ArticlePDF Available

Exploring the Untapped Potential of Neuromarketing in Online Learning: Implications and Challenges for the Higher Education Sector in Europe

Authors:
  • Institute for Neuromarketing & IP
  • Oxford Business College
  • OXFORD BUSINESS COLLEGE. GEORGE STREET. OXFORD. OXFORDSHIRE. OX1 2BQ. www.oxfordbusinesscollege.ac.uk

Abstract

This research investigates the impact of applying neuromarketing techniques to three practical examples of higher education (HE) branding: an official college website page, an official college Facebook page, and recorded online video lectures used for teaching at HE institutions. The study was conducted in three different HE institutions with a representative sample of 720 participants, with n = 529 used for testing the CARE college website, n = 59 for testing the HAZEF Facebook page, and n = 132 for testing the emotional response of students studying online. To assess the participants’ emotional responses, this study utilized automated facial coding through a webcam (15 Hz) and considered mood intensities. Additionally, a sentiment analysis was employed to verify the survey results and determine any discrepancies in the cognitive response. By analyzing gaze activity, movement patterns, and emotional responses, valuable insights were gained into students’ behaviors and preferences. This study recommends incorporating neuromarketing research into HE branding and online teaching to enhance students’ learning experiences. Overall, this study contributes to the understanding of human expectations and behaviors in response to online teaching and provides valuable insights for HE institutions in Europe.
Citation: Šola, H.M.; Qureshi, F.H.;
Khawaja, S. Exploring the Untapped
Potential of Neuromarketing in
Online Learning: Implications and
Challenges for the Higher Education
Sector in Europe. Behav. Sci. 2024,14,
80. https://doi.org/10.3390/
bs14020080
Academic Editor: Gemma
Anne Calvert
Received: 20 November 2023
Revised: 11 January 2024
Accepted: 18 January 2024
Published: 23 January 2024
Copyright: © 2024 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
behavioral
sciences
Article
Exploring the Untapped Potential of Neuromarketing in Online
Learning: Implications and Challenges for the Higher Education
Sector in Europe
Hedda Martina Šola 1,2,* , Fayyaz Hussain Qureshi 3and Sarwar Khawaja 3
1Oxford Centre for Applied Research and Entrepreneurship (OxCARE), Oxford Business College, 65 George
Street, Oxford OX1 2BQ, UK
2Institute for Neuromarketing, Jurja Ves III Spur no 4, 10000 Zagreb, Croatia
3Oxford Business College, 65 George Street, Oxford OX1 2BQ, UK;
fayyaz.qureshi@oxfordbusinesscollege.ac.uk (F.H.Q.); advice@oxfordbusinesscollege.ac.uk (S.K.)
*Correspondence: hedda.martina@oxfordbusinesscollege.ac.uk
Abstract: This research investigates the impact of applying neuromarketing techniques to three prac-
tical examples of higher education (HE) branding: an official college website page, an official college
Facebook page, and recorded online video lectures used for teaching at HE institutions. The study
was conducted in three different HE institutions with a representative sample of
720 participants
,
with n= 529 used for testing the CARE college website, n= 59 for testing the HAZEF Facebook
page, and n= 132 for testing the emotional response of students studying online. To assess the
participants’ emotional responses, this study utilized automated facial coding through a webcam
(
15 Hz
) and considered mood intensities. Additionally, a sentiment analysis was employed to verify
the survey results and determine any discrepancies in the cognitive response. By analyzing gaze
activity, movement patterns, and emotional responses, valuable insights were gained into students’
behaviors and preferences. This study recommends incorporating neuromarketing research into
HE branding and online teaching to enhance students’ learning experiences. Overall, this study
contributes to the understanding of human expectations and behaviors in response to online teaching
and provides valuable insights for HE institutions in Europe.
Keywords: neuromarketing in HE; cognitive control on Facebook design; eye tracking on website;
human–robot interaction; online learning
1. Introduction
The field of neuromarketing, which combines neuroscience and marketing to under-
stand consumer behavior and decision-making processes, has garnered significant interest
in recent years. One area of application that has emerged is the use of neuromarketing
techniques for online learning in higher education. This approach aims to enhance the
effectiveness of online learning platforms by leveraging insights from brain and cognitive
sciences. The design of learning materials is a potential area in which neuromarketing can
significantly influence online learning. Traditional online learning materials often require
greater engagement and interactivity to capture students’ attention and promote effective
learning. Neuromarketing can provide valuable insights into the design of more engaging
and effective online learning materials. By measuring neural and physiological signals such
as facial coding and eye movement, neuromarketing can identify patterns and responses in
learners that indicate their level of engagement, interest, and comprehension [1].
Research conducted by Smith and Seitz suggests that technology-enhanced learning
has significantly contributed to higher education quality [
2
]. It has provided students
with access to various resources and interactive learning opportunities, facilitating a more
engaging and personalized learning experience. Moreover, the adoption of e-learning has
Behav. Sci. 2024,14, 80. https://doi.org/10.3390/bs14020080 https://www.mdpi.com/journal/behavsci
Behav. Sci. 2024,14, 80 2 of 27
expanded access to education, allowing learners from diverse backgrounds and geographi-
cal locations to pursue educational goals.
However, despite these advancements, online learning still needs to improve student
engagement and retention. According to [
3
], neuromarketing can address these challenges
by analyzing brain activity to gain insight into students’ motivation and learning pref-
erences. By understanding neural responses to different learning stimuli, educators can
design online learning experiences that are more engaging and tailored to individual
learners’ needs. Furthermore, using neuromarketing in online learning can help improve
memory retention and knowledge transfer.
The practice of employing neuromarketing in online learning involves measuring
neural and physiological signals to gain insight into learners’ decisions, motivation, learn-
ing, and preferences [
4
]. Physiological tracking, which encompasses facial coding and
eye movement measurements, is a frequently used method in this field. Neuromarket-
ing is a marketing technique that aims to understand consumers’ unconscious responses;
comprehend consumer preference, expectancy, motivation, and behavior prediction; and
evaluate the effectiveness of advertising [
5
]. This interdisciplinary field combines the
fields of psychology, neuroscience, and economics. The core objective of neuromarketing
research is to eliminate subjectivity and ambiguity through the direct measurement of
observable brain behavior, thereby enhancing behavioral predictions and comprehension
of consumer behavior. Neuromarketing has also shed light on the neural variance observed
in individuals in the absence of behavioral variance [
6
]. Technological advances have en-
abled neuromarketing to progress beyond traditional quantitative and qualitative research
tools by focusing on consumers’ brain reactions to marketing stimuli [
7
] Neuromarketing
research aims to connect neural activity with consumer behavior and has a wide range
of applications for brands, products, packaging, advertising, and marketing for stores in
determining the intention to buy, level of novelty, awareness, and emotions generated. As
neuromarketing is a relatively underdeveloped discipline, theoretical research relies on
neuroimaging methods to assess hypotheses, improve existing knowledge, and test the
effects of marketing stimuli on consumer brains [
8
]. Recently, educational institutions have
typically relied on conventional techniques, such as surveys and focus group discussions, to
gather insights from students. However, the increasing use of technology has led to greater
reliance on telemarketing and social media communication to obtain information from stu-
dents and devise marketing strategies for prospective students. As a result, there has been
growing interest in the application of neuromarketing techniques in the education sector.
This literature review examines the use of neuromarketing in higher education, focusing
on studies employing technology and neuroimaging to investigate students’ cognitive
responses to various stimuli from social media and online learning environments.
1.1. Neuromarketing Research on Online Learning
Online learning is a self-regulated learning method that higher education has gained
increased reliance on because of the COVID-19 pandemic [
9
]. The reasons for this shift to
online learning include increased accessibility, advancements in communication technolo-
gies, and the need for institutions to remain competitive by offering flexible and varied
learning platforms ([
10
]). Recently, the COVID-19 pandemic has necessitated the immediate
suspension of in-person teaching in higher education institutions in the UK and beyond
and its replacement with online learning [
11
,
12
]. Therefore, it is crucial to consider the
factors that affect students’ motivation and learning performance in online classrooms.
Students in higher education in Europe and the UK have become heavily reliant on online
learning, and there are documented concerns about student engagement during unsu-
pervised learning, the lack of functional group engagement, and its impact on learning
capacity or mental and emotional responses to presented materials [
13
]. Consequently,
most colleges depend on conversations with their students through interviews, surveys,
and focus groups. However, researchers have documented that traditional data acquisition
methods are more effective [14].
Behav. Sci. 2024,14, 80 3 of 27
Ref. [
4
] conducted a study aimed at understanding the potential of online classrooms
to enhance learning performance with a sample of 297 students from Oxford Business
College. This study utilized stimuli-based gaze analytics to assess the impact of short
and long video lectures on student motivation and learning. In the short video lecture,
students’ gaze behavior was actively traced to gather data on their attention distribution,
and their emotions were assessed using facial coding software. In contrast, the long video
lecture assessed emotional involvement. The researchers found that students lost focus
during a 90 s lecture when no accompanying visual material was used. Moreover, the
results indicate that the students experienced higher levels of sadness and diminished
attention when exposed to a single shareable content page for more than 5.24 min. However,
when the visual material was changed and student discussions were introduced, students
experienced an improvement in their mood. This study utilized several neuroscience
metrics, including eye tracking, facial coding, and emotional analysis, to demonstrate
the significance of using neuromarketing to enhance students’ learning performance and
motivation in an online classroom. The findings of this study have practical implications
for improving the learning experience and motivation of lecturers, teachers, and students
in a virtual learning environment. This evidence could be leveraged to enhance online
learning by boosting students’ engagement when inattentive moods are recognized.
