Motivational Strategies to Support Engagement of Learners in
Serious Games
Ramla Ghali, Maher Chaouachi, Lotfi Derbali and Claude Frasson
Département d’informatique et de recherche opérationnelle, Université de Montréal,
2920 Chemin de la Tour, Montréal, Canada
Keywords: Motivation, Engagement, Serious Games, EEG, ARCS Model.
Abstract: The use of Video Games as learning tool is becoming increasingly widespread. Indeed, these games are well
known as educational games or serious games. They mainly aim at providing to the learner an interactive,
motivational and educational environment at the same time. In order to better study the necessary
characteristics for the development of an effective serious game (both motivational and educational), we
evaluated the physiological responses of participants during their interaction with our serious game, called
HeapMotiv. We essentially measured a physiological index of engagement through an EEG wifi headset
and studied the evolution of this index with the different missions and motivational strategies of HeapMotiv.
Focusing on the gaming aspects, the analysis of this engagement index behavior showed the significant
impact of motivational strategies on skills acquisition and motivational experience. An agent-based
architecture is proposed as a methodological basis for serious games conception.
1 INTRODUCTION
The success of Computer-Based Education (CBE)
over the past decades, established a trend towards
the development of new engaging, immersive and
effective environment. In this context, the
development of Serious Games (SGs) intended to
train and educate learners within an enjoyable and
challenging environment represents a new attractive
approach for technology-mediated learning (Garris
et al. 2002, Prensky, 2001, Johnson et al. 2008).
However, like in the other CBE systems, the
interaction in SGs can be entertaining and
motivating for the learners, or annoying and
frustrating (Malone et Lepper 1987). SGs should
support learners during interaction for instance with
motivational strategies such as actions (or tactics)
suitable to scaffold learners’ motivation towards
tasks and goals. Resulting learning would be easier,
faster, more enjoyable, more self-directed, and more
effective. What is really surprising is that very few
studies concern motivational strategies.
From a conceptual point of view, aligning
learning and fun may be a difficult challenge for the
designers. Prioritizing playful aspects over learning
content in order to motivate and engage learners,
risks to make them more focused on the gameplay
and less concentrated on the learning content.
However, a game design which is based on intensive
learning phases, can be in contradiction with
learners’ expectations and may rapidly annoy them
(Gunter et al. 2006). Thus, striking an appropriate
balance between a right learning mode and
entertaining aspects in the game constitutes an
important challenge for designing effective SGs.
Hence, the assessment of learners’ experience with
SGs may be a key factor in the success (or failure) of
such systems, as it will allow the designers to
improve and adjust adequately the game.
Moreover, the idea of supporting SGs with
motivational strategies can also be a promising
target for SGs designers. These strategies are based
on the idea of giving learners more control in
adapting and adjusting the pace and the game
components to their skill level. The aim of these
strategies is to enhance SGs’ capabilities to maintain
learners in an appropriate level of engagement and
motivation (derbali et al. 2012, Huang et al. 2010).
Nonetheless, measuring the impact and effectiveness
of such interventions can also be an important task
for the SGs designers.
To that end, we propose a sensor-based approach
using an electroencephalogram device (EEG) to
analyze the behavior and the engagement of learners
518
Ghali R., Chaouachi M., Derbali L. and Frasson C..
Motivational Strategies to Support Engagement of Learners in Serious Games.
DOI: 10.5220/0004823305180525
In Proceedings of the 6th International Conference on Agents and Artificial Intelligence (ICAART-2014), pages 518-525
ISBN: 978-989-758-015-4
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
while interacting with a SG which embeds
motivational strategies. We define a computed
engagement index able to identify trends in learners’
behavior. The objective of this research is twofold:
first we investigate whether the physiological
engagement index is effective to provide valuable
information about learners’ behavior. Second, we
study the impact of SGs design and motivational
strategy on learners’ motivation and engagement.
The organization of the paper is as follows: the
first section presents previous works in similar
fields. The second section presents our SG
environment. In the third section, we present the
details of the experimental procedure. Finally, the
fourth presents the results and a discussion about the
impact of our findings in the field of SGs design.
