New Didactic Models for MOOCs
Léon Rothkrantz
Czech Technical University in Prague, Faculty of Transport, Konviktskastreet 20, Prague, Czech Republic
Delft University of Technology, Department of Intelligent Interaction, Mekelweg 4, Delft, The Netherlands
Keywords: Massive Open Online Courses, Didactic Models, Emotion, Dropout Rate.
Abstract: In recent years we observed an enormous rise of Massive Open Online Courses (MOOCs). A problem of
MOOCs is the high dropout rate. This caused by the lack of an appropriate didactic model, low interaction
students teacher, poor feedback mechanism. In this paper we propose some improvements. Our proposed
didactic model is a description of the learning interaction process in the course of the time. Emotions play
an important role in this model. It proves that emotions have a great impact on the study behaviour of
students. Some emotions as happiness, can stimulate students to go on with the study, other emotions as fear
for exams, anger, disappointment about results of exams can block the study behaviour of students. Next we
present some educational actions grounded on our model and proposed to decrease the dropout rates. One of
the actions is based on verbal and nonverbal feedback of students about their emotional state. Our proposed
actions are tested on a small scale using a MOOC implemented under Moodle.
Massive Open Online Courses (MOOCs) are a
recent phenomenon in distant learning. Consortia as
edX, Coursera and Udacity offer hundreds of
courses. Students have free access to courses taught
by excellent teachers from renowned Institutes.
MOOCs are attractive for unprivileged students
because usually no entrance exams and tuition fees
are required. One big disadvantage of MOOCs is the
low success rate. On average only 7% of the
students finish a MOOC successfully. One of the
reasons is that an appropriate didactic model for
MOOCs is still lacking, interaction students teacher
is difficult to realise and poor feedback mechanism.
Nowadays the focus is on training 21
skills, as cooperation, problem solving, creativity
and critical thinking. MOOCs are well suited for
training such abilities. But a great problem is that
students easily drop-off as soon as the learning
material starts to be boring, difficult to understand or
too much cognitive processes are required. These
learning problems are not new. Many students have
the experience that they lose their attention regularly
during oral lectures or presentations. But a good
lecturer is able by giving summaries, examples to
get students again on-board. And a big advantage is
that students are not allowed to leave the lecture
room and can give the teacher the impression that
they are still listening. But as soon students starts
consulting their smartphones or laptop, the attention
for the oral lectures is dropping. In case of MOOCs
students can also lose their attention, motivation and
interest and can easily drop-out. The challenge is
how to keep students on board during the MOOCs
learning process.
In this paper we introduce a new didactic model
for MOOCs. Important components in this model
are emotions, attractiveness, social support and of
course the learning material. In the student-teaching
interaction model the components can be positively
or negatively stimulated. A sequence of negative
stimuli will result in drop-off from the MOOC
course. We will present some didactic educational
actions focussed on positive stimuli to increase the
interest, motivation and learning activities of
students. The new didactic approach has been
implemented in a prototype of a MOOC on crisis
management during flooding disaster developed at
Czech Technical University in Prague. We
performed some small scale experiments to test our
approach. The preliminary results will be reported in
this paper.
The outline of the paper is as follows. In chapter
two we present some related work. And in chapter
three we present our new didactic model. In the rest
of the paper we present some methods how to assess
Rothkrantz, L.
New Didactic Models for MOOCs.
DOI: 10.5220/0006362805050512
In Proceedings of the 9th International Conference on Computer Supported Education (CSEDU 2017) - Volume 1, pages 505-512
ISBN: 978-989-758-239-4
Copyright © 2017 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
the emotional state of students and how the
information can be used to improve the learning
One online survey lists down the top 10 reasons for
the low completion rates, and lecture fatigue and
poor course design figure at numbers 4 and 5
respectively in this list (Colman, 2013). Many
researchers emphasizes on the lack of professional
instructional design for MOOCs. Especially
disturbing is that none of the major MOOC
providers have hired anyone trained in instructional
design, the learning sciences, educational
technology, course design, or other educational
specialties to help with the design of their courses.
