ArticlePDF Available

Neuromarketing, Subliminal Advertising, and Hotel Selection: An EEG Study


Abstract and Figures

This study aims to understand how hotel videos embedded with a smiling face emoji as a subliminal message affect consumers’ selection of hotels, with their brain activities measured and collected while they watched the videos. Data was collected from sixteen participants who completed two rounds of experiments. A chi-square test of homogeneity, paired sample t-test, and Bayes factor were performed to address the two proposed research questions. The results of this study reveal that participants’ selection of hotels would be significantly affected by the subliminal stimuli of a smiling face emoji. Meanwhile, neuroscientific data identifies significant differences between participants’ two (theta and beta) out of five bands of brainwaves while they were viewing hotel videos with and without the subliminal message. Suggestions for future studies and practical operations are also discussed.
Content may be subject to copyright.
JID: AMJ [m5G; May 22, 2020;0:55 ]
Australasian Marketing Journal xxx (xxxx) xxx
Contents lists available at ScienceDirect
Australasian Marketing Journal
journal homepage:
Neuromarketing, subliminal advertising, and hotel selection: An EEG
Liwei Hsu, Yen-Jung Chen
Graduate Institute of Hospitality Management, National Kaohsiung University of Hospitality and Tourism, No. 1, Song-he Rd., Hsiao-kang District, Kaohsiung
City, Taiwan
a r t i c l e i n f o
Article history:
Received 1 February 2020
Revised 27 April 2020
Accepted 28 April 2020
Available online xxx
Keywo rds:
Subliminal advertising
Hotel selection
Electroencephalography (EEG)
a b s t r a c t
This study aims to understand how hotel videos embedded with a smiling face emoji as a subliminal
message affect consumers’ selection of hotels, with their brain activities measured and collected while
they watched the videos. Data was collected from sixteen participants who completed two rounds of
experiments. A chi-square test of homogeneity, paired sample t -test, and Bayes factor were performed to
address the two proposed research questions. The results of this study reveal that participants’ selection
of hotels would be significantly affected by the subliminal stimuli of a smiling face emoji. Meanwhile,
neuroscientific data identifies significant differences between participants’ two (theta and beta) out of
five bands of brainwaves while they were viewing hotel videos with and without the subliminal message.
Suggestions for future studies and practical operations are also discussed.
©2020 The Author(s). Published by Elsevier Ltd on behalf of Australian and New Zealand Marketing
This is an open access article under the CC BY-NC-ND license.
( )
1. Introduction
Neuromarketing (or ‘consumer neuroscience’, as termed by
Smidts et al. 2014 ) is an emerging field of marketing that com-
bines perspectives of marketing, neuroscience, economics, deci-
sion theory, and psychology. Neuromarketing employs brain imag-
ing technology to effectively reveal the underlying reasons for
consumer behaviour and predict consumers’ decision-making pro-
cesses ( Morin, 2011 ; Nilashi et al., 2020 ; Yoon et al., 2006 ). It has
drawn considerable attention in the business world ( Hafez, 2019 ;
Plassmann et al., 2012 ) and undoubtedly, neuromarketing plays
an important role in the theoretical development of consumer
decision-making ( Spence, 2019 ). Humans are susceptible to com-
mercial advertisements ( Koenigs and Tranel, 2007 ), and the main
goal of any type of marketing is to influence consumers’ pur-
chase decisions ( Bargh, 2002 ; Shrum et al., 2012 ). Under the um-
brella of neuromarketing, subliminal advertising has evolved as a
skill and technique adopted by many marketers to influence po-
tential consumers’ impressions ( Soomro, 2018 ), and affect their
behaviour ( Sofi et al., 2018 ). In hospitality and tourism market-
ing, subliminal advertising has also been exploited to appeal to
Corresponding author.
E-mail addresses: (L. Hsu),
(Y.-J. Chen).
more consumers and tourists (
McKercher and Koh, 2017 ), as hu-
man cognition can be affected by subliminal stimuli that en-
able one to activate his/her existing knowledge or belief toward
that specific stimuli ( Bargh, 2002 ; Smarandescu and Shimp, 2015 ;
Verwijmeren et al., 2011 ).
Subliminal advertising means presenting the message below
the conscious threshold in visual or auditory modals ( Piwowarski,
Shankar Singh, & Nermend, 2019 ; Smarandescu and Shimp, 2015 ).
For example, the famous (or notorious) experiment conducted by
James Vicary in 1957 claimed to increase the sales of Coca-Cola
and popcorn through subliminally conveying Piwowarski et al.,
2019 the messages “Eat Popcorn” and “Drink Coca-Cola” while
viewers watched a movie ( Atasoy, 2015 ; Karremans et al., 2006 ;
Smarandescu and Shimp, 2015 ; Sofi et al., 2018 ), which is a case
of advertising marketing using visual language stimulation as a
subconscious message. However, subliminal messages can also be
delivered nonverbally, such as through facial expressions, which
are the most significant stimuli in socialisation ( Andrews et al.,
2011 ). A large body of research supports the claim that the fa-
cial expressions of individuals can be perceived unconsciously
and alter cortical structures ( Huang et al., 2019 ). Facial expres-
sion stimuli are processed faster than any other stimuli in the
environment, and emotional facial expressions are more effective
if perceived unconsciously rather than consciously ( Palermo and
Rhodes, 2007 ). among all facial expressions, a smile is deemed as
1441-3582/© 2020 The Author(s). Published by Elsevier Ltd on behalf of Australian and New Zealand Marketing Academy. This is an open access article under the CC
BY-NC-ND license. ( )
Please cite this article as: L. Hsu and Y.-J. Chen, Neuromarketing, subliminal advertising, and hotel selection: An EEG study, Australasian
Marketing Journal,
2 L. Hsu and Y. -J. Chen / Australasian Marketing Journal xxx (xxxx) xxx
JID: AMJ [m5G; May 22, 2020;0:55 ]
a social sign that plays a critical role in social decision-making be-
haviour as they can add value and shape human decision-making
behaviour ( Heerey and Gilder, 2019 ). Furthermore, Heerey and
Gilder (2019) have discerned the value and function of sincere and
polite smiles.