2 PREVIOUS WORK
Recently, a large body of research was directed
towards improving learners’ experience and
interaction in learning environments. Affective and
social dimensions were considered in these
environments to provide learners with intelligent and
adaptive interaction (Picard et al. 2001, D’Mello et
al. 2009). These approaches were either based on
empirical observations and studies of learners’
behavior or on correlation between physiological
cues and interaction data. They have been used to
feed different models aimed to improve the learning
environment design by giving real-time adaptive
interaction or adjustments according to learners’
states.
Various ranges of sensors and devices, such as
skin conductance, heart rate, electromyogram,
camera and respiration, combined with machine
learning models were generally used to build up
sensor-based models capable of detecting learners’
reactions. They also support educative content with
appropriate interventions (Conati 2002, Predinger et
Ishizuka 2005).
From a motivational standpoint, the review of
literature demonstrates that several tools and
frameworks for motivational assessment and support
were provided for learning environments (Boyer et
al. 2008, Vincente and Pain 2002, Rebolledo et al.
2011, Johnson et al. 2005, Ryan et al. 2006). For
example, games appeared as the most appropriate
tool to motivate people (Whitton 2007, Tychsen et
al. 2008). Besides, millions of people are captivated
by games. They spend their time and money to play.
Therefore, the potential for combining games and
learning becomes ever more significant. Many
experimental studies state that computer games can
provide new ways of learning (Coles et al. 2007). In
fact, they show that educational games or serious
games are capable of helping players to learn.
Johnson and colleagues (2005) reported that the
designers of educational games employ a range of
artificial intelligence techniques, (controlling the
behavior of non-player characters, providing
performance feedback, etc.) to promote long-term
user engagement and motivation (Johnson et al.
2005). Ryan and colleagues (2006) stated that the
motivational pull of computer games is attributed to
the combination of optimal challenge and
informational feedback (Ryan et al. 2006). However,
few studies tackled to what extent these strategies
impacted learner’s motivation or the way they
impacted learner’s objectively. In the present work,
we propose to assess learner’s engagement using an
engagement index measure (EEG mental
Engagement Index) combined with subjective self-
reporting estimation of motivation, to analyze how
learners reacted to motivational strategies. For that
we developed a game in which it was possible to
assess learner’s reactions in different missions
without and with motivational strategies.
3 MOTIVATIONAL STRATEGIES
AND HEAPMOTIV
3.1 ARCS Model and Motivational
Strategies
In his ARCS model (Keller 2010), John Keller used
existing research on psychological motivation to
identify four categories of motivation: Attention,
Relevance, Confidence, and Satisfaction. Keller’s
model has been used in learning, training and games
(Gunter et al. 2006, Dempsey et Johnson 1998).
Therefore, it is of particular interest in our study.
Keller also, defines four different motivational
strategies associated to each category of his ARCS
model (Keller 2010): Attention getting strategies,
Relevance producing strategies, Confidence building
strategies, and Satisfaction generating strategies.
These theoretical strategies tend to (1) find the right
balance between consistency and novelty; (2) find
out which tactics to use and how to adjust them for
the learners; (3) build relevance in the instruction by
connecting it to the learners’ backgrounds, interests,
and goals; enhance learners’ confidence by allowing
them to control some situations; etc.
In this paper, we use the ARCS model as a basis
MotivationalStrategiestoSupportEngagementofLearnersinSeriousGames
519
to implement different motivational strategies in our
game HeapMotiv as described hereafter, more
precisely in the second version of the game,
HeapMotivV2. For example, an Attention getting
strategy is based on submitting challenges as time
and errors constraints: (1) a time constraint for each
level of difficulty: unlimited, 90 seconds, and 45
seconds for easy, normal, and hard level
respectively, and (2) wild cards representing the
number of accepted errors committed by the player:
unlimited, 3 wild cards, and 1 wild card for easy,
normal, and hard level respectively. A Relevance
producing strategy has been designed before the
beginning of each mission by presenting an
instructional video to explain and inform learners of
the main goal of the mission and its relation to the
binary heap data structure. Then, this version of
HeapMotiv integrates a Confidence building
strategy which allows learners to control the level
of each mission (easy, normal, and hard) and to
possibly repeat the mission with the same or a
different level (at most six trials). Finally, a virtual
companion “Sinbad” applies a Satisfaction
generating strategy by providing feedback on
learners’ performance when they find a way out of
the labyrinth and meet “Sinbad”. A detailed
description of these motivational strategies is
contained in (Derbali et al. 2013).
