AN ANALYSIS OF EDUCATIONAL GAMES DESIGN FRAMEWORKS FROM SOFTWARE ENGINEERING PERSPECTIVE

Game-based learning has dominantly embedded itself into a tool of education in the 21st century. In developing educational games, many researchers have proposed frameworks to de ﬁ ne elements of an educational game. This paper presents a survey of the different frameworks for educational games and analyzes these frameworks against several criteria for effective video games, well-designed games and key elements of educational games. The authors will also look at the frameworks support towards learning theories. In addition, the analysis continues in the context of software engineering practices to develop effective educational games.


INTRODUCTION
Learning is an ultimate practice of the interpretation and evaluation of knowledge with one's experience which is subsequently translated into skills, values and comprehension (Ge, 2012).There are many ways in which the knowledge can be delivered such as scaffolding, project-based or collaborative learning (Garris, Ahlers, & Driskell, 2002).
Game-based learning (GBL) is described as the involvement of suitable game mechanics' and recreation of the scenario-based, problem-based, problemoriented learning processes found in games in order to accomplish learning purposes (Kovačević, Minović, Milovanović, De Pablos, & Starčević, 2013;Lester et al., 2014;Poulsen, 2011).Squire and Jenkins (2003) highlighted that a large number of scholars have confi dence on computer games to become a learning tool that will inspire the students and enhance their focus on learning in subjects such as chemistry (Romli, Shiratuddin, & Hashim, 2003), physics (Abang Abu Bakar;2002), health education (Barker, Quennerstedt, & Annerstedt, 2013;Mellecker, Witherspoon, & Watterson, 2013) and mathematics (Ke, 2013), and initiate a pioneering educational paradigm (Tan, Ling, & Ting, 2007).Furthermore, GBL allows a rich learning context to assist learners to build a higher-level of knowledge through imprecise and confronting opportunities (Van Eck, 2006).In addition, there has been a research which declared that games are capable of tendering 16 learning principles (Gee, 2005).
Conversely, there have been many confl icts on how to design educational games that is entertaining and at the same time sustaining the learning aspects.For this reason, researchers argue that an educational game should include good game design and suitable learning pedagogy (Amory & Seagram, 2003;Becker, 2009;Moreno-Ger, Burgos, Martínez-Ortiz, Sierra, & Fernández-Manjón, 2008).A recent study has conducted a case study based approach where the authors implemented design-based learning through educational computer game for math subject and how the novice developer's designing thinking and the contents usage for knowledge development can be collaborative in game-making process (Ke, 2014).The research concluded that there was a great improvement for students to experience activities actively; however, the time was consumed for the development and content deliverance for learners.Hirumi and Stapleton (2009) stated that the research expansion in educational video game production is moving more quickly in design and effi cacy, however, not much is known of methodologies that could be used to enhance the game design processes to achieve a better outcome (Hirumi & Stapleton, 2009).Table 1 summarizes the frequently discussed trends in GBL.

Table 1
Selected Trends in Education learning (De Freitas & Oliver, 2006) Key Trends "Widespread use of games technologies and serious games 1 movement.""Authoring and development of immersive worlds (e.g.development of content creation tools).""Growth of online gaming and online gaming communities.""New forms of gaming (e.g.mobile games) and convergent forms of gaming emerging (e.g.social and alternate reality gaming)." A systematic review of 117 selected literatures was conducted to investigate the educational design frameworks.They were analyzed to investigate and confi rm with the criteria of well-designed and effective video games.The key words used to fi nd the literature were game design (2000 -2014), learning theories (1913 -2013), frameworks and models (1999 -2013), subjects or courses of games (2000 -2014), and other key words such as software engineering in GBL, learning environments, motivational aspects, educational games in case studies, game developing components and key elements of designing educational game (1999 -2014).In addition to that, literature types were mainly books (9), book section (4), conference proceedings (23), generic (1), journal articles (78), report (1) and thesis (1).