In the era of Web 2.0, emojis are widely used in computer-
mediated communication (CMC) as a substitute for real facial ex-
pressions in virtual environments ( Hof, 2016 ; Marengo et al., 2017 ;
Walthe r and D’addario, 2001 ; Wang et al., 2014 ), and emojis are
prevalent in marketing campaigns ( Gantiva et al., 2019 ; Leung and
Chan, 2017 ). Thus, it is necessary to determine whether consumers
process emojis as subliminal stimuli, as they do with human fa-
cial expressions; this study then discusses the following research
RQ1: Are there significant effects in using a smiling face emoji as
the subliminal stimuli for influencing consumers’ decision making?
In marketing research, three approaches are usually used
to measure consumers’ responses to stimuli including be-
havioural measures, self-reports, and psychophysiological measures
( Wang and Minor, 2008 ). Shortcomings of the traditional ways of
marketing research have been reported including the validity of
self-report surveys and scepticism about interviews, leading to em-
ploying brain-based approaches in marketing research ( Ariely and
Berns, 2010 ; Hsu, 2017 ; Hsu and Chen, 2019 ; Piwowarski et al.,
2019 ; Wang and Minor, 2008 ). The study aims to understand “the
relation between mental and bodily processes” ( Andreassi, 20 0 0 ,
p.1) by employing neuroscientific methods to increase our current
knowledge on the ability of subliminal messages to influence con-
sumers’ selections. Lakhani (2008) postulated that subliminal mes-
sages can persuade and covertly lead an individual to take certain
actions. In the field of marketing, subliminal messages can affect
consumers’ decisions on their product selection unconsciously. To
attest this inference and understand the mechanism of its subcon-
scious effect on consumers’ decision-making behaviour, this study
employs the method of neuromarketing and hereby proposes the
second research question:
RQ2: How do consumers’ brain activities respond when they are
watching video commercials with and without subliminal stimuli?
By addressing these two questions, this study will improve our
current understanding of how emojis and subliminal advertising
work. The results may provide valuable insights to marketers and
scholars of hospitality marketing since we believe that pertinent
research is scarce in the field of hospitality marketing and man-
agement; therefore, we adopted commercial videos of hotels as
the research context. This study makes the following contributions:
firstly, it provides further confirmation of the social value of emo-
jis that are widely used in the digital age; secondly, it enhances
our current knowledge on how emojis are being used as subliminal
messages to influence consumers’ brand choices ( Karremans et al.,
2006 ); thirdly, it applies the concept of neuromarketing to reveal
how it impacts consumer decisions, by using the theoretical basis
of neuroscience. This study bridges the research gap between the
effectiveness of subliminal advertising and consumers’ purchase
behaviours ( Smarandescu and Shimp, 2015 ). We also believe that
this study is the first of its kind that combines ‘emoji’ visual cues,
subliminal stimuli, consumer behaviour, and brainwave assessment
within the subject of neuromarketing.
The remainder of this paper is structured as follows. In
Section 2 , we briefly review prior studies on this theme. In
Section 3 , the details of the methods used for this study are ex-
plained, including the description of participants, experiment pro-
cedure, apparatus, and subliminal stimuli with the emoji senti-
ment. In Section 4 , we present our empirical results. In Section 5 ,
we discuss the findings of this study. In Section 6 , we discuss con-
clusion, the implications of this research on theory and in practice,
and limitations of our study.
2. Prior studies
2.1. Emojis and consumer behaviour
An Emotion Icon refers to the metacommunicative visuals
( Manganari and Dimara, 2017 ) represented by human facial expres-
sions, which is divided into "typographic or text-based emoticons”
(e.g., :-) to represent a smiling face) and "graphic emoticons” (e.g.,
). This is also known as an Emoji ( Rodrigues et al., 2018 ). An
Emoji is a graphic symbol widely used in computer-mediated com-
munication (CMC). It is an emerging form of pictographic language
used very frequently; 92% of online users use it as an intuitive and
informal way to express their emotions and attitude ( Hof, 2016 ;
Marengo et al., 2017 ), with its usage evidently increasing in social
media conversations ( Eisner et al., 2016 ). It is an important part
of digital-based communication and it becomes a catalyst for ex-
pressing written emotions, as emojis can fill the gaps in natural
language processing, and enable dialogue between users of differ-
ent languages ( Li et al., 2019 ; Komrsková, 2015 ). Emojis can even
solve the problem of insufficient nonverbal cues (such as facial ex-
pressions or body language) in CMC, and have become an alterna-
tive to nonverbal messages in CMC ( Walt her and D’addario, 2001 ;
Wang et al., 2014 ). The use of emojis has become very common in
CMC ( Wang et al., 2014 ), which is also why graphic emojis will be
used as the stimulus in this study.
In recent years, an increasing number of marketing adver-
tisements have used emojis in promotions ( Das et al., 2019 ).