The study of learning capacity in neuromarketing has advanced significantly using
eye tracking, as demonstrated by various researchers. For instance, ref. [
15
] found that
the integration of eye tracking and content tracking technologies in an adaptive e-learning
platform allows for the delivery of relevant, accurate, and reliable information to students,
tailored to their level of knowledge and behavior in real-time. Additionally, ref. [
16
]
provided evidence for the value of eye tracking technologies in higher education, as they
were found to promote strategic processing and enhance integrated image and text analysis
and learning.
1.2. Neuromarketing Research on Social Media and Education
The concepts of user engagement and participation have become central to neuro-
marketing and transcending transactional contexts. In a study conducted by [
17
], the
researchers examined the influence of content characteristics communicated by a company
on Facebook on user behavior. Focusing on the medium’s form, content type, publication
time, number of likes, comments, shared actions, and interaction period on the brand page,
the study found that entertainment content was the most influential, whereas posts with
data related to the brand increased engagement through likes and comments. Additionally,
photos were the most attractive form of publication medium, and the number of comments
was higher on weekdays.
Similarly, ref. [
18
] analyzed the impact of cultural differences on social networks and
users’ commitment, loyalty, and brand recommendations. The researchers documented
videos as influential in improving the number of likes, comments, and shares. The number
of comments was reported to be associated with video content, and posts by the brand
remained at the top of the page for an extended time.
In a study conducted by [
19
], the marketing techniques utilized by leading global
brands on Facebook were investigated to determine the qualitative influences of messages
that are most likely to elicit a consumer response. The findings of this study are consistent
with previous research, indicating that images are more effective in attracting consumer
responses than text-based content and, in many cases, are associated with increased engage-
ment compared to video content. The study also found that increased content publication
on social network pages was associated with elevated levels of interaction, which may
have implications for fostering long-term customer–brand relationships. In 2019, ref. [
2
]
conducted research on the rectification of neuroscience myths through Facebook, reporting
that readers were more likely to positively evaluate articles that were consistent with their
preexisting beliefs. In contrast, ref. [
20
] found that posts on Facebook, Instagram, and
Behav. Sci. 2024,14, 80 4 of 27
Snapchat were primarily used to share positive emotions, whereas Twitter and Messenger
were primarily used to share negative emotions.
The advantages of neuromarketing research in the social media realm have been
demonstrated in the higher education sector. The utilization of social media in academia has
significantly increased, as indicated by a study revealing that almost 66% of the surveyed
academics employed Twitter for information retrieval, event organization, status updates,
and networking [
21
]. Despite this, none of the academics reported using Twitter to support
assignments or engage in structured class debates. On the other hand, other studies have
shown that certain academics have begun incorporating social media into their classes.
A recent investigation revealed that more than 75% of students utilized social media in
their professional endeavors; however, only 40% reported using it in the classroom. The
primary reasons for this disparity were concerns regarding privacy and the preservation of
academic integrity. Approximately half of the academics in this study concurred that social
media can be beneficial in fostering collaborative learning [22].
Recent scholarly investigations conducted across European universities have disclosed
the utilization of Twitter as an instrument to involve students in course materials and facili-
tate collaboration among peers [
23
,
24
]. Specific examples of Twitter’s applications include
the establishment of a communication channel for students to submit inquiries, obtain as-
sistance with academic tasks, and receive timely reminders about class activities. Twitter
was employed to facilitate interactions between students and professors, fostering a more
connected and cohesive learning environment. Research on Twitter-based assignments has re-
vealed that activities such as posing questions, responding to statements derived from course
readings, posting reactions to peers’ comments, and discussing project-related experiences
can significantly contribute to enhancing students’ engagement and academic performance.
Ref. [
24
] found that first-year students who incorporated Twitter into their seminar courses
demonstrated higher levels of engagement than students in a control group who did not use
Twitter. Moreover, the students who used Twitter exhibited significantly higher cumulative
Grade Point Averages (GPAs) than their non-Twitter counterparts. Another study conducted
by [
24
] revealed that although students were given the option to use Twitter, there were no
significant differences in GPAs or engagement levels between students who chose to use
Twitter and those who did not. Collectively, these studies underscore the potential benefits of
employing neuromarketing research in the realm of social media in higher education. They
also emphasized the importance of providing clear guidelines on how to utilize social media
platforms, such as Twitter, for fostering interactions with professors and collaborating with
peers. Incorporating social media into an educational curriculum in a purposeful and strategic
manner can yield positive outcomes when implemented effectively.
1.3. Levying Neuromarketing Research to Higher Education
The educational process involves students’ active engagement in the intellectual,
emotional, and physical aspects of learning through various methods and techniques.
Neuromarketing research in education can offer continuous improvements by gathering
feedback from students to meet their requirements.
Several concerns have been raised regarding the effectiveness of higher education in
enhancing students’ learning outcomes and its long-term impact [
25
]. In response, some
educators are adopting technology to transform teaching and create more interactive learn-
ing environments in both online and in-person classrooms [
25
]. For example, researchers
have used methods such as galvanic bracelets to measure the levels of engagement through
skin responses, magnetic resonance tomography to explore the relationship between brain
biology and cognitive processes, and biofeedback techniques to provide direct improve-
ments for both learners and educators [
26
]. These techniques have been widely applied to
various educational tools, including website design, the creation of engaging educational
materials and handouts, ebooks, and multimedia, which are of interest to contemporary
educators. Based on the findings of Sola’s [
4
] study utilizing facial coding and eye tracking
software, it is suggested that academics incorporate visual shareable content every 4–5 min,
Behav. Sci. 2024,14, 80 5 of 27
utilize a visual board, and introduce class discussions to maintain student attention during
long lectures. This is because prolonged lectures, exceeding 5.24 min, without shareable
content result in decreased attention and poor recall and information retention. Historically,
research methods employed by higher education institutions have focused on describing
and predicting the effectiveness of advertising campaigns aimed at influencing student
minds. Understanding and modeling students’ cognitive responses using neuroimaging
techniques allows institutions to gather valuable data on subconscious processes, which
can be utilized to develop effective strategies.
Neuromarketing can be a valuable tool for higher education institutions, enabling
them to investigate students’ cognitive processes and the corresponding changes that
occur during decision making. This can facilitate a better prediction of student behavior,
both inside and outside the classroom, for example, in online learning. Additionally,
neuromarketing can analyze the permanence of communication engraved in cognition
regarding the scientific aspects of advertising. By utilizing this tool, institutions can
devise more effective, student-driven strategies. In conclusion, the insights gained from
neuromarketing research on online learning and social media can be applied to the higher
education sector to increase engagement in teaching activities and promote community
engagement through teaching activities that leverage social media.
2. Material and Methods
The current study aims to demonstrate the potential of neuromarketing methods for
enhancing university branding at a subconscious level through three practical examples.
These examples include the official website page, the official Facebook page, and recorded
online video lectures of higher education institutions. By utilizing neuromarketing tech-
niques, precise and profound measurements can be made to better understand student
reactions and identify enrollment behavior. This approach can be particularly useful, as
conventional methods for testing and predicting the effectiveness of these investments have
generally failed because of consumers’ unwillingness and inability to describe their feelings
when exposed to an advertisement. Although there are challenges associated with the
translation of academic research into practical applications in commercial neuromarketing,
such as cost and timing, their potential benefits are significant. Therefore, neuromarketing
should be considered for future implementation in higher education.
In our study on all three examples which we tested, we opted to employ the advanced
quantitative research and neuromarketing research platform Tobii Sticky to measure elicited
emotions. This platform enabled us to conduct remote neuromarketing research, allowing
us to test a broader audience without being constrained by the geographical limitations of
laboratory research. Additionally, using Tobii Sticky, we were able to capture subtle emotions
through automatic facial coding with the webcam (15 Hz) modality, which traditional mar-
keting methods cannot measure. To ensure accuracy, we tested the Tobii Sticky software and
found that its average gaze error in a real-world (non-lab) environment was
1.6–1.8 degrees
(approximately 5% of the screen width and 7% of the screen height) on a laptop, which
provides us with reliable results regardless of the low spatial resolution of the webcam.
Moreover, a sentiment analysis was performed on two instances pertaining to the
website and Facebook page trials. Regrettably, we encountered technical constraints that
prevented us from executing a sentiment analysis on the archived online video lecture, as
discussed in Section 6.
A sentiment analysis employs text mining and computational linguistics to determine
the affective nature of written materials. The aim of a sentiment analysis is to evaluate
overall consumer sentiment towards neuromarketing. By examining consumer opinions
on blogs and social media, researchers have been able to identify the positive aspects of
neuromarketing and its impact on consumer behavior [27].
Several studies have explored sentiment analyses on social media platforms, such
as Facebook and Twitter. Ref. [
28
] proposed a method for ranking Facebook fan pages
that considers both user engagement and comment polarity, finding it to be more accurate
Behav. Sci. 2024,14, 80 6 of 27
than traditional methods. Refs. [
29
,
30
] emphasized the potential of sentiment analysis in
understanding consumer attitudes and behaviors, and ref. [
30
] achieved 85.25% accuracy
in a sentiment analysis using a natural language processing (NLP)-based pre-processed
data framework [
29
,
30
]. Refs. [
31
,
32
] further enhance the accuracy of sentiment analysis
by proposing new models. Ref. [
31
] achieved 91.2% accuracy using a morphological
sentence pattern model, while [
32
] experimented with a feedforward neural network for a
sentiment analysis of tweets. These studies collectively underscore the growing importance
and potential of sentiment analysis in the field of neuromarketing [
31
,
32
]. The sentiment
analysis tool assigns a score ranging from
100 to +100, where a score of
100 indicates a
very negative or serious tone and +100 suggests a very positive or enthusiastic tone.