3.2 HeapMotiv
For the purpose of experimentations we have built
HeapMotiv; a serious game intended to teach binary
heap data structure. This SG is a 3D-labyrinth that
has many routes with only one path that leads to the
final destination (Fig. 1). Along the paths of the
labyrinth, several information signs are placed to
help the learners to find the correct destination.
Learners have to play different 2D missions aiming
to entertain and educate them about some basic
concepts of binary heap, before obtaining
information signs.
In order to study the impact of motivational
strategies, we have implemented two versions of this
game: HeapMotivV1 and HeapMotivV2, which are
intended respectively to control group (CTR) and
experiment group (EXP) during the experiment. In
HeapMotivV1, players interact with the game
without introducing the motivational strategies.
However, in HeapMotivV2, the game have been
reproduced based on the ARCS model (Keller
2010), and incorporated mostly some motivational
strategies as described in the previous section.
In its current implementation, HeapMotiv is
Figure 1: HeapMotiv environment.
composed of three missions: the first two missions
(Tetris and Shoot) are designed to build a binary
heap and maintain the heap property, whereas the
third mission (Sort) is designed to show basic
operations for a binary heap (insertion and deletion)
and the heap-sort algorithm. An overview of these
missions is presented in figures 2, 3 and 4.
Tetris is based on traditional Tetris game. A
learner has to move nodes during their falling using
the arrows to fill a binary tree without violating the
heap property. In the first version HeapMotivV1,
Tetris is over when the tree is completely filled. In
the second version HeapMotivV2, players are
penalized (time constraint or loss of wild cards)
when they make mistakes and Tetris may be over
without filling the whole tree.
Figure 2: Tetris mission.
Shoot is based on shooter games. A learner has
to spot violations of shape and heap properties, and
then has to fix these violations by shooting
misplaced nodes. Shoot is over when all errors are
detected or balls are exhausted. In addition, the
mobility of nodes is an additional constraint in
HeapMotiV2.
Figure 3: Shoot mission.
ICAART2014-InternationalConferenceonAgentsandArtificialIntelligence
520
Sort begins by building a binary heap out of the data
set, and then removing the largest item and placing it
at the end of the partially sorted array. It is a
comparison-based sorting algorithm to create a
sorted array. Sort mission involves the discovery of
rules to insert and delete a node correctly.
Figure 4: Sort mission.
4 EXPERIMENT
An experimental protocol was established where
participants were invited to play our serious game
HeapMotiv. Following the signature of a written
consent form, each participant was placed in front of
the computer monitor to play HeapMotiv. During
the experiment, the participant was equipped with an
EEG headset. EEG recordings were adopted using
the multi-channel wireless portable device, called
EMOTIV EPOC. This device is a high resolution,
neurone-signal acquisition and processing wireless
neuroheadset. It produces a reliable and valid EEG
data collected from 14 channels, each based on
saline sensors. The EEG monitoring using this
device is as accurate as other conventional EEG
systems (Stytsenko et al. 2011).
In the interest of measuring the learner’s
engagement index from his brainwaves and studying
the evolution of these measures in different
situations of HeapMotiv game, a baseline was
computed before starting the game. This technique
consists of calculating the average of all the EEG
channels during a fixed period of time (5 minutes).
10 pre-test and 10 post-test quizzes about general
knowledge of the binary tree and the heap data
structure were also administered to compare
learners’ performance regarding the knowledge
presented in HeapMotiv. The pre-test and post-test
questions were different and balanced. Besides, an
Instructional Materials Motivational Survey (IMMS)
and a Self-Report Engagement (SRE) were
administrated after each mission to assess learner’s
motivation and engagement, respectively. IMMS is
derived from four categories of ARCS model of
motivation (Keller 1987). An illustration of the
experimental process is shown in the following
Figure.
Figure 5: Experimental protocol.
