Advances in Game-Based Learning

Fengfeng Ke
Valerie Shute
Kathleen M. Clark
Gordon Erlebacher

Interdisciplinary
Design of Game-
based Learning
Platforms
A Phenomenological Examination of the
Integrative Design of Game, Learning,
and Assessment
Advances in Game-Based Learning

Series Editors
Dirk Ifenthaler
Scott Joseph Warren
More information about this series at http://www.springer.com/series/13094
Fengfeng Ke • Valerie Shute
Kathleen M. Clark • Gordon Erlebacher

Interdisciplinary
Design of Game-based
Learning Platforms
A Phenomenological Examination
of the Integrative Design of Game,
Learning, and Assessment
Fengfeng Ke Valerie Shute
Department of Educational Psychology Department of Educational Psychology
and Learning Systems and Learning Systems
Florida State University Florida State University
Tallahassee, FL, USA Tallahassee, FL, USA

Kathleen M. Clark Gordon Erlebacher

School of Teacher Education Department of Scientific Computing
Florida State University Florida State University
Tallahassee, FL, USA Tallahassee, FL, USA

ISSN 2567-8086     ISSN 2567-8485 (electronic)

Advances in Game-Based Learning
ISBN 978-3-030-04338-4    ISBN 978-3-030-04339-1 (eBook)
https://doi.org/10.1007/978-3-030-04339-1

Library of Congress Control Number: 2018962489

© Springer Nature Switzerland AG 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims
in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Acknowledgment

The E-Rebuild project is funded by the National Science Foundation, grants

1318784 and 1720533. Any opinions, findings, and conclusions or recommenda-
tions expressed in this book are those of the authors and do not necessarily reflect
the views of the National Science Foundation.
This work would not have been possible without the participation and support of
our E-Rebuild design and research team members. We are especially indebted to
Danial A. Smith, an E-Rebuild lead programmer; Dr. Peyman Fazian, an E-Rebuild
level and assessment programmer; Dr. Anne Taylor, an architecture consultant;
Sungwoong Lee, an E-Rebuild learning support designer; and Biswas Parajuli, an
E-Rebuild assessment programmer. We are also grateful to all teachers and students
who have participated in the design-based research of E-Rebuild.

v
Contents

1 Introduction and Prior Research Review ����������������������������������������������   1

2 Chronicle of Designing a Game-Based Learning Platform ������������������ 15
3 Interdisciplinary Design Activities and Patterns ������������������������������������ 51
4 Design of Gameplay for Learning ���������������������������������������������������������� 75
5 Interweaving Task Design and In-Game Measurement ������������������������ 99
6 Designing Dynamic Support for Game-Based Learning ���������������������� 119
7 An Evolving Design Framework for Game-Based Learning
Platforms���������������������������������������������������������������������������������������������������� 141

Index ������������������������������������������������������������������������������������������������������������������ 153

vii
Chapter 1
Introduction and Prior Research Review

Abstract Games are not just a vehicle to enhance learning but a new way of
understanding and organizing learning. The starting point of an optimal integration
of learning and play in the game setting is to identify the salient elements of learn-
ing itself and the inherent learning processes in gameplay. In this introductory
chapter, we discuss relevant theoretical perspectives on the nature of knowledge
and learning that guide the exploration of playful elements in learning and conse-
quently the opportunities for learning that games offer. We then provide a concep-
tual review of prior research of game design and discuss the challenges and the
interdisciplinary nature of educational game development. In the end, we intro-
duce the overall goal, structure, and chapters of this book.

1.1 Games for Learning

The best way to motivate students to embrace learning is to “make it fun”, and
hence making learning fun seems the most well-known claim on the educational
power of games (Kim, Song, Lockee, & Burton, 2018). Games are structured or
organized play, requiring the player to follow a specific set of rules to tackle obsta-
cles and attain a goal (Klopfer, Osterweil, & Salen, 2009; Suits, 1978). Although
games are typically considered as a high-quality multisensory rendering environ-
ment, the main reasons for the interest in game for learning are not the medium’s
success but the motivation of players and their deep engagement while playing
(Denis & Jouvelot, 2005).
While learning and play are often associated with opposite connotations
(Mitgutsch, 2009), learning and play share a key attribute: Both are challenging,
time-intensive, and interactive processes that require cognitive effort and willing-
ness to acquire new knowledge or skills. Yet people tend to avoid and dislike chal-
lenging learning experiences in the educational setting while engaging in and
enjoying gameplay as a form of hard fun. This observation, along with the interest
in creating intrinsically motivating learning, has triggered the examination of
games as a promising tool and environment for learning. Particularly, prior

© Springer Nature Switzerland AG 2019 1

F. Ke et al., Interdisciplinary Design of Game-based Learning Platforms,
Advances in Game-Based Learning, https://doi.org/10.1007/978-3-030-04339-1_1
2 1 Introduction and Prior Research Review

research has focused on defining the nature and circumstances of “pleasantly

frustrating” learning experiences and how these are and can be aligned with the
gaming experience (e.g., Gee, 2003; Rieber, 1996). Some scholars have also
examined the phenomena of game engagement from both intrinsic motivation and
well-internalized extrinsic motivation perspectives (e.g., Csikszentmihalyi, 1997;
Deci & Ryan, 1985; Malone & Lepper, 1987).
The idea of using games for learning purposes was discussed early by Clark
C. Abt in 1975 and further formulated by other scholars of serious games (e.g.,
Breuer & Bente, 2010; Charsky, 2010; Sawyer, 2003; Susi, Johannesson, &
Backlund, 2007). The educational purpose can be assigned to a game in the design,
or by the context used. Games designed primarily for entertainment, like commer-
cial off-the-shelf (COTS) games, can be used in a teaching and learning context,
though its learning effect is an exogenous one.
Despite the innate connection between gameplay and learning, not all forms of
games are equally suitable for learning purposes; simply adding educational mate-
rial to an enjoyable game or presenting it in a game-like setting will not create an
engaging and effective game for learning purposes (Breuer & Bente, 2010; Ke,
2016). A frequent cause of a segmentation between play and learning in a game is
that win criteria for a game do not necessarily equate with the things players are
supposed to learn through playing the game (Van Staalduinen & de Freitas, 2011).
Specifically, there are variant categories of learning in games, including things that
we can learn and are deliberately designed in a game (or intentional learning), things
that we can learn but not necessarily designed into the game (or collateral learning),
and things that we must learn in a game to successfully finish the game (Becker,
2008). The desirable situation is that the first two categories overlap with the third
one, in that both intended and incidental learning are enabled and assured by
gameplay. The starting point of an optimal integration of play and learning in the
game setting should be to identify the potential enjoyment (or the playful elements)
of learning itself and the inherent learning processes in gameplay (Rodriguez,
2006). Hence, games are not just a vehicle to enhance learning but “a new way of
understanding and organizing learning” (Breuer & Bente, 2010, p. 14).
In this introductory chapter, we describe relevant theoretical perspectives on the
nature of knowledge and learning that help us to explore playful elements in learn-
ing and consequently the opportunities for learning that games offer. We then pro-
vide a review of prior research of game design and discuss the challenges and the
interdisciplinary nature of educational game development. We introduce the overall
goal, structure, and chapters of this book at the end.