Such an examination was not undertaken in the evaluation of the online learning ma-
terial utilized in the video lecture, given the technical constraints associated with analyzing
the video content. The application of neuroscience techniques in marketing, commonly
known as neuromarketing, has been a subject of significant ethical debate. This includes
concerns regarding the potential erosion of consumer autonomy, privacy, and control [
33
],
as well as the capacity for manipulation and violation of autonomy and privacy. Despite
these qualms, some argue that the current capabilities and implementation of neuromar-
keting research do not present meaningful ethical issues [
33
]. However, there is growing
interest in the incorporation of ethical principles in neuromarketing research, as evidenced
by the participation of various stakeholders in the field [
34
]. Ultimately, the field of neuro-
marketing has the potential to positively impact both society and consumers. However,
ethical concerns must be addressed to ensure its responsible use [35].
We adhered to ethical standards and obtained written informed consent from all
subjects on a digital form prior to their participation in the study. Participation was
voluntary, and no incentives were provided. All studies were conducted in accordance with
the European Code of Ethics for Research, and subject data were handled in accordance
with standard practices and the General Data Protection Regulation (GDPR). The Ethics
Committee of the Institute for Neuromarketing approved our research and supervised
the study to ensure compliance with local and international ethical guidelines, which are
publicly available on the institute’s official website.
2.1. Methodology
2.1.1. CARE Website Page
When crafting web content, it is crucial to carefully select words to ensure optimal
visibility on Google and attract students. The behavior of students on college websites and
the information they obtained from the content were examined using eye tracking sensors.
This research revealed that the eye movements and attention of the students differed based
on their academic backgrounds. Science major students outperformed non-science major
students in online scientific literacy assessments [
36
]. Eye tracking technology has also been
used to explore how viewers process web-based multimedia information, revealing that
students’ eyes are more fixated on text than on illustrations [
37
]. Another study focused on
learning objects and found that students paid more attention to headlines and illustrations,
leading to better learning performance [
38
]. Additionally, eye tracking was used to track
the eye movement process of college students while surfing websites with different levels
of complexity. The results showed that task complexity could moderate the effect of website
complexity on users’ visual attention and behavior [39].
Experiment Setup
In our methodology, we employed a combination of eye tracking technology and ques-
tionnaires. In our study, 529 subjects from Oxford Business College (OBC), comprising both
genders and spanning the age range of 18 to 50 years with various occupations, including
students, professors, and employed professionals from the UK, participated in our experi-
ments. We aimed to analyze the recently added “CARE page on the official OBC website,
as it was the newest addition to the site, and sought to gain deeper insights into how it was
Behav. Sci. 2024,14, 80 7 of 27
perceived at a subliminal level. The content of the page was evaluated by comparing the text’s
search engine optimization (SEO) position and emotional distribution through sentiment and
neuromarketing analyses, including facial coding. Using G*power [
40
], we determined that
a sample size of n= 35 was required to detect an effect with 90% power and a two-sided
significance level of 5%. Our research sample of 529 subjects surpassed this merit in terms of
the tested sample size (see Appendix A). All subjects were provided with an HTML” link to
the subpage of the Oxford Business College website and were instructed to browse the Center
for Applied Research and Entrepreneurship (CARE) page on their mobile device for 90 s. Our
software captured various metrics, such as heat maps, seen maps, gaze plots (Figure 1), mouse
clicks, and all areas of interest (AOIs), performed for each section of the text on the website
page. Heatmaps are graphical representations of complex data that utilize color to facilitate
visualization and comprehension. They are commonly employed for post hoc analyses of user
behavior and can reveal the most clicked area, the point at which individuals cease scrolling,
general website navigation patterns, and areas of interest. Warm tones signify higher data
values, while cooler tones indicate lower values. In the instance of Figure 1, the heatmap
employs shades of blue, green, yellow, and red, with red being indicative of increased visibility.
The seen map, also depicted in Figure 1, utilizes shades of grey, blue, and white, with lighter
colors indicating higher visibility. The gaze plot, also shown in Figure 1, is a visualization that
showcases individual data points connected by lines for each participant. This visualization is
most useful when examining the gaze patterns of a limited number of participants, as it can
become cluttered with an increasing number of individuals. The Media Only visualization,
also displayed in Figure 1, presents the raw media file. It is crucial to evaluate the performance
of various contexts, which is why we conducted tests on heat maps, seen maps, and gaze
plots (as demonstrated in the example in Figure 1.
The participants were instructed that they could proceed to the question section only
if they had read the page text earlier and were familiar with it. Statistical indicators of
OBC revealed that the greatest number of visits to websites originated from mobile devices.
Consequently, we conducted an experiment using mobile phones. However, owing to
the reduced optical resolution and sensitivity of the software to slight changes, a larger
sample size was required. The use of mobile devices in neuromarketing presents obstacles
in terms of sample size and data quality. Ref. [
41
] found that decreasing the sample size can
significantly impact research outcomes, with the threshold for comparable results increasing
with the task duration. This is particularly relevant in the context of mobile device use,
where the quality of image reviews can be affected by factors such as ambient light [
42
].
In addition, participants’ emotional reactions to the website were assessed through facial
coding using their mobile phone camera to obtain insights into their emotional activity. The
mood intensity measure reflects the intensity of the elicited positive and negative emotions
and ranges from 0 to 1. A negative mood score indicates the intensity of negative emotions
such as anger, puzzlement, fear, and sadness, while a positive mood score reflects the
intensity of joy. Text from the CARE webpage was analyzed using a sentiment analysis.
Finally, the participants were asked to answer three multiple-choice questions and one
open-ended question. Each question had a 10 s time limit (Appendix D). Along with
neuromarketing testing, we performed a sentiment analysis to understand and leverage
user-generated content on a website.
Sentiment analysis is a critical tool for comprehending the opinions expressed by
individuals on various websites, including social media platforms and product review
sites [
43
]. Among the different types of sentiment analysis, aspect-based sentiment analysis
(ABSA) is particularly effective for extracting sentiment features from web comments [
44
].
It is essential to develop an accurate website sentiment for artists’ websites, as it can
heighten interest, set appropriate expectations, and ultimately lead to increased sales [
45
].
Researchers have also paid significant attention to analyzing online word of mouth and
sentiment in online consumer reviews [
46
]. A novel approach that utilizes a conditional
random field algorithm and support vector machine classifier has been proposed for the
sentiment classification of online reviews, achieving high accuracy and eliminating the
need for relying on domain dictionaries [47].
Behav. Sci. 2024,14, 80 8 of 27
Behav. Sci. 2024, 14, x FOR PEER REVIEW 9 of 31
Figure 1. The CARE page was analyzed in this study.
Behav. Sci. 2024,14, 80 9 of 27
2.1.2. HAZEF Facebook Page
Neuroscientific methods, such as eye tracking devices, have been utilized by re-
searchers to track unconscious responses to visual stimuli and elucidate human behavior
on Facebook pages. These devices offer valuable insights into users’ attention and emo-
tional reactions when viewing Facebook pages [
48
]. Through the analysis of eye movements
and fixations, researchers can explore the interrelationship between individual differences
in personality, mental well-being, and the focus of users’ visual attention on Facebook [
49
].
Furthermore, the application of eye tracking-based brain–computer interface (BCI) sys-
tems has enabled real-time analyses of brain activity in response to visual and auditory
stimuli, providing an additional understanding of human behavior [
50
]. By integrating
these approaches, researchers can gain insights into users’ cognitive perceptions, emotional
responses, and behavioral reactions to visual stimuli on Facebook pages, ultimately helping
marketers to optimize content and capture user attention [51,52].
The aim of our research was to examine human behavior on Facebook and enhance the
content to capture students’ attention. To achieve this objective, we employed neuromar-
keting methods in conjunction with a semantic analysis, which is a conventional approach
in traditional marketing. This study aimed to ascertain why users follow the Facebook
page of the Croatian Academic Union of the Faculty of Economics (HAZEF). Through an
assessment of the present content, this investigation offers insight into the potential for
attracting new followers who demonstrate sustained interest and engage with the content
on an emotional level.
Experiment Setup
A sample of 190 individuals, equally consisting of male and female participants, aged
between 20 and 55 years, were chosen to participate in a neuromarketing experiment
conducted on the official Facebook page of the Croatian Academic Union of the Faculty of
Economics (HAZEF). Of the total number of participants, 59 successfully completed the
study (the number of participants that completed the entire experiment), yielding 45 eye
tracking recordings that were deemed usable (the number of participants whose gaze
and/or emotion were trackable during their session, i.e., had proper lighting and did not
move), with zero instances of screen-out (participants that ended their session based on a
screen-out question or did not meet technical requirements) and zero partial participants
(participants who started the experiment but closed the browser or had timed out before
reaching the end). This study was conducted in Croatia. According to the output from
G*Power, a sample size of 43 was required to detect the desired effect, with a power of
95% and a two-sided significance level of 5% (Appendix B). All subjects were presented
with a pre-recorded 70.01 s video of three Facebook posts from 12 May to 13 June 2021 and
were instructed to browse through them on their mobile devices as they typically would.
Based on Facebook statistics, mobile devices are the primary means of visiting the HAZEF
site, hence the requirement for mobile phone access in the study. The video was recorded
to allow subjects to scroll through each post, from top to bottom, and vice versa, but the
technical limitation of the software prevented subjects from clicking “See more” on the
posts. In this study, the Facebook posts depicted in Figure 2were analyzed. These posts
included Post 1, Post 2, and Post 3, each accompanied by selected areas of interest (AOIs)
and heatmaps. The attention heat map illustrates how customers direct their gaze while
using the three Facebook posts, with warmer hues signifying extended periods of focus.