Armellini et al (Armellini, Padilla Rodriguez,
2016) researched the question if MOOCs are
pedagogically innovative. They surveyed different
stakeholders. Their conclusion was that claims
linking pedagogic innovation to MOOCs are largely
unfounded. xMOOCs in particular seem to rely on
strategies that have been used in online and distance
learning for decades. The authors state that MOOCs
provide good examples of technological innovation
but also of highly debatable approaches to
pedagogy. They may be deemed valuable as
resources (MOORs), but far less so in terms of being
pedagogically innovative courses.
Tak-ming Wong, (Tak-ming Wong, 2016)
summarises the factors leading to effective teaching
of MOOCs at the various stages of course delivery,
as reported in relevant studies. The preparation stage
involves understanding the various aspects of the
development of a MOOC. For attraction, how to
draw and arouse the interest of target learners in
the course is discussed. The emphasis in
participation is on the ways to make learners engage
in learning activities and interact with the course
contents. Interaction centres on encouraging learners
to interact with each other to foster learning. In
consolidation, the assessment issues are addressed.
Finally, post-course support examines monitoring
and analysis of student data for the continuous
improvement of teaching.
Dillon et al (Dillon, 2016) measured a range of
self-reported student emotions in a MOOC context.
They found that hope and enjoyment were the most
frequently reported emotions, followed by
contentment, anxiety and pride, while shame,
disappointment, isolation, anger and sadness were
rarely reported. Maybe students don’t like to report
that they experienced negative emotions. It proved
that the results depend of the context. Self-reported
emotions related to specific segments of content, or
study activities may elicit different emotional
responses. The emotions anxiety, confusion,
frustration and hope were reported to dropout
Leony et al (Leony, 2015) discovered
correlations between emotions such as frustration,
confusion, boredom and happiness and evaluation
metrics as percentage of exercises solved well, time
spent in exercises, video abandon and video
avoidance. Their hypothesis is that learner’s
emotions can be inferred by analysing their actions.
We present similar ideas in this paper. The authors
used the affective information to personalise the
MOOC experience to the learner’s skills, objectives
and profile. The goal is to create an adaptive
educational system with personalised action plan.
The learner profile can include information like
current learning skills. Learning style, learning goals
and accessibility needs. The inclusion of affective
information in MOOCs can be used to customise the
learning experience.
“Emotions, technology and learning” is a volume
of research papers on emotions in education edited
by Tettegah et al (Tettegah, 2016). According to the
editor research suggests two important roles of
emotion related to learning and technology. First,
emotion can be the key factor that is being learned
or taught through technological means. Second,
emotional responses with and through technology
can alter what is being learned or how the content is
learned. Both perspectives are discussed in the
volume by focusing on the relationship between
emotion and learning as facilitated by technology.
The book is divided into four sections to represent
the specific interest related to emotion and learning,
with an interesting section on theory and emotions
and Learning Online.
In (Rothkrantz, 2016b) we presented new didactic
models for regular courses and distant learning. The
first model was focussed on finding the impact of
different components on the academic performance
measured as the number of exams passed
successfully. Because we used a longitudinal model
the impact of the components is time dependent. To
understand the dropout process of students we have
to study the interaction process of students with the
CSEDU 2017 - 9th International Conference on Computer Supported Education
learning material. What is the causality of the abrupt
stop of the interaction process? We stress the
dynamic aspect of the interaction process over time.
It is impossible to measure the interaction process
continuously over time by surveys. Another
assessment methodology is requested. Next we
developed a model describing the individually based
interaction process of a student with the learning
material. We like to measure the activity level of
students, his emotional state and to understand the
impact of these components on the interaction
process of student-learning material.
To visualise the dynamic interaction process we
take a metaphor from fluid dynamics. We assume a
system of connected liquid reservoirs. In the basic
reservoir the interaction process is fuelled by a
stream of study liquid. Several components of the
model as presented in Figure 1, are represented as
sub processes of the interaction process and can
generate additional liquid or consume available
liquid. Such a model is needed to assess at some
moment which resources consume too much energy
and which resources don’t produce any energy
anymore. Based on that assessment we are able to
compute if the level of study liquid is approaching a
critical level. If a student passes a minimal threshold
he will be at risk dropping-out. In that case some
educational actions are needed to fuel the interaction
process again. In our experiments we will discuss
some examples.
