Received: 28 October 2019 | Accepted: 1 March 2020

DOI: 10.1002/cae.22227

RESEARCH ARTICLE

Adapting gamified learning systems using educational

data mining techniques

Lamya F. Daghestani1 | Lamiaa F. Ibrahim2,3 | Reem S. Al‐Towirgi1,4 |

5,6
Hesham A. Salman

1
Computer Science Department, Faculty of Computing and Information Technology, King Abdulaziz University Jeddah, Jeddah, Saudi Arabia
2
Head of Computer Science Department, Faculty of Graduate Studies for Statistical Research, Cairo University, Giza, Egypt
3
Information Technology Department, Faculty of Computing and Information Technology, King Abdulaziz University Jeddah, Jeddah, Saudi Arabia
4
Department of Computer Science, Faculty of Computing and Information Technology, Taif University, Taif, Saudi Arabia
5
Department Management Information Systems, Higher Institute of Computer and Information Technology, Alshrouk Academy, Cairo, Egypt
6
Information System Department, Faculty of Computing and Information Technology, King Abdulaziz University Jeddah, Saudi Arabia

Correspondence
Lamiaa F. Ibrahim, Head of Department Abstract
of Computer Science, Faculty of Graduate Artificial intelligence (AI) provides opportunities to improve the effectiveness of
Studies for Statistical Research, Cairo
University, 5 Ahmed Zewail St. Tharwat
e‐learning by increasing students' engagement. Adaptive e‐learning uses AI to
previously Orman, Giza 12613, Egypt. support individual learners by responding to their different learning needs which
Email: [email protected] can be determined by analyzing their navigation history of e‐learning systems using
data mining methods. Educational data mining (EDM) discovers new patterns of
learning and teaching to facilitate the process of decision‐making to serve education
improvement. Gamification is another way of increasing students' engagement by
using game elements in a nongame context. In this paper, the gamification
technique and EDM methods were used in combination with adaptive learning to
increase the students' engagement and learning performance. An adaptive gamified
learning system (AGLS) was developed which combines gamification, classification,
and adaptation techniques to increase the effectiveness of e‐learning. This paper
studies the impact of gamification and adaptive gamification on the effectiveness of
e‐learning through increasing students' engagement and learning performance.
AGLS was applied to the data structure course. Results showed that adaptive
gamification has a positive effect on students' engagement and learning
performance compared to just gamification.

KEYWORDS
adaptive system, data structure, e‐learning, engagement, gamification, gamified course

1 | INTRODUCTION learning applications. AI serves education by providing an

adaptive e‐learning environment to support individual
The development of artificial intelligence (AI) techniques in learners and increase their engagement. This is done by
the last few years made it suitable for computer‐assisted responding to different types of needs of the students [47].

568 | © 2020 Wiley Periodicals, Inc. wileyonlinelibrary.com/journal/cae Comput Appl Eng Educ. 2020;28:568–589.
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 569

Students' needs can be determined by analyzing their fun and effective and applies them to any learning en-
navigation through e‐learning systems. This information vironment. Gamification was used in traditional learning
creates massive datasets, which need to be analyzed using such as developing a leaderboard containing the best
computational and statistical methods that can intelligently students and giving a ribbon or badge to excellent
assist in discovering new patterns of learning and teaching students. In e‐learning, gamification was applied to
[33] and [56]. This leads to a new field that applies data courses by adding some elements of video games to make
mining methods on an educational dataset called educational them more motivating.
data mining (EDM) [8]. For motivating students through gamification, a
One of the most successful methods used to increase critical issue emerges from the fact that people do
students' engagement is games which can create enga- not have the same expectations and responses to
ging learning for students [23]. gamification elements [65]. To achieve the desired goal,
Werbach and Hunter [64] provided a game element list the gamification elements need to be adapted to
and organized them into three levels of abstraction, that fit distinct types of students. Therefore, adaptive
is, dynamics, mechanics, and components. Dynamics re- learning and EDM can be used in cooperation with
present the big picture aspects such as constraints, emo- gamification to increase the effectiveness of e‐learning
tions, narratives, progression, and relationships. Game and raise student motivation which is the objective of
dynamics also designed rules that described how to collect this paper.
elements and reach achievements [27,54]. Adaptive e‐learning platforms provide personalized,
Gamification has been widely employed in the edu- flexible learning via monitoring of student progress and
cational domain over the past 8 years since the term performance, and subsequent provision of an individualized
became a trend [62]. However, the literature states that learning path containing various resources [12].
gamification still lacks a formal definition to support the Although many researchers have used gamification in
design and analysis of gamified strategies [22,53,58]. e‐learning to increase students' engagement and learning
Mechanics are the processes that drive the action performance, they did not consider the differences
forward such as challenges, chance, competition, between students' personalities and their responses to
cooperation, feedback, and reward. Components are the gamification elements. This study focused on whether
specific instantiations of mechanics and dynamics such classifying students based on their navigations through
as points, levels, badges, avatar, achievement, content gamified learning systems and adapting the gamification
unlocking, and leaderboards. The main goal of gamifi- elements based on their playing type could be used
cation is to increase user engagement with these systems, to increase the effectiveness of e‐learning through
without undermining its credibility, by combining enhancing the desired goal of gamification.
intrinsic and extrinsic motivations [61]. Intrinsic moti- The main objective of this study is to investigate the
vations come about when the user decides whether to impact of gamification and adaptive gamification on the
make an action such as competition, altruism, coopera- effectiveness of e‐learning through increasing students'
tion, sense of belonging, love, and aggression. Extrinsic engagement and learning performance for data structure
motivations occur when something leads the user to course students at King Abdulaziz University in
make an action such as points, levels, badges, classifica- Saudi Arabia.
tions, awards, and missions. The success of gamification This paper is organized as follows: Section 2 presents
on education depends on its potential to engage students the literature review of using data mining in e‐learning
in learning activities because engagement has been and the related studies on using gamification in
proved to be positively correlated with outcomes of stu- e‐learning. Section 3 presents the methodology used in
dent success including satisfaction, persistence, and the design of the proposed system. Discussion of results
academic achievements [28]. Game mechanics and and performance evaluation are presented in Section 4.
dynamics are used to trigger the player's emotions. Section 5 discusses the conclusions, contributions,
Gamification has recently emerged to increase user limitations, and future work.
engagement by using these mechanics and dynamics
in a serious context [25]. It also promotes student
engagement by incorporating competition, leaderboards, 2 | LITERATURE REVIEW
leveling opportunities, achievement rewards, and other
game mechanics in classroom activities. This section is divided into two subsections. The first
To improve knowledge in different fields, gamifica- subsection discusses the related work in EDM. The
tion can be used instead of creating an actual game. second subsection discusses the related work in
Gamification infuses elements that make video games gamification.
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
570 | DAGHESTANI ET AL.