As mentioned previously, EMOTIV can record
14 EEG channels based on the International 10-20
locations (AF3, F7, F3, FC5, T7, P7, O1, O2, P8,
T8, FC6, F4, F8, AF4). The EEG recordings were
then managed in real-time by an EEG-capture tool
developed in our lab using the EMOTIV software
development kit. The developed tool provided
temporal measurements of the user’s signals and
collected data was pre-processed and synchronized
with HeapMotiv log file. An artefact rejection
technique based on a threshold on epoch power was
employed in the capture software to remove noise
and data contaminated from body-movement or eye
blinks. EEG data were decomposed into 1-second
length segment overlapped by 0.5 second. The
resulting segments were multiplied by a Hamming
window function to decrease spectral leakage. A
real-time Fast-Fourier-Transform (FFT) was used to
extract 1-Hz bin power data segment for each EEG
site location.
4.1 Computing EEG Engagement
Index
In this study we used an EEG engagement index
developed by Pope and colleagues at NASA (Pop et
al. 1995). This index showed a great reliability in
switching between piloting mode. It was also, used
as criteria for adaptive and automated task allocation
(Chaouachi et al. 2010). In assessing users’
engagement within educational context, this index
showed to provide an efficient assessment of
learners’ mental vigilance and cognitive attention
(W et al. 2010). This index uses three EEG bands:
Theta (4–8 Hz), Alpha (8–13 Hz) and Beta (13–22
Hz). It has the following equation:
⁄
(1)
MotivationalStrategiestoSupportEngagementofLearnersinSeriousGames
521
To compute this index, we proceeded as follows.
First, we applied a Fast Fourier Transformation to
convert the EEG signal from each active site into a
power spectrum. Then, we produced the EEG bands
by summing Bin powers (the estimated power over 1
Hz) together with respect to each band. Next, we
mean the sum of band power computed from each
measured scalp site. Finally, we obtain the EEG
mental engagement index at instant T by applying a
moving average technique. This technique consists
of averaging each engagement index within a 40
seconds sliding window preceding instant T. This
procedure was repeated every 2s and a new 40s
sliding window were used to update the index.
4.2 Participant’s Repartition
A total of 20 participants (8 females and 12 males)
were invited to play our SG HeapMotiv. The sample
mean age was 24.8 ± 5.94 years. Participants were
recruited from the University of Montreal and had
no prior knowledge about heap data structure.
Participants were randomly distributed to the control
group (CTR: n=10, HeapMotivV1), or to the
experimental group (EXP: n=10, HeapMotivV2).
5 STATISTICAL RESULTS
5.1 Motivation and Learning
A paired sample t-test was conducted to compare the
reported motivational scores in the experimental and
control groups. Results showed a significant
difference in the motivational scores reported by the
participants in the experimental group (EXP:
M=61.86, SD=7.01) and the control group (CTR:
M=51.1, SD=8.39); t(18)=-3.115, p=0.006. The
previous result highlights the positive impact of the
motivational strategies embedded in HeapMotivV2
on the subjective measure of motivation of the
experimental group by comparison with learners in
the control group who did not benefit from the
motivational strategies in HeapMotivV1.
A paired sample t-test was also conducted to
compare learners’ scores in the pre-test and the post-
test. There was a significant difference between all
learners in the pre-test (M=9.5, SD=3.1) and the
post-test (M=13.1, SD=3.6); t(38)=-3.335, p=0.02.
After playing and finishing the game, the number of
correct answers is significantly higher. This result
evidences that learners can improve their knowledge
even in a complex field (i.e. heap data structure).
Although a paired sample t-test result was not
significant between the correct answers of the two
groups (EXP: M=14.2, SD=2.2; CTR: M=12,
SD=4.61; t(18)=-1.36, p=n.s.), participants of the
EXP group have, in general, outperformed those of
CTR group. The addition of motivational strategies
in HeapMotivV2 could explain the considerable
increase of the number of correct answers after
finishing the game.
5.2 EEG Engagement Index Evaluation
Next, we conducted statistical tests to study the
behavior of the computed EEG engagement index
with regard to self-reported engagement and ARCS
motivation. A positive significant correlation
(r=0.54, p=0.03) was found between learners’ self-
reported engagement and the respective mean EEG
engagement index. This preliminary result
confirmed that the measure of engagement used in
this experiment can reflect reliably the learners’
perception of their own engagement during the
game. In addition, the correlation run between the
self-reported engagement and the motivational
scores reported after each mission has been
significant (Tetris: r=0.743, p=0.000; Shoot:
r=0.446, p=0.049; Sort: r=0.488, p=0.029).