The different components of the model interact
via the basic reservoir. A lack of cognitive
capabilities can be compensated by strong
motivation, hardworking and control of study
behaviour. The basic fuel of the process is a melting
pot of study liquid. Several components provide new
fuel to this process, other components consume fuel.
It is important to research which components are
main resources for new study liquid. In this section
we describe some didactic measurements stimulating
several components to trigger more study liquid.
Important to note that the amount of liquid depends
of the time, student characteristics and context
Before a student starts a MOOC his basin should
be filled by sufficient study liquid. His perception of
his capabilities, abilities, motivation, goal and
evaluation of a preview of the study generate some
fuel. Once a student enrolled in the course the
learning interaction process provides additional
liquid but also costs some liquid. Viewing
interesting movies or mastering assignments
generates additional liquid via the component
motivation. Asking questions, special alerts or
Figure 1: Didactic model.
switching to more interesting topics stimulate the
component attention. But over time attention is
decreasing, student can lost his interest and liquid is
leaking away via the component physics.
We stress the fact that the impacts of
components are not static but can change over time.
We also consider interaction of components and
summation of liquid contribution and liquid leakage
of different components.
Attraction: An important component in the
interaction student learning material is interest of
students. As long as the learning material is
interesting a student keeps on board. But as soon the
learning material gets boring, student lose their
interest and the dropout process has been started. In
case of regular courses there is a lot of pressure of
peers, Institute, teaching schemas to follow lectures
the other day. New topics provide opportunities for a
new start. Usually this is missing in case of MOOCs.
By the huge amount of students individual tutoring
is difficult to realize. A special didactic approach is
needed. The span of attention or control is very
short, only some minutes. A varied way of
presentation of learning material is needed by
showing movies, video lectures, simulations and
interesting applied assignments for students. After
positive experiences of the students with the learning
material the attraction will be increased. For
example after solving assignments, understanding
the learning material, or after positive feedback of
peers or tutor.
Capabilities: A student assumes at start he has
sufficient capabilities to complete the course
successfully. But when he interacts with the learning
material and it proves that it is far beyond his
capabilities, the dropout process has been started. In
regular face to face courses a student is not allowed
to leave the teaching hall. In case many students lose
their interest an experienced teacher starts a
summary, a clarifying example to get students back
in the teaching-learning process. In regular courses
there is support of peers during the breaks or after
the lectures. In MOOCs social support of peers is
New Didactic Models for MOOCs
wanted but usual less developed. There is a trend to
develop MOOCs as self-paced courses for individual
students. That makes these students vulnerable for
negative interactions.
Activity: Attractive learning material and
required capabilities are prerequisite of a positive
interaction process of a student with his learning
material. Next a student is assumed to play an active
role in this process. Many students read the
description of the offered courses, they even enrol in
the courses but the next step is to start the study
activity. Students have to set themselves into action.
Wandering about, missing adequate study abilities
can take a lot of energy but without the expected
result. Most MOOCs are not designed for passive
students, a lot of activity is required varying from
posing and answering questions, making
assignments and involvement in project activities.
After positive feedback from the interaction process
of students with the learning material the activity
can be increased
Distraction: Interest and sufficient capabilities
are the positive drives of the interaction students
with the learning material. But there are also two
negative drives. Distraction is the first negative
drive. In case of MOOCs students usually study in a
stimulus rich environment. The computer used for
taking the course offers a lot of alternatives for
distraction especially in case the MOOC material
gets boring or is beyond the capabilities of students.
Surfing on the web is a favourite activity, takes a lot
of energy but doesn’t contribute to successful
completion of the course.
Physiology: A second negative drive is the
physiological state of the students. If a student gets
tired, continuation of the learning activities take a lot
of additional energy. If a student gets hungry, sleepy
he can start a break. A lot of discipline is needed for
a restart and to keep the length of the break under
control. In case of regular courses there are social
rules, institutional rules regulating breaks.
Self-motivation: In a classroom setting is able to
motivate his students and to push him to activities.
In distant learning this is more complicated and can
only be realised by the learning material. A student
in distant learning courses is assumed to motivate
himself. In some cases fellow students can take the
role of tutor. A low degree of self-motivation in self-
paced courses is a main factor to explain the dropout
rates of students in MOOCs (Rothkrantz, 2016a). A
strong stimulating role should be played by the
learning material. This material should contain a lot
of variation in learning activities and variation in
energy needed to process the offered material.