2.1 | Educational data mining 2.1.2 | Clustering

EDM can be used to analyze data from any interactive Romero et al. [51] grouped students based on their
learning environment. This study focuses on EDM activities in the LMS. These activities include the number
research applied to higher education. These pieces of of assignments and quizzes and participation in discus-
research are presented in the following section and sions. Clustering was performed using the k‐means
categorized according to the method used. algorithm to group students into three clusters: Very
active, active, and nonactive students.
Abdous et al. [1], categorized question themes using a
2.1.1 | Regression (numeric prediction) hierarchical cluster analysis into four categories: Class
check‐in, deadline/schedule, evaluation/technical, and
In [36], Macfadyen and Dawson analyzed the learning learning/comprehension.
management system (LMS) student usage of an online Valsamidis et al. [60] categorized courses based on
course. They use correlation analysis, multiple regres- usage rate using the k‐means algorithm. They found two
sion analysis, and logistic regression to predict the final clusters: High‐activity courses and low‐activity courses.
grade. They found that the final grade was predicted Researchers proposed a way for measuring the quality of
depending on the total number of posted topics in the the course based on the number of pages viewed by
forum, the total number of sent emails, and the total students for each session. They also investigated the
number of completed assignments. relationship between the quality of the course and the
Abdous et al. [1], studied the relationship between total course score.
the question type asked by a student in the forum and Barata et al. [6] analyzed the students' data from a
his final grade. In this paper, researchers classified gamified course. They clustered students according to
the questions into four themes: Class check‐in, their performance using the expectation–maximization.
deadline/schedule, evaluation/technical and learning/ Three types of students were found, the achievers, the
comprehension. Results showed that students that disheartened, and the underachievers. Researchers
asked about learning/comprehension had a higher described each student type and derived guidelines to
final grade. improve their experience.
Thakur et al. [57] used the logistic regression and In [66], Yildiz et al. clustered the students depending
neural network model to predict the students' success. on usage data (such as quiz score, exam score) using
The number of completed assignments was the basic three clustering methods to find the best approach for
predictor of the students' success and failure, followed by estimating the final grade: k‐means, fuzzy c‐means, and
the completed quizzes and the submitted post. subtractive clustering. They found that fuzzy c‐means is
In [31] Jo et al. used multiple regression analysis and the best method for predicting student performance.
found that the regularity of learning intervals was an Lee et al. [34] applied expectation–maximization and
important predictor of the final grade, while total time k‐means to cluster courses depending on instructor and
and login frequency were not important predictors. students' usage data. Clusters differ in terms of assign-
Lee et al. [34] used multinomial logistic regression to ments, quizzes, discussion topics, and wiki pages. They
predict the final grade based on the used features. Results found three students clusters and three instructor
showed that assignments are the strong predictor of the clusters based on the LMS features used.
final grade in face‐to‐face courses, and quizzes are The results of clustering studies showed that k‐means
a strong predictor of the final grade in online courses. was the most commonly used algorithm because it is one
This review found that all prediction studies are of the most popular and simplest algorithms used in data
similar in predicting the final grade but differ in the mining. Regarding the clustering object, students
type of predictor variable. For example, studies were the most used clustering object [6,17,35,51,61].
[34,35,56] found that the assignment is the strongest Another clustering objects were courses [60] and course
predictor of the final grade whereas in the contents [1].
studies [1,31] the strongest predictors are the types of
questions asked in the forum and regularity of learning
intervals, respectively. Regarding the methodology 2.1.3 | Classification
used, most of the prediction studies used logistic
regression [1,34,36,57] and multiple regression [31,36] Chen et al. [11] used a decision tree based on student
because of the ordinal nature of the predicted variable behavior to discover student groups with the same char-
(the final grade). acteristics and the same reaction to a specific strategy.
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 571

The decision tree will help the instructors to analyze A framework of a recommender system was proposed
student historical data in a weblog and to estimate the by Lu [35] to help students find the needed learning
effectiveness of a certain strategy to achieve the desired materials. Using association rule mining, this re-
pedagogical goals. commender system can help the students to learn more
Minaei‐Bidgoli and Punch [39] proposed an approach effectively by providing personalized learning materials.
for classifying students based on extracted data from the Ramli [50] used association rule mining to optimize
educational system log to predict their final grades. They the course content. By applying the Apriori algorithm to
used a combination of four classifiers to increase the log data, they mined the relationship between visited
accuracy of the classification. Results showed that the pages. This will be helpful for the instructor to increase
total number of correct answers was the most important the performance of web services by improving the
feature needed for the classification. website in terms of its contents, structure, presentation,
Kotsiantis et al. [32] used a decision tree, naïve and delivery.
Bayesian, and k‐nearest neighbor to detect learner's Markellou et al. [38] analyzed the weblog data using
performance based on demographic information and association rules to produce recommendations to users in
written assignments. This will help the instructor to the learning system. They used the Apriori algorithm on
construct the software support tool for enhancing learning materials to determine the most suitable one for
students' performance. the student.
Cocea and Weibelzahl [13] investigated if the log data Romero et al. [51] used the Apriori algorithm for
can estimate the motivational level of the student. Using association rule mining. On the basis of the student's data
the decision tree, the researchers found that the reading from the weblog, the authors found some interesting
time was an important factor to detect motivation. rules that may help the instructor to pay attention
Hamalainen and Vinni [24] used classification to to students who are prone to failure and help them to
detect student success in the middle of the course. They succeed.
used naïve Bayes based on exercise points to detect the All studies used association rule mining for optimiz-
performance as early as possible. ing the e‐learning contents according to the student's
Romero et al. [51] used the decision tree to classify interest. Content such as activities [67], learning
students into different groups according to final marks materials [35,38], web pages [49], messages [51] will be
based on their activities in the LMS. The results can be recommended to the student for providing a personalized
used to make decisions to improve course activities or to learning environment.
detect learners with learning problems.
In summary, the review found that all classification
studies are oriented toward the students' behavior. For 2.1.5 | Distillation of data for human
example, studies [24,31,39] classified students according judgment
to their learning performance whereas [11,12,51] classi-
fied students based on their characteristics. Shen et al. [55] present a Data Analysis Center system to
Regarding the algorithm used, most of the classifica- present students' data such as solved assignments and
tion studies used the decision tree [11,12,39,51] and naïve quizzes marks. This system can show the rank of the
Bayes [24,32]. The decision tree is the easiest method in learner in the class and define the weakness of the
interpreting and explaining results and needs little effort learning process. It can help teachers to analyze learners'
for data preparation. Naïve Bayes is simple, fast, and easy patterns and construct the materials efficiently.
to implement. Romero et al. [51] used the GISMO tool to visualize
data from the modular object‐oriented dynamic learning
environment (Moodle) log, such as graphs represent-
2.1.4 | Association rule mining ing course access, resource access, and discussion
participation. This will allow the instructor to be aware of
Zaiane [67] used association rule mining to build an what is happening in the learning process.
agent that can recommend activities on the learning Valsamidis et al. [60] used a Markov clustering
system. On the basis of the history of learner navigation, algorithm, which combines cluster analysis and graphical
the material of the course will be improved. Re- representation to represent the relationship between
commendations will help the learner to navigate the students. Similar students were grouped visually to help
materials by facilitating the finding of relevant resources interpreting results.
and assist the learner to choose activities that improve Arnold and Pistilli [3] developed a Course Signals
their performance. application that used the dashboard to provide a real‐time
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
572 | DAGHESTANI ET AL.