Das et al. (2019) have pointed out that including emoticons in
advertisements can generate a positive impact on target con-
sumers, and increase their willingness to purchase hedonic prod-
ucts. Manganari and Dimara (2017) researched the impact of emo-
jis on online consumer reviews (OCRs) of online travel agencies
(OTAs), and revealed that emojis in negative reviews would en-
hance the credibility and practicality of these reviews, and re-
duce the customer’s willingness to book. Li et al. (2019) further
reported that when service providers use emojis for communica-
tion, customers tend to appreciate the service more, and the level
of their satisfaction increased. Riordan (2017) argues that emojis
seem to be able to change the intensity of the customers’ emo-
tions towards the product or service as Wang et al. (2014) elab-
orated; using positive emojis would enhance cognitive communi-
cation that would reinforce its positive image. However, Walther
and D’Addario (2001), and Wang et al. (2014) have warned that
the use of emojis may also backfire. They suggested that its use
should involve the process of selection and embedding, as it is
easy for consumers to think that the emojis are equal to invol-
untary uptake, and they can also perceive emojis to be symbols of
marketers’ intentional control. Hof (2016) and Lebovits (2015) have
also mentioned that in some situations, the use of emojis may be
considered informal and unprofessional. In summary, emojis seem
to be double-edged swords; they may increase the positive impact
and affect business behaviour positively, but may also cause con-
sumers to have negative psychological concerns. In this study, we
use emojis as subliminal stimuli in an attempt to reduce their pos-
sible negative effects.
In addition, although emojis are widely used in CMC, they
have become a common stimulus in today’s scientific research
( Riordan, 2017 ; Rodrigues et al., 2018 ). Until now, there has been
little academic research on emojis ( Li et al., 2019 ), especially in
the context of consumers’ responses to emojis in advertisements
( Das et al., 2019 ) to better understand any effects and underly-
ing reasons of any impact ( Das et al., 2019 ; Wang et al., 2014 ).
The information extracted through neuromarketing can better re-
flect the potential preferences of consumers, reduce the biases that
tend to occur in traditional surveys and studies ( Meyerding and
Mehlhose, 2018 ), and capture subtle differences in consumer be-
Please cite this article as: L. Hsu and Y.-J. Chen, Neuromarketing, subliminal advertising, and hotel selection: An EEG study, Australasian
Marketing Journal,
L. Hsu and Y. -J. Chen / Australasian Marketing Journal xxx (xxxx) xxx 3
JID: AMJ [m5G; May 22, 2020;0:55 ]
haviour ( Stasi et al., 2018 ). Hence, this study will explore the im-
pact of emojis as subliminal stimuli on consumers’ brain activity
through experimental research design. This study may put forward
new concepts, actions, and significance to the research and role of
emojis in marketing.
There is extensive research on how human cognition can be
affected by subliminal stimuli which enable one to activate ex-
isting knowledge or beliefs toward that specific stimuli. Influence
of such stimuli would be stronger while they map one’s current
needs and/or desires ( Bargh, 2002 ; Smarandescu & Shimp, 2015 ;
Verwijmeren et al., 2011 ), but the subliminal stimuli of these stud-
ies were semantic and none of them discussed the effects of
emoji facial expressions. Quite a large body of research have al-
ready stated that others’ facial expressions can be perceived un-
consciously and alter cortical structures ( Huang et al., 2019 ). The
literature mentioned above can help postulate that both emojis
and subconscious stimulation may have an impact on consumer
behaviour. Nevertheless, how the facial expressions (particularly
when the stimuli are not real human faces, but emojis) as sublim-
inal stimuli influence customers’ hotel selection is still unknown.
Hence, the first research hypothesis is proposed as follows:
RH1: Using an emoji as a subliminal stimulus can have a signifi-
cant effect on influencing consumers’ selection
2.2. Subliminal stimuli and brainwaves
To understand human cognition, it is advised to begin with
the neural mechanism of one’s emotion ( Hsu, 2017 ). The affec-
tive primacy hypothesis suggests affective information has priority
to be accessed than conscious cognition by the message receivers
( Huang et al., 2019 ; Lai et al., 2012 ). Therefore, emotion plays such
a pivotal role in consumers’ purchasing decisions ( Schaefer et al.,
2006 ) which triggered activity in many brain regions ( Hsieh, 2018 ).
One’s purchase behaviour is found to correlate with activating the
reward system (e.g., nucleus accumbens, NAcc) in the mesolimbic
region of brain triggered by preferred brands ( Deppe et al., 2005 ;
Erk et al., 2002 ; Knutson et al., 2007 ), and the behaviour is en-
hanced by the dorsolateral prefrontal cortex (DLPFC) through mak-
ing the most equitable decisions to optimize personal gain ( Knoch
and Fehr, 2007 ).
Moreover, one’s responses to affective stimuli are different with
or without faces being presented ( Huang et al., 2019 ). A neu-
rological explanation can be that the orbitofrontal cortex (OBFC)
is part of the valuation circuit which predicts one’s preferences.
While making purchasing decisions, it has been argued that our
ventromedial prefrontal cortex (VMPC) is responsible for inter-
preting the messages conveyed in the commercial advertisements
( Koenigs and Tranel, 2007 ). Further studies ( Hsu and Chen, 2019 ;
Kawasaki and Yamaguchi, 2012 ; Khushaba et al., 2013 ) have ad-
dressed consumers’ cognitive and affective responses to mar-
keting stimuli through the lens of brainwaves ( Aprilianty and
Purwanegara, 2016 ; Khushaba et al., 2013 ; Malik et al., 2018 ;
Mostafa, 2012 ). This present study aims to extend our current un-
derstanding on subliminal stimuli towards how emoji facial ex-
pressions affect consumers’ neural activities in the brain, through
examining changes in spectral bands of five brainwaves in the pre-
frontal lobe.