This visualization pinpoints the content or sections that capture the viewer’s interest the
most, as indicated by the concentrated red area. Conversely, green areas represent regions
that require less cognitive effort. These heat maps, which provide a static representation
of gaze distribution, were complemented by facial coding techniques that assessed the
emotional responses of subjects based on their facial expressions. This analysis aimed
to evaluate the influence of HAZEF’s Facebook page, as shown in Figure 2. The three
Facebook posts displayed relevant announcements regarding HAZEF’s Entrepreneurial
Academy at the Medias Res conference. Along with eye tracking testing, these posts
Behav. Sci. 2024,14, 80 10 of 27
were analyzed using a sentiment analysis tool after being translated into English, as the
sentiment analysis platform conducted analyses on written English texts. The sentiment
analysis was performed only on the text of the Facebook posts, while a neuromarketing
analysis was conducted on both text and images.
Behav. Sci. 2024, 14, x FOR PEER REVIEW 11 of 31
three Facebook posts displayed relevant announcements regarding HAZEF’s Entrepre-
neurial Academy at the Medias Res conference. Along with eye tracking testing, these
posts were analyzed using a sentiment analysis tool after being translated into English, as
the sentiment analysis platform conducted analyses on wrien English texts. The senti-
ment analysis was performed only on the text of the Facebook posts, while a neuromar-
keting analysis was conducted on both text and images.
Figure 2. Facebook posts analyzed in this study. Post 1, Post 2, Post 3 with selected AOIs and heat
maps.
2.1.3. Online Lecture at Higher Education Institution
The integration of eye tracking technology in online lectures at higher education in-
stitutions holds great potential for enhancing learning experiences. Eye tracking enables
researchers to monitor students reading and learning behaviors, oering valuable in-
sights into their cognitive processing and topic interests [53]. By tracking students eye
movements, researchers can assess their aention to various stimuli, such as the instruc-
tors gaze, wrien information on the board, and crucial sentences in the text [54–56]. This
information can be used to design video lectures that optimize students learning experi-
ences and engagement. Furthermore, eye tracking data can be combined with articial
intelligence algorithms to predict student performance and provide personalized feed-
back [57]. In summary, the implementation of eye tracking in online lectures presents a
promising approach for enhancing the eectiveness of higher education instruction.
Experiment Setup
The impact of online teaching in higher education on students’ emotional responses
was investigated in Zagreb, Croatia, using a sample of 132 students (both genders) stud-
ying German at the Faculty of Teacher Education at the University of Zagreb. A priori
calculations were performed using G*Power to ensure that the required sample size was
Figure 2. Facebook posts analyzed in this study. Post 1, Post 2, Post 3 with selected AOIs and heat maps.
2.1.3. Online Lecture at Higher Education Institution
The integration of eye tracking technology in online lectures at higher education in-
stitutions holds great potential for enhancing learning experiences. Eye tracking enables
researchers to monitor students’ reading and learning behaviors, offering valuable insights
into their cognitive processing and topic interests [
53
]. By tracking students’ eye move-
ments, researchers can assess their attention to various stimuli, such as the instructor’s
gaze, written information on the board, and crucial sentences in the text [
54
56
]. This infor-
mation can be used to design video lectures that optimize students’ learning experiences
and engagement. Furthermore, eye tracking data can be combined with artificial intelli-
gence algorithms to predict student performance and provide personalized feedback [
57
].
In summary, the implementation of eye tracking in online lectures presents a promising
approach for enhancing the effectiveness of higher education instruction.
Experiment Setup
The impact of online teaching in higher education on students’ emotional responses
was investigated in Zagreb, Croatia, using a sample of 132 students (both genders) study-
ing German at the Faculty of Teacher Education at the University of Zagreb. A priori
calculations were performed using G*Power to ensure that the required sample size was
achieved. A sample size of n= 42 (n= 21 per device condition) was required to detect
the effect with a power of 80% and a two-sided significance level of 5% (Appendix C).
Students were randomly assigned to two conditions: computer (n= 68) and mobile phone
(n= 64). The main objective of this study was to determine whether there is a difference
in the elicited emotional responses between the two devices during online learning. The
Behav. Sci. 2024,14, 80 11 of 27
two video lectures were pre-recorded and uploaded to Tobii Sticky, with students receiving
the “HTML” link to one of the video lectures according to their condition. The lecture
viewed on the computer device lasted five minutes, while that viewed on the mobile phone
lasted 69 s. Although a sentiment analysis could have been theoretically implemented
in this study, it was not performed because of the technical restrictions of the available
sentiment analysis tools. To perform a sentiment analysis using free online sentiment
analysis software, videos must be publicly posted on YouTube. Unfortunately, a sentiment
analysis could not be conducted on the lecture material, as it was subject to copyright and
intended exclusively for students enrolled in the course. Nevertheless, even if conducted,
its findings would have been limited, as they would only have provided insight into the
sentiment of the lecture but not the student satisfaction score. In contrast, a neuromarketing
analysis can reveal subconscious reactions to visual, auditory, and textual stimuli.
3. Results
3.1. CARE Website Page
The sentiment analysis of the CARE webpage yielded a sentiment score of
1.5, indicat-
ing a predominantly negative or profound sentiment or tone in the text (Figure 3), which
was substantiated by the results of the neuromarketing analysis. This analysis revealed that
the subjects mostly experienced neutral and sad emotions (Figure 4) and that the general
mood intensity was relatively negative (Figure 5). Furthermore, the eye tracking behavioral
neurometrics, emotion analysis, and survey answers suggest that a redesign is necessary, as
the information of interest for visitors is not optimally positioned on the website, and much
of it goes unnoticed. It is important to note that eye tracking analyses on websites can be
instrumental in capturing emotions and can be used to explore information acquisition,
emotional experience, and behavioral intention on different information displays [
58
].
Additionally, eye tracking can be employed to offer interactive and personalized online
shopping experiences on mobile smartphones [
59
]. The detection of emotions using eye
tracking is a relatively novel approach, but it is gaining popularity in affective comput-
ing [
60
]. Finally, gaze direction can play a vital role in understanding and processing large
volumes of image and video content, personalization in human–media interaction, visual
content design, and affective analysis [61].
Behav. Sci. 2024, 14, x FOR PEER REVIEW 13 of 31
Figure 3. Sentiment analysis.
Figure 4. Elicited emotions during the CARE page visit.
-100
-80
-60
-40
-20
0
20
40
60
80
100
CARE Page FB Post 1 FB Post 2 FB Post 3
Sentiment score
Text type
Sentiment Analysis
0
20
40
60
80
100
Puzzlement Disgust Fear Joy Neutral Sadness Surprise
Probability Distribution (%)
Emotion
Elicited Emotion during CARE Page Visit
Figure 3. Sentiment analysis.
Behav. Sci. 2024,14, 80 12 of 27
Behav. Sci. 2024, 14, x FOR PEER REVIEW 13 of 31
Figure 3. Sentiment analysis.
Figure 4. Elicited emotions during the CARE page visit.
-100
-80
-60
-40
-20
0
20
40
60
80
100
CARE Page FB Post 1 FB Post 2 FB Post 3
Sentiment score
Text type
Sentiment Analysis
0
20
40
60
80
100
Puzzlement Disgust Fear Joy Neutral Sadness Surprise
Probability Distribution (%)
Emotion
Elicited Emotion during CARE Page Visit
Figure 4. Elicited emotions during the CARE page visit.
Behav. Sci. 2024, 14, x FOR PEER REVIEW 14 of 31
Figure 5. Mood intensity during the CARE page visit.
3.2. HAZEF Facebook Page
Upon examining the sentiment scores for the HAZEF Facebook posts, it is evident
that the results were inconsistent. Specically, Facebook Post 1 aained a total score of
100, indicating a predominantly positive and zealous sentiment. A similar overall tone
was observed for Facebook Post 2, which achieved a score of 82.1. Conversely, the senti-
ment score for Facebook Post 3 was 93.0, suggesting a negative and solemn sentiment.
Furthermore, the assessment of emotional responses through eye tracking revealed that
the subjects exhibited heightened levels of negative emotions (as shown in Figure 6) and
more intense feelings of neutral and sad emotions (as illustrated in Figures 79), regard-
less of the type of post. Eye tracking analysis on Facebook is important for capturing emo-
tions [50,60,62] because it allows researchers to explore the relationship between individ-
ual dierences in personality, mental well-being, SNS usage, and the focus on Facebook
users’ visual aention [63]. Eye tracking technology can be used as a primary sensor mo-
dality for emotion detection, alongside other methods such as EEG, facial image pro-
cessing, and speech inections [61]. The preliminary results from an eye tracking study
indicate that dynamic body features, such as torso and arm movements, are aended to
most often and longest, suggesting their importance in decoding emotions. Eye tracking
can also be used to analyze the distribution of visual aention on Facebook pages, provid-
ing insights into users’ visual aention and preferences for dierent types of advertise-
ments. Overall, eye tracking analyses on Facebook can contribute to the understanding of
emotions, personalization, visual content design, and aective analysis.
0
0.1
0.2
0.3
0.4
0.5
0 5 10 15 20 25 30
Average mood intensity
Time (s)
Mood during CARE Page Visit
Negative
Positive
Figure 5. Mood intensity during the CARE page visit.
3.2. HAZEF Facebook Page
Upon examining the sentiment scores for the HAZEF Facebook posts, it is evident
that the results were inconsistent. Specifically, Facebook Post 1 attained a total score of
100, indicating a predominantly positive and zealous sentiment. A similar overall tone
was observed for Facebook Post 2, which achieved a score of 82.1. Conversely, the senti-
ment score for Facebook Post 3 was
93.0, suggesting a negative and solemn sentiment.