Emotions: We stressed already that affective
components as feelings and emotions have a huge
impact on successful study behaviour of students.
Students should like their study, being proud of their
study and have progressive learning results.
Successive negative emotions can result in a
negative mood of students and in that state they are
very sensitive to dropout. In the next section we
focus on the assessment of the emotional state of the
students and how to transform negative mood to a
positive mood. An important role is played by
variation in learning material and learning activities
and affective support by peers.
3.1 Validation of Learning Model
A fluid dynamic approach of psycho-somatic
processes has its roots in ancient time. The model
enables us to relate different processes involved in
learning. The idea that physic somatic processes
share simultaneous resources has been researched
recently by the brain scientist Scherder (Niet, van
der). He claims that exercises and making music
have healthy effect on the brains.
In (Rothkrantz, 216b) we presented a didactic
model describing students as open systems. This
implies an ecological attitude with respect to
causality. Context and evolution have to be
considered. Success of study is not only dependent
of student’s characteristics as capabilities,
knowledge, used study methods, but is the result of a
complex interaction process. To validate the model
we performed a survey study. All students
Informatics (n=160) were requested to fill in a
questionnaire of 190 items on a 3 point Likert scale.
In these items students were questioned about their
opinions, motivations, and assessments, experiences
with respect to the study, study environment and
study conditions with a focus on the interactive
aspect. It proves that 80% of the variables show a
significant difference between the groups of students
passing, 0, 1, 2, 3 or 4 exams successfully at the first
exam session of the academic year. Students with
bad academic results had low scores on motivation,
capabilities, learning activities. Students who
decided to stop the study had an exit interview with
the student’s counsellor and again it proves that
during the first weeks/months of study the
motivation, activity, appreciation and pleasure
decreased and finally resulted in the drop-off
We assumed that factors underlying the drop-off
process during regular courses and e-learning play a
role in learning via MOOCs. Based on the survey
CSEDU 2017 - 9th International Conference on Computer Supported Education
research we were able to detect factors having an
impact on study-success or failure of the whole
cohort of students. Analysis of the data of individual
students obtained by questionnaires on succeeding
moment we were able to confirm our results. To
evaluate MOOCs studies more advance learning
analytics methods are needed. The response rate via
questionnaires is usual low in MOOCs studies. In
our first preliminary assessments we researched the
feedback and interaction of students via special
interfaces presented in the next chapters of this
paper. The impact of these special educational can
be understood and explained using our didactic
In regular classroom teaching a teacher is able to
perceive the level of attention, participation and
other learning activities of his students by
observation of their body language. If students start
reading their e-mail, interacting with their
smartphone, closing their eyes, looking at their
watch, yawning or starting a discussion with peers,
the teacher may assume he is losing contact with his
students and start special actions as giving a
summary, giving an example, making a joke or
starting a discussion with the students.
Observing students during our MOOC course is
much more complicated. Many students are
supposed to study the course, remote in place and
time. Individual tutoring, personalised interaction is
difficult to realise because of the massive character
of the course. As a consequence all kind of
interaction has to be automated. Observation should
be unobtrusive and not disturb but support the
learning process. Sending e-mail at regular times to
check if students are still alert was not an option for
us because expected response rate will be low. We
preferred observation and interaction integrated in
the learning process. We implemented the following
different kind of interactions. Some of these
interactions will be discussed in more detail in the
next section.
Keyboard Interaction, Mouse Klicks, and
Button Interaction. If there is no user response for
a longer time, probably the user is not active
anymore. Requests of a new learning material
nugget of a user have been logged and the time
between requests is computed. If the time lap
between requests surpasses a threshold, a special
alert will be generated, calling attention for the next
attractive part of the course such as a movie, or
simulation. If the student doesn’t show any reaction,
the assumption is that he left for a longer or shorter
time and an automated break will be generated. We
realise that a tailor made reaction could be possible
but then additional information is needed.
To go from one learning nugget to the other
students have to click buttons. But we implemented
much more interactions in our module such as replay
a simulation or a movie or requesting an example or
additional information or help. One of the reasons is
to keep the student active without boring him. These
are ad-on actions to the mainstream of the course.