Learning via Educational Games
Games are more synonymous to entertainment and fun than learning (Baranowski, Buday, Thompson, & Baranowski, 2008;Whitton, 2012).Therefore, using games as an effective learning tool is only possible when the games are developed properly and incorporates established education theories.
Educational games are typically developed based on learning theories, pedagogical components, gender preferences and game design components (Baranowski et al., 2008;Gress, Fior, Hadwin, & Winne, 2010;Hirumi & Stapleton, 2009;Kruger, 2012;Mikropoulos & Natsis, 2011).Adapted from Smith (Smith, 1999), the four major learning theories in GBL are behaviorism, cognitivism, humanism and constructivism, which include 16 1 Serious games are also known as educational games; video games; game-based learning; instructional games learning principles.Constructivism theory (Wu, Hsiao, Wu, Lin, & Huang, 2012) deals with human constructing their knowledge subjectively through experience and learning in active process.Behaviorism theory developed by (Thorndike, 1913) deals with learning through a change in behavior or the environment shape behavior and the principle of reinforcement.Cognitivism theory deals with how the memory and prior knowledge plays an important role in learning (Merriam, 1999).Humanism theory deals with factors such as self-determination, the value and potential of the user depending on their dignity (Combs, 1981).A study by Wu, Hsiao, Wu, Lin & Hang (Wu et al., 2012) shows that only around 14% of GBL research includes learning theories and the most used is experiential learning and situated learning.
From the learning theories, researchers adopted some of their principles for GBL and some of the adopted theories are:


According to the experiential learning theory, instructional environment must be able to assess learner's knowledge, be interactive, provides ongoing feedback and able to sustain attention by adapting levels of challenges that the learner encounters (Dalgarno & Lee, 2012;Hays, 2005;Herrington & Oliver, 2000;Kruger, 2012;Smythe, 2012).It is recently proven that instructional game-based learning environment has assisted learner's in improving their learning (Wouters & Van Oostendorp, 2013).


A learning environment must be able to be controlled, fundamentally inspiring, and involves active memory with knowledge exchange and/or physical commitment towards learning for the learners to gain knowledge and experience through failure and retry (Klopfer, Osterweil, & Salen, 2009).


A learning environment should allow four types of actions: 1) Experimentation: the establishment of learning objectives and the activities to be carried out, 2) Refl ection: the comparison of results obtained from groups, 3) Activity: the design of the game, which may involve other educational materials as reference such as books, internet etc., and 4) Discussion: the refl ection on the actual process of learning, and the joint discussion related to the proposed activities.


Effective learning must balance between challenges and the learners' skills in completing the challenges.This requirement is based on the Flow theory (Admiraal, Huizenga, Akkerman, & Dam, 2011; Csikszentmihalyi, 1990;Csikszentmihalyi, 1997;Finneran & Zhang, 2003;Kiili, 2005).The fl ow theory states the symbiotic relationship between challenges and the skills needed to meet those challenges.

Effective Educational Game Designs
Researchers have proposed the requirements/guidelines for effective educational game design.Figure 1 illustrates these requirements.
Other than Prensky Key Elements, Van Staalduinen and de Freitas (2010) also listed 12 game elements from the literature and grouped them into four dimensions, which are shown in Figure 2. Some of the game elements appearing in the four dimensions are also mentioned in the core elements of a well-designed game and elements of an effective game (De Freitas & Jarvis, 2009;De Freitas & Oliver, 2006).

Figure 2. Game elements grouped in a four dimensional framework
This framework is used to choose an appropriate educational game and learn to earn knowledge correctly (De Freitas & Jarvis, 2009;De Freitas & Oliver, 2006).The four dimensions are explained in Table 2.
Table 2 The Four-Dimensional Framework Domains and Their Description Domain Description

Context
Game and its utilization are central to the effectiveness of how the game consumes it.
Learner Specifi c Action The learner or learner group is also central to which game is chosen and utilized.

Representation
The game itself has a level of immersion and devotion, familiarity of interface with the learner group and the internal reality and narrative of the game for effective learning.

Pedagogic Model
The utilization of games in the learning context, rather than for leisure time activities.