display of student performance to instructors and students. • There is no gamified learning environment to adapt its
They got a positive result on students' grades when using gamification elements to match the personal needs of
this application. distinct groups of users.
Thakur et al. [57] examined LMS usage data to
investigate constancy in students' marks over the course
by using heatmaps. The result of the investigation was 3 | METHODOLOGY
that the senior students' grades were stable, whereas the
freshmen grades tended to be unstable. The proposed system is an adaptive gamified learning
Lee et al. [34] used a cluster gram of the features used system (AGLS). It combines techniques of gamification,
by students in LMS to help visualization of student's classification, and adaptation to increase the effectiveness
behavior in the course and relate it to the final grade. of e‐learning. The architecture of the proposed system is
As a result, all studies of visualization are focused on presented in Figure 1. In this section, we will discuss
students' performance in the learning process to get a the methodology and architecture of the proposed
better understanding of students' behavior and detect the system AGLS.
students who are at‐risk. In Figure 1, self‐determination theory (SDT) [52] is a
Studies on EDM and their applications are listed in broad theory of human personality and motivation
Table 1. dealing with how the individual interacts with and
From the related work mentioned in Table 1, some depends on the social environment. SDT defines intrinsic
limitations and disadvantages are found: and several types of extrinsic motivations and outlines
how these motivations influence situational responses
• All studies used EDM on a classical e‐learning in different domains, as well as social and cognitive
platform, no study used a gamified e‐learning system development and personality.
(except [66]). SDT is centered on the basic psychological needs of
• All data mining operations were performed based on autonomy, competence, and relatedness and their
educational attributes, such as assignments, quizzes, necessary role in self‐determined motivation, well‐being,
marks, to categorize students based on their educa- and growth. Finally, SDT describes the critical impact of
tional characteristics, but no study tended to categorize the social and cultural context in either facilitating or
students according to their personality or learning thwarting people's basic psychological needs, perceived
style. sense of self‐direction, performance, and well‐being.
• Most of the EDM studies did not consider how EDM Fogg's [19] behavioral model has three principal
contributes to increasing the effectiveness of the factors: Motivation, ability, and triggers. In brief, the
learning process. They just apply data mining methods model asserts that for a target behavior to happen, a
and present results without mentioning how to get person must have enough motivation, sufficient ability,
benefited from it in making a decision about the and an effective trigger. All three factors must be present
development of learning. at the same instant for the behavior to occur.
With these two theories, we determine the gamifica-
tion element. The AGLS app was implemented to adapt
2.2 | Gamification the gamified system based on the student's player type.
For each player type, a different version of the gamified
Gamification can be used in any system to engage its system with suitable gamification elements was designed.
users. Table 2 shows some related work using gamifica- AGLS consists of three methodologies.
tion in e‐learning, especially in higher education.
From the related studies mentioned in Table 2, some
limitations and disadvantages were found: 3.1 | Gamification

• Some gamified systems tend to be games rather than A gamified learning system was developed by incorporating
gamification [7] or just animation [30]. gamification elements into the learning courses in Moodle
• Game elements are designed without considering the LMS [40]. Students had to study the fundamentals of the
different personalities of different students. learning courses by navigating the gamified learning system
• No study classifies learners or groups them according and its material while using gamification features. The
to any factor (except [6] which clusters students of the navigation of students through a gamified learning system
gamified course in [5] according to their accumulated was traced to collect their interaction data with the LMS
points over time along with the course duration). variables to be used in the classification subsystem.
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 573

TABLE 1 EDM methods and applications

Data mining method Paper Year Application

Prediction [36] 2010 Predicting the final grade based on posted topics in the forum, sent mail and
completed assignment
[1] 2012 Predicting the final grade based on the online question theme
[57] 2014 Predicting the final grade based on the number of the completed assignment
[31] 2015 Predicting the final grade based on the regularity of learning intervals
[34] 2016 Predicting the final grade based on assignment participation in face‐to‐face courses
and on quiz participation in the online course
Clustering [51] 2008 Clustering students into three clusters: Very active students, active students, and
nonactive students
[1] 2012 Clustering question theme into four clusters: Class check‐in, deadline/schedule,
evaluation/technical, and learning/comprehension
[60] 2012 Clustering course into two courses clusters: High activity courses and low activity
courses
[6] 2014 Clustering students into three clusters: Achievers, disheartened, and underachievers
[66] 2015 Clustering students into 8 clusters using k‐means, 9 clusters using fuzzy c‐means,
and 11 clusters using subtractive clustering
[34] 2016 Clustering instructors into three instructor clusters. Clustering students into three
student clusters
Classification [11] 2000 Discovering students group with the same characteristics and same reaction to a
specific strategy to estimate the suitable strategy
[39] 2003 Predicting student performance and final grade
[32] 2004 Detecting learners performance and determining the support tool
[13] 2006 Identifying learners motivation based on students behavior to make an intervention
[24] 2006 Detecting the performance of a learner as early as possible
[51] 2008 Estimating the suitable course activities based on learner behavior
Association rule mining [67] 2002 Recommending activities on the learning system
[35] 2004 Recommender system providing personalized learning materials
[49] 2005 Optimizing the course content for the instructor to increase the performance of web
services
[38] 2005 Producing recommendation to users in the learning system
[51] 2008 Paying attention to students that prone to failure
Distillation of Data for human [55] 2002 Analyzing learner's patterns and constructing the materials efficiently
judgment [51] 2008 Helping the instructor to be aware of what is happening in the learning process
[60] 2012 Represent the relationship between students
[57] 2014 Investigating the constancy in students marks over the course
[34] 2016 visualizing of student's behavior among the course

Abbreviation: EDM, educational data mining.

3.2 | Educational data mining 3.3 | Adaptation

Data mining techniques were used to classify students On the basis of the player type of a student, the gamified
based on their collected interaction data into different learning system will be adapted to provide a personalized
player types. Classification models are created and design with the preferred gamification features for each
evaluated to be used for classifying new students. Five student.
classifiers were built using different techniques to Different tools are used to implement AGLSs which
select the best classifier for predicting the player type are listed in the following section with a brief description
of new student. and reasons for using each of them.
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
574 | DAGHESTANI ET AL.

TABLE 2 Comparison of related studies in gamification

Paper, Gamified Mechanics Effect of Study

year Course component used Methodology Collected data gamification Outcome results
[5], 2013 Multimedia E‐learning Points Quantitative Usage data No difference Behavioral Mixed
content platform Levels question-
Increased Psychological
production Challenges naire
satisfaction
Badges
Leaderboard
[30], 2013 Data structure E‐learning Ribbon Quantitative Quiz mark High mark Behavioral Positive
and platform Learning time Less quiz time Behavioral
algorithms Questionnaire High acceptance Psychological
[7], 2013 Software Web Points Quantitative Questionnaire No engagement Psychological Negative
engineering application Levels
Rewards
Progress bar
Leaderboard
[16], 2013 Qualification E‐learning Badges Quantitative Marks High score in Behavioral Mixed
for users platform Leaderboard practical
of ICT assignment
Low score in
written
assignment
Questionnaire High motivation Psychological
[59], 2013 Electronics E‐learning Points Qualitative Observation Improved Psychological Positive
platform Avatar competence
Progress Increased
Feedback interesting
Rewards
Badges
Leaderboard
[2], 2014 Software E‐learning Weekly Quantitative Survey High rated Behavioral Positive
engineering platform challenge learning
cup Usage of High interesting Psychological
High program- High
score list ming engagement
practices Friendly
competition
[28], 2014 C programming E‐learning Points Quantitative Pretest Increased Behavioral Positive
platform Badges Posttest compre-
Leaderboard hension
High points
Questionnaire High Psychological
engagement
[4], 2015 Data Structure Online Badges Quantitative Exercise points High exercise Behavioral Positive
and learning Visualized points
algorithm envir- feedback Survey Increased Psychological
onment Achievement engagement
[14], 2016 Qualification LMS Trophies Quantitative Pretest No difference Behavioral Mixed
for ICT Badges Assignment High marks Behavioral
users Challenges marks
Leaderboard Final Low marks Behavioral
exam mark
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 575

TABLE 2 (Continued)

Paper, Gamified Mechanics Effect of Study

year Course component used Methodology Collected data gamification Outcome results
[26], 2016 Designing LMS Points Quantitative Pretest No difference Behavioral Mixed
question- Badges Posttest No difference Behavioral
naire Leaderboard Questionnaire High motivation Psychological
[45], 2017 Computer LMS Badge Quantitative Programming No impact on Behavioral Mixed
programm- Web‐based test learning
ing application performance
No. of optional High Psychological
activities engagement
done
[17], 2017 Economics Mobile Points Qualitative Observation Increased Psychological Positive
application Levels engagement
Leaderboard

Abbreviations: ICT, information and communication technologies; LMS, learning management system.