A special type of buttons are emotion related
buttons to enable nonverbal emotional
communication to be discussed in the next section.
Students can express their emotional state by
pressing buttons with emotional words or visualizing
their emotional state by an icon or facial expression,
as displayed in Figure 2. Nonverbal expressions
reduce the ambiguity of words. The verbal label to
emotional representation can be used to compute the
valence and arousal score using DAL The possible
emotions can also be represented by their valence
and arousal coordinates on a two dimensional plane
as displayed in Figure 3. In case a student selects a
strong negative or passive emotional state, an
attractive intermezzo by movie or simulation will be
offered. In our experiments we found that students
have no problems to select icons showing negative
facial expressions. In face to face communication
students usually are not willing to show such
negative emotions. In case the student is studying on
an advanced level with assignments or further
readings the option to return to the basic level will
be offered to the student.
Text based Interaction. Natural language
processing is usually rather complicated. We found
that text input in chat session or tweets is usually not
grammatically correct, includes a lot of out-of-
vocabulary words, restarts and corrections. In our
case we process only one-liners and keywords. We
searched for emotional keywords, negations and
adjectives of tweets. In (Fitrianie, 2007) we created
lists of emotional words with valence and arousal
scores using the Dictionary of Affect in Language
(DAL) and Affective Norms for English Words
(ANEW). We preferred ANEW because the list of
emotional words is limited. To maximize the
matching between our words and the words from the
ANEW list, we have applied stemming for all
words. The score of each utterance is initialized to
neutral valence and arousal values All the words
from the utterance are looked up in the ANEW list.
If matches are found, a new score is computed by
New Didactic Models for MOOCs
averaging over the valence and arousal scores of the
matching words. We find that only 34% of the
utterances contained words from the ANEW list,
therefore the majority of the scores still indicated
neutral, valence and arousal.
Peers Discussion. Students are stimulated to use
social media for their discussion. At this moment
there is a focus on individual based, self-paced
learning systems. To stimulate cooperative learning
some modules of our MOOC is based on group
work. In one of the assignments of the module of the
management of the flooding disaster in Prague
students have to play different roles in the crisis
team. In (Rothkrantz, 2015) we designed a special
procedure to assign individual students to a team.
Different tasks are assigned to the group members
and the group assignments end with a common
written report.
One student is supposed to play the role of the
mayor of the city and head of the crisis team. In our
experiments it proves that the head of the crisis team
feels responsible for the performance of the crisis
team. Communication between team members was
stimulated; some team members were pushed to
activities. From a didactical point of view it proves
that peers played the role of tutor and mentor in face
to face learning reserved to the teacher.
Questioning-answering Systems. In the past we
developed an Elisa-like Q-A system (Fitrianie,
2003). The system enables a dialogue between
student-tutor. The focus of the system was to
simulate a human dialogue about users with mental
problems (student counsellor). To help students with
study problems (tutor) requires much more
knowledge of problematic interaction with the
learning content of the course. At this moment a first
prototype has been developed of a study information
system to help students to choose the right study. A
digital tutor is postponed to future work.
At start the learning material was grouped together
in in interactive e-learning nuggets ordered in a
linear way. Usually a module starts with an
introductory movie, or video or a combination,
followed by a short video lecture, assignments
followed by a new learning cycle. Many MOOCs
using Moodle, edX or similar learning management
tools or designed in this way. The characteristic
feature of these MOOCs is that the designer has full
control over the learning model. The students have
to follow the nuggets, learning modules according to
his design. He has the freedom to start or dropout of
the MOOCs
In recent e-learning design students have more
control of the order of the learning nuggets. We also
ordered the nuggets in a layered design. At the
bottom we have the basic nuggets (labelled by green
buttons). Every student should be able to follow
these nuggets composed of light way learning
material as movies, simulations and short pitch talks.
In the second layer we have the more complicated
learning topics and the assignments (labelled by blue
buttons). At the third level (orange buttons) we have
advanced readings, mathematical models and finally
we have the red buttons composed of group
In the navigation interface of the lessons the
coloured buttons are displayed and students have the
freedom to select the next buttons. In some test
experiments it proves that students not only want a
freedom of choice but also wants to express their
valuation and emotional state. We implemented
several other buttons, as facial expression and with a
text label as displayed in Figure 2 (Desmet, 2012),
(Fitrianie, 2007).