Furthermore, the assessment of emotional responses through eye tracking revealed that
the subjects exhibited heightened levels of negative emotions (as shown in Figure 6) and
more intense feelings of neutral and sad emotions (as illustrated in
Figures 79
), regardless
of the type of post. Eye tracking analysis on Facebook is important for capturing emo-
tions
[50,60,62]
because it allows researchers to explore the relationship between individual
Behav. Sci. 2024,14, 80 13 of 27
differences in personality, mental well-being, SNS usage, and the focus on Facebook users’
visual attention [
63
]. Eye tracking technology can be used as a primary sensor modality
for emotion detection, alongside other methods such as EEG, facial image processing, and
speech inflections [
61
]. The preliminary results from an eye tracking study indicate that
dynamic body features, such as torso and arm movements, are attended to most often
and longest, suggesting their importance in decoding emotions. Eye tracking can also be
used to analyze the distribution of visual attention on Facebook pages, providing insights
into users’ visual attention and preferences for different types of advertisements. Over-
all, eye tracking analyses on Facebook can contribute to the understanding of emotions,
personalization, visual content design, and affective analysis.
Behav. Sci. 2024, 14, x FOR PEER REVIEW 15 of 31
Figure 6. Mood intensity during Facebook page visits.
Figure 7. Elicited emotion during HAZEF FB page scrolling.
0
0.1
0.2
0.3
0.4
0.5
012345
Average Mood Intensity
Time (s)
Mood intensity
Negative - Post 1 Positive - Post 1 Negative - Post 2
Positive - Post 2 Negative - Post 3 Positive - Post 3
0
20
40
60
80
100
Puzzlement Disgust Fear Joy Neutral Sad Surprise
Probability Distribution (%)
Emotions
Elicited Emotion during HAZEF FB page scrolling
Figure 6. Mood intensity during Facebook page visits.
Behav. Sci. 2024, 14, x FOR PEER REVIEW 15 of 31
Figure 6. Mood intensity during Facebook page visits.
Figure 7. Elicited emotion during HAZEF FB page scrolling.
0
0.1
0.2
0.3
0.4
0.5
012345
Average Mood Intensity
Time (s)
Mood intensity
Negative - Post 1 Positive - Post 1 Negative - Post 2
Positive - Post 2 Negative - Post 3 Positive - Post 3
0
20
40
60
80
100
Puzzlement Disgust Fear Joy Neutral Sad Surprise
Probability Distribution (%)
Emotions
Elicited Emotion during HAZEF FB page scrolling
Figure 7. Elicited emotion during HAZEF FB page scrolling.
Behav. Sci. 2024,14, 80 14 of 27
Figure 8. Emotional analysis of HAZEF FB page scrolling.
Behav. Sci. 2024, 14, x FOR PEER REVIEW 16 of 31
Figure 8. Emotional analysis of HAZEF FB page scrolling.
Figure 9. Emotional valence chart.
3.3. Online Lecture at Higher Education Institution
The results of the neuromarketing assessment conducted on the emotional reactions
of the participants during the rst 70 s of the lecture revealed that, regardless of the type
of device used, elevated levels of neutral and sad emotions were experienced (as depicted
in Figure 10), and the intensity of negative emotions was heightened (illustrated in Figure
11). The eye tracking analysis of online lectures in higher education is crucial for capturing
Figure 9. Emotional valence chart.
3.3. Online Lecture at Higher Education Institution
The results of the neuromarketing assessment conducted on the emotional reactions of
the participants during the first 70 s of the lecture revealed that, regardless of the type of device
used, elevated levels of neutral and sad emotions were experienced (as depicted in Figure 10),
and the intensity of negative emotions was heightened (illustrated in Figure 11). The eye
tracking analysis of online lectures in higher education is crucial for capturing emotions and
improving the learning process [
55
]. By monitoring students’ eye movements, researchers can
gather valuable data on their behavior and attention levels during online learning [
57
,
64
,
65
].
This information can be used to understand students’ content coverage, reading patterns, and
Behav. Sci. 2024,14, 80 15 of 27
attention at both the perceptual and conceptual levels [
54
]. Moreover, eye tracking can help
identify students’ levels of processing and topic interest, allowing for personalized learning
experiences that spark their interest. Furthermore, eye tracking can be used to maintain
student attention and vigilance throughout an entire online lecture, not just in the first few
minutes. In conclusion, eye tracking analyses of online lectures can provide insights into
students’ emotions, attention, and learning behaviors, enabling the development of adaptive
learning strategies and personalized feedback to enhance the learning process.
Behav. Sci. 2024, 14, x FOR PEER REVIEW 17 of 31
emotions and improving the learning process [55]. By monitoring students’ eye move-
ments, researchers can gather valuable data on their behavior and aention levels during
online learning [57,64,65]. This information can be used to understand students’ content
coverage, reading paerns, and aention at both the perceptual and conceptual levels [54].
Moreover, eye tracking can help identify students’ levels of processing and topic interest,
allowing for personalized learning experiences that spark their interest. Furthermore, eye
tracking can be used to maintain student aention and vigilance throughout an entire
online lecture, not just in the rst few minutes. In conclusion, eye tracking analyses of
online lectures can provide insights into students’ emotions, aention, and learning be-
haviors, enabling the development of adaptive learning strategies and personalized feed-
back to enhance the learning process.
Figure 10. Elicited emotions during lectures.
0
20
40
60
80
100
Puzzlement Disgust Fear Joy Neutral Sadness Surprise
Probability Distribution (%)
Emotion
Elicited Emotion during Lecture
Mobile Phone
Computer
Figure 10. Elicited emotions during lectures.
Behav. Sci. 2024, 14, x FOR PEER REVIEW 18 of 31
Figure 11. Elicited emotions during online lectures.
4. Experimental Analysis
To assess the ecacy and implications of neuromarketing techniques in European
higher education, it is necessary to conduct an experimental analysis using statistical tech-
niques. This study aims to broaden our understanding and develop measurements for
branding neuromarketing methods through the examination of emotions and mood in-
tensities experienced on social media websites. The statistical methods employed in this
study included a linear regression, MANOVA, and Pearson’s correlation.
The key ndings of this study are as follows:
1. A linear regression analysis was used to examine the relationship between and the
eectiveness of mood intensities, specically positive and negative moods.
2. The MANOVA provided insight into the variation in emotions, such as puzzlement,
fear, disgust, neutrality, joy, sadness, and surprise, observed by the subjects.
3. Pearson’s correlation was used to analyze the strength and direction of the emotions.
4.1. Demographic Analysis of CARE Website Page
The survey was conducted with a sample of 529 students, professors, and profession-
als who were randomly selected from the Oxford Business College. Following the exami-
nation and ltering of the data, 529 questionnaires were used for the analysis. The survey
respondents were diverse and had a wide range of characteristics. Among the surveyed
respondents, the percentage of male participants was 40%, which was lower than that of
the female respondents (60%). Many of the respondents were students, while professors
and professionals comprised equal proportions. When considering the sample popula-
tion, 40% fell within the age range of 18–38 years, 30% were under 28 years of age, and
30% were above 38 years of age (Table 1).
Table 1. Demographic analysis.
0
0.1
0.2
0.3
0.4
0.5
0 10203040506070
Average Mood Intensity
Time (s)
Elicited Emotion during Online Lecture
Negative - mobile phone
Positive - mobile phone
Negative - computer
Positive - computer
Figure 11. Elicited emotions during online lectures.
Behav. Sci. 2024,14, 80 16 of 27
4. Experimental Analysis
To assess the efficacy and implications of neuromarketing techniques in European
higher education, it is necessary to conduct an experimental analysis using statistical
techniques. This study aims to broaden our understanding and develop measurements
for branding neuromarketing methods through the examination of emotions and mood
intensities experienced on social media websites. The statistical methods employed in this
study included a linear regression, MANOVA, and Pearson’s correlation.
The key findings of this study are as follows:
1.
A linear regression analysis was used to examine the relationship between and the
effectiveness of mood intensities, specifically positive and negative moods.
2.
The MANOVA provided insight into the variation in emotions, such as puzzlement,
fear, disgust, neutrality, joy, sadness, and surprise, observed by the subjects.
3.
Pearson’s correlation was used to analyze the strength and direction of the emotions.
4.1. Demographic Analysis of CARE Website Page
The survey was conducted with a sample of 529 students, professors, and profes-
sionals who were randomly selected from the Oxford Business College. Following the
examination and filtering of the data, 529 questionnaires were used for the analysis. The
survey respondents were diverse and had a wide range of characteristics. Among the
surveyed respondents, the percentage of male participants was 40%, which was lower
than that of the female respondents (60%). Many of the respondents were students, while
professors and professionals comprised equal proportions. When considering the sample
population, 40% fell within the age range of 18–38 years, 30% were under 28 years of age,
and 30% were above 38 years of age (Table 1).
Table 1. Demographic analysis.
Variables Options Percentages (%)
Gender Male 40
Female 60
Profession Students 40
Professors 30
Professionals 30
Age 18–28 30
28–38 40
38–50 30
4.2. Linear Regression for CARE Website Page and HAZEF Facebook Page
A linear regression measures the extent to which the independent variables impact
the dependent variables and, consequently, the direction of influence, whether positive
or negative. The data provided pertain to the relationship between time and mood and
whether it has a positive or negative effect.
The analysis of variance (ANOVA) model was instrumental in comprehending the
variability in the model. The regression model revealed that the dataset exhibited 11% vari-
ability, signifying minor fluctuations in the dependent variable. The t- and p-values yielded
substantial and statistically significant outcomes, demonstrating a correlation between the
dependent and independent variables. Notably, the p-value was below the threshold, indi-
cating the model’s confidence level and verifying that the model was statistically significant
(Table 2).