1.2 Nature of Knowledge and Learning Pertinent to Games

Dewey defined knowledge as “not information, but a mode of intelligent practice

and habitual disposition of mind” (1910, p. 124). This definition highlights two
salient elements of knowledge that are apt to frame the design of game-based
1.2 Nature of Knowledge and Learning Pertinent to Games 3

learning: a practice that crafts and enacts intelligence and intelligent habits or
ways of thinking constituted in the practice. The intelligent practice, called reflec-
tive inquiry by Dewey, comprises mainly the processes of identifying a problem
(i.e., a quandary that needs to be resolved), studying the problem through active
engagement, and reaching cognitive conclusions as the problem is resolved. The
central facet is the method by which a cognitive conclusion is reached; hence, the
process/action of identifying, actively studying, and resolving a problem, more
than the cognitive conclusion itself, is the item of value (Dewey, 1929; Hiebert
et al., 1996). During the action of inquiry or problem-solving, scientific attitude—
“willingness to endure a condition of mental unrest and disturbance” or that is
capable of enjoying the doubtful—should be deliberately constructed and is alone
worthy of the title of knowledge (Dewey, 1910, p. 13). In view of that, perfecting
the method of inquiry and constructing scientific attitude are more significant to
realize the meaning of knowledge than that of acquiring and utilizing information
and technical procedures.
Derived from Dewey’s perspective is the argument for problematizing subject
matter knowledge (Hiebert et al., 1996). “Allowing the subject to be problematic
means allowing students to wonder why things are, to inquire, to search for solu-
tions, and to resolve incongruities” (Hiebert et al., 1996, p. 12). The conventional
conception of problem-solving typically focuses on the application of acquired
knowledge in realistic situations, whereas problematizing a topic emphasizes knowl-
edge acquisition. Problematizing mathematics, for example, involves treating expe-
riences as problems, examining or manipulating them via overt doing, and
consequently constructing structural understanding that includes insights into the
structure of mathematical objects, strategies for solving problems (both procedures
used for particular problems and general ways of thought needed to construct the
procedures), and dispositions toward mathematics (e.g., seeing mathematics as an
intellectual activity in which they can participate). Notably, real-life problems pro-
vide a legitimate context for problematizing mathematics, though it is only one
context for it. The value of a real-life problem depends on whether students prob-
lematize the situation and whether it offers the chance for students to acquire struc-
tural understanding after problem solving. In other words, the knowledge to be
acquired depends as much on the mathematical ideas embedded in the task (e.g., a
game-based inquiry) as on the way it is packaged. Therefore, the focus of the design
research of learning in games is not only on what subject content or information is
integrated but how core game tasks will legitimize subject-problematizing actions to
generate structural understanding.
In alignment with the aforementioned learning perspectives is the proposition to
establish the learning activity or task as an epistemic practice (Eriksson & Lindberg,
2016). With a similar assumption that knowledge is constituted in people’s actions,
an epistemic learning practice is characterized by open and question-generating
objects (or tasks) with which learners work. Apart from contextualizing a learning
task in relation to students’ everyday interests, it is important to choose and shape a
task so that the students discern that the previous tools and solutions they are famil-
iar with are restricted. When working with the learning task, students get to test
4 1 Introduction and Prior Research Review

previously developed methods and tools and (re)develop theoretical generalizations

that are characteristic for a specific subject. In other words, the learning task should
be developed to enhance such kinds of theoretical thinking, by requiring that “core
principles” or “conceptual relations” constituting a specific knowing are discerned
and understood through learning actions in a content-rich epistemic practice.
In summary, the above theoretical underpinnings for learning in games suggest
that the acquisition of discipline-specific, structural understanding (or theoretical
thinking), compared with the application or acquisition of static content or informa-
tion, is more crucial for the goal of the learning-play integration. Subsequently,
designing gameplay as an epistemic learning practice that delineates subject-­
problematizing actions is a fundamental means to establishing game-based learn-
ing. Educational game design is to create meaningful, interactive, and challenging
experiences involving the user as the conductor of his own intellectual practice and
development.

1.3 Designing Games for Learning

1.3.1 Research of Game Design

The content of a game is its behavior, not the media that streams out of it toward the
player (Hunicke, LeBlanc, & Zubek, 2004). The difference between games and
other design artifacts is that their consumption (by players) is relatively unpredict-
able, because the interaction between players and the coded game subsystems cre-
ates dynamic and often unpredictable behaviors (Hunicke et al., 2004). Gameplay is
hence defined as the structures of player interaction with the game system and inter-
action with other players (Björk & Holopainen, 2005).
Although game design is a frequent theme of textbooks and articles, a unified
design theory diagnosing and predicting the structural relations between desirable
gameplay and game subsystems is still lacking. The mechanics, dynamics, and aes-
thetics (MDA) framework by Hunicke et al. (2004) is an earlier effort contributing
to this aim and describes three sectors of a game system. Mechanics are various
actions, behaviors, and control mechanisms afforded to the player within a game
context. Dynamics refers to the run-time behavior of the mechanics acting on player
inputs over time or gameplay. Aesthetic represents the desirable emotional responses
or experiences evoked in the player. From the designer’s perspective, the mechanics
give rise to dynamic system behavior, which in turn leads to particular aesthetic
experiences. From the player’s perspective, their respective player or aesthetic expe-
riences, such as sensation, fantasy, challenge, fellowship, discovery, and expression,
are brought forth by observable dynamics and, eventually, operable mechanics.
Even though the MDA framework lacks instruction detailing the combination and
proportion of mechanics that will result in variant aesthetic goals, it helps to describe
how and why different games appeal to different players or to the same players at
1.3 Designing Games for Learning 5