METHODOLOGY
Developing an educational game is certainly not an effortless job.It engages various viewpoints into consideration, which may consist of the major learning theories with their principles, game design components as well as the learning content.A few researches have specifi ed clearly that educational games must merge a good pedagogy structure and a known game design to guarantee education effi ciency (Amory & Seagram, 2003;Fisch, 2005;Garzotto, 2007;Ibrahim & Jaafar, 2009).
However, developing an educational game is not only about learning theories.
If the development is only in the hands of education expert, the outcome will not be fun or attractive, whereas, if game designers take over, the games may be unsuccessful in applying the key pedagogical principles.A study stated that to certify game development, the content experts of the game should work constantly with game designers (Hirumi & Stapleton, 2009).However, educational game development is not limited to the involvement of game designers and education experts.Other stakeholders can also contribute to the development such as programmers, teachers, instructional designers and students (Ibrahim & Jaafar, 2009).
A few frameworks have been discussed in literature for educational game development.12 game elements were analyzed and divided into four dimensions (De Freitas & Jarvis, 2009;De Freitas & Oliver, 2006).It is extended further to be used in educational game development and is known as the Four Dimensional Framework (FDF).The extended framework, as being shown in Figure 3, maintains the four dimensions, and extends and adds more elements that should be in an educational game such as instructions, feedback, challenges and interactions.FDF is improved further by including the cognitive and instructional approaches into the game process (Staalduinen & De Freitas, 2010).The framework, as shown in Figure 4, is able to map the design of existing games, serves as a checklist for designers of serious games and provide replicable and scalable immersive game content.The framework also establishes relationships between the elements in the four dimensions.Learning objectives and player goals defi nes users' engagement in the game and guides their learning.Learner specifi c game elements are customized to give positive infl uence towards users learning and behaviours.Appropriate debriefi ng and system feedback will be given by the game according to the users' achievements to assist the students learning.The students learn from the system feedback and improve (player feedback).The players' feedbacks will infl uence the learning contents (maintain the current content, move to a more advance topic or go deeper into the same topic) and this will infl uence learning objectives and player goals.
Figure 4. Van Staalduinen and De Freitas (2010) Framework FDF mainly specifi es the elements of educational games and the inclusion of these elements are based on learning theories such as experiential learning.
Other researchers have also included the process of learning in their framework.Learner-centered design (LCD) framework has been proposed with the experiential learning at its core.This framework has an adjustable feature to enhance student's learning performance for games that require data input from keyboard (Wang, Chen, & Tseng, 2010).Three key components of the framework are: 1) learner's personal assortment is created according to their learning progressions, 2) an adjustable/adaptive learning mechanism where playing the game plus providing documentation of experiments that performs a comparison between learning techniques for fi ne-tuning, and 3) a confi gurable mechanism which provides a cooperative act to adjust content of the game according to changes made.Learner-centered design (LCD) approaches, such as contextual inquiry and participatory design, have been implemented to propose new tools for learners in emerging nation (Lalji & Good, 2008).
Another research proved that adaptive learning in educational games does not depend on student's cognitive or learning styles, thus, they choose educational games based on personal preferences (Hwang, Sung, Hung, & Huang, 2013).
A framework that is based on active learning theory by Garris, Ahlers & Driskell (Garris et al., 2002) allows education expert to defi ne their own games (Pivec & Dziabenko, 2004a).Figure 5 shows the model of GBL by Garris et al. (2002).
Figure 5. Model of game-based learning by (Garris et al., 2002;Pivec & Dziabenko, 2004a, 2004b) Experiential learning does not necessitate teachers as it is exclusively depending upon the useful process accomplished by individual's experience with the game itself.Game play involves interaction with a game through its rules, the connection between the player and the game, challenges and solutions, the plot and the players emotional connection with the plot (Egenfeldt-Nielsen, Smith, & Tosca, 2008).The experiential gaming model allows game play to link with experiential learning (Kiili, 2005).This shows the process of learning through experiential loop where it is adapted to the outcome of the fi rst problem's solution, then it provides a challenge and continues the loop to experience problems and solve throughout the loop.However, a new aspect is that the fl ow of learning is an important fact to understand and overcome the challenges to achieve the objective (Kiili, 2005(a); Kiili, 2005(b)).Although the skill level increases and the player can control the game with subject matter, the player endeavors to test different kinds of solutions in order to expand knowledge.Figure 6 shows the experiential framework by Kiili (Kiili, 2005(a); Kiili, 2005(b)).1999;Hoffman & Novak, 1996).Flow antecedents consist of educational game elements that are the pre-requisite to effective learning.Flow experience includes the conditions that the users experience when playing the game and fl ow consequence is the result obtained from the learning experience.A game based learning model by (Osman & Bakar, 2012) was proposed for Malaysian chemistry classroom and a few challenges were faced when it was implemented.The authors argued that availability and sustainability of the games in the market and in classroom is fl uctuating, adopting learner's requirement and designing a computer game for classroom is hectic and barriers that hinder the implementation of educational computer games is typical and vague.Therefore, the authors introduced a model with fi ve factors: pedagogical factor (teaching strategy), game-play design factor (rules integrated in the game), teacher factor, student factor, and other factors such as availability of technology.
Yue, Zin & Jaafar (2009) proposed a framework for history education that also includes the country's curriculum, moral values and psychology in addition to the learning theories and educational elements mentioned in FDF. Figure 9 shows the framework by Yue et al.The change of perspective changes the elements that researchers included in the framework from learning elements to gaming elements and software components.Furthermore, the representation of the frameworks starts to look like a game architecture than a learning framework.Research is even performed to fi nd ways of lowering game development cost such as 1) developing simple and cost-effective game with less technical specifi cations such as (online toolkits), 2) learning from games rather than learning with them, 3) providing learner's agency with creators rather than simple players (Whitton, 2012)   Figure 11 is a framework that combines learning elements (pedagogy and learning contents) with development elements (game design) (Ibrahim & Jaafar, 2009).The authors have emphasized on the usability that comprises of effi ciency, competence, and satisfaction in both the game design and learning content of the game as compared to the FDF.The Game-Object Model II (GOM II) (see Figure 12), based on objectoriented programming concepts, provides a theoretical basis for the design of educational games.GOM II is able to create tension between game elements and pedagogical dimensions (Amory, 2007).This model is an enhancement from the Game Object Model I (GOM I) developed in 2003 by Amory.GOM I does not have the collaboration between the spaces of the framework whereas GOM II provides the collaboration.Objects are interconnected into three major spaces, namely challenges, narrative, and conversation.Components of the model categorized as concrete or interface.The abstract interfaces will conclude the pedagogical fi tting of a game while the concrete interfaces will show if the game is able to achieve the educational objective.Social space that includes relationship among the equipment or social factors is not having any connectivity to the visualization space, which has more or less the same factors as representation dimension of FDF.This shows that, although the models are able to support development of educational games and provide the mechanism to review the game on how well a game achieves educational objective; it is also agreed that the model can be easily misconceived to suit different viewpoints.