Even though human beings have conscious minds that are re-
sponsible for making all kinds of decisions including purchasing
decisions, the subconscious mind has been programmed to receive
subliminal stimuli as hints that affect choices, preferences, and
behaviour ( Karam et al., 2017 ). Subliminal stimuli may also by-
pass the conscious mind, which may have developed resistance to
change. Understanding consumers’ minds is a concept that is gain-
ing traction with the advancement of neuroscientific techniques to
explore target audiences’ responses to advertising, from both emo-
tional and cognitive perspectives ( Sung et al., 2019 ). With the pos-
sibility of affordable neuroimaging technologies such as electroen-
cephalogram (EEG) and magnetic resonance imaging (MRI), mar-
keting researchers are able to detect consumers’ brain activities
(i.e., neural mechanism) while they are receiving marketing stim-
uli ( Solnais et al., 2013 ). Lin et al. (2018) suggest that the appli-
cation of neuroscientific approaches, particularly EEG, may be able
to improve marketing research by revealing the hidden informa-
tion embedded in consumers’ brains. EEG can be used to mea-
sure a person’s brainwave oscillations at the surface of the scalp
through electrodes, which enables researchers to understand how
neurons in the brain communicate with each other. Previous adver-
tising research concerning brainwave measurement had commonly
identified the following frequency bands: delta ( < 4 Hz), theta (4–
8 Hz), alpha (8–12 Hz), beta (10–30 Hz), and gamma (30–40 Hz)
( Khushaba et al., 2013 ; Youn g, 2002 ).
Regarding the relationship between an individual’s brainwaves
and responses to external stimuli, prior empirical studies have
summarized that different brainwave bands indicate neural activ-
ities that can be influenced by external stimuli (Piwowarski, Singh,
& Nermend, 2019). Morin (2011) revealed that alpha brainwaves
can be generated in the left prefrontal cortex while an individ-
ual is experiencing positive emotion, which could increase the
chance to purchase. Gamma bands activated in the temporal lobe
can be attributed to high levels of visual stimuli ( Tan ak a, 1996 ),
and the theta bands in the left frontal lobe increase along with
an individual’s memory storage capacity ( Davidson, 2004 ). More-
over, theta brainwaves in the frontal lobe can be activated by the
pleasure gained when watching certain TV commercials ( Vecchiato
et al., 2011 ). Additionally, a correlation between beta brainwaves
and product preferences has also been found by Boksem and
Smidts (2015) . An increasing number of empirical studies have
used EEG signals to identify emotions ( Nie et al., 2011 ) or affective
responses to marketing messaged ( Khushaba et al., 2013 ). However,
Sung et al. (2019) argue that psychophysiological measures on con-
sumers’ brain activities are still limited. Deeper insights into con-
sumers’ minds can be analysed through the adoption of neuromar-
keting techniques to better understand the effectiveness of sublim-
inal stimuli toward consumers’ purchase decisions. Therefore, this
study proposes the following research hypothesis:
RH2: Consumers’ brainwave oscillations will be significantly af-
fected when they are watching video commercials of hotels with emo-
jis embedded as subliminal stimuli.
3. Methods
3.1. Participants
As suggested by Bastiaansen et al. (2018) , and Hsu and
Chen (2019) , we recruited participants based on the screening cri-
teria for subjects in brain activity-related studies and the needs of
this study. All participants are at least 20 years old; right-handed;
have normal physical and mental conditions; have no significant
head trauma, brain tumours, or other neurological diseases; and no
history of major organ diseases. Additionally, participants have not
used therapeutic drugs in the past three months and have not par-
taken in any brainwave-related research in the past. Furthermore,
they were expected to have had enough sleep and avoid drinking
alcohol or caffeinated drinks (coffee, tea, chocolate, cola, etc.) be-
fore the experiment.
Consistent with the views of Cha et al. (2019) ,
Wongtadat (2019) states that due to financial, technical, and
experimental design constraints, the subjects must meet specific
requirements in order to avoid factors other than the expected
stimuli that may cause a bias in neural responses. Studies using
neuroscience imaging techniques tend to use fewer research
Please cite this article as: L. Hsu and Y.-J. Chen, Neuromarketing, subliminal advertising, and hotel selection: An EEG study, Australasian
Marketing Journal,
4 L. Hsu and Y. -J. Chen / Australasian Marketing Journal xxx (xxxx) xxx
JID: AMJ [m5G; May 22, 2020;0:55 ]
participants and also limit the age of participants; that is, the
number of participants should be between 12 and 44, and par-
ticipants’ age should be between 18 and 43 years old. Several
neuroscience-related studies, such as those by Cuesta et al. (2018) ,
Meyerding and Mehlhose (2018) , and Pileliene and Grigaliunaite
(2017) have followed this suggestion, as has this current study.
There were 20 participants that passed the first round of selec-
tion and subsequently joined this study. Every participant went
through both control and experimental conditions. In total, they
were asked to view 16 videos with and without the subliminal
stimuli. After the data were collected, some data contained too
much noise and were thus removed from dataset. Consequently,
data of 16 participants ( n = 16, female = 9, male = 7, aged 24–32)
were valid for subsequent statistical analyses.
3.2. Experiment procedure & stimuli
Before the commencement of the experiment, each participant’s
baseline brainwaves were measured and recorded by NeuroSky. At
this phase, participants were asked to close their eyes and re-
main in a calm state. Afterwards, all participants went through two
rounds/conditions (with and without subliminal stimuli) which
contained the same eight blocks of 20-second stimuli which in-
cluded 5 s of relaxation followed by eight different hotel pro-
motion videos. Additionally, the video clips of experimental con-
ditions also included the subliminal message of a smiling face
emoji. Based on past research on stimuli, compared to negative
stimuli, positive stimuli are more effective and rapid in promot-
ing behavioural responses ( Ma and Han, 2010 ). By a similar to-
ken, positive expressions are more influential to an individual’s
decision-making process ( Leppänen et al., 2003 ). Furthermore, in
order to avoid memory issues, the interval between the two rounds
of watching the movie was at least 4 h, and the sequence of the
movie was randomly arranged to reduce the possibility of the sub-
ject being affected by memory. The schematic diagram of the pro-
cess of watching the hotel room video is as follows:
The stimuli were eight videos of major five-star hotel chains
(e.g., Marriott International, Sheraton Hotels and Resorts, W Hotels)
showing their rooms (selected from YouTube). The videos of exper-
imental condition were presented with an emoji of a smiling face
showing for 1 millisecond at 3-second intervals. Participants were
asked to rate their selected hotels on a scale of 1–10 (10 being the
highest score, and 1 being the lowest). Afterwards, we ranked the
participants’ selections based on their ratings of hotel room videos
(the lowest score was ranked 1 and the highest score was ranked
8). This standardized score was to correct the individual differences
in the score criteria between the participants.