Figure 2: Personalized button interface.
The valence and arousal value of every button and
corresponding verbal label, can be computed using
the Whissel or DAL dictionary. In that dictionary
every emotional word has two coordinates
corresponding with valence and arousal. In Figure 3
we plot all the buttons by their valence and arousal
In case the student is in the advanced levels and
presses a sad expression, the system concludes that
the level is too advanced or the learning material is
boring, so a nugget on the basis level with
interesting visual material is offered.
In case a student on the basic level choses a sad
smiley, the system assumes the learning material is
boring a new topic or more challenging nuggets are
CSEDU 2017 - 9th International Conference on Computer Supported Education
Figure 3: Buttons plotted on a valence-arousal map,
valence corresponds with the horizontal axis and arousal
with the vertical axis.
offered, depending of the recent learning history. As
default a new topic will be offered because the
student didn’t select a nugget on a higher level. The
system tries to keep the student on-board.
FeedBackFruits a start-up company at the campus of
TUDelft developed educational tools. One of those
tools allows students to give feedback during
lectures via their smartphone. Later the company
generated a plugin for edX one of the MOOCs
consortia (Rothkrantz, 2015). We developed a
similar tool for our special MOOC. The tool allows
students to add comments via one-liners or
emotional words to the teaching material. The idea
was that students can provide in this way
information about their current emotional state and if
they like or dislike the study. Users of Facebook are
familiar with this (dis-)like comment.
In one of our small scale test experiments with
the current MOOC, 25 students Mathematics from
Delft University of Technology were requested to
provide at least one emotional word for every
learning nugget. This resulted in a total of 750
emotional words. Using the Whissel database, it was
possible to compute the coordinates on the valence
and arousal dimension. A plot of the used emotional
words can be seen in Figure 4. The similarity
measure of words corresponds with the Euclidean
distance measure. If users use words from the 2, 3
quadrant, the system starts a stimulating action as
stated before.
In (Fitrianie, 2007), we developed an icon-based
communication interface to represent concepts and
Figure 4: Plot of used emotional words during e-learning
on a valence-arousal scale.
ideas. Users can create messages to communicate
with others using a spatial arrangement of visual
symbols. We even developed a context free
grammar for our icon language based on an adapted
version of the Backus-Naur Form. If we restrict the
set of icons to the set of emoticons we can
communicate our motional state by using emoticons,
as displayed in Figure 5. The use of emoticons is
rather popular under users of social media.
Emoticons can be expressed by a picture or by one
or more characters. In (Rothkrantz, 2000, 2009,
2010) we researched the relationship between facial
expressions and emoticons in e-learning sessions.
Figure 5: Examples of emoticons.
In this paper we presented a new didactic model for
MOOCs. Most current didactic models are modified
copies of didactic models for face to face teaching-
learning. But the dropout rate of MOOCs is very
high and to our opinion an inappropriate didactic
model is one of the causalities.
The learning material is supposed to attract the
attention of students and is supposed to stimulate
students to learning activities. In face to face
learning a teacher has an important role to keep
students focussed and can use many didactic tricks
New Didactic Models for MOOCs
to reactive students in case they lose their attention
to the lessons. The impact of teachers in the
interactive learning process in MOOCs is reduced to
a minimum. Everything should be included in the
learning material.
We discussed our layered ordering of the
learning material to enable students to select their
individual learning paths of different difficulty
degree. But the learning material should include
much more stimuli to attract and keep the attention
of students and stimulate them to learning activities.
We presented a new didactic approach modelling the
learning activities and study interaction students
with the MOOC. The basic idea is that the learning
process of students is fuelled by a study liquid.
Several resources consume this energy liquid and
other resources provide new liquid. When the liquid
drops below some threshold, the learning process
will stop and probably a student drop-out. In face to
face learning it can happen that students lost their
attention and are not able to process the presented
material anymore. But students are not able to leave
the classroom and a didactically gifted teacher is
able to get students back in the learning process.
During MOOCs learning the situation is different.
We stressed that peers are important because they
are able to support students and because of social
interaction in peer groups they take part in the
Emotions play an important role in our model.