However, non-significant correlation was found
between the motivational scores and the EEG
physiological measures. This result is not very
surprising as the relationship between motivational
state and mental engagement is complex and
difficult to estimate by the learners themselves. We
can however extract a clear trend, as we will detail
hereafter with the measures obtained from the EEG.
5.3 Learners’ Engagement and
Motivational Strategies
Figures 6 (a), (b), and (c) depict the behavior of
EEG engagement index of two learners of different
groups. According to respective performance scores
obtained we distinguish a mean player in the CTR
group and a mean player in the EXP group. We
show their respective engagement index for the
Tetris mission (Fig. 6(a)), the shoot mission (Fig.
6(b)), and the Sort mission (Fig. 6(c)). A closer look
to the figure shows that the EEG engagement index
of EXP player is clearly above the baseline value
throughout all missions. This result confirms also
our first goal which was validating such an index for
assessing learners’ performance progression. The
following finding confirms also, the positive impact
of motivational strategies when comparing an
average player in the EXP group to an average
ICAART2014-InternationalConferenceonAgentsandArtificialIntelligence
522
player in the CTR group.
Figure 6: Engagement index (solid line) and Baseline
values (dashed line) for EXP learner and CTR learner
while trying Tetris (a), shoot (b) and sort (c).
For each one of the three missions, we compared
the difference of learners’ engagement index
behavior in the two groups. The learner’s average
engagement index was extracted and the difference
between this value and the corresponding learner’s
baseline was computed. Results of an independent
samples t-test showed that, for each mission, the
difference between the engagement index and the
baseline was significantly higher in the EXP group:
Tetris (t(18)=-2.262, p=0.036), Shoot (t(18)=-2.819,
p=0.011) and Sort (t(18)=-2.496, p=0.023).
This result highlights the significant impact of
the motivational strategies on the objective measure
of engagement which seems to enhance learners’
mental alertness and vigilance. However, a general
decrease in engagement index differentiates in the
last mission (we can see this trend in Figure 6(c)).
Our research aims at determining the role of
motivational strategies in supporting learners’
motivation and engagement through the different
missions. As a matter of fact, the three missions of
Table 1: Mean (Standard Deviation) of the difference
between the engagement index and the baseline for all the
learners.
Mission
Group
CTR EXP
Tetris 0.08 (0.2) 0.27 (0.15)
Shoot 0.03 (0.14) 0.28 (0.22)
Sort 0.02 (0.17) 0.17 (0.06)
HeapMotiv were designed differently: Tetris and
Shoot missions had playful aspects which are
theoretically and intrinsically attractive for the
player. However, the Sort mission had non-game-
like characters and involved more reflection effort. It
also, required a certain level to master previous
educative content. Repeated ANOVA measures
determined that average engagement index differed
statistically and significantly between the three
mission for both groups (F(2, 38) = 3.35, p=0.042).
Post hoc tests using the Bonferroni correction
revealed that learners’ engagement index was
slightly but, not significantly reduced between the
Tetris and Shoot mission (M=0.54, SD=0.17 vs.
M=0.53, SD= 0.19, respectively, p=n.s). However,
Sort mission reduced significantly the engagement
index (M=0.44, SD=0.02, p=0.013) from the first
two missions. This result highlights the fact that the
game missions’ design might have affected the
learners’ engagement. The variation of the index
showed that it had significantly higher value with
game-like missions and lower value in the last
mission. In terms of motivational strategies,
participants of EXP group have been more engaged
than those of CTR group during Sort mission. Used
strategies seemed to have then, a slight role in
maintaining learners’ engagement when playful
aspects are almost inexistent.
In the next section we discuss the implication of
our findings in a more generic methodology to build
SGs using an Agents-Based architecture.