3.3. Apparatus
The electroencephalograph (EEG), a non-invasive brainwave ap-
paratus, is a nonconventional technique used in marketing research
( Wang et al., 2016 ) that can appropriately measure consumers’
neural activities in the brain, for example, cognitive and affective
states ( Wang and Minor, 2008 ), to further predict their purchase
decisions ( Aprilianty and Purwanegara, 2016 ). It is considered to
be the most portable and affordable method of detecting physio-
logical signals ( Nguyen et al., 2017 ), which is suitable for measur-
ing the relationship between physiology and psychology ( Xu and
Zhong, 2018 ). Moreover, it will be more effective and accurate than
traditional self-reporting survey methodology ( Sun, 2014 ).
According to Peters et al. (2009) , the forehead and frontal cor-
tex are at the centre of high cognitive signals and consciousnesses,
allowing researchers to effectively capture the required brainwaves,
and measure high-resolution EEG signals. This study was based
on the international 10–20 brain wave electrode configuration
method (international 10–20 system) that has been recommended
by the International Federation of Societies for Electroencephalog-
raphy and Clinical Neurophysiology (IFSECN). The Fp1 point at
the frontal lobe and the A1 point at the left ear lobe were used
as reference points for the collection and elimination of physi-
ological signals of brain waves ( Ismail et al., 2016 ). The brain-
wave data were collected using a non-invasive brainwave mea-
surement instrument, MindWave Mobile, developed by NeuroSky
(San Jose, CA, USA, ). This uses a single-
electrode signal processor to capture signals generated by neural
activity ( Rebolledo-Mendez et al., 2009 ), with real-time electroen-
cephalogram data being recorded with an electrode on the fore-
head ( Chen and Wu, 2015 ). The EEG, with a frequency of 1 Hz, is
used to collect brainwave bands in seconds and convert raw data
to standardized t scores.
Despite the fact that NeuroSky EEG has only a single chan-
nel, it has been prevalently used in many neuromarketing studies
( Goode and Iwasa-Madge, 2019 ; Hsu and Chen, 2019 ) for its relia-
bility has been confirmed by scholars such as Rogers et al. (2016) ,
Sezer et al. (2015) , Chen and Huang (2013) , and Rebolledo-
Mendez et al. (2009) . It has also been proven to have sufficient
accuracy in collecting brainwave oscillations, and the participants
may not feel as awkward when wearing it ( Sun, 2014 ). The stress
or distraction that may occur during the experiment is also mini-
mized; therefore, using NeuroSky EEG to analyse the participant’s
brainwave activity should be preferred ( Rebolledo-Mendez et al.,
2009 ).
3.4. Data analysis
A chi-square analysis was carried out on the collected data to
examine the first research hypothesis. A paired t -test was admin-
istered to investigate whether the research hypothesis would be
accepted at 5% level of significance. Effect sizes of significant t-
tests would also be reported. In addition to frequentist statistics
(p-value and null hypothesis), the alternative Bayesian theorem
was also adopted to extract more insights ( Van De Schoot et al.,
2017 ) and overcome problems of traditional frequentist statistics,
such as failing to differentiate the null hypothesis and the alternate
hypothesis when p > .05 ( Beard et al., 2016 ). In Bayesian statis-
tics, the Bayes Factor (BF) is a robust way to quantify evidence of
null hypothesis over alternate hypothesis ( Han et al., 2018 ). More-
over, the deficit of a small sample size can also be solved with
this form of statistics ( Biel and Friedrich, 2018 ). Based on the rea-
sons discussed above, we used BFs, which were calculated from the
ratio of Bayesian information criterion (BIC) as an approximation
to gauge the posterior probability ( Kass and Raftery, 1995 ), and to
measure the marginal likelihood in favour of a given hypothesis
among two hypotheses (the alternate hypothesis and null hypoth-
esis) ( Lavine and Schervish, 1999 ). All the statistical analyses were
performed using R, ( Figs. 1 and 2 ).
4. Results
For the first research question, the chi-square test of homogene-
ity was performed to examine whether the participants’ ranking of
their selection of hotels would remain the same with or without
subconscious stimulation. Results of the analyses are shown in the
Table 1 and Fig. 3 .
According to Table 1 and Fig. 3 , compared with the first round
of unstimulated hotel room videos, the rankings of hotels with
smiling face emojis as subliminal stimuli embedded in the second-
round video were significantly moved up (63.6%, p < 0.01); mean-
while, the rankings of video without subliminal stimuli dropped
significantly (63.2%, p < 0.01). Results of the Chi-Square test re-
ported that the effect of subliminal messages toward participants’
Please cite this article as: L. Hsu and Y.-J. Chen, Neuromarketing, subliminal advertising, and hotel selection: An EEG study, Australasian
Marketing Journal,
L. Hsu and Y. -J. Chen / Australasian Marketing Journal xxx (xxxx) xxx 5
JID: AMJ [m5G; May 22, 2020;0:55 ]
Fig. 1. The Round 1 schematic diagram of the process of watching the hotel room video.