Behav. Sci. 2024,14, 80 17 of 27
Table 2. Linear regression—model summary—start time.
Model Summary—Start Time
Model R R2Adjusted R2RMSE R2Change F Change df1 df2 p
H00 0 0 20.315 0 0 280
H10.32 0.103 0.099 19.28 0.103 31.885 1 279 <0.001
ANOVA
Model Sum of Squares df Mean Square F p
H1Regression 11,852.201 1 11,852.201 31.885 <0.001
Residual 103,709.05 279 371.717
Total 115,561.25 280
4.3. Relationship and Effectiveness of Mood Intensities between the Variables
The scatter plots illustrate the linear relationship between variables in the linear
regression analysis, which considers the interdependence between variables. The findings
indicate that a 10% change in the online dependent variable explains 90% of the variation
in the independent variable (see Figure 12). This also implies that the remaining 90% of the
variability can be attributed to other factors (Figure 13).
Behav. Sci. 2024, 14, x FOR PEER REVIEW 19 of 31
Variables Options Percentages (%)
Gender Male 40
Female 60
Profession Students 40
Professors 30
Professionals 30
Age 18–28 30
28–38 40
38–50 30
4.2. Linear Regression for CARE Website Page and HAZEF Facebook Page
A linear regression measures the extent to which the independent variables impact
the dependent variables and, consequently, the direction of inuence, whether positive or
negative. The data provided pertain to the relationship between time and mood and
whether it has a positive or negative eect.
The analysis of variance (ANOVA) model was instrumental in comprehending the
variability in the model. The regression model revealed that the dataset exhibited 11%
variability, signifying minor uctuations in the dependent variable. The t- and p-values
yielded substantial and statistically signicant outcomes, demonstrating a correlation be-
tween the dependent and independent variables. Notably, the p-value was below the
threshold, indicating the model’s condence level and verifying that the model was sta-
tistically signicant (Table 2).
Table 2. Linear regression—model summary—start time.
Model Summary—Start Time
Model R Adjusted R² RMSE R² Change F Change df1 df2 p
H 0 0 0 20.315 0 0 280
H 0.32 0.103 0.099 19.28 0.103 31.885 1 279 <0.001
ANOVA
Model Sum of Squares df Mean Square F p
H Regression 11,852.201 1 11,852.201 31.885 <0.001
Residual 103,709.05 279 371.717
Total 115,561.25 280
4.3. Relationship and Eectiveness of Mood Intensities between the Variables
The scaer plots illustrate the linear relationship between variables in the linear re-
gression analysis, which considers the interdependence between variables. The ndings
indicate that a 10% change in the online dependent variable explains 90% of the variation
in the independent variable (see Figure 12). This also implies that the remaining 90% of
the variability can be aributed to other factors (Figure 13).
Figure 12. Residuals vs. dependent variable.
Behav. Sci. 2024, 14, x FOR PEER REVIEW 20 of 31
Figure 12. Residuals vs. dependent variable.
Figure 13. Standardized residuals histogram.
4.4. MANOVA
The multivariate analysis of variance was conducted to test the signicance of one or
more independent variables in a set of two or more dependent variables. The dependent
variable in this study was gender, with two categories: male and female. The test allowed
for the simultaneous analysis of two or more samples. The normality of the data was as-
sessed based on skewness, which was calculated to be within the acceptable range for a
MANOVA. The Pillai trace statistic (1.000) indicated a signicant relationship between
gender and the response variables, suggesting that the volunteers who participated in the
study established a signicant association between gender, dependent variables, and
emotions. The data showed that each volunteer had a substantial paern of feelings and
relativity (Table 3). The residuals in the multivariate analysis refer to the unexplained data.
Approximately 278 degrees of freedom were associated with the residuals, which were
the dierences between the observed values and those predicted by the MANOVA model.
The residual variability was used to measure how well the data t into the model.
Table 3. MANOVA: Pillai.
Test Cases df Approx. F Trace
Pillai
Num df Den df p
(Intercept) 1 152,503.9 1 9 270 <0.001
Gender 2 19.017 0.774 18 542 <0.001
Residuals 278
4.5. Assumption Checks
Box’s M-test was used to assess the homogeneity of two or more covariance matrices.
The value of Box’s M was 937.883, which adhered to a chi-square distribution with 90
degrees of freedom. The resulting p-value was less than 0.001, indicating that a multivar-
iate normal distribution was followed, and the variance–covariance matrices were not
equivalent across the cells (Table 4).
Table 4. Boxs M-test for the homogeneity of covariance matrices.
Box’s M-Test for Homogeneity of Covariance Matrices
χ² df p
937.883 90 <0.001
4.6. Pearson Correlation
The Pearson correlation table provides a correlation constant, its signicance, and p-
values, which indicate the strength and signicance of the relationships between the
Figure 13. Standardized residuals histogram.
4.4. MANOVA
The multivariate analysis of variance was conducted to test the significance of one or
more independent variables in a set of two or more dependent variables. The dependent
variable in this study was gender, with two categories: male and female. The test allowed
for the simultaneous analysis of two or more samples. The normality of the data was
assessed based on skewness, which was calculated to be within the acceptable range for
Behav. Sci. 2024,14, 80 18 of 27
a MANOVA. The Pillai trace statistic (1.000) indicated a significant relationship between
gender and the response variables, suggesting that the volunteers who participated in
the study established a significant association between gender, dependent variables, and
emotions. The data showed that each volunteer had a substantial pattern of feelings and
relativity (Table 3). The residuals in the multivariate analysis refer to the unexplained data.
Approximately 278 degrees of freedom were associated with the residuals, which were the
differences between the observed values and those predicted by the MANOVA model. The
residual variability was used to measure how well the data fit into the model.
Table 3. MANOVA: Pillai.
Test Cases df Approx. F TracePillai Num df Den df p
(Intercept) 1 152,503.9 1 9 270 <0.001
Gender 2 19.017 0.774 18 542 <0.001
Residuals 278
4.5. Assumption Checks
Box’s M-test was used to assess the homogeneity of two or more covariance matri-
ces. The value of Box’s M was 937.883, which adhered to a chi-square distribution with
90 degrees of freedom. The resulting p-value was less than 0.001, indicating that a multi-
variate normal distribution was followed, and the variance–covariance matrices were not
equivalent across the cells (Table 4).
Table 4. Box’s M-test for the homogeneity of covariance matrices.
Box’s M-Test for Homogeneity of Covariance Matrices
χ2df p
937.883 90 <0.001
4.6. Pearson Correlation
The Pearson correlation table provides a correlation constant, its significance, and
p-values, which indicate the strength and significance of the relationships between the
variables in question. These values are of great importance in quantifying the strength
and direction of linear relationships. It is worth noting that the majority of the p-values
associated with these pairs of variables are below 0.05, indicating that their relationships
are meaningful. However, it should be noted that the correlation between time and the
other variables was not substantial, as indicated by the p-values (p> 0.05). The relation-
ship between puzzlement and various emotions was also investigated. It was found to
be positively correlated with sadness (r = 0.184) and negatively correlated with disgust
(r = 0.094
,p= 0.116), fear (r =
0.210, p< 0.001), joy (r =
0.105, p= 0.087), neutral
emotions (
r = 0.103
,p= 0.084), and surprise (r =
0.084, p= 0.159). However, this rela-
tionship was not statistically significant. Disgust displayed distinct relationships with the
variables, showing a significant negative correlation with fear (r =
0.125, p= 0.036), neutral
emotions (r =
0.134, p< 0.001), sadness (r =
0.032, p= 0.596), and surprise (r =
0.002,
p: 0.978). It also exhibited a positive correlation with joy (r = 0.195, 0 < 0.001). Fear was
found to have a significant negative correlation with joy (r:
0.224, p< 0.001) and sadness
(
r: 0.101
,p: 0.092) and a significant positive correlation with neutral emotions (r: 0.010, p:
0.863) and surprise (r: 0.148, 0: 0.013). Joy showed a significant positive correlation with
surprise (
r = 0.141
,p= 0.018) and a significant negative correlation with neutral emotions
(r =
0.493, 0 < 0.001) and sadness (r =
0.164, p= 0.006). Neutral emotions displayed
significant positive and negative correlations with sadness (r = 0.310, p< 0.01) and surprise
(r =
0.129, p= 0.031), respectively. Sadness exhibited a significant negative correlation
with surprise (r:
0.633, p< 0.001). All these relationships were statistically significant. The
directions of the correlations are represented by negative and positive signs, which also
Behav. Sci. 2024,14, 80 19 of 27
indicate whether they tend to move in the same or opposite direction when plotted in a
graph (Table 5).
Table 5. Pearson correlation constants and significance.
Variable Time
Puzzlement
Disgust Fear Joy Neutral
Emotions Sad Surprise
1. Time Pearson’s r
p-value
2. Puzzlement Pearson’s r 0.089
p-value 0.136
3. Disgust Pearson’s r 0.373 0.094
p-value <0.001 0.116
4. Fear Pearson’s r 0.166 0.21 0.125
p-value 0.005 <0.001 0.036
5. Joy Pearson’s r 0.555 0.105 0.195 0.224
p-value <0.001 0.078 <0.001 <0.001
6. Neutral emotions Pearson’s r 0.875 0.103 0.314 0.01 0.493
p-value <0.001 0.084 <0.001 0.863 <0.001
7. Sad Pearson’s r 0.378 0.184 0.032 0.101 0.164 0.31
p-value <0.001 0.002 0.596 0.092 0.006 <0.001
8. Surprise Pearson’s r 0.325 0.084 0.002 0.148 0.141 0.129 0.633
p-value <0.001 0.159 0.978 0.013 0.018 0.031 <0.001
5. Conclusions
In this study, we conducted three separate neuromarketing investigations, employing
eye tracking to uncover subtle emotional responses, survey questionnaires to compare
subconscious shifts with cognitive reactions, and a sentiment analysis to scrutinize varia-
tions in the outcomes. Deemed worthy of merit, we decided to examine three commonly
utilized domains within higher education: an official college website, an official college
Facebook page, and recorded online video lectures that are fundamental to teaching at a
higher education institution. For our research, we utilized a highly representative sample
size of n=720 participants, where we used n= 529 to test the CARE college website, n= 59
to test the HAZEF Facebook page, and n= 132 to test the emotional response of students
studying online. With the intention of gaining meritorious results, we employed three
different higher educational institutions (Oxford Business College, Croatian Academic
Union of the Faculty of Economics, University of Zagreb, Faculty of Teacher Education),
as each institution possesses distinct branding strategies that are associated with the three
measured segments.