different times. It should be noted that in games for learning purposes, the desirable
player experiences are not just emotional responses but more cognitive involve-
ment. However, inadequate game design research addresses the mechanics and
dynamics in relation to the cognitive facet of player experience.
Extending the MDA framework, Salen and Zimmerman (2005) discussed mean-
ingful play as an emotional and psychological experience in a game that emerges
from the relationship between player action and system outcome. Specifically, they
described that meaningful play occurs when the relationships between gaming
actions and their outcomes in a game are both discernable and integrated into the
larger context of the game. Discernable means that the result of the game action is
communicated to the player in a perceivable way. Integrated means that an action a
player takes not only has immediate significance in the game but also affects the
play experience at a later point in the game. In other words, game design is the pro-
cess by which a designer creates an interactive and engaging system to be encoun-
tered and proactively investigated by a player, from which meaning emerges.
Importantly, interactivity afforded by the coded/designed system comprises not
only functional interactivity (or the usability of the interaction interface) but also
cognitive interactivity—psychological, emotional, and intellectual participation
between a participant and a system.
Prior research of game design tends to depict play experience as the product of
coding operable mechanics and observable interactivity. But there is still a debate
between ludology and narratology in the game design literature. In comparison with
the argument for ludology—systems of puzzles, rules, and interactivity designed to
frame play or fun (e.g., Adams & Dormans, 2012; Koster, 2013; Salen, Tekinbaş, &
Zimmerman, 2004)—the narratology proposition (e.g., Jenkins, 2004) emphasizes
the design of narrative structure as the anchor of play experience, describes the
importance of designing environmental storytelling via spaces and artifacts in a
game, and argues that gameplay enacts a personally meaningful story or life experi-
ence whether story is a defining game feature or not.
Recent discussions of game design (e.g., Adams, 2014; Schell, 2014) manage to
coordinate the perspectives of ludology and narratology by arguing that mechanics
(or interactivity) and environmental storytelling (or embedded narrative) in a game,
jointly and ultimately, serve the purpose of experience design. Play or fun is co-­
framed by players’ prior experiences, the designed game system, and the emergent
interactivity (dynamics) between the players and the system. Even though compo-
nents of the experience to be designed are well specified in the game literature, the
paths toward the goal are dynamic and murky. Due to such a dynamic and participa-
tory nature of play experience design, mostly game design is based on the intuition
and experience of designers with the infield experimentation and iterative refining.
Descriptive or analytical analyses of the game elements, design rationales, and
experimented design strategies, however, will help to guide the experimentation.
Particularly, empirical and longitudinal research on variant processes of design cre-
ation and experimentation in diverse game design projects will help us construct the
collective knowledge as well as a unified while scalable theory of game design.
6 1 Introduction and Prior Research Review

1.3.2 Prior Research of Educational Game Design

Multiple conceptual or design models of educational games, or game-based

learning platforms, have been proposed in the literature of instructional design
and learning technologies. For example, Kiili (2005) proposed an experiential
gaming model that reflects the perspective that learning is a construction of cog-
nitive structures through action or practice in the game world. The model depicts
game challenges (or problems) as the center of game-based learning flow by
initiating active experimentation, reflective observation, and schemata construc-
tion. Kiili’s model outlines the process of experiential game-based learning with
its implications for educational game development but lacks a design description
of the game systems. Amory (2007) described a theoretical framework, called
the game object model, for educational game research and development. The
model sets game space, problem space, and social space as the three components
of the educational game system and prescribes a list of defining features for each
space. For example, drama, role models, interaction, and gestures are listed as
the game space features; explicit knowledge, literacy, and information commu-
nication are considered design elements of the problem space, while the social
space includes computer-mediated communication along with social network
analysis. In general, Amory’s game object model presents a visionary integration
of related theoretical constructs of gaming, learning, and social community. It
has not explained the actual design mechanism and lacked technical definitions
of the theoretical constructs for designers. Similarly, de Freitas and Oliver (2006)
used a four-dimensional framework to summarize the key analytical or evalua-
tion constructs (i.e., learner specifics, pedagogy, representation, and context) of
a virtual-world and gaming-based learning environment. These aforementioned
generic models are all aimed to outline and categorize related constructs under-
lying the integration of gaming and learning processes. They focus on a conjec-
tural or an interpretive analysis of the relationships and structure of these
constructs as a system, from more of an educational researcher perspective.
Other works, such as that of Calvillo-Gámez, Cairns, and Cox (2015) and Dickey
(2007), intend to explore educational game design from the player or user per-
spective. They highlight an interpretive and evaluative analysis of the salient
elements with corresponding design norms of an optimal gaming experience and
then discuss how such design norms can be extended with or transferred to the
instructional design principles.
Empirical studies examining educational game design approaches and fea-
tures, compared with those evaluating a game’s learning effectiveness, are lim-
ited. Our recent review of the literature with the search phrase of “game
design” + “learning” or “education” generated mostly articles that examined
game design as a learning-­by-making activity by participants (e.g., Kafai, 2012;
Robertson & Howells, 2008) or used game design elements in nongame educa-
tional settings (e.g., Deterding, Dixon, Khaled, & Nacke, 2011). Published design
research of educational games is frequently a conjectural or an evaluative analysis
1.3 Designing Games for Learning 7

with existing COTS or serious games (e.g., Annetta, 2010; Dickey, 2007;
Mitgutsch & Alvarado, 2012). Pinelle, Wong, and Stach (2008), for example, dis-
cerned 10 principles of game usability—“the degree to which a player is able to
learn, control, and understand a game”—after reviewing 108 different games
from six major game genres (p. 1453). These usability principles, applicable for
game design in general, are: allowing users to customize video/audio settings,
difficulty, and game speed, allowing users to skip non-playable and frequently
repeated content, and providing (a) unobstructed views appropriate for the user’s
current actions; (b) intuitive and customizable input mappings; (c) controls that
are easy to manage and that have an appropriate level of sensitivity and respon-
siveness; (d) users with information on game status, providing training and help;
and (e) visual representations that are easy to interpret and that minimize the need
for micromanagement. All of these game usability principles are transferrable to
the design practice of educational games and are highly consistent with the infield
experimentation findings of our current project.
On the other hand, there is still a lack of studies, especially design-based ones,
that present infield testing or design experimentation findings governing the
development processes and strategies of educational or serious games (Torbeyns,
Lehtinen, & Elen, 2015; Warren & Jones, 2017). In particular, a rich, data-driven,
and theory-contributing description of educational game-design exploration and
findings should help us develop structural and in situ knowledge of game design.
The studies by Andersen, Butler, O’Rourke, Popović and their colleagues (2011,
2014, 2015), for example, examined variant design features and development
mechanisms, including game objective, level progression strategy, incentive
structure, and scaffolding, of a math learning game (called “refraction”). Based
on the play data of online gamers and the perspectives of human-computer inter-
action and educational psychology, their research efforts and findings helped to
illustrate diverse design challenges, the framework of design solution exploration
and generation, and solution validation/refining strategies during the educational
game design process. In another example, Denis and Jouvelot (2005) conducted
a case study on design strategies that were aimed to reconcile learning and fun in
a video game dedicated to music education. Building on the self-determination
motivation theory (Deci & Ryan, 1985), they explored game design patterns or
features that helped to reinforce competence, autonomy, and relatedness in game-
based learning. Their design-based research contributed tangible heuristics of
educational game mechanics design, such as reifying values into rules to convey
knowledge in interactions rather than static data, providing the player expressive
ways to confront with and test rules, tuning usability so that entry barriers (e.g.,
technical difficulty or the game’s gender bias) that go against the player’s urge to
practice gameplay will be leveled, sequential embedding of novel challenges and
information in game levels to present a positive slope of the learning curve, and
a natural and stealth assessment mechanism of in-game performance. All of these
previous learning integration game design suggestions have shed light on our
design efforts in this current project and have been corroborated by our project
findings.
8 1 Introduction and Prior Research Review