6 AGENT-BASED SERIOUS
GAMES
Our above results showed that motivational
strategies have a positive impact not only in
supporting overall motivation and engagement, but
also, in attaining high performance during the third
mission. Indeed, players of EXP group performed
MotivationalStrategiestoSupportEngagementofLearnersinSeriousGames
523
better than those of CTR group when they played
Sort mission. The design of serious games is
complex and expensive. It involves a lot of
resources: humans, techniques, funds, ergonomics,
gameplay, etc. However, special attention should be
given to aspects which trigger motivation and
engagement of the player. Motivational strategies
are essential, and in order to benefit from strategies
which can not only dynamically adapt to the game
but also improve themselves, it is useful to have a
structure of agents.
Agents provide a flexible framework for SGs
design. Agents can be used to respond quickly and
autonomously to variable game situations according
to the learner, and trigger appropriate strategy. They
can also learn from the behavior of the learner and
complete or improve existing strategies. The
following architecture (Figure 7) shows the role of
each component.
Figure 7: Motivational agent.
- The EEG acquisition module is in charge to
measure the different brainwaves and extract
the engagement index. It includes also personal
information about the player (age, level of
dexterity, historic)
- The motivational strategies agent contains a
variety of parameterized strategies associated
with information provided by EEG. Each
strategy is evaluated and weighted according to
the degree of performance obtained by the
learner after the interaction with the game. The
agent computes also a type of learner
associated with the strategy.
- The parameterized game module consists of
different characteristics of a game such as
environments (scenes), artefacts, periods of
reactions, difficulty level, re-initialization
procedures. It selects the environment (game
scenes) to be presented to the learner according
to the selected motivational strategy.
- The motivational agent receives the
engagement index from the EEG module and
selects an adequate strategy. The selection
integrates the IMMS evaluation and the self-
report questionnaire (cf section 4).
- The evaluation module is in charge to control
the evolution of learner’s performance
resulting from a given strategy. The weight of
the motivational strategy is updated after
analysis of this performance.
The advantage of this framework is that the
motivational strategies can be improved with the
time and also multiple new strategies can be
introduced to complete the efficiency of SGs.
7 CONCLUSIONS
This paper introduced a new assessment metric for
learners’ engagement in SGs. Results obtained from
an experimental study, showed promising results in
assessing learners’ engagement and motivation.
Deeper analysis of our results showcased also, the
importance of motivational strategies to enhance
learning outcomes and to support the lack of playful
aspects in some tasks of our SG, which impacts
learners’ engagement. Our future work will involve
measuring the impact of SGs design as well as,
players’ profile on the cognitive reasoning level.
Further detailed strategies will also, be considered to
distinguish motivational factors and situations in our
SG that support learners’ performance.
REFERENCES
Boyer, K., Phillips, R., Wallis, M., Vouk, M., and Lester,
J., 2008. ‘Balancing Cognitive and Motivational
Scaffolding in Tutorial Dialogue’. Proc. of the 9th
International Conference on Intelligent Tutoring
Systems.
Coles, C. D., Strickland, D. C., Padgett, L., and Bellmoff,
L., 2007.‘Games that Work: Using computer games to
teach alcohol-affected children about fire and street
safety’, Research in Developmental Disabilities, 28,
(5).
Conati, C., 2002. ‘Probabilistic Assessment of User’s
Emotions in Educational Games’, Applied Artificial
Intelligence, 16, pp. 555-575.
ICAART2014-InternationalConferenceonAgentsandArtificialIntelligence
524
Chaouachi, M., Frasson, C. 2010. ‘Exploring the
relationship between learner EEG mental engagement
and affect’. International Conference on ITS, pp.291-
293.
Dempsey, J. V., and Johnson, R. B., 1998. ‘The
development of an ARCS gaming scale’, Instructional
Psychology, 25, (4), pp. 215-221.
Derbali, L., Frasson, C. 2012. ‘Exploring the effects of
Prior Video-Game Experience on Learner's Motivation
during Interactions with HeapMotiv’. International
conference on ITS, Greece. pp129-134.
Derbali, L., Ghali, R., Frasson, C. 2013. ‘Assessing
Motivational strategies in Serious Games using
Hidden Markov Models’, The 26th International
FLAIRS conference.
D'Mello, S., and Graesser, A., 2009. ‘Automatic Detection
of Learner's Affect From Gross Body Language’,
Applied Artificial Intelligence, 23, (2), pp. 123-150.