Learning Needs Analysis
Post-analysis Game Assessment Tool

Post-analysis Game Selection Tool
Learner Questionnaire in designing educational games is frameworks that not only specify elements but also incorporates the learning process specifi ed in the learning theories.The learning theories being incorporated are experiential learning (Wang et al., 2010), fl ow theory (Finneran & Zhang, 2003) and active learning (Garris et al., 2002).The second movement is recognition that an effective educational game should incorporate the learning process and not just the learning elements.The third movement includes in the framework other aspects of education such as national curriculum, school hours, and budget and teacher competencies.This movement highlights the importance of local factors/issues/constraints that may infl uence effectiveness of GBL, and considers the application of GBL in a formal education setting.The fourth movement is the frameworks evolution from the learning perspective to development perspective, and highlights that educational game effectiveness requires the combination of game design and learning process.This evolution results in the framework adopting a software model like notations (Amory, 2007), becoming more like a software architecture (Mills & Dalgarno, 2007), including software properties such as usability (Ibrahim & Jaafar, 2009) and becoming part of a software development methodology (Jarvis & de Freitas, 2009).

RESULTS
The frameworks and models were analyzed based on fi ve major categories: 1) Developing game design, 2) Well-design game, 3) Effective video game, 4) Four major learning theories, and 5) Key elements of a game.These components have been discussed in the previous section in details and the analyses' results are illustrated in Table 3.
The EGM framework has all the components under the developing game design category.Furthermore, 11 frameworks (FDF, SF, CMBD, LCL, DGBL, EGDM, OBDM, GOM2, EGM, FFCM, and SGDM) uses pedagogical effectiveness component.Nine frameworks (FDF, SF, LCL, GG, OBDM, GOM 2, EGM, FFCM, and SGDM) have implemented evaluation of game component.There are fi ve frameworks (LCL, DGBL, OBDM, GOM 2, EGM) who use the gamifi cation component and only two frameworks (CMBD and EGM) used cost-effectiveness component.This shows that two most critical elements, pedagogical effectiveness and evaluation of game, should be highly considered when the game design is being developed.
Second category, well-designed games, there are eleven frameworks (FDF, CMBD, LCL, DGBL, EGDM, GG, OBDM, GOM 2, EGM, FFCM and SGDM) which have used interactive problem solving component and ten frameworks (FDF, SF, LCL, DGBL, EGBL, OBDM, GOM 2, EGM, FFCM, and SGDM) used specifi c goals/rules components for well-designed games.Ongoing feedback component has been used in eight frameworks (FDF, CMBD, LCL, GG, GOM 2, EGM, FFCM, and SGDM).Adaptive challenge is also an important component that is considered by six frameworks (SF, CMBD, GG, OBDM, GOM 2, and EGM).There are fi ve frameworks (FDF, LVL, GOM 2, EGM, and FFCM) that have used sensory stimuli, where they have added fun to the game design, whereas, control of design is also involved in the fi ve frameworks (SF, LCL, GG, EGM, and FFCM).The least involved component, uncertainty to gain engagement, is implemented in two frameworks (DGBL and GOM 2).With this analysis, we can easily understand that many researchers have considered the interactive problem solving design of the game and the games must have specifi c goals/rules where the learners are aware of.Ongoing feedback must also be considered while designing the game as it provides and update the learners as they play the game.Only six frameworks used the adaptive challenges, which is indeed a serious problem.If the game is adaptive, it is more interactive, and with constant challenges provided to the learners, the learners are able to achieve the goals or objectives easily.The fourth category is the learning theories.There are six frameworks (DGBL, EGDM, GG, FFCM, EGM, and SGDM) that have used humanism theory while six other frameworks (FDF, SF, LCL, DGBL, GOM 2, and EGM), used cognitivism.This shows that the authors who have developed the framework mostly consider these theories.Furthermore, fi ve frameworks (SF, CMDB, GG, OBDM, and FFCM) have used behaviorism theory.Finally yet importantly, four frameworks (FDF, GOM 2, EGM, and SGDM) have implement constructivism theory.The authors have showed humanism and cognitivism to be the most crucial theory; however, four major theories must be revised before or while developing the frameworks.