1.3.3  nduring Challenges and Interdisciplinary Nature

E
of Game-Based Learning Design

The design of games for learning resembles the innate challenge of entertainment
game design: a definite formulation of gaming behaviors or consequences is diffi-
cult because interactions between players and the coded game system are often
complex, dynamic, and unpredictable. Moreover, play is voluntary and intrinsically
motivating and involves proactive cognitive engagement that allows for the freedom
of effort and interpretation, whereas these characteristics do not automatically apply
to game-based learning (Shute & Ke, 2012). Therefore, existing approaches and
tools for developing entertainment games cannot simply be transferred to educa-
tional games or game-based learning platforms. There are multiple enduring and
distinctive challenges associated with the design of games for learning.
First, problematizing a subject and reinforcing an investigative or epistemic
practice with problem-solving are believed to be crucial to game-based learning, as
discussed above. But not all domain knowledge or topics share the same approach
or advantage in being problematized. Certain learning processes are by nature
abstract and sign-mediated and hence cannot always be situated or contextualized.
Reality isn’t always “fun”—mechanics that impact the dynamics to create the desir-
able emotional responses may negatively impact the game’s ability to recreate the
reality of epistemic practices to be simulated. In particular, we lack understanding
of how to design a game to reinforce purposeful development of abstract or sym-
bolic domain knowledge.
Second, a game is typically an open-ended learning environment that empha-
sizes self-regulated experiential learning as well as reflective inquiry. However, not
all students are by default self-regulated learners or capable of performing sense-­
making during game-based problem-solving. Students also differ in their prior
domain knowledge, game skills, motives for gaming, and preferences in relation to
modes of play. Their attention span with textual information in the game world can
be vastly reduced due to the multisensory rendering environment. Given such a
heterogeneous player group and the goal to engage and educate all students with a
single game-based learning platform, designing a motivating while learning-­
constructive gaming experience is particularly challenging.
Third, prior research and practice of game-based learning tends to underline a
stealth (or incidental) learning idea. Such a perspective is sensible especially when
only generic scientific attitudes and generic thinking skills are the focus of learning
outcomes. On the other hand, learning by nature is a conscious activity. The desir-
able player responses in game-based learning should not be just emotional but cog-
nitive, metacognitive, and domain-specific. Inclusion of the idea of informed and
intentional learning in games (Annetta, 2010) will entail the integration of gameplay-­
coherent learning supports in the game that promote purposeful knowledge con-
struction while maintaining motivation and the game flow. A general game design
framework (e.g., the MDA framework) has to be extended to address these addi-
tional facets of aesthetics or player experience. Corresponding with the inclusion of
1.4 Goal and Structure of the Current Book 9

intentional learning to game-based learning platform design is the challenge to

build and coordinate functional usability/interactivity and cognitive usability/inter-
activity. Rather than creating an intuitive interface, one should also design the inter-
face to be intermediary between gaming actions and action-based knowledge
construction.
Ultimately, another unique change of educational game design is to develop
embedded assessment for game-based learning outcomes that assess performance-­
based competency development. Even though recently there is increasing research
on the methods of educational data mining in game-based learning assessment,
designing and using games as both a learning and an assessment tool are challenging
in that gradual learning progression can be in conflict with evidence accumulation
for assessment in the game setting. How to achieve an integral design and develop-
ment of learning tasks and the tasks that discern competency status remains murky.
Interdisciplinary Nature of Game Design Games, acting as an interactive micro-
world, a digital toy, and an environmental storytelling device simultaneously, call
for the integration of multiple disciplinary perspectives and approaches. A game
artifact involves the application of communication/information design, visual
design, human-computer interaction, and software engineering. Designing games as
an effective learning platform, given all of the abovementioned challenges, entails
interdisciplinary knowledge and methods of additional disciplines, such as the sub-
ject domain knowledge, learning system design, psychology, and educational mea-
surement. Interdisciplinary collaboration among the design team members from
these different fields, by itself, is a challenging process. It has to balance and incor-
porate their often conflicting values, ways of thinking, domain-specific language,
and methods, especially those of game designers, instructional designers, content
experts, and psychometricians. Although cross-cultural communication and inter-
disciplinary design are recurrent topics in workplace studies and engineering educa-
tion research, the functional processes and patterns of interdisciplinary design for
game development are generally absent in the literature.

1.4 Goal and Structure of the Current Book

In spite of the plethora of research on game-based learning, the design descriptions

or operational conjectures on how to effectively and intrinsically design intentional,
domain-specific learning in gameplay are scarce. Multiple enduring challenges asso-
ciated with educational game design remain unaddressed. The previous efforts to
explain the framework of game-based learning typically focus on an evaluative anal-
ysis of existing games or present a general and theoretical level of reasoning, thus
lacking tangible guidance for the actual educational game design. Game design is
multidisciplinary in nature, yet few efforts have contributed to describing the inter-
disciplinary design process, techno-pedagogical design knowledge, or accessible
design language.
10 1 Introduction and Prior Research Review