Freeman, F. G., Mikulka, P. J., Prinzel, L. J., and Scerbo,
M.W., 1999. ‘Evaluation of an adaptive automation
system using three EEG indices with a visual tracking
task’, Biol Psychol, 1999, 50, (1), pp. 61-76.
Garris, R., Ahlers, R., Driskell, J. E., 2002. ‘Games,
motivation, and learning: a research and practice
model’, Simulation & Gaming, 33, (4), pp. 441-467.
Gunter, G. A., Kenny, R. F., and Vick, E. H., 2006. ‘A
case for a formal design paradigm for serious games’,
International Digital Media & Arts Association, 3, (1),
pp. 93-105.
Huang, W. H., Huang, W. Y., and Tschopp, J. 2010.
‘Sustaining iterative game playing processes in
DGBL: The relationship between motivational
processing and outcome processing’, Comput. Educ.,
55, (2), pp. 789-797.
Johnson, W. L., Wu, S., 2008. ‘Assessing Aptitude for
Learning with a Serious Game for Foreign Language
and Culture’, in Editor (Eds.): ‘Book Assessing
Aptitude for Learning with a Serious Game for
Foreign Language and Culture’, pp. 520-529.
Johnson, W. L., Vilhjalmsson, H., and Marsella, S.,
2005.‘Serious Games for Language Learning: How
Much Game, How Much AI?’. Proc. of the conference
on AIED.
Keller, J. M., 1987. ‘Development and use of the ARCS
model of motivational design’, Journal of
Instructional Development, 10, (3), pp. 2-10.
Keller, J. M., 2010. ‘Motivational Design for Learning and
Performance: The ARCS Model Approach’ (Springer,
2010. 2010).
Malone, T. W., Lepper, M. R., 1987. ‘Making learning
fun: A taxonomy of intrinsic motivations for learning’,
in Snow, R. E., and Farr, M. J. (Eds.): ‘Aptitude
learning and instruction’, pp. 223-253.
Picard, R. W., Vyzas, E., and Healey, J., 2001. ‘Toward
machine emotional intelligence: analysis of affective
physiological state’, Pattern Analysis and Machine
Intelligence, IEEE Transactions on, 23, (10), pp.
1175-1191.
Pope, A. T., Bogart, E. H., and Bartolome, D. S.,
1995.‘Biocybernetic system evaluates indices of
operator engagement in automated task’, Biological
Psychology, 1995, 40, (1–2).
Prensky, M., 2001: ‘Digital Game-Based Learning’ , New
York: McGraw Hill.
Prendinger, H., and Ishizuka, M., 2005. ‘The Empathic
Companion: A Character-Based Interface That
Addresses Users' Affective States’, Applied Artificial
Intelligence: An International Journal, 19, (3), pp.
267-285.
Rebolledo-Mendez, G., Luckin, R., and Boulay, B., 2011.
‘Designing Adaptive Motivational Scaffolding for a
Tutoring System’, in Calvo, R.A., and D'Mello, S.K.
(Eds.): ‘New Perspectives on Affect and Learning
Technologies’, Springer New York.
Ryan, R. M., Rigby, C., and Przybylski, A., 2006. ‘The
motivational pull of video games: A self-
determination theory approach’, Motivation &
Emotion, 30, (4), pp. 344-360.
Stytsenko, K., Jablonskis, E., Praham, C., 2011.
‘Evaluation of consumer EEG device Emotiv EPOC’,
MEi:CoSci Conferences 2011, Ljubljana.
Tychsen, A., Hitchens, M., and Brolund, T., 2008.
‘Motivations for play in computer role-playing
games’: ‘Proceedings of the 2008 ACM Conference on
Future Play’, pp. 57-64.
Vicente, A., and Pain, H., 2002. ‘Informing the detection
of the students' motivational state: an empirical
study.’. Proc. Proceedings of the 6th International
Conference on ITS, Biarritz, France and San Sebastian,
Spain, June 2-7.
Whitton, N., 2007.‘Motivation and computer game based
learning’, Proceedings of the Australian Society for
Computers in Learning in Tertiary Education,
Singapore.
MotivationalStrategiestoSupportEngagementofLearnersinSeriousGames
525