Moodle is an open‐source LMS to help educators its features that support gamification and because of its
create effective online learning (Moodle—Open‐source behavior tracking feature and its detailed log information
learning platform | Moodle.org, 2017). Its modular design which helps in the analysis of the course content and
makes it easy to create new courses, adding content and students.
support student's cooperation. Moodle was used in this After getting a registered domain name and web host,
study to develop the gamified learning system because of Moodle 3.1.2 was installed on a Linux web server with

FIGURE 1 Adaptive gamified learning system architecture. LMS, learning management system
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
576 | DAGHESTANI ET AL.

PHP 5.4 version. MySQL 5.1.73 was used to create a appearance. This has been achieved through a clear
database for holding Moodle data. phpMyAdmin was user interface, using themes, different fonts, colors,
used to execute the SQL queries. and a global layout [37]. The persuasive dimension is
HTML is a hypertext markup language used to concerned with the game elements to be incorporated into
describe the structure of web pages [29]. After setting the system. Game elements used in our study and the
up Moodle, HTML5 was used to design the gamified reasons for its inclusion in the course are listed in Table 2.
learning system interface to be attractive and legible The human–computer interaction guidelines that
while keeping its professional and formal appearance. have been followed in the gamification subsystem
Canva is a free graphic‐design tool website. It uses implementation are as follows [9,10,43]:
a drag‐and‐drop format and provides access to over a
million photographs, graphics, and fonts [20]. Canva was 1. Getting the users' attention through the following:
used to design the course badges where their criteria • Creating an interface that is enjoyable to use by
were identified in Moodle. using different fonts, colors, and pictures.
• The consistency of colors, fonts, layout across the
sections of the system
3.4 | System architecture • Logical grouping of similar and related content.
• Organizing the interface through spacing elements
The data structure course was used as a prototype, its evenly and aligning them well.
lectures were uploaded to the system, some assignments 2. Putting the users in control by giving them choices, for
and quizzes were constructed, and accounts for students example, give the student the ability to choose the
were created. preferred view, enabling, or disabling system blocks.
To make the learning framework more captivating 3. Making the interface memorable to be easy to learn
and intriguing, it was gamified by adjusting between and use through applying the following:
graphical outline and persuasive design [37]. The gra- • Placing a particular object in a consistent location,
phical plan was focused on the visual parts of the for example, putting the search box in the upper
interface is attractive and legible while keeping its expert right‐hand corner of the system interface,
and formal appearance. Persuasive design concerns with navigation block at the right side of the interface,
the gamification elements need to be incorporated into the announcement in the front‐page ticker.
the learning system [38]. These two dimensions of • Using conventional symbols that are easier to
gamified system design are described in detail in the remember, for example, symbols of the home page,
following subsections. search, social media, chatting, forum, and calendar.
Rather than achieving the goals of the course and its 4. Increasing the predictability of the system to make the
students through extrinsic motivation, such as grades and user able to determine the results of actions ahead of
points, we tried to amplify the intrinsic motivation time. This includes:
through gamification. To achieve this, we had to improve • Designing pages consistently in a way to enable the
the students' sense of competence, autonomy, and user to remember and predict the outcomes.
relatedness, which are three basic elements of intrinsic • Using familiar menu names and options to help
motivation. According to SDT, satisfying these three users find objects more easily.
psychological needs will enhance self‐motivation. 5. Making the system interface simple, so it will be easy
Przybylski et al. [48] suggested that the attractiveness of to understand and thus easy to learn and remember.
video games comes from their ability to satisfy these three This is done through:
psychological needs. They proposed a motivational model • Showing the user only what is necessary without
based on SDT to discover how video games satisfy these unnecessary details. For example, the student can-
psychological needs. For example, (a) using feedback and not see the details of the different sections on the
showing progress will satisfy the competence feeling, main page unless entering the wanted section.
(b) providing choices of strategies and opportunities will • Limiting the actions that can be performed in a
satisfy the sense of autonomy, and (c) competition in the design. For example, when in the binary tree
leaderboard and cooperation in the forum will satisfy section, the student cannot do any activity of
relatedness. To make the course more engaging and stack topic.
interesting, we gamified it by balancing between graphical 6. Providing clear rules for interface navigation through
and persuasive designs [37]. The graphical dimension the following:
focused on the visual aspects of the interface is attractive • Standardizing task sequences through allowing
and legible while keeping its professional and formal users to perform tasks in the same sequence and
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 577

TABLE 3 Learning system features and gamification elements used to achieve work features

System feature Gamification elements

Autonomy Avatars, options in selecting topics, options in selecting
badges, external resources, videos tutorial
Competence Levels, points, badges, feedback, progress, leaderboard,
challenges
Relatedness Leaderboard, forum, chatting
Small objectives Splitting course content into smaller topics
Fun External resources, videos tutorial
Flow Levels (open and closed)
Loss and confusion avoidance Feedback

manner across similar conditions, for example, a 2. After collecting the interaction data of students, nine
student will download and read lecture notes first, variables of Moodle resources and activities were selected.
then submit the homework, do the quiz, and finally, These variables were selected because they represent
the student can read external sources or view the gamification elements that can identify the player type
video. according to BrainHex player typology Table 4 [42]. Player
• Using descriptive links text and headings so they typologies are considered as another form of personality.
will accurately describe the contents. Categorizing players in video games into player types to
understand their behaviors appeared during playing.
To be persuasive, the learning system was designed to 3. Gamification elements associated with each player
satisfy some features that support students' intrinsic type are listed in Table 5.
motivations in addition to the extrinsic ones (marks). 4. A table was created to summarize the information
These features were achieved by incorporating needed for the classification process. This table has a
different gamification elements based on various summary of all the interactions between students (in a
studies [19,41,44,46,50,52]. Learning system features row) and LMS variables (in columns) during the
and gamification elements used to achieve these course, in addition to the player type obtained by the
features are listed in Table 3. student through the BrainHex questionnaire. Student
The classification process consists of the following steps: data and their definition are listed in Table 6.
5. The objective of the classification subsystem is to
1. Students' interaction data when they navigate through classify students according to player types based on
the gamified learning system was used. This data is their interaction data when they are navigating the
extracted from the Moodle log. Also, the player type gamified learning system. The classification task was
for each student from the BrainHex questionnaire [63] applied using student data (Table 6) in which the class
was obtained. label is the player type.

TABLE 4 Player topology

Player type Description

Conqueror Likes challenges and enjoys defeating impossibly difficult foes and beating
other players
Achiever Likes achieving goals and completing tasks, focus on themselves before others
Socializer Likes talking to people and helping them
Mastermind Enjoys solving puzzles, advising strategies and decision making
Seeker Likes to explore new things, do not like the linear and structured environment
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
578 | DAGHESTANI ET AL.