Based on the four-year design and research of an architecture simulation game

for mathematical learning (called E-Rebuild), this book is aimed to present a phe-
nomenological examination and explanation of a functional design framework for
games in education. It aims to provide a rich description of the experiences and
perceptions of designing a digital learning game and of performing interdisciplinary
collaborative design among experts of different fields.
Specifically, it will address both practical and conceptual issues about the design
of play-based learning systems that align and interweave game and learning mechan-
ics, evidence-centered design of learning tasks and assessment, and in-game learn-
ing support. It also aims to explicate the process of coordinating the interdisciplinary
language and perspective differences in design communication and negotiation for
the future development of learning games.
In Chap. 2, we provide an overview of the integrative, interdisciplinary design
process of E-Rebuild and the adopted phenomenological research approach. Core
phases of designing a game-based learning platform are illustrated with a longitu-
dinal log of the iterative design, experimentation, and refining of E-Rebuild. In
Chap. 2 we also present a positioning statement and design reflection notes of
each co-­author (and interdisciplinary design team member). In Chap. 3, we
explain salient interdisciplinary design activities and patterns that emerged from
the E-Rebuild project. The description of the interdisciplinary design patterns in
this chapter consists of a contextualized design narrative/account of core design
processes and an analytical synthesis of core design pattern elements—a design
problem statement with its context and specifics, the solution or technique (with
its key structure or mechanism) to solving the stated problem, and the pattern of
transferring or scaling this design solution or design move. In Chap. 4, we report
design challenges associated with the core components of gameplay and review
the gameplay design propositions and infield test findings of E-Rebuild. We dis-
cuss how domain-specific learning is integrated in and activated by core game
actions, rules, game objects, and the game world design. In Chap. 5, we describe
an integrative-design approach that interweaves game-based task design with in-
game assessment of learning. We discuss core design processes and functional
conjectures on the approaches of task generation and evidence accumulation, with
support of infield observations on the implementation feasibility and outcomes of
the design assumptions. In Chap. 6, we explore the design of dynamic game-
based learning support. We review prevalent practices and prior research of scaf-
folding and support in game-based learning, share our observations of the
obstacles that learners experienced in game-based learning processes, and report
the corresponding learning support strategies with infield testing findings of
E-Rebuild. In the conclusive Chap. 7, we discuss an emerged, functional design
framework for game-based learning platforms, with a summary of the design
problem structuring process for the interdisciplinary, integrative design of game-
based learning. We conclude the chapter and the book by discussing potential
directions of future design and research efforts of game-based learning design,
with a set of core design concepts highlighted.
References 11

References

Abt, C. C. (1975). Serious games. New York: Viking Compass.

Adams, E. (2014). Fundamentals of game design. New York: Pearson Education.
Adams, E., & Dormans, J. (2012). Game mechanics: Advanced game design. Berkeley, CA: New
Riders.
Amory, A. (2007). Game object model version II: A theoretical framework for educational game
development. Educational Technology Research and Development, 55(1), 51–77.
Andersen, E., Liu, Y. E., Snider, R., Szeto, R., Cooper, S., & Popović, Z. (2011, June). On the
harmfulness of secondary game objectives. In Proceedings of the 6th International Conference
on Foundations of Digital Games (pp. 30–37). New York: ACM.
Annetta, L. A. (2010). The “I's” have it: A framework for serious educational game design. Review
of General Psychology, 14(2), 105–112.
Becker, K. (2008). The invention of good games: Understanding learning design in commercial
video games (Doctoral dissertation, University of Calgary).
Björk, S., & Holopainen, J. (2005). Patterns in game design. Boston, MA: Charles River Media.
Breuer, J., & Bente, G. (2010). Why so serious? On the relation of serious games and learning.
Journal for Computer Game Culture, 4(1), 7–24.
Butler, E., Andersen, E., Smith, A. M., Gulwani, S., & Popović, Z. (2015, April). Automatic game
progression design through analysis of solution features. In Proceedings of the 33rd annual
ACM conference on human factors in computing systems (pp. 2407–2416). New York: ACM.
Calvillo-Gámez, E. H., Cairns, P., & Cox, A. L. (2015). Assessing the core elements of the gam-
ing experience. In Game user experience evaluation (pp. 37–62). New York: Springer, Cham.
Charsky, D. (2010). From edutainment to serious games: A change in the use of game characteris-
tics. Games and culture, 5(2), 177–198.
Csikszentmihalyi, M. (1997). Finding flow: The psychology of engagement with everyday life.
New York: Basic Books.
Deci, E., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior.
New York: Springer Science & Business Media.
De Freitas, S., & Oliver, M. (2006). How can exploratory learning with games and simulations
within the curriculum be most effectively evaluated? Computers & Education, 46(3), 249–264.
Denis, G., & Jouvelot, P. (2005, June). Motivation-driven educational game design: Applying
best practices to music education. In Proceedings of the 2005 ACM SIGCHI International
Conference on Advances in computer entertainment technology (pp. 462–465). New York:
ACM.
Deterding, S., Dixon, D., Khaled, R., & Nacke, L. (2011, September). From game design elements
to gamefulness: Defining gamification. In Proceedings of the 15th international academic
MindTrek conference: Envisioning future media environments (pp. 9–15). New York: ACM.
Dewey, J. (1910). Science as subject-matter and as method. Science, 31, 121–127.
Dewey, J. (1929). The quest of certainty: A study of the relation of knowledge and action.
New York: G.P. Putnam’s Sons.
Dickey, M. D. (2007). Game design and learning: A conjectural analysis of how massively multiple
online role-playing games (MMORPGs) foster intrinsic motivation. Educational Technology
Research and Development, 55(3), 253–273.
Eriksson, I., & Lindberg, V. (2016). Enriching ‘learning activity’ with ‘epistemic practices’–
enhancing students’ epistemic agency and authority. Nordic Journal of Studies in Educational
Policy, 2016(1), 32432.
Gee, J. P. (2003). What video games have to teach us about learning and literacy. Computers in
Entertainment (CIE), 1(1), 20–20.
Hiebert, J., Carpenter, T. P., Fennema, E., Fuson, K., Human, P., Murray, H., et al. (1996).
Problem solving as a basis for reform in curriculum and instruction: The case of mathematics.
Educational Researcher, 25(4), 12–21.
12 1 Introduction and Prior Research Review