The fi fth category is the key elements of game.There are seven key elements in the Table 3.All the seven elements were implemented in two frameworks (SF and GOM 2).There are nine frameworks (SF, CMBD, LCL, EGDM, GG, OBDM, GOM 2, EGM, and FFCM) that discuss outcome element importance and include this element into their framework.Feedback is also implemented in nine frameworks (FDF, SF, CMBD, LCL, DGBL, GG, GOM 2, EGM, and FFCM).Interaction element is also emphasized in nine frameworks (FDF, SF, DGBL, EGDM, OBDM, GOM 2, EGM, FFCM, and SGDM).Six frameworks (FDF, SF, DGBL, EGDM, OBDM, and GOM 2) emphasized on the confl ict component where they believed that competition and challenge is very important to be considered.Representation element is also implemented in six frameworks (FDF, SF, CMBD, LCL, GOM 2, and EGM).Last but not least, there are fi ve frameworks (FDF, SF, DGBL, GOM 2 and EGM) have implemented the rules that need to be followed by the learners as they play the educational game.
Frameworks have been proposed from the developers' perspective and these frameworks covers some development concerns such as development process and usability.However, the research from developers' perspective is still immature, as they lack in interpreting the frameworks in perspective of methods or technique that can be adopted for the game developers to develop their games effi ciently.
Research from the development perspective is important because educational games are also a type of software and as such, software engineering practices should be practice in its development.According to (Hirumi & Stapleton, 2009), developing educational games require the involvement of people from education and game development.Therefore, communication between both parties is essential in ensuring a successful development.Furthermore, the importance of communication between stakeholders is well established in the fi eld of software engineering.A proof of this fact is the creation of Agile development model (Sommerville, 2006) where one of the main principal of agile development is active communication between stakeholders.In educational game development, the method of communication between stakeholders, the roles played by each stakeholder and the information that is provided by each stakeholder are among the topics that are still under research.
The information (also known as requirements in software engineering) given by clients (i.e.teachers, students and educationist) needs to be captured and the method to do this is not well established for educational games.Some of the common methods to record the requirements for other types of software are formal methods, modelling, and natural language documentation.The research on adopting these methods to educational game development, to the best of the authors' knowledge, is still unheard of.
As has been discussed in this paper, there exist frameworks and modeling languages for designing educational games.However, from software engineering perspective, research in the design of educational games is still in its infancy level.The translation from the requirements to the model and framework is yet to be elaborated.Furthermore, there is no discussion on how to build the educational games from the models or use the models for verifying effectiveness of the educational games.Model-Driven Engineering (MDE) (Meyers & Vangheluwe, 2011;Molina & Toval, 2009;Nodenot, Caron, Le Pallec, & Laforcade, 2007;Tang & Hanneghan, 2010;2011)  leverages on the use of well-defi ned models of the game.However, for the development of educational games to use MDE, a domain-specifi c modeling language with well defi ne semantics is required.From our survey, GOM2 can be used as a modeling language but it only covers the structural elements of the game.Modeling language for the behaviors of educational game is to our knowledge not available.Another fi eld that may be of benefi t to educational game development is model-based testing (MBT).MBT is a method to improve the testing process of software by referring to the models for guidance (Apfelbaum & Doyle, 1997;Broy, 2005;Dalal et al., 1999;Hemmati, Arcuri, & Briand, 2013;Prensky, 2005;Shafi que & Labiche, 2010;Timmer, Brinksma, & Stoelinga, 2011).
Verifying effectiveness of the game has been explored many times (Omar & Jaafar, 2010) and the most common method is a combination of acceptance testing and quantitative/qualitative research methods.Other methods such as unit testing, white-box testing and verifi cation using models are yet to be well known.Further research is required for these areas: 1) Communications between stakeholders to effectively capture all requirements.