TABLE 5 Gamification elements for different player types “information gain” measure (93.00%). Decision‐tree
classifiers with “gain ratio” and “Gini index” measures
Player type Gamification elements in Moodle variables
have the same accuracy (92.00%). The lowest accuracy is
Conqueror Challenges and leaderboard
of naïve Bayesian classifier (91.00%).
Achiever Levels and badges The best value of sensitivity is of the k‐nearest
Socializer Forum post and chatting neighbor classifier (87.62%), followed by a decision‐tree
classifier with an “information gain” measure (83.25%).
Mastermind Problems
Decision‐tree classifiers with “gain ratio” and “Gini
Seeker Navigation, external reading, and videos index” measures have the same sensitivity (80.40%). The
Abbreviation: Moodle, modular object‐oriented dynamic learning lowest sensitivity is of the naïve Bayesian classifier
environment. (76.90%). Therefore, the k‐nearest neighbor classifier is
selected to predict the player types of new students
because it has the best values of accuracy and sensitivity
Five classifiers were built using different techniques than the other classifiers.
to select the best classifier for predicting the player type For making an adaptive gamified system, the
of new student. Classifiers are: relationships between player types and the appropriate
game elements need to be understood. This is because
1. Decision tree classifiers with information gain attri- the choice of the appropriate game elements will lead to
bute splitting measure. a personalized experience that assists the students to be
2. Decision tree classifiers with a gain ratio attribute more engaged and motivated.
splitting measure. Authors in [18] proposed a theoretical model to
3. Decision tree classifiers with Gini index attribute identify the relationship between player types and game
splitting measure. elements. On the basis of Ferro's model, some adaptation
4. Naïve Bayesian classifier. forms were suggested, and different versions of the
5. k‐nearest neighbor classifier. gamified system were designed. Examples of different
versions of the gamified system based on student's player
To select the best classifiers for predicting the player type are presented in the following sections.
type of new students, the values of accuracy for all For the Conqueror player type, difficult challenges
classifiers have been compared. The k‐nearest neighbor were added and divided into levels according to their
classifier has the best accuracy (95.00%) among others, difficulty. Also, some challenging quotes were added
followed by the decision‐tree classifier with the to motivate students. The leaderboard is shown always
in the interface and some motivating statements
are provided to the student. Challenges and leader-
TABLE 6 Student data and their definition
board of Conqueror player‐type version are shown in
Student data Data definition Figures 2 and 3.
student_id Identification number of the student For the Achiever player‐type, the course topics were
divided into more levels with attractive names rather
forum_participated Number of post or replies sent to the
forum
than just numbers. Also, new badges were added
continuously to motivate students. Levels and badges
chatting Number of messages sent to the chat
of the Achiever player‐type version are shown in
levels_completed Number of completed levels Figure 4.
badges_taken Number of taken badges For the Socializer player‐type student, group activities
such as assignments and workshops were added. Also,
challenges_completed Number of completed challenges
forums and chatting are presented always in the interface.
leaderboard_viewed Number of times the leaderboard Group assignments and teamwork of the Socializer player‐
viewed
type version are shown in Figures 5 and 6.
navigation Type of navigation, linear = 1, For Mastermind player type, the course topics were
random = 0 divided into many levels with attractive names rather
external_reading_viewed Viewed = 1, not_viewed = 0 than just numbers. Also, difficult challenges were added
videos_viewed Viewed = 1, not_viewed = 0
and divided into levels according to their difficulty. Also,
some challenging quotes were added to motivate
player_type Player type based on BrainHex player
students. Levels and challenges of the Mastermind player
typology
type version are shown in Figures 7 and 8.
10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
|579

Leaderboard and motivating statement in Conqueror version of adaptive gamified learning system
Challenges and challenging quotes in Conqueror version of adaptive gamified learning system
ET AL.

FIGURE 2

FIGURE 3
DAGHESTANI
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
580 | DAGHESTANI ET AL.

FIGURE 4 New badges added continuously in Achiever version of adaptive gamified learning system

For the Seeker player‐type, course contents and activities all pages. Many links to external resources were added.
were distributed to many links to make students motivated The Seeker player‐type version is shown in Figure 9.
to open them. These links associated with ambiguous names AGLS app is an application of an AGLS that adapts
to get students' attention. The navigation block appeared on gamification elements based on a student's player‐type.

FIGURE 5 Forum, chatting, and group assignments in the Socializer version of adaptive gamified learning system
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 581

FIGURE 6 Group workshops in Socializer version of adaptive gamified learning system

The interface of the application is shown in Figure 10. classify the students to the appropriate player‐type.
The student enters her ID and the system will forward On the basis of the player‐type, the application will
her to the suitable version of the gamified learning automatically forward the student to the gamified
system that fits her player‐type. The application can be system version suitable for her player‐type. The archi-
used for predicting the player‐type of new students. tecture of the application is shown in Figure 11. Due to
When a new student login to the application for the a student's interaction pattern changing over time each
first time, the application will present the original time the student logs into the system, the last interac-
gamified system, which contains all gamification tion data collected by the last login is used and sent to
elements. After the first session, student interaction the classifier to classify the student to the appropriate
data will be collected and sent to the classifier to player‐type.

FIGURE 7 Attractive levels in Mastermind version of adaptive gamified learning system

10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ET AL.

Ambiguous links to course contents and external resources, and navigation block in Seeker version of adaptive gamified
DAGHESTANI

Challenges and challenging quotes in Mastermind version of adaptive gamified learning system

learning system
FIGURE 8

FIGURE 9
|582
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 583

Pretest and posttest that covered the basic topics of

the data structure course were developed. The goal of the
pretest was to measure students' initial knowledge about
data structure concepts. The goal of the posttest was to
measure the students' knowledge after the experiment.
Pretest and posttest were compared to measures the
FIGURE 10 Adaptive gamified learning system application
effectiveness of gamification on students' learning
interface
performance.

4 | DISCUSSION AND R ESULTS 4.1 | Comparison of the effectiveness of

gamification and adaptive gamification on
This study was applied to Computer Science students student engagement
at King Abdulaziz University in Saudi Arabia. Three
groups were used and compared for proving the work Based on the Online Student Engagement Scale [15] a
hypotheses: Control group, and two distinct groups 5‐point Likert scale questionnaire was conducted [21] to
from two different experiments. The control group measure the students' engagement.
had 23 students, who studied the data structure course To determine which technique, gamification or
using classical methods without any intervention. The adaptive gamification, has the greatest effect on students'
gamification group comprised those who studied the engagement, the results of the engagement questionnaire
data structure course using the gamified learning for the gamification group and adaptive gamification
system (40 students). The adaptive gamification group needed to be compared with the engagement
group studied the data structure course using AGLS questionnaire on the control group.
(13 students). Table 7 summarizes the data of the The students were asked to answer the engagement
three groups. questionnaire to measure their engagement using the

FIGURE 11 Adaptive gamified learning system application architecture

 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
584 | DAGHESTANI ET AL.