Hunicke, R., LeBlanc, M., & Zubek, R. (2004, July). MDA: A formal approach to game design
and game research. In Proceedings of the AAAI workshop on challenges in game AI (Vol. 4(1),
p. 1722–1726).
Jenkins, H. (2004). Game design as narrative architecture. In P. Harrington & N. Frup-Waldrop
(Eds.), First person (pp. 44–53). Cambridge, MA: MIT Press.
Kafai, Y. B. (2012). Minds in play: Computer game design as a context for children’s learning.
New York: Routledge.
Ke, F. (2016). Designing and integrating purposeful learning in game play: A systematic review.
Educational Technology, Research and Development, 64(2), 219–244.
Kiili, K. (2005). Digital game-based learning: Towards an experiential gaming model. The Internet
and Higher Education, 8(1), 13–24.
Kim, S., Song, K., Lockee, B., & Burton, J. (2018). Engagement and fun. In Gamification in
Learning and Education (pp. 7–14). Cham, Switzerland: Springer.
Klopfer, E., Osterweil, S., & Salen, K. (2009). Moving learning games forward. Cambridge, MA:
The Education Arcade.
Koster, R. (2013). Theory of fun for game design. Sebastopol, CA: O’Reilly Media.
Malone, T. W., & Lepper, M. R. (1987). Making learning fun: A taxonomy of intrinsic motivations
for learning. In R. E. Snow & M. J. Farr (Eds.), Aptitude, learning, and instruction volume
3: Conative and affective process analyses (pp. 223–253). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Mitgutsch, K. (2009). Passionate digital play-based Learning, (re)Learning in computer games like
Shadow of the Colossus. Eludamos, 3(1), 9–22.
Mitgutsch, K., & Alvarado, N. (2012, May). Purposeful by design?: A serious game design assess-
ment framework. In Proceedings of the International Conference on the foundations of digital
games (pp. 121–128). New York: ACM.
O'Rourke, E., Andersen, E., Gulwani, S., & Popović, Z. (2015, April). A framework for automati-
cally generating interactive instructional scaffolding. In Proceedings of the 33rd annual ACM
conference on human factors in computing systems (pp. 1545–1554). New York: ACM.
O'Rourke, E., Haimovitz, K., Ballweber, C., Dweck, C., & Popović, Z. (2014, April). Brain
points: A growth mindset incentive structure boosts persistence in an educational game. In
Proceedings of the SIGCHI conference on human factors in computing systems (pp. 3339–
3348). New York: ACM.
Pinelle, D., Wong, N., & Stach, T. (2008, April). Heuristic evaluation for games: Usability prin-
ciples for video game design. In Proceedings of the SIGCHI conference on human factors in
computing systems (pp. 1453–1462). New York: ACM.
Rieber, L. P. (1996). Seriously considering play: Designing interactive learning environments
based on the blending of microworlds, simulations, and games. Educational Technology
Research & Development, 44(2), 43–58.
Robertson, J., & Howells, C. (2008). Computer game design: Opportunities for successful learn-
ing. Computers & Education, 50(2), 559–578.
Rodriguez, H. (2006). The playful and the serious: An approximation to Huizinga’s Homo Ludens.
Game Studies, 6(1), 1604–7982.
Salen, K., Tekinbaş, K. S., & Zimmerman, E. (2004). Rules of play: Game design fundamentals.
Boston: MIT press.
Salen, K., & Zimmerman, E. (2005). Game design and meaningful play. In J. Raessens &
J. Goldstein (Eds.), Handbook of computer game studies (pp. 59–79). Cambridge, MA: MIT
Press.
Sawyer, B. (2003) Serious games: Improving public policy through game-based learning and
simulation. Woodrow Wilson International Center for Scholars. Retrieved from http://wwics.
si.edu/foresight/index.htm.
Schell, J. (2014). The art of game design: A book of lenses. New York: AK Peters/CRC Press.
Shute, V. J., & Ke, F. (2012). Games, learning, and assessment. In D. Ifenthaler, D. Eseryel, &
X. Ge (Eds.), Assessment in game-based learning: Foundations, innovations, and perspectives
(pp. 43–58). New York: Springer.
References 13

Suits, B. H. (1978). The grasshopper: Games, life and utopia. Toronto, ON: University of Toronto
Press.
Susi, T., Johannesson, M., & Backlund, P. (2007). Serious games: An overview. Retrieved from:
https://www.diva-portal.org/smash/get/diva2:2416/FULLTEXT01.pdf
Torbeyns, J., Lehtinen, E., & Elen, J. (2015). Describing and studying domain-specific serious
games. Switzerland, Cham: Springer.
Van Staalduinen, J. P., & de Freitas, S. (2011). A game-based learning framework: Linking game
design and learning. In M. S. Khine (Ed.), Learning to play: Exploring the future of education
with video games (pp. 29–53). New York: Peter Lang.
Warren, S. J., & Jones, G. (2017). Learning games: The science and art of development. Cham,
Switzerland: Springer.
Chapter 2
Chronicle of Designing a Game-Based
Learning Platform

Abstract Our phenomenological examination of learning game design is situated

in a four-year, longitudinal design-based research project that encompasses iterative
design processes to develop, refine, and study a game-based learning platform
called E-Rebuild. This chapter presents an introductory overview of the four facets
of the interdisciplinary educational game design—interdisciplinary collaboration,
learning-play integration, integrative task and assessment design, and game-based
learning support. We then provide a design chronicle of E-Rebuild as the key setting
of the phenomenon examined, by explaining its iterative design, testing, and refin-
ing processes. The authors’ researcher positionality and reflective summaries of
design experiences are presented as well.

2.1 Introduction

Adopting a phenomenological perspective, we aim to describe as richly and accu-

rately as possible the phenomenon of designing a game for the learning purpose (or
a game-based learning platform), from the perspectives of people involved while
refraining from pre-given frameworks (Giorgi, 2009; Groenewald, 2004). The
examination evolves around two overarching questions: What is it like to design a
game for the learning purpose? What salient design patterns, strategies, and under-
standings have been derived from the design effort?
Our phenomenological examination is situated in a four-year, longitudinal
design-­based research project that encompasses iterative design processes to
develop, refine, and study the game-based learning platform in situ. Specifically, we
have been designing and developing an architecture simulation game-based mathe-
matics learning platform (called Earthquake Rebuild or E-Rebuild) that also incor-
porates evidence-centered learning assessment via an unobtrusive collection and
analysis of performance data. A major conjecture of E-Rebuild is that architectural
design and modeling via this simulation game-based learning platform will engage
middle-­school students in problem-centered learning and epistemic practice of
mathematics.

© Springer Nature Switzerland AG 2019 15

F. Ke et al., Interdisciplinary Design of Game-based Learning Platforms,
Advances in Game-Based Learning, https://doi.org/10.1007/978-3-030-04339-1_2
16 2 Chronicle of Designing a Game-Based Learning Platform

Informants or participants are nine members of the design team of E-Rebuild.

They include the four authors of this book, in addition to one learning scientist,
two game programmers who are graduate students who majored (or are majoring)
in scientific computing and computer science, one student instructional designer,
and one architecture consultant. We have examined the data derived from archived
design and meeting notes, open-ended interviews, observation notes, and an arti-
fact analysis with the game that was iteratively designed and refined during the
four project years.
We analyzed the qualitative data in an attempt to identify themes and make func-
tional generalizations regarding the nature and perceived experience of collabora-
tively designing a game-based learning platform by an interdisciplinary team.
Following a general framework of phenomenological inquiry (Giorgi, 2009;
Groenewald, 2004), we have bracketed our positionality (as described in a later sec-
tion), delineated and clustered units of meaning from the data to form the themes,
and then extracted general and unique themes to make a composite description of
every facet of the phenomenon. These thematic descriptions of the interdisciplinary
game design are presented in Chaps. 3, 4, 5, and 6.
In this chapter, we present an introductory overview of the four facets of the
interdisciplinary educational game design that will be discussed in details in later
chapters. We then provide a design chronicle of E-Rebuild as the key setting of the
phenomenon examined, by explaining its iterative design, testing, and refining pro-
cesses. Ultimately, the authors’ researcher positionality and reflective summaries of
design experiences are presented.