We realised that students should be attracted by the
learning material, they should like the interaction
and their motivation to follow the MOOCs should
not decrease. Based on our new didactic model we
designed some didactic stimuli to re-activate
students during the learning activities. The
communication of students with the Learning
Management System is modelled by using emotional
buttons, keywords, one-liners and by facial
expression. The recognition of a positive or negative
emotional state of the student will result in special
adaptation in the flow of the learning material.
Students can go back to lower level or more
attractive study activities as movies, simulations etc.
Next future we envision a digital tutor at the end
who is able to supervise and improve the learning
process of students. Such a tutor requires
improvement of the needed technology for e-
learning. We discussed already the option of
multipath, streaming and circular ways of learning.
But unfortunately many public domain tools of e-
learning are still based on a linear approach. But
recent developments around e-learning and MOOCs
are promising.
Armellini, A., Padilla Rodriguez, B.C. (2016). Are
massive Open Online Courses pedagogically
innovative? Journal of Interactive Online Learning,
Volume 14, Number 1.
Colman, D. (2013). MOOC interrupted: top 10 reasons our
readers didn’t finish a Massive Open Online Courser.
User-Centered Design Strategies for Massive Open
Online Courses (MOOCs). IGI Global.
Desmet, P. (2012). Designing emotions PhD thesis Delft
University of Technology.
Dillon, J., Bosch, N., Chetlur, M. (2016). Student emotion,
co-occurrence and dropout in a MOOC context. in:
Proceedings of the 9th International Conference on
educational data mining. June 29-July 2. Raleigh,
North Carolina. 1-5.
Fitrianie, S., Rothkrantz, L. (2007) A visual
communication language for crisis management.
International Journal of Intelligent Control and
Systems 12, 2.
Fitrianie,S., Wiggers, P., Rothkrantz, L.J.M. (2003). A
multi-modal Eliza using natural language processing
and emotion recognition. in: International Conference
on Text, Speech and Dialogue, 394-399.
Leony, D., Munoz-merino, P.J., Ruiperez-Valiente, J. A.,
Pardo, A. Kloos, C. D., (2015). Detection and
evaluation of emotions in Massive Open Online
courses. Journal of Universal Computer Science, vol.
21, no 5, 638-655.
Niett van der, A. G., Smith, J., Scherder, E. J. A.,
Oosterlaan, J., Hartman, E., & Visscher, C. (2015).
Associations between daily physical activity and
executive functioning in primary school-aged children.
Journal of science and medicine in sport, 18(6), 673-
Rothkrantz, L. Wojdel, A. (2000). A text based talking
face. in: Proceedings Conference on Text, Speech and
Dialogues, pp 231-2140.
Rothkrantz, L., Datcu, D., Chiriacescu, I., Chitu, A.
(2009). Assessment of the emotional states of students
during e-Learning, in: Proceedings of the
International Conference on e-Learning and
Knowledge Society, pp. 77-82.
Rothkrantz, L., Datcu, D. (2010). Assessment of Emotion
States During e-Learning. Communication &
Cognition, 43 (1), 81.
Rothkrantz, L.J.M., (2015). How Social Media Facilitate
Learning Communities and Peer Groups around
MOOCS. International Journal of Human Capital and
Information Technology Professionals, 6, 1-13.
Rothkrantz, L. (2016a). Dropout rates of regular courses
and MOOCs, in: 8th International Conference on
Computer Supported Education. 21-23 April, Rome.
Rothkrantz, L. (2016b). Didactic Model for e-Learning
and Regular Courses. in: International Conference on
e-Learning. 8-9 September, Bratislava. pp. 156-161.
Tak-ming Wong, B. (2016). Factors leading to effective
teaching of MOOCs. Asian Association of Open
Universities Journal, Vol. 11 Iss 1 pp. 105 118.
Tettegah, S. Y., McCreery, M.P.. (EDS), (2016).
Emotions, Technology, and Learning. Elsevier.
Weiss, J., Nolan, J., Hunsinger, J., Trifonas, P., (2006).
The international handbook of virtual learning
environments, vols. I & II, Springer, The Netherlands.
CSEDU 2017 - 9th International Conference on Computer Supported Education