TABLE 7 Summarized data of the three groups

Groups Control group Gamification group Adaptive gamification group

No. of participants 23 students 40 students 14 students
System used No system used Gamified learning system Adaptive gamified learning system (AGLS)

TABLE 8 Engagement questionnaire results of classical learning

Strongly Neither agree Strongly

Question statement disagree Disagree nor disagree Agree agree
I enjoyed the course 0.00% 0.00% 73.91% 17.39% 8.70%
I feel that the course was attractive 0.00% 21.74% 52.17% 21.74% 4.35%
The course content was presented effectively 0.00% 0.00% 69.57% 30.43% 0.00%
Using the course material was easy for me 0.00% 8.70% 43.48% 34.78% 13.04%
I learned about the course topics easily 4.35% 0.00% 56.52% 30.43% 8.70%
I feel engaged in this course 8.70% 13.04% 52.17% 17.39% 8.70%
The course enhanced my motivation 0.00% 8.70% 65.22% 17.39% 8.70%
I like the method of teaching (classical) 4.35% 17.39% 30.43% 30.43% 17.39%
I like this teaching method (classical) in other courses 4.35% 8.70% 60.87% 8.70% 8.70%

classical learning method. Question statements and Figure 12 shows the average points of the
students' answers are summarized in Table 8. The engagement questionnaire results. It can be shown
average of these answers is 3.24 on the 5‐point scale that using adaptive gamification techniques has the
which indicates that the students have a moderate highest effect on student's engagement.
attitude to classical learning.
It can be shown that 73.91% of students neither
agreed nor disagreed that they enjoyed classical learn- 4.2 | Comparison of the effectiveness of
ing. About 52.17% neither agreed nor disagreed that the gamification and adaptive gamification on
course was attractive, 30% of students believed that the learning performance
course content was presented effectively, and 34% found
that its material was easy to use. Figure 13 shows the percentage of gain points average for the
gamification and adaptive gamification groups. The gamifi-
cation group had an average increase in points of 30% in the
posttest, whereas the adaptive gamification group had an
average increase in points of 46% in the posttest. The result
shows that the adaptive gamification group had a higher
point average in the posttest than the gamification group.

F I G U R E 1 2 Engagement questionnaire results for the control FIGURE 13 Percentage of gain points average for
group, gamification group, and adaptive gamification group gamification group and adaptive gamification group
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 585

FIGURE 14 The pretest and posttest points average for the

gamified learning system group and the adaptive gamified learning
system group
FIGURE 15 Average of points of posttest for the control
group, gamification group, and adaptive gamification group

Figure 14 shows the pretest and posttest points for

the gamified learning system group and the AGLS
group. To compare the pretests of the two group, the group and adaptive gamification group. The result indicated
independent sample t test with a confidence level of that there is a significant difference between the learning
0.05 was used. The result indicated that there is no gain of the two groups after using the gamified learning
significant difference between the initial knowledge of system and the AGLS (p = .0033 < .05).
the gamified learning system group and the AGLS group Figure 15 shows the average of points of posttest for
(p = .33 > .05). the control, gamification, and adaptive gamification
The independent‐sample t test with a confidence level of groups. The analysis of variance test with a confidence
0.05 was used to compare the posttests of the gamification level of 0.05 was used to compare the posttests of the

TABLE 9 Comparison between the results of the proposed gamification system and related studies

Paper Study objective Experiment Learning gain Study results

[5] Using gamification for engaging students Control group NA No statistically
significant changes
Gamification group NA
[30] Comparing the effect of GBL and Control group 40% Higher points
gamification
Gamification group 70%
[7] Using gamification for engaging students NA NA Negative
[16] Using gamification for engaging students Control group 86.53% Fewer points
Gamification group 75.19%
[59] Using gamification for engaging students NA NA NA
[2] Enhancing student engagement NA NA NA
NA NA
[28] Engaging students in learning activities Pretest 30.45% Increased points
Posttest 40.73%
[4] Increasing students' awareness of their Control group NA No statistically
behavior significant changes
Gamification group NA
[14] Comparing educational game and Pretest 76% Decreased points
gamification
Posttest 74%

(Continues)
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
586 | DAGHESTANI ET AL.

TABLE 9 (Continued)

Paper Study objective Experiment Learning gain Study results

[26] Using gamification for engaging students Pretest 22.35% Decreased points
Posttest 20.93%
[45] Using gamification for engaging students Control group 70.28% Higher points
Gamification group 90.64%
[17] Using gamification for engaging students NA NA NA
The proposed Evaluating the effect of gamification Pretest 41% Increased points
system AGLS
Posttest 71%
Control group 24.78 Higher points
Gamification group 71%
Evaluating the effect of adaptive Pretest 38.57% Increased points
gamification
Posttest 85%
Control group 24.78% Higher points
Adaptive Gamification 85%
group
Comparing the effect of gamification and Gamification group 71% Higher points
the effect of adaptive gamification
Adaptive gamification 85%
group

three groups. The result indicated that there is a study which combines gamification, classification, and
significant difference between the learning gain of the adaptation techniques to increase the effectiveness of
three groups after using the gamified learning system and e‐learning. A gamified learning system was developed to
the AGLS (p = .0000 < .05). increase student's engagement and learning outcomes. A
A comparison between the knowledge gain using the classification was used to classify students into different
proposed system and the knowledge gain of the related player types based on their interaction with the gamified
works using gamification is presented in Table 9. As shown system. The adaptation technique was used to adapt to the
from the table, the enhancement of points between the gamified system according to the student's player‐type.
pretest and posttest of the proposed system is higher than AGLS app was implemented to adapt the gamified
the other studies [13,26,28]. This is due to the knowledge system based on the student's player‐type. For each
gained from using the gamified learning system. player‐type, a different version of the gamified system
Also, the difference between the points of the control with suitable gamification elements was designed.
group and the gamification group of the proposed system is An engagement questionnaire was used to measure
higher than the other studies [5,17,30,45]. the students' engagement. Results showed that gamifi-
Also, the adaptive gamification was evaluated, and its cation and adaptive gamification have a positive effect on
effect was compared with the gamification effect, students' engagement and learning performance. This
whereas there is no study using adaptive gamification. difference was due to the knowledge gained from the
data structure course taught using gamification and
adapted gamification Also, the results showed that
5 | CONCLUSION students using adaptive gamification performed better
than students using gamification, which means that
Gamification is the use of game elements in nongame adaptive gamification has a higher effect on student' en-
systems. It is used in e‐learning to increase the effective- gagement and learning performance than gamification.
ness of the learning process and make it more interesting. Some limitations were encountered in this study such
After reviewing the literature, it was found that most of as the limited time and the small number of participant
the gamified learning systems were designed without students.
considering the different personalities of different In the future, more students would be required to
students. For this reason, the AGLS was proposed in this confirm the results of gamification and to increase the
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 587