2.2  ore Facets of Interdisciplinary Educational Game

C
Design

Four core facets of game design, governing the heuristics of interdisciplinary col-
laboration, learning-play integration, integrative task and assessment design, and
game-based learning support, have emerged from the thematic analysis of quali-
tative data. The first facet depicts general patterns framing the allocation and
negotiation of collaborative design efforts during every phase of game design.
The second facet sets the tone for the rest of game design with a specification of
core gameplay —what, in which way, and when learning is legitimized by basic
game actions, rules, and settings. The third facet involves a systematic develop-
ment of game tasks that exemplify and assemble core actions in meaningful con-
texts, to not only enable subject-specific learning progressions but also discern
and accumulate performance evidence of competency development. The fourth
facet captures the design endeavors to reinforce players’ motivated and inten-
tional cognitive efforts contributed to game action-based knowledge application
and acquisition.
2.2 Core Facets of Interdisciplinary Educational Game Design 17

2.2.1 Managing Interdisciplinary Collaborative Design

The challenges of collaborative design by an interdisciplinary team relate to the

management of unshared languages, values, and habitual methods ascribed to dif-
ferent forms of design, technological, and educational endeavor, within which is
a mutual lack of awareness of each other’s theoretical stance and motives about
game-­based learning. As Weingart (2000) identified, the way in which we see the
world and approach novel problems is affected by a host of subcultures that are
specific to individual disciplines, making truly interdisciplinary working difficult
to achieve. The disciplines of education, mathematics, architecture, and scientific
computing, in this case, have their own theoretical or applied stance toward edu-
cation or design, criteria for appraising quality (e.g., of an effective learning
experience or a design artifact), and their own ways of working (e.g., with design
and teamwork).
For example, in spite of a general consensus among the interdisciplinary
design team on the importance of engaging students in authentic and constructive
learning, we differed in our original understanding of the underlying design con-
structs, such as the definition of a game, the nature of mathematics, the dynamics
between learning and play (or fun), and interdisciplinary subject matters (e.g.,
mathematics in architecture). These different theoretical stances are exacerbated
by our differences in epistemic frames (e.g., being designer, programmer, educa-
tor, or evaluator), experience of gaming technologies as both a consumer and a
developer, understanding of the target users (e.g., middle school students and the
school setting), as well as our prioritized motives (e.g., research of game design
or learning assessment, mathematics education, or design education). All of these
differences are reflected in language, design ideas or concepts, and preferred rep-
resentation of concepts.
Given such a heterogeneity, the interdisciplinary collaborative design is never
a simply division of work or a sum of ideas or expertise. Instead, there is an
intentional and ongoing effort contributed to building a common design culture
among the interdisciplinary team, including the specification of common con-
cepts and language, the explanation and selection of the theoretical and design
conjectures underlying the aforementioned design constructs, and the negotiation
of an overarching design goal with a tangible design agenda that reflects the
dynamics of different epistemic perspectives on learning, game, assessment, and
the embedded subject matters. Last but not least, the team has explored and grad-
ually established an informal protocol or routine for collaborative working and
decision-making. This protocol deals with the balance between collaborative
brainstorming and efficient decision-making, the integration of self-governing
obligations and overlapped responsibilities, and the approach of information
sharing and archiving.
18 2 Chronicle of Designing a Game-Based Learning Platform

2.2.2 Designing Gameplay for Learning

Integrating the design perspective about an optimal integration of play and learning
in gaming (e.g., Breuer & Bente, 2010; Rodriguez, 2006) and the argument for
constituting knowledge in actions (Dewey, 1929; Hiebert et al., 1996), coding game
mechanics (or interactivity) evolves around the effort of understanding and organiz-
ing mathematics as inquiry or subject-problematizing practices. As the data illus-
trated, the exploration of core gameplay involves two dimensions: (a) specifying a
series of game actions that embody mathematical learning as an intellectual inquiry
or practice and (b) designing coherent game objectives, obstacles (constraints), and
rewarding rules that legitimize an active performance of these actions by the play-
ers, thus creating the desirable, learning-constructive interactivity between the play-
ers and the game system.
Along with the core game action and rule set are the interactivity interface (or the
player input/output controls) and the game settings (or environmental storytelling)
that should work coherently to foster and motivate emergent functional and cogni-
tive interactivity between the players and the game system. In a game-based learn-
ing platform, a user interface should not only promote an “intuitive” action but also
intermediary, discipline-specific theoretical thinking during the action. Designing
game settings (or a game world) for the learning purpose is more than environmen-
tal storytelling. Game objects and scenarios in E-Rebuild, for example, embody the
external representation of game-based, interactive mathematical problems for the
players to investigate and solve. Therefore, designing gameplay for learning is to
design a series of subject-problematizing actions that are structured by a set of rules
and obstacle-tackling objectives, an intermediary action interface, and a coherent
game world as both the milieu and drive of the actions.

2.2.3 I ntegrative Design of Game-Based Learning Task

and Assessment

A fundamental theoretical stance toward the nature of game-based learning (Eriksson

& Lindberg, 2016; Hiebert et al., 1996), as discussed in Chap. 1, is that the knowl-
edge to be acquired depends as much on the mathematical ideas embedded in the
task as on the way it is packaged. Therefore, the focus of the design is not only on
what subject content or information is integrated but how core game tasks will exem-
plify subject-problematizing actions to generate understanding. In other words, the
game-based learning task should be developed to enhance sign-­mediated, subject-
related theoretical thinking (Eriksson & Lindberg, 2016), by requiring that “core
principles” or “conceptual relations” constituting a specific knowing are discerned
and understood through learning actions in a game-based task.
2.2 Core Facets of Interdisciplinary Educational Game Design 19

Apart from designing inquiry- and understanding-generating tasks, another core

design challenge related to intentional learning in gaming is stealth assessment that
enables the real-time capture and analysis of performance-based competency devel-
opment data. With the recent methodology development in educational data mining
and learning analytics, stealth learning assessment based on the dynamic perfor-
mance of players could better capture process and performance-oriented evidence
on competency development while not being intrusive to distract players’ gameplay
or state of flow.
Using the evidence-centered design approach (Almond, Mislevy, Steinberg, Yan,
& Williamson, 2015; Mislevy & Haertel, 2006), the project of E-Rebuild will illus-
trate an integrative design approach that intertwines learning task development with
assessment construction. The selection and construction of game-based learning
tasks will elicit the core gaming/learning actions that operationalize the targeted
competencies and serve as both the source and evidence of competency develop-
ment. The game log is developed to capture these actions, with the logged data
parsed as the values of observables (variables) in a competency-based statistical
model (e.g., Bayesian networks). The sequencing and accumulation of a collection
of tasks (or game levels) then need to support both learning progression and the
accumulation of evidence for stealth assessment.

