and
cognition. For example, do people believe that smelling coffee scent
would increase the length of time they could remain vigilant? What
other types of mental performance enhancement would people
expect to experience when smelling coffee scent? Here we operationalized people’s beliefs about coffee scent effects in a limited
manner, using math problems from a standardized test. Future
research could examine whether people also believe they could
process verbal information more efficiently (e.g., comprehend advertisements seen in a shopping mall faster), or calculate market
transactions more accurately (e.g., purchase totals, expected
change).
This research has several limitations. Study 1 was conducted
among students and thus would benefit from replication with other
types of participants in other situations. For example, a study in a
more natural coffee-scented environment such as a coffee shop or a
work place would provide a more realistic setting. Study 2, while
helpful in terms of understanding the basis for people’s expectations regarding the effect of coffee scent on arousal and performance, was nevertheless conducted without the emission of actual
scent.
In conclusion, this research demonstrates that a coffee-like
scent (vs. no scent) had a positive effect on performance on an
analytical reasoning task, driven by people’s expectations. Given
the prevalence and popularity of coffee-scented environments in
people’ daily lives, the demonstrated effect has multiple practical
implications and might suggest that people may benefit from being
immersed in coffee-scented environments without the consumption of caffeine. In addition, our results also suggest that architects,
building developers, and retail space managers should be cognizant
of the potential impact of exposure to incidental ambient odors
such as coffee scent, on people’s beliefs and behavioral patterns.
Exposure to these and other types of odors might elicit unanticipated placebo effect type responses that designers may wish to
mitigate via heating, ventilation and air conditioning systems.
References
Beedie, C. J., & Foad, A. J. (2009). The placebo effect in sports performance. Sports
Medicine, 39(4), 313e329.
Beedie, C. J., Stuart, E. M., Coleman, D. A., & Foad, A. J. (2006). Placebo effects of
caffeine on cycling performance. Medicine & Science in Sports & Exercise, 38(12),
2159.
Cheng, A. (2014, March 28). Uniqlo’s new partnerships with Starbucks, MoMa are
signs of a new retail trend. Market Watch. August 5, 2016 http://blogs.
marketwatch.com/behindthestorefront/2014/03/28/uniqlos-new-partnershipswith-starbucks-moma-are-signs-of-a-new-retail-trend/.
, L., Poels, K., Janssens, W., & De Backer, C. (2013). Smelling the books: The
Douce
effect of chocolate scent on purchase-related behavior in a bookstore. Journal of
Environmental Psychology, 36, 65e69.
€ther, S. J. L., & Giesbrecht, T. (2013). Caffeine as an attention enhancer:
Eino
Reviewing existing assumptions. Psychopharmacology, 225(2), 251e274.
Geers, A. L., Kosbab, K., Helfer, S. G., Weiland, P. E., & Wellman, J. A. (2007). Further
evidence for individual differences in placebo responding: An interactionist
perspective. Journal of Psychosomatic Research, 62(5), 563e570.
Glade, M. J. (2010). Caffeinednot just a stimulant. Nutrition, 26(10), 932e938.
guen, N. (2012). The sweet smell of… courtship: Effects of pleasant ambient
Gue
fragrance on women’s receptivity to a man’s courtship request. Journal of
Environmental Psychology, 32(2), 123e125.
Hayes, A. F., & Preacher, K. J. (2013). Conditional process modeling: Using structural
equation modeling to examine contingent causal processes. Structural Equation
Modeling: A Second Course, 2, 217e264.
Ilyuk, V., & Block, L. (2015). The effects of single-serve packaging on consumption
closure and judgments of product efficacy. Journal of Consumer Research, 42(6),
858e878.
Krishna, A., Elder, R. S., & Caldara, C. (2010). Feminine to smell but masculine to
touch? Multisensory congruence and its effect on the aesthetic experience.
Journal of Consumer Psychology, 20(4), 410e418.
Madzharov, A. V., Block, L. G., & Morrin, M. (2015). The cool scent of power: Effects
of ambient scent on consumer preferences and choice behavior. Journal of
Marketing, 79(1), 83e96.
Morrin, M., & Ratneshwar, S. (2003). Does it make sense to use scents to enhance
brand memory? Journal of Marketing Research, 40(1), 10e25.
Oldenburg, R. (1989). The great good place: Cafes, coffee shops, bookstores, bars, hair
salons, and other hangouts at the heart of a community. New York: Marlowe & Co.
Price, D. D., Finniss, D. G., & Benedetti, F. (2008). A comprehensive review of the
placebo effect: Recent advances and current thought. Annual Review of Psychology, 59, 565e590.
Rescorla, R. A. (1988). Pavlovian conditioning: It’s not what you think it is. American
Psychologist, 43(3), 151.
Smith, A. (2002). Effects of caffeine on human behavior. Food and Chemical Toxicology, 40(9), 1243e1255.
Vinitzky, G., & Mazursky, D. (2011). The effects of cognitive thinking style and
ambient scent on online consumer approach behavior, experience approach
behavior, and search motivation. Psychology and Marketing, 28(5), 496e519.
Walton, A. G. (2012, May 29). Starbucks’ power over us is bigger than coffee: It’s
personal. Forbes. August 5, 2016 https://www.forbes.com/sites/alicegwalton/
2012/05/29/starbucks-hold-on-us-is-bigger-than-coffee-its-psychology/
#159961b34aed.