Fig. 2. The Round 2 schematic diagram of the process of watching the hotel room video.
Fig. 3. Results of Chi-square test of homogeneity on the participants’ selection of
hotels with and without subliminal stimuli.
hotel selection was significant (Pearson χ2 = 10. 21, p < .01); in
other words, consumers’ selections of hotel were significantly in-
fluenced by the subliminal message of emoji smiling face. We ran a
Bayesian test of association, and the resulting Bayes factor of 18.15
to 1 in favour of the alternate hypothesis indicates that there is
moderately strong evidence for the effect of subliminal stimuli.
The second research question was addressed with the adminis-
tration of the sample paired t- test to examine changes in the par-
ticipants’ brain activity when watching hotel room videos; com-
paring the first and second round of videos. Five brainwave data
( δ, θ, α, β, and γ) were calculated as standardised percentages to
be contrasted ( Hsu and Chen, 2019 ). The analysis results are shown
in Table 2 and Fig. 4:
Tabl e 1
Chi-square test of homogeneity and Bayes factor on the participants’ selection of
hotels with and without subliminal stimuli.
Declined No Change Increased Tota l
No Subliminal
Number 24
a 16
a 24
b 64
Percentage 63.2% 66.7% 36.4%
Number 14
a 8
a 42
b 64
Percentage 36.8% 33.3% 63.6%
Tota l Number 38 24 66 128
Percentage 100% 100% 100%
= 10.21
∗∗, cc = 0.27
∗∗, p = .006 Bayes Factor = 18.15
∗∗p < .01.
Note: According to Jeffreys (1961) Bayes factor cut-offs, 10 < BF < 30 indicates strong
evidence for the alternate hypothesis .
Fig. 4. Results for participants’ brain activities in two rounds of experiments.
Based on the data reported in Table 2 and Fig. 4 , significant
differences were found between the participants’ two (theta and
beta) out of five bands of brainwaves while they were viewing ho-
tel videos with and without a subliminal message. Specifically, par-
ticipants’ theta waves significantly increased when watching the
video with subliminal stimuli ( t = 5.65, p < .001); conversely,
their beta brainwaves significantly decreased ( t = 2.93, p < .01).
Effect sizes (Cohen’s d) of theta and beta were 0.63 and 0.56, re-
spectively, which indicate that medium effects did exist between
two rounds of experiments ( Sawilowsky, 2009 ) . The resulting t-
tests were further triangulated by Bayesian statistics, as BF pro-
vides evidence of about 35,801.12:1 and 6.58:1 in favour of the al-
ternate hypothesis. Therefore, the null hypothesis can be rejected
with some confidence.
5. Discussions
This experimental study proposed the examination of two re-
search hypotheses. The first hypothesis is found to be significant
through chi-square analysis, and the Bayes factor also indicates
strong evidence. Results show that subliminal messages did have
significant influence on consumers’ selection of hotels, even if the
message is simply a smiling face emoji embedded in the video.
Regarding the impact of emojis, this study echoes the findings of
Das et al. (2019) and Riordan (2017) , confirming that positive emo-
jis can indeed exert a positive effect on consumer behaviour. The
effect of incorporating subliminal messages in advertising supports
the findings of Brown (2008) and Hemp (2006) , confirming that
subliminal messages in advertising are likely to have an influence
on consumers. The findings of this study contrast with those of
Kotler et al. (2008) who claimed that “numerous studies by psy-
chologists and consumer researchers have found little or no link
between subliminal messages and consumer behaviour” (p. 259).
Please cite this article as: L. Hsu and Y.-J. Chen, Neuromarketing, subliminal advertising, and hotel selection: An EEG study, Australasian
Marketing Journal,
6 L. Hsu and Y. -J. Chen / Australasian Marketing Journal xxx (xxxx) xxx
JID: AMJ [m5G; May 22, 2020;0:55 ]
Tabl e 2
Results of t -test for participants’ brain activities in two rounds of experiments (with and with-
out subliminal stimuli embedded).
Brainwave Round M Mean difference SD t Cohen’s d Bayes Factor
Delta Round 1 17.29 .66 9.19 .57 .16
Round 2 16.63
Theta Round 1 8.54 4.72 6.68 5.65
.63 35,801.12
Round 2 13.26
Alpha Round 1 23.33 1.74 12.02 1.16 .26
Round 2 25.07
Beta Round 1 26.25 5.04 13.75 2.93
.56 6.58
Round 2 21.21
Gamma Round 1 24.58 2.01 9.58 1.68 .52
Round 2 22.57
∗∗p < .01,
∗∗∗p < .001.
Note: According to Jeffreys (1961) Bayes factor cut-offs, BF > 100 indicates extremely strong evi-
dence for the alternate hypothesis; 3 < BF < 10 indicates moderate evidence for the alternate hy-
pothesis; BF < 1 indicates evidence for the null hy Fig. 4 . pothesis .
As a new topic in marketing research, this study integrates two
kinds of marketing stimuli that have always been presented sepa-
rately in the past, using emojis as the form of subconscious stimuli.
These results demonstrate that emojis delivered as subliminal mes-
sages can possibly affect consumers’ purchase decisions. This study
suggests that hotels are encouraged to embed more smiling faces
in their promotion videos as even subliminal smiling face emojis
can have positive results in consumers’ selection of hotels.
As with the second research hypothesis, brainwave analysis fur-
ther bridges the research gap between consumers’ brain activi-
ties when viewing videos with and without subliminal stimuli.
This study has provided possible insights into the underlying rea-
sons behind the effects of emojis on the consumption of adver-
tisements, proposed by Das et al. (2019) and Wang et al. (2014) .