Our findings from the HAZEF Facebook research revealed inconsistencies in the senti-
ment analysis and eye tracking results. Regrettably, the sentiment analysis produced false
positive outcomes, while the eye tracking analysis revealed profound negative emotions,
particularly capturing sadness. This research emphasizes the significance of incorporating
neuromarketing research into social media marketing research to uncover the actual be-
havioral patterns and emotional states of participants. Facebook eye tracking analysis can
contribute to comprehending emotions, personalization, visual content design, and affec-
tive analysis. We believe that the results where participants may have paid less attention
than other posts on the page are due to the color combination utilized for the visuals. The
analysis of the participants’ facial expressions suggests that visiting the HAZEF Facebook
page evokes emotions of neutrality and sadness, which is consistent with the muted colors
that are prevalent on the page. The possible implications of the elicited negative emotions
as a reflection of the conference are discussed. Research in the fields of color psychology
and marketing has consistently demonstrated that color can exert a substantial influence on
mood and attitude. A study conducted by [
66
] revealed that cool background colors tend to
engender more positive attitudes and behavioral intentions, particularly in positive-mood
and low-involvement conditions. Similarly, ref. [
67
] emphasized the importance of colors
in shaping moods and feelings, with the potential to attract a larger customer base. Ref. [
68
]
further investigated the relationship between color attributes and emotional dimensions,
Behav. Sci. 2024,14, 80 20 of 27
suggesting that the cognitive quantity of color, as well as the presentation medium, can
impact emotional responses.
Our CARE website research confirms the results of the sentiment and neuromarketing
analyses. By comparing these results with those of the cognitive survey testing, sadness
exhibited a significant negative correlation with surprise. However, this does not provide
substantial insights into why these emotions persist among visitors on college websites.
Employing neuromarketing research is essential to gain deeper insights into the persistence
of these emotions and what a college can modify in the future to elicit more favorable
emotions from its visitors. This research has validated the utility of eye tracking technology
in college website testing and highlights the limitations of traditional marketing methods
when seeking to understand the visitors who access our website and their genuine feelings
towards it. The analysis of eye movements on websites in higher education is essential for
gaining insight into the emotions of students. This method provides valuable information
on students’ behavior, preferences, and needs, which can be used to design customized
e-learning environments that cater to each student’s requirements [
27
,
37
,
69
,
70
]. By tracking
the gaze of students, researchers can understand how they interact with the website and
what content they focus on, which can be used to intervene and support students who may
struggle with attention or performance issues [
71
]. Moreover, the relationship between eye
movements and emotional states has been found to be highly correlated, making it possible
to gain insight into the psychological state of students during online learning [
15
]. Thus,
the analysis of eye movements is critical for understanding the emotions of students in
higher education and enhancing their learning experience.
Our investigation at a higher education institution, which focused on online lectures,
was the sole study among the other two that we conducted because of copyright restrictions.
This research was unique in that it exclusively utilized eye tracking for the semantic analysis,
as detailed in Experiment Setup. Our findings are insightful and highlight the need for this
type of research to be conducted more frequently in higher education. The study revealed
that, regardless of the device used for online lectures (mobile or desktop), the participants
experienced heightened levels of neutral and sad emotions. The findings of our study on
online learning demonstrate how advanced quantitative research based on eye tracking
data and facial coding can provide insight into students’ attention levels and emotional
engagement during online lectures. The expansion of online learning is driven by advances
in communication technologies and the need for flexibility [
3
,
67
,
69
]. Emotional engagement
is another important factor influencing students’ online learning [
72
]. Facial coding was
used in our study to obtain details about immediate emotional responses and general mood,
and the results showed that a neuromarketing analysis can provide time-based insights
into mood intensity and emotional engagement, which can inform adjustments to lectures
and pinpoint areas of weakness during the teaching process.
6. Limitations of Study
It is imperative to acknowledge the limitations of the sentiment analysis and the
Tobii Sticky platform. Conducting research simultaneously with both computers and
mobile phone devices using Tobii Sticky is not possible. Moreover, the software has a
technical limitation that precludes clicking “See more.” Sentiment analysis is a valuable
tool; however, it has certain limitations. It can only be applied to content published on
YouTube and requires Python for prerecorded videos, which restricts its use in copyrighted
or exclusive videos. Additionally, a sentiment analysis does not provide insights into
student satisfaction or dissatisfaction with online learning. The use of neuromarketing in
online learning offers a means of measuring student satisfaction and mood during lectures,
which directly affects learning outcomes. By monitoring a student’s mood throughout a
lecture, valuable feedback can be obtained to improve the lectures and identify areas of
weakness. Although sentiment analysis has limitations, the statistical analysis conducted
in this study using a linear regression, Pearson correlation, and MANOVA demonstrated
a significant model for understanding human expectations and behaviors in response to
Behav. Sci. 2024,14, 80 21 of 27
online teaching. The incorporation of neuromarketing techniques in online learning can
enhance user experience and engagement by identifying students’ subconscious preferences
and cognitive load, leading to increased satisfaction, better learning outcomes, and an
improved institutional reputation. The implications of incorporating neuromarketing
techniques for online learning in higher education in Europe are far-reaching and significant,
but more research is needed in this field.
7. Recommendation for Future Research
Future research utilizing eye tracking technology in higher education should concen-
trate on several key areas. First, it is crucial to investigate the potential of eye tracking
in studying self-regulated learning processes in university students [
73
]. This includes
examining the judgments of learning, metacognitive monitoring, meta-comprehension,
and learning strategies. Second, researchers should consider employing eye tracking to
evaluate and refine the textual and visual elements of educational presentations [
74
]. This
approach can provide valuable insights into tailoring educational content for different types
of students [
55
]. Furthermore, eye tracking can be incorporated into lectures and classes to
promote student engagement in research activities within their field of specialization. Lastly,
future research on eye tracking on social media in higher education should concentrate on
exploring how eye tracking can enhance self-regulated learning processes when students
are learning from multimedia materials [
71
]. Additionally, research should examine the
utilization of social media by higher education academics, including the advantages and
challenges that they encounter. This can involve studying how academics use social media
to disseminate research findings, advance their careers, and teach [73,75].
Author Contributions: H.M.Š.: conceptualization, methodology, writing—original draft, formal
analysis, and supervision. F.H.Q.: resources, writing the original draft, and funding acquisition. S.K.:
writing, reviewing, editing, and funding instruction. All authors contributed to article finalization,
review, and approval of the submitted version. All authors have read and agreed to the published
version of the manuscript.
Funding: This work was supported by the Institute for Neuromarketing, Zagreb, Croatia (research
activities including designing and conducting research utilizing neuromarketing equipment and
analyzing the data) and the Oxford Business College (paying the article processing charges for this
publication).
Institutional Review Board Statement: The study was conducted in accordance with the Declaration
of Helsinki and approved by the Institutional Review Board of the Institute for Neuromarketing
(IRBJ2023, 15 January 2023).
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: The data supporting this study’s findings are available in Figshare at
DOI 10.6084/m9.figshare.25035098. These data were published under CC BY 4.0. Deed Attribution
4.0. International license.
Acknowledgments: The authors would like to acknowledge the support of Syeda Umme Marya,
from the Institute for Neuromarketing, for conducting all SPSS analyses for this study and the support
of Rashmi Dhake for her assistance in contributing to the formatting and editing of this article.
Conflicts of Interest: The authors declare that the research was conducted in the absence of any
commercial or financial relationships that could be construed as potential conflicts of interest.
Behav. Sci. 2024,14, 80 22 of 27
Appendix A
Behav. Sci. 2024, 14, x FOR PEER REVIEW 25 of 31
Appendix A
Figure A1. G*Power output.
Appendix B
Figure A1. G*Power output.
Behav. Sci. 2024,14, 80 23 of 27
Appendix B
Behav. Sci. 2024, 14, x FOR PEER REVIEW 26 of 31
Figure A2. G*Power output.
Appendix C
Figure A2. G*Power output.
Behav. Sci. 2024,14, 80 24 of 27
Appendix C
Behav. Sci. 2024, 14, x FOR PEER REVIEW 27 of 31
Figure A3. G*Power output.
Appendix D
1. What is your general impression on the CARE webpage?
a. Very positive
b. Somewhat positive
c. Neutral
d. Somewhat negative
e. Very negative
2. How often do you visit the CARE page?
a. Never
b. Very rarely (once per month)
c. Rarely (2–3 times per month)
d. Occasionally (2–3 times per week)
e. Frequently (1–2 times per day)
f. Very frequently (more than 3 times per day)
3. How likely are you to return to the CARE webpage to search for more information
and to read the latest updates?
a. Very likely
b. Somewhat likely
Figure A3. G*Power output.