accuracy of classification. Also, more game elements need 12. R. M. Clark, and A. Kaw, Adaptive learning in a numerical
to be included in the system to increase the accuracy of the methods course for engineers: Evaluation in blended and flipped
classification process. A multilevel adaptation is needed to classrooms, Comput. Appl. Eng. Educ. 28 (2020), no. 1, 62–79.
https://doi.org/10.1002/cae.22175
build a more interesting and motivating experience.
13. M. Cocea and S. Weibelzahl, Can log files analysis estimate
Also, students' engagement can be measured by
learners' level of motivation?, In Proceedings of the workshop
tracing their behavior through the system and not just week Lernen – Wissensentdeckung – Adaptivitate t, Hilde-
from their self‐reported questionnaire. Analyzing student sheim, 2006, pp. 32–35.
engagement can be done by using more scientific 14. L. de‐Marcos, E. Garcia‐Lopez, and A. Garcia‐Cabot, On the
methods for measuring engagement, such as sequential effectiveness of game‐like and social approaches in learning:
pattern mining. Comparing educational gaming, gamification & social net-
working, Comput. Educ. 95 (2016), 99–113.
15. M. D. Dixson, Measuring student engagement in the online
ORCID
course: The online student engagement scale (OSE), Online
Lamya F. Daghestani http://orcid.org/0000-0002-
Learn. 19 (2015), 561.
4776-7626 16. A. Domínguez et al., Gamifying learning experiences: Prac-
Lamiaa F. Ibrahim http://orcid.org/0000-0001- tical implications and outcomes, Comput. Educ. 63 (2013),
5671-8941 380–392.
17. J. C. Fernández‐Zamora and D. Arias‐Aranda, Implementation
REFERENCES of a Gamification Platform in a Master Degree (Master in Eco-
nomics), Working Papers on Operations Management, 2017,
1. M. Abdous, W. He, and C. J. Yen, Using data mining for
vol. 8, pp. 81–190.
predicting relationships between online question theme and final
18. L. S. Ferro, S. P. Walz, and S. Greuter, Towards personalized,
grade, J. Educ. Technol. Soc. 15 (2012), no. 3, 77–88.
gamified systems: An investigation into game design, personality
2. B. Akpolat and W. Slany, Enhancing software engineering
and player typologies, In Proceedings of The 9th Australasian
student team engagement in a high‐intensity extreme programming
Conference on Interactive Entertainment: Matters of Life and
course using gamification, IEEE 27th Conference on Software
Death, New York, NY, 2013, pp. 7:1–7:6.
Engineering Education and Training (CSEE&T), 2014, p. 149.
19. B. J. Fogg, A behavior model for persuasive design, Proceedings
3. K.E. Arnold and M.D. Pistilli, Course signals at Purdue:
of the 4th International Conference on Persuasive Technology,
Using learning analytics to increase student success, In
Claremont, CA, 2009.
Proceedings of the 2nd International Conference on Learn-
20. Free Online Badge Maker‐Canva, 2017, available at https://
ing Analytics and Knowledge, ACM, Vancouver, BC, 2012,
about.canva.com/create/badges/
pp. 267–270.
21. Gamified Course Evaluation, 2018, available at https://docs.
4. T. Auvinen, L. Hakulinen, and L. Malmi, Increasing students'
google.com/forms/d/1mdaxpLwEgmKb0tC5rwZHvL5bxhMur
awareness of their behavior in online learning environments with
ONhwYufC16j3cE/
visualizations and achievement badges, IEEE Trans. Learn.
22. P. Garone, and S. Nesteriuk, Gamification and learning: A
Technol. 8 (2015), no. 3, 261–273.
comparative study of design frameworks, Digital human mod-
5. G. Barata et al., Engaging engineering students with gamification an
eling and applications in health, safety, ergonomics and risk
empirical study, 5th International Conference on Games and Vir-
management. healthcare applications. HCII 2019. Lecture
tual Worlds for Serious Applications (VS‐GAMES), Poole, 2013,
Notes in Computer Science (V. Duffy, ed.), 11582, Springer,
pp. 1‐8. https://doi.org/10.1109/VS‐GAMES.2013.6624228
Cham, 2019, pp. 473–487.
6. G. Barata et al., Identifying student types in a gamified learning
23. J. P. Gee, What video games have to teach us about learning and
experience, Int. J. Game‐Based Learn. 4 (2014), no. 4, 19–36.
literacy, Palgrave Macmillan, Hampshire, UK, 2003, p. 256.
7. K. Berkling and C. Thomas, Gamification of a Software En-
24. W. Hamalainen and M. Vinni, Comparison of machine learning
gineering Course‐ and a detailed analysis of the factors that lead
methods for intelligent tutoring systems, In Proceedings of the
to its failure, International Conference on Interactive Colla-
Eighth International Conference in Intelligent Tutoring Sys-
borative Learning (ICL), 2013, p. 525.
tems, Taiwan, 2006, pp. 525–534.
8. M. Berland, R. Baker, and P. Blikstein, Educational data
25. J. Hamari, J. Koivisto, and H. Sarsa, Does gamification
mining and learning analytics: Applications to constructionist
work?—A literature review of empirical studies on gamification,
research, Technol. Knowl. Learn. 19 (2014), no. 1‐2, 205–220.
47th Hawaii International Conference on System Science, 2014.
9. A. Chakraborty, Human computer interaction (HCI) guidelines,
26. K. F. Hew et al., Engaging Asian students through game me-
2018, available at https://www.linkedin.com/pulse/human‐
chanics: Findings from two experiment studies, Comput. Educ.
computer‐interaction‐hci‐guidelines‐amitabha‐chakraborty
92 (2016), 221–236.
10. N. Chapman, and J. Chapman, Digital multimedia, 2nd ed.,
27. R. Hunicke, M. Leblanc, and R. Zubek, MDA: A formal approach to
John Wiley & Sons, Chichester, UK, 2004.
game design and game research, Discovery 83 (2004), no. 3, 1–5.
11. G. Chen et al., Discovering decision knowledge from web log
28. M. Ibanez, A. Di‐Serio, and C. Delgado‐Kloos, Gamification for
portfolio for managing classroom processes by applying decision
engaging computer science students in learning activities: A case
tree and data cube technology, J. Educ. Comput. Res. 23 (2000),
study, IEEE Trans. Learn. Technol. 7 (2014), no. 3, 291–301.
no. 3, 305–332.
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
588 | DAGHESTANI ET AL.