The results of the brainwave analyses indicated that the theta band
would be significantly different while viewing hotel videos with
subliminal messages. The results of this study echoes the previ-
ous study by Vecchiato et al. (2011) , which asserts that an individ-
ual’s theta band would be increased due to the pleasure of view-
ing commercials on TV. Furthermore, Davidson (2004) proposes
that when an individual’s information storage is activated, the
theta brainwaves would also increase. Videos with smiling faces as
subliminal stimuli have positive effects on participants’ memory.
However, contrary to the findings of Boksem and Smidts (2015) ,
the participants’ beta brainwaves were reduced in this study due
to the subconscious stimulation through emojis. This study at-
tempts to explain this phenomenon from another point of view:
the lower beta brainwaves when the participants watched videos
with an embedded smiling face emoji as subliminal stimuli,
meant that their level of anxiety decreased ( Ismail et al., 2016 ;
Satapathy et al., 2019 ). This empirical research can encourage hotel
operators to consider including subliminal messages in their pro-
motional videos, which may be able to generate positive consumer
emotions and lower consumers’ levels of anxiety, stress or scepti-
cism when watching marketing advertisements.
However, Choong et al. (2019) asserted that subliminal advertis-
ing may not lead to purchase behaviour, and the reasons may be
attributed to different industries as smiling faces are important for
hotels to establish and maintain good relations with their guests
( Baker and Kim, 2018 ; Islam and Kirillova, 2020 ). This view is also
concurrent with the results of this study, since the alpha brain-
waves did not increase the likelihood of consumer purchasing op-
portunities, as Morin (2011) had suggested. The subconscious mar-
keting stimulation method used in this research can help explain
the issues raised by Hof (2016) and Lebovits (2015) , who asserted
that emojis appear too intentional or unprofessional to consumers,
and thus would negatively affect consumers’ purchasing decisions.
This study found that using emojis as subliminal stimuli would re-
duce consumer psychological pressure, and this finding has impli-
cations for marketing research and practice.
Additionally, this study can serve as a link to existing mar-
keting theories. As pointed out by Shrum et al. (2012) , the main
goal of any marketing campaign is to influence consumers’ pur-
chase decisions. In light of extensive marketing theories related to
consumer behaviour, the dual-process theory (the central route of
deliberate processes versus the peripheral route of intuitive pro-
cesses) was proposed to shed light on consumers’ mental process-
ing while reaching their purchase decisions ( Dhar and Gorlin, 2013 ;
Evans, 1984 ; Sanjari et al., 2017 ). Our findings extend the current
understanding of the dual-process theory by adding to the study
of Wang et al. (2013) . Our study discovered that participants’ theta
and beta brainwave activities in the prefrontal cortex were signif-
icantly different with and without the stimuli, because their me-
dial prefrontal cortex (MPFC) may be subconsciously activated by
emojis (subliminal stimuli). As such, consumers’ behaviour appears
to be influenced by including subliminal stimuli in advertisements.
Beta brainwaves are triggered in the prefrontal cortex, and the in-
crease of beta bands indicates when one’s level of attention rises.
In our study, participants’ beta brainwaves were lower when they
viewed videos with subliminal stimuli and this result echoed the
dual-process theory by showing that consumers perceived sublim-
inal stimuli through the peripheral route of intuitive processes.
Participants processing subliminal stimuli through the peripheral
route can also be confirmed by their higher theta bands which in-
dicate that an individual is more susceptible to subconscious in-
formation. This finding also implies that subliminal stimuli may be
able to enhance participants’ memory, which affects their purchase
6. Conclusion, implications and limitations
6.1. Conclusion and implications
Consumers’ limited attention spans are one of the scarce re-
sources for companies in the information age. How to compete
with other businesses in this situation is a major marketing chal-
lenge ( Wang et al., 2016 ). We are currently at the first decade of
neuromarketing or consumer neuroscience research ( Smidts et al.,
2014 ). Neuromarketing uses consumers’ psychophysiological re-
sponses as variables in the decision-making process, which are
being used successfully in the field of marketing ( Cuesta et al.,
2018 ). It can also provide hidden information that is not available
in other marketing-related studies, which increases the value of
neuromarketing research ( Meyerding and Mehlhose, 2018 ). How-
ever, conducting neuromarketing research itself is expensive and
time-consuming; therefore, this kind of research is still very lim-
Please cite this article as: L. Hsu and Y.-J. Chen, Neuromarketing, subliminal advertising, and hotel selection: An EEG study, Australasian
Marketing Journal,
L. Hsu and Y. -J. Chen / Australasian Marketing Journal xxx (xxxx) xxx 7
JID: AMJ [m5G; May 22, 2020;0:55 ]
ited in practice ( Cha et al., 2019 ). This study adopted an experi-
mental design with 16 participants who went through two con-
ditions (control and experimental condition). Each condition had
eight blocks of hotel video presentations and a break. The only
difference between the two conditions was that some videos of
experimental conditions included a smiling face emoji as a sub-
liminal message. The results of this study showed that subliminal
messages can have significant influence on consumers’ hotel selec-
tions, which was demonstrated by their altered neural activities.
The theoretical implications of this study are two-fold: first, this
study provided empirical evidence from the perspective of neu-
romarketing to show that emoji, as subliminal stimuli, can influ-
ence consumers’ purchase decisions. Drawing on the results of this
study, participants’ theta and beta bands were significantly differ-
ent in the two rounds of experiments; embedding emoji as sub-
liminal stimuli in commercials may trigger consumers’ neural ac-
tivities. Nevertheless, the data was based on evidence related to
participants’ prefrontal cortexes, and while this region is consid-
ered the decision-making centre of the brain, neural activities in
other regions should also be explored. The second theoretical im-
plication of this relates to the statistical technique. To the best of
our knowledge, most previous studies used the concept of frequen-
tist statistics to address research questions and Bayesian statistics
have been posited to be able to predict the way people make de-
cisions. This study triangulated the frequentist statistics with the
Bayesian theorem to provide more robust results. Future marketing
studies are encouraged to include the Bayesian theorem to increase
the generalisability of their results.
This study provides important managerial implications for ho-
tels, which can use the findings of this study to design more ef-
fective websites or video commercials to boost prospective guests’
intentions to make reservations. As the late Steve Jobs stated, Peo-
ple don’t know what they want until you show it to them. That’s why
I never rely on market research. Our task is to read things that are
not yet on the page .” Specifically, hotels or other e-commerce com-
panies can embed emojis or any other genres of smiling faces as
subliminal stimuli in their official website or video commercials
because the target audiences’ emotions may be affected positively
by such stimuli, which could influence purchase decisions. Such
subconscious stimuli will not be noticed by consumers but can ef-
fectively reduce or even avoid the negative effects of advertising
when the stimulus is clearly indicated. The use of subliminal stim-
uli can also reduce the psychological pressure, anxiety, and suspi-
cion of consumers, and also increase their awareness on products
preferences. With slight tweaks to websites or commercial videos,
the consequences may be advantageous to reservation or sales de-
partments. In addition, in terms of technical cost considerations,
such a design will not bring a substantial increase in the expen-
diture of promotional planning as consumers will not perceive the
existence of such stimulus, so it has significant meaning and prac-
ticality for marketing management.
6.2. Limitations and future research directions
The major limitation of this study is the sample size. The is-
sue of small sample size has been discussed in neuromarketing
research, that is, how accurately can a small number of partici-
pants represent the entire population ( Spence, 2019 ). Another limi-
tation of this study is based on gender differences within the sam-
ple. Previous studies have pointed out that females are more ex-
pressive ( Komrsková, 2015 ), which infers that females may be af-
fected more by emojis than males. However, due to the low sample
size of the experimental study design, this topic has not been dis-
cussed in depth. Future studies can increase the sample size and
also diversify their research’s demographic group to enhance its
generalisability. Furthermore, as NeuroSky EEG has only one elec-
trode, this may lower the validity of brainwave data; sophisticated
equipment with more electrodes can be used to avoid this prob-
lem. We advise studies future research using EEG on event-related
potential in the field of hospitality and tourism to gain more in-
sights into the minds of consumers when pushing the buy button
in their brains, as 80% of buying behaviours are driven on an un-
conscious or subconscious level ( Cuesta et al., 2020 ; Hafez, 2019 ;
Morin, 2011 ). However, another fact remains that the effects en-
gendered by subliminal stimuli can only last for a very short pe-
riod of time ( Ruch et al., 2016 ), and the long-term effects of these
techniques are still in need of more empirical evidence. It has been
suggested that EEG can be used in conjunction with other tech-
niques such as facial electromyography (fMEG) ( Li et al., 2015 ) or
eye-tracking ( Khushaba et al., 2012 ). Finally, because this study was
conducted in an experimental setting, the results of this study can-
not directly indicate that consumers will book hotel rooms only
because emojis have been embedded in advertisements as sublim-
inal stimuli. Other factors, such as budget issues and hotel loca-
tions, would also influence consumers’ purchase decisions regard-
ing hotel bookings. As Weeks and Weber (2019) put forth, sublim-
inal advertising may not lead to actual purchase behaviour; this
must be verified by further relevant research in the future.
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
Declaration of Competing Interest
The authors have NO affiliations with or involvement in any or-
ganization or entity with any financial interest (such as honoraria;
educational grants; participation in speakers’ bureaus; member-
ship, employment, consultancies, stock ownership, or other equity
interest; and expert testimony or patent-licensing arrangements),
or non-financial interest (such as personal or professional relation-
ships, affiliations, knowledge or beliefs) in the subject matter or
materials discussed in this manuscript.
Andreassi, J.L. , 20 0 0. Psychophysiology: Human Behavior and Physiological Re-
sponse. Lawrence Erlbaum Associates, Mahwah, NJ .
Andrews, V. , Lipp, O.V. , Mallan, K.M. , König, S. , 2011. No evidence for subliminal
affective priming with emotional facial expression primes. Motiv. Emot. 35 (1),
33–43 .
Aprilianty, F. , Purwanegara, M.S. , 2016. Effects of colour towards underwear choice
based on electroencephalography (EEG). Aust. Market. J. 24 (4), 331–336 .
Ariely, D. , Berns, G.S. , 2010. Neuromarketing: the hope and hype of neuroimaging in
business. Nat. Rev. Neurosci. 11, 284–292 .
Atasoy, S.N., 2015. Visual perception management
in advertisements. Glob. J. Hum.
Soc. Sci. 3, 397–404. Available from: .
Baker, M.A. , Kim, K. , 2018. The role of language, appearance, and smile on percep-
tions of authenticity versus rapport. Int. J. Hosp. Manag. 74 , 171–179 .
Bargh, J.A. , 2002. Losing consciousness: automatic influences on consumer judg-
ment, behavior, and motivation. J. Consum. Res. 29 (2), 280–285 .
Bastiaansen, M. , Straatman, S. , Driessen, E. , Mitas, O. , Stekelenburg, J. , Wang, L. ,
2018. My destination in your brain: a novel neuromarketing approach for evalu-
ating the effectiveness of destination marketing. J. Destination
Market. Manage.
7, 76–88 .
Beard, E. , Dienes, Z. , Muirhead, C. , We st, R. , 2016. Using Bayes factors for testing hy-
potheses about intervention effectiveness in addictions research. Addiction 111
(12), 2230–2247 .
Biel, A.L., Friedrich, E.V., 2018. Why you should report Bayes factors <