2.2.4 Designing Game-Based Learning Supports

A main challenge of designing games is that their consumption (by players) is rela-
tively unpredictable (Hunicke, LeBlanc, & Zubek, 2004). For an educational game,
the interaction between the player and the coded game system is even more dynamic
due to the anticipated learning interactivity. Differing in their prior competency
status, learning and gaming preferences, and dispositions toward the subject matter
and gameplay, players will differ in their reactions and behaviors in a game-based
learning system. Apart from learning game mechanics and user interface, the play-
ers may be involved in shortcut solutions rather than expected learning actions
(Butler, Andersen, Smith, Gulwani, & Popović, 2015), or they may fail to be purpo-
sive or mindful to generate theoretical thinking from game tasks and actions.
Therefore, support for game-based learning should be designed to not only
reduce the entry barriers (e.g., via training) but also enhance the opportunities of
intentional learning and knowledge acquisition during gaming for diverse learn-
ers. Game-based learning support can be tool- or material-mediated support
embedded in the game or external, human agent-mediated support arranged as
part of the game-based learning environment. The design efforts focus on explor-
ing what, when, and how learning supports are integrated and presented during
gameplay to foster task and action engagement and performance for knowledge
acquisition.
20 2 Chronicle of Designing a Game-Based Learning Platform

2.3  hronicle of Iterative Design, Testing, and Refinement

C
of E-Rebuild

E-Rebuild creation is a heavily iterative design process. There are more than 100
published versions of the game, with either major or minor revisions in between
these versions. Below we present a design chronicle recording major developments
and revisions associated with E-Rebuild. Iterative player testing was conducted dur-
ing and across the following design phases.
1. Starting to explore the action of building: In the original version of a build, drag-
ging objects around with the mouse in a top-down view was the only control
mechanism. The objective was to fill a certain percentage of the given space with
shipping containers. The containers were little more than rectangular prisms.
After a given time limit, the space would shake to simulate an earthquake. If the
same amount of containers remained, the level was passed. The objectives were
to quickly simulate an earthquake, track occupied space, and enable container
manipulation in a three-dimensional space.
2. Designing interface of site surveying and collection: In our next iteration, we
decided a fixed view offered limited flexibility and experimented with a first-­
person view that would offer the freedom to explore the world and give the
player a sense of presence. Moving items would happen by picking them up. The
player would approach an object, press a key, and carry the item with them. We
also introduced a “stamina” meter to stop the user from carrying items forever.
This was a great improvement from the previous version but not without
­compromise. Where the previous iteration required only mouse movement and a
single click, the new control scheme was inspired by the WASD keys control
scheme. This added a total of six keys and constant mouse control. For experi-
enced gamers, this may be second nature, but experienced gamers do not com-
prise the totality of our target demographic (Fig. 2.1).
3. Designing alternative modes of gameplay, controls for building, and initial game
world: After switching to keyboard and mouse movement, we doubled down on
the controls. Items could be held from a distance, and then they moved following
the player at the distance they were picked up. Using another pair of keys, the
player could move the object closer to or farther from the player. An earthquake
was simulated on the platform shown in the mini-map. To complete the level,
players should move a certain number of containers to the platform and maintain
a given height after the earthquake (Fig. 2.2).
A toolbar was added to control many options. A number of tools were available.
These included marking tools (such as freehand, line, rectangle, and ellipse) and
shaping tools, (such as cutting and scaling). We began working on the beginnings of
the building mode of E-Rebuild. The view was static with an unmoveable camera.
This became an issue when working in three dimensions (Fig. 2.3).
2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 21

Fig. 2.1 Earliest prototype with the basic item-movement control and a stamina meter

Fig. 2.2 Alternative perspective of play and basic controls

22 2 Chronicle of Designing a Game-Based Learning Platform

Fig. 2.3 A tool bar was added

At the same time, we started to explore game world design by experimenting

with larger environments that appeared to overwhelm players with the opportunity
to explore. We settled on a somewhat middle ground—a smaller island. This island
remained the basis for the island episode for a number of iterations, which was also
the beginning set of the adventure mode of E-Rebuild. In the adventure mode, the
player was to explore the area and collect items via a magic wand that made a green
“poof” as the item was collected and subsequently disappeared. Collected contain-
ers were marked on the bottom-left corner, while additional resources collected
were given as credits on the bottom-right corner (Figs. 2.4 and 2.5).
4. Designing the allocation action, creating the initial game level with the reward
mechanism: A family allocation objective was added. Blue capsules were used
as placeholders for people (families). The island people had three states: lost,
found, and assigned. Collecting and allocating families into the shelter increased
happiness credits, which gave the game three classes of credits: time, resource,
and happiness. These can be seen on the bottom left (Fig. 2.6).
5. Refining the design of the initial game level: To enable (family) allocation
inside a structure (e.g., a container), we refined the collection interface so that
players could switch their wand from collecting a whole structure to only col-
lect its parts (e.g., panels). This also allowed the player to make windows and
entryways in the structure. The building and adventure modes were integrated
in a single level: Players could toggle the mode by pressing 0 (zero). In the
initial level, the player would collect and place containers to rebuild a row
house and then assign the families into the house (Fig. 2.7).
2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 23

Figs. 2.4 and 2.5 Initial island for the adventure mode of gameplay

6. Developing additional in-game learning supports: An in-game scratchpad was

designed to allow users to make quick notes and conduct calculations. Two
dimensions (2D) of the target structure (e.g., the row house) and site landmarks
were added to assist the player in building the structure. An issue remained
unaddressed: it was difficult for the players to get out of the water when they
were distracted and explored the underwater land (Figs. 2.8 and 2.9).
24 2 Chronicle of Designing a Game-Based Learning Platform

Fig. 2.6 The first game level

Fig. 2.7 Adding an allocation action

2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 25

Fig. 2.8 Adding in-game learning supports

Fig. 2.9 Getting stuck under the water

7. Chunking the task in the game level: The task is a complex multistep problem
that has many steps that could stop the player from progressing. We tried chunk-
ing the task or game level into sub-goals to guide the player, with a progress bar
showing the distance to the goal. A tutorial level was also created to train the
player on game controls (Figs. 2.10, 2.11, and 2.12).
26 2 Chronicle of Designing a Game-Based Learning Platform

Figs. 2.10