29. Introduction to HTML, W3schools.com, 2018, available at International Science Conference e‐Learning Software Educa-
https://www.w3schools.com/html/html_intro.asp tion, 2012, pp. 345–350.
30. U. Jayasinghe and A. Dharmaratne, Game‐based learning vs. 47. E. Popescu, P. Trigano, and C. Badica, Adaptive educational
gamification from the higher education students' perspective, hypermedia systems: A focus on learning styles, The Inter-
IEEE International Conference on Teaching, Assessment and national Conference on Computer as a Tool, Warsaw, 2007.
Learning for Engineering (TALE), 2013, p. 683. 48. A. K. Przybylski, C. S. Rigby, and R. M. Ryan, A motivational
31. I. H. Jo, D. Kim, and M. Yoon, Constructing proxy variables to model of video game engagement, Rev. Gen. Psychol. 14 (2010),
measure adult learners' time management strategies in LMS, no. 2, 154–166.
Educ. Technol. Soc. 18 (2015), no. 3, 214–225. 49. A. A. Ramli, Web usage mining using apriori algorithm: UUM
32. S. B. Kotsiantis, C. J. Pierrakeas, and P. E. Pintelas, Predicting learning care portal case, In International Conference on
students' performance in distance learning using machine learning Knowledge Management, Malaysia, 2005, pp. 1–19.
techniques, Appl. Artif. Intell. 18 (2004), no. 5, 411–426. 50. R. Raymer, Gamification: Using game mechanics to enhance
33. A. Krumm et al., A learning management system‐based early e‐learning, eLearn (2011), no. 9, 3.
warning system for academic advising in undergraduate 51. C. Romero, S. Ventura, and E. García, Data mining in course
engineering, Lear. Anal. (2014), 103–119. management systems: Moodle case study and tutorial, Comput.
34. J. E. Lee et al., Applying data mining methods to understand Educ. 51 (2008), no. 1, 368–384.
user interactions within learning management systems: 52. R. M. Ryan, and E. L. Deci, Self‐determination theory and the
Approaches and lessons learned, J. Educ. Technol. Dev. facilitation of intrinsic motivation social development, and
Exchange 8 (2016), no. 2, 99–116. well‐being, Am. Psychol. 55 (2000), no. 1, 68–78.
35. J. Lu, Personalized e‐learning material recommender system. In 53. M. Sailer, and L. Homner, The gamification of learning: A meta‐
International Conference on Information Technology for Ap- analysis, Educ. Psychol. Rev. 32 (2020), 77–112.
plication, Utah, 2004, pp. 374–379. 54. J. Schell, The art of game design: A book of lenses, Ann. Phys. 54
36. L. P. Macfadyen and S. Dawson, Mining LMS data to develop (2008), 489.
an early warning system for educators: A proof of concept, 55. R. Shen, F. Yang, and P. Han, Data analysis center based on
Comput. Educ. 54 (2010), no. 2, 588–599. e‐learning platform, In Workshop The Internet Challenge: Tech-
37. C. Marache‐Francisco, and E. Brangier, Perception of gamification: nology and Applications, Berlin, Germany, 2002, pp. 19–28.
Between graphical design and persuasive design, Design, user ex- 56. G. Siemens and P. Long, Penetrating the fog: Analytics in
perience, and usability. Health, learning, playing, cultural, and learning and education, Educause Rev. 46 (2011), no. 5,
cross‐cultural user experience (A. Marcus, ed.), Springer, Berlin, 30–32.
Heidelberg, 2013, pp. 558–567. 57. G. S. Thakur et al., Towards adaptive educational assess-
38. P. Markellou et al., Using semantic web mining technologies ments: Predicting student performance using temporal
for personalized e‐learning experiences. In Proceedings of the stability and data analytics in learning management systems,
Web‐Based Education, Grindelwald, Switzerland, 2005, In Proceedings 20th ACM SIGKDD Conference on
pp. 461–826. Knowledge Discovery and Data Mining, New York, NY:
39. B. Minaei‐Bidgoli and W. Punch, Using genetic algorithms for ACM, 2014.
data mining optimization in an educational web‐based system, 58. A. M. Toda et al., Analysing gamification elements in educa-
In Genetic and Evolutionary Computation Conference, tional environments using an existing Gamification taxonomy,
Chicago, IL, 2003, pp. 2252–2263. Smart Learn. Environ. 6 (2019), 16. https://doi.org/10.1186/
40. Moodle ‐ Open‐source learning platform | Moodle.org, 2017, s40561‐019‐0106‐1
available at https://moodle.org/?lang=ar 59. V. Todor and D. Pitică, The gamification of the study of elec-
41. C. I. Muntean, Raising engagement in e‐learning through ga- tronics in dedicated e‐learning platforms, In 36th International
mification, The 6th International Conference on Virtual Spring Seminar on Electronics Technology, 2013, p. 428.
Learning ICVL, 2011. 60. S. Valsamidis et al., A clustering methodology of web log data for
42. L. E. Nacke, C. Bateman, and R. L. Mandryk, BrainHex: a learning management systems, J. Educ. Technol. Soc. 15 (2012),
neurobiological gamer typology survey, Entertain. Comput. 5 no. 2, 154–167.
(2014), 55–62. 61. F. Viola, Gamification I videogiochinella vita quotidiana,
43. H. Nguyen, Human computer interaction in game design, Arduino Viola, 2011.
Bachelor's Thesis, Business Information Technology, Oulu 62. T. E. J. Vos et al., IMPRESS: Improving engagement in
University of Applied Sciences, 2012. software engineering courses through gamification, Product‐
44. S. Nicholson, A user‐centered theoretical framework for mean- focused software process improvement. PROFES 2019.
ingful gamification, In Proceedings of Games Learn. Soc. 8.0, Lecture notes in Computer Science (X. Franch, T. Männistö,
2012, pp. 223–230. and S. Martínez‐Fernández, eds.), 11915, Springer, Cham,
45. M. Ortiz‐Rojas1, K. Chiluiza, and M. Valcke, Gamification in 2019, pp. 613–619.
computer programming: Effects on learning, engagement, self‐ 63. Welcome to the BrainHex questionnaire!, 2018, available at
efficacy and intrinsic motivation, European Conference on http://survey.ihobo.com/BrainHex/
Game Based Learning, 2017. 64. K. Werbach, and D. Hunter, For the win: How game thinking
46. V. Petrovic and D. Ivetic, Gamifying education: A proposed can revolutionize your business, Wharton Digital Press,
taxonomy of satisfaction metrics, in Proceedings of 8th Philadelphia, PA, 2012.
 10990542, 2020, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/cae.22227 by Capes, Wiley Online Library on [12/05/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
DAGHESTANI ET AL. | 589

65. N. Yee, Motivations for play in online games, Cyberpsychol

techniques in wired and wireless network planning.
Behav. 9 (2006), 772–775.
66. O. Yildiz, A. Bal, and S. Gulsecen, Statistical and clustering She has published papers in many international
based rules extraction approaches for fuzzy model to estimate journals and international conferences in the areas
academic performance in distance education, Eurasia J. Math. of networks, data mining, and wired and mobile
Sci. Technol. Educ. 11 (2015), no. 2, 391–404. network planning.
67. O. Zaiane, Building a recommender agent for e‐learning systems,
In Proceedings of the International Conference in Education, Reem S. Al‐Towirgi is a teaching
Auckland, New Zealand, 2002, pp. 55‐59.
assistant in the Faculty of Computing
and Information Technology at Taif
University, Taif, Saudi Arabia. Currently,
AUTHOR BIOGRAPHIES
she is a master's student, Faculty of
Computing and Information Technology, King Abdu-
Lamya F. Daghestani is an assistant laziz University, Jeddah, Saudi Arabia. She obtained a
professor in the Faculty of Computing BSc from the Computer Science Department, Faculty
and Information Technology at King of Science, King Abdulaziz University in 2007. Her
Abdulaziz University. She obtained research interest includes e‐learning, gamification, and
her bachelor's degree from Computer data mining.
Science, Faculty of Science, King Abdulaziz
University 1989. She obtained a master's degree from Hesham A. Salman received a mas-
the Department of Informatics/Digital Media, Com- ter's degree in Engineering from Ain
puting and Engineering, University of Huddersfield Shams University in 1996, and he
2007. She obtained a doctorate degree from Digital received a PhD degree in the Faculty
Media, Computing and Engineering, University of of Computing and Information Sys-
Huddersfield, 2013. Her research interests include tems from Ain Shams University. He is currently
computer graphics, virtual reality, and virtual envir- working in the Higher Institute of Computer and
onments, e‐learning technologies, e‐library manage- Information Technology, Alshrouk Academy, Cairo,
ment, search engines, and the semantic web. Egypt. Previously, he was in the Information System
Department, Faculty of Computing and Information
Lamiaa F. Ibrahim is a Professor and Technology, King Abdulaziz University. He was the
Head of the Department of Computer General Manager of the Technology Competency
Science, Faculty of Graduate Studies for Center, Ministry of Commerce and Industry, in
Statistical Research, Cairo University. Cairo, Egypt. Throughout his working life, he has
She obtained a PhD from the Faculty participated in many studies and works related to
of Engineering, Cairo University in 1999, a master's Information Systems. He has over 37 years of
degree from the Computer and Systems Engineering experience in the fields of network engineering
Department in the Faculty of Engineering, Ain Shams and programming applications. He has published in
University in 1993, a master's degree from the Ecole the area of networks, data mining, and wire and
National Superieur de Telecommunication, ENST mobile network planning.
Paris in 1987, and a BSc from the Computer
and Automatic Control Department, Faculty of
Engineering, Ain Shams University in 1984.
Previously, she was in the Information Technology How to cite this article: Daghestani LF, Ibrahim
Department, Faculty of Computing and Information LF, Al‐Towirgi RS, Salman HA. Adapting gamified
Technology, King Abdulaziz University. She has learning systems using educational data mining
over 33 years of experience in the fields of network techniques. Comput Appl Eng Educ. 2020;28:
design engineering and artificial intelligence, focusing 568–589. https://doi.org/10.1002/cae.22227
on applying knowledge base and data mining