2)
Modelling languages to be used in all parts of development from requirements to detailed design.

3)
Verifi cation of educational games effectiveness from models and designs.

4)
Traceability from requirements to code.

5)
Testing educational games other methods other than acceptance testing.

CONCLUSION
This paper provides a review on several frameworks for the development of educational games.The authors have identifi ed four different types of movement in the educational game design framework research.The fi rst movement is the framework demonstrates the key elements of an educational game.The second movement produced frameworks that are built around learning theories and include the process of learning.The third movement starts producing frameworks that go beyond learning theories and add elements important in GBL implementation in a formal education setting.The fourth movement is a change from learning centric framework to game development centric framework.
The authors have also analyzed the compliance of these design frameworks with several effective educational games criteria and their support of learning theories.The learning theories most supported by the design frameworks are experiential learning and cognitivism.On the other hand, the effective educational games criteria that researchers focused more are pedagogical effectiveness, interactive, specifi c objectives and goals, and feedback.The authors have also discussed and recommended a few research topics that are important to educational games development but are not emphasized enough by the research community.
For the future work, the authors will focus their research towards modeling educational games with high emphasis on verifying the effectiveness of the games from its models.Towards this end, the authors will work on improving the existing frameworks to highlight the relationships between the elements of an educational game.These relationships are crucial in establishing how one element collaborates with other elements to ensure effectiveness of the game.Without these relationships, effectiveness of the design cannot be measured.

Figure 7 .
Figure 7. Framework of fl ow in computer-mediated environments

Figure 9 .
Figure 9. Proposed components in DGBL Model for History Educational Games Design

Figure 10 .
Figure10.A conceptual model for 3D games-based learning intelligent tutoring system using multi-agent and machine learning technologies byMills & Darlgarno (2007).

Figure 11 .
Figure 11.Educational Game Design Model

Figure 12 .
Figure 12.Game Object Model II

Figure 13 .
Figure13.Serious-games design model: process, principles, and tools and techniques for supporting serious-games design and evaluation(Jarvis & de Freitas, 2009) .

Table 3
Summary Matrix of Game-based learning Frameworks and Models