Appendix D
1. What is your general impression on the CARE webpage?
a. Very positive
b. Somewhat positive
c. Neutral
d. Somewhat negative
e. Very negative
2. How often do you visit the CARE page?
a. Never
b. Very rarely (once per month)
c. Rarely (2–3 times per month)
d. Occasionally (2–3 times per week)
Behav. Sci. 2024,14, 80 25 of 27
e. Frequently (1–2 times per day)
f. Very frequently (more than 3 times per day)
3.
How likely are you to return to the CARE webpage to search for more information
and to read the latest updates?
a. Very likely
b. Somewhat likely
c. Neither likely nor unlikely
d. Somewhat unlikely
e. Very unlikely
4. What do you dislike about the CARE website?
______________________________________
References
1.
Alsharif, A.H.; Salleh, N.Z.M.; Al-Zahrani, S.A.; Khraiwish, A. Consumer Behaviour to Be Considered in Advertising:
A Systematic Analysis and Future Agenda. Behav. Sci. 2022,12, 472. [CrossRef] [PubMed]
2.
Smith, C.N.; Seitz, H.H. Correcting Misinformation about Neuroscience via Social Media. Sci. Commun. 2019,41, 790–819.
[CrossRef]
3.
Shah, M.; Goode, E.; West, S.; Clark, H. Widening Student Participation in Higher Education through Online Enabling Education.
Widening Particip. Lifelong Learn. 2014,16, 36–57. [CrossRef]
4.
Šola, H.M.; Qureshi, F.H.; Khawaja, S. Enhancing the motivation and learning performance in an online classroom with the use of
neuromarketing. Eur. J. Manag. Mark. Stud. 2021,7, 1–25. [CrossRef]
5.
Mansor, A.A.; Isa, M.S. Development of Neuromarketing Model in Branding Service. In Proceedings of the 8th International
Conference on Education and Information Management (ICEIM-2015), Penang, Malaysia, May 2015; pp. 1–10. Available
online: https://www.researchgate.net/publication/306396646_Development_of_Neuromarketing_Model_in_Branding_Service
(accessed on 1 November 2023).
6.
Venkatraman, V.; Clithero, J.A.; Fitzsimons, G.J.; Huettel, S.A. New scanner data for brand marketers: How neuroscience can help
better understand differences in brand preferences. J. Consum. Psychol. 2012,22, 143–153. [CrossRef]
7.
Reimann, M.; Schilke, O.; Weber, B.; Neuhaus, C.; Zaichkowsky, J. Functional magnetic resonance imaging in consumer research:
A review and application. Psychol. Mark. 2011,28, 608–637. [CrossRef]
8.
Karmarkar, U.R.; Plassmann, H. Consumer Neuroscience: Past, Present, and Future. Organ. Res. Methods 2019,22, 174–195.
[CrossRef]
9.
Schacter, D.L.; Szpunar, K.K. Enhancing attention and memory during video-recorded lectures. Sch. Teach. Learn. Psychol. 2015,1,
60–71. [CrossRef]
10.
Skordis-Worrall, J.; Haghparast-Bidgoli, H.; Batura, N.; Hughes, J. Learning Online: A Case Study Exploring Student Perceptions
and Experience of a Course in Economic Evaluation. Int. J. Teach. Learn. High. Educ. 2015,27, 413–422.
11.
Adedoyin, O.B.; Soykan, E. COVID-19 pandemic and online learning: The challenges and opportunities. Interact. Learn. Environ.
2023,31, 863–875. [CrossRef]
12.
Qureshi, M.I.; Khan, N.; Qayyum, S.; Malik, S.; Hishan, S.S.; Ramayah, T. Classifications of Sustainable Manufacturing Practices
in ASEAN Region: A Systematic Review and Bibliometric Analysis of the Past Decade of Research. Sustainability 2020,12, 8950.
[CrossRef]
13.
Ferguson, J.M.; DeFelice, A.E. Length of online course and student satisfaction, perceived learning, and academic performance.
Int. Rev. Res. Open Distrib. Learn. 2010,11, 73–84. [CrossRef]
14.
Berman, F.; Crosas, M. The Research Data Alliance: Benefits and Challenges of Building a Community Organization. Harv. Data
Sci. Rev. 2020,2, 1–11. [CrossRef]
15.
Barrios, V.M.G.; Gütl, C.; Preis, A.M.; Andrews, K.; Pivec, M.; Mödritscher, F.; Trummer, C. AdELE: A Framework for Adaptive
E-Learning through Eye Tracking. In Proceedings of IKNOW, (IICM); Faculty of Computer Science: Graz, Austria, 2004; pp. 609–616.
16.
Mason, L.; Pluchino, P.; Tornatora, M.C. Using eye-tracking technology as an indirect instruction tool to improve text and picture
processing and learning. Br. J. Educ. Technol. 2016,47, 1083–1095. [CrossRef]
17.
Cvijikj, I.P.; Michahelles, F. Online engagement factors on Facebook brand pages. Soc. Netw. Anal. Min. 2013,3, 843–861.
[CrossRef]
18.
Khan, I.; Dongping, H.; Wahab, A. Does culture matter in effectiveness of social media marketing strategy? An investigation of
brand fan pages. Aslib J. Inf. Manag. 2016,68, 694–715. [CrossRef]
19.
Kim, D.-H.; Spiller, L.; Hettche, M. Analyzing media types and content orientations in Facebook for global brands. J. Res. Interact.
Mark. 2015,9, 4–30. [CrossRef]
20.
Vermeulen, A.; Vandebosch, H.; Heirman, W. #Smiling, #venting, or both? Adolescents’ social sharing of emotions on social
media. Comput. Hum. Behav. 2018,84, 211–219. [CrossRef]
Behav. Sci. 2024,14, 80 26 of 27
21.
Knight, C.G.; Kaye, L.K. ‘To tweet or not to tweet?’ A comparison of academics’ and students’ usage of Twitter in academic
contexts. Innov. Educ. Teach. Int. 2016,53, 145–155. [CrossRef]
22.
Moran, M.; Seaman, J.; Tinti-Kane, H. Teaching, Learning, and Sharing: How Today’s Higher Education Faculty Use Social Media.
Pearson Learning Solutions: Boston, MA, USA, 2011; pp. 1–26. Available online: https://files.eric.ed.gov/fulltext/ED535130.pdf
(accessed on 1 November 2023).
23.
Junco, R.; Heiberger, G.; Loken, E. The effect of Twitter on college student engagement and grades. J. Comput. Assist. Learn. 2011,
27, 119–132. [CrossRef]
24.
Junco, R.; Elavsky, C.M.; Heiberger, G. Putting twitter to the test: Assessing outcomes for student collaboration, engagement and
success. Br. J. Educ. Technol. 2013,44, 273–287. [CrossRef]
25. Arum, R.; Roksa, J. Limited Learning on College Campuses. Society 2011,48, 203–207. [CrossRef]
26. Norman, G. Education and neuroscience. Adv. Health Sci. Educ. 2016,21, 919–920. [CrossRef] [PubMed]
27.
Luna-Nevarez, C. Neuromarketing, Ethics, and Regulation: An Exploratory Analysis of Consumer Opinions and Sentiment on
Blogs and Social Media. J. Consum. Policy 2021,44, 559–583. [CrossRef]
28.
Ngoc, P.T.; Yoo, M. The lexicon-based sentiment analysis for fan page ranking in Facebook. In Proceedings of the International
Conference on Information Networking 2014 (ICOIN2014), Phuket, Thailand, 10–12 February 2014; pp. 444–448.
29.
Neri, F.; Aliprandi, C.; Capeci, F.; Cuadros, M.; By, T. Sentiment Analysis on Social Media. In Proceedings of the 2012 IEEE/ACM
International Conference on Advances in Social Networks Analysis and Mining, Istanbul, Turkey, 26–29 August 2012; pp. 919–926.
30.
Hasan, R.; Maliha, M.; Arifuzzaman, M. Sentiment Analysis with NLP on Twitter Data. In Proceedings of the 2019 International
Conference on Computer, Communication, Chemical, Materials and Electronic Engi-neering (IC4ME2), Rajshahi, Bangladesh,
11–12 July 2019; pp. 1–4.
31.
Han, Y.; Kim, K.K. Sentiment analysis on social media using morphological sentence pattern model. In Proceedings of the 2017
IEEE 15th International Conference on Software Engineering Research, Management and Applications (SERA), London, UK, 7–9
June 2017; pp. 79–84.
32.
Duncan, B.; Zhang, Y. Neural networks for sentiment analysis on Twitter. In Proceedings of the 2015 IEEE 14th International
Conference on Cognitive Informatics & Cognitive Computing (ICCI*CC), Beijing, China, 6–8 July 2015; pp. 275–278.
33.
Stanton, S.J.; Sinnott-Armstrong, W.; Huettel, S.A. Neuromarketing: Ethical Implications of its Use and Potential Misuse. J. Bus.
Ethics 2017,144, 799–811. [CrossRef]
34.
Pop, N.A.; Dabija, D.C.; Iorga, A.M. Ethical Considerations regarding Stakeholders in Neuromarketing Research. Empirical
Insights from NMSBA Corporate Members, TAAN Advertising Agencies and Romanian Companies. In Ethics and Neuromarketing;
Springer International Publishing: Cham, Switzerland, 2017; pp. 123–146. [CrossRef]
35.
Isa, S.M.; Mansor, A.A.; Razali, K. Ethics in Neuromarketing and its Implications on Business to Stay Vigilant. KnE Soc. Sci.
Volume 2019, 687–711. [CrossRef]
36. Tsai, P.-Y.; Yang, T.-T.; She, H.-C. Explore College Students’ Cognitive Processing during Scientific Literacy Online Assessments
with the Use of Eye Tracking Technology. In Proceedings of the 2015 IEEE 15th International Conference on Advanced Learning
Technologies (ICALT), Hualien, Taiwan, 6–9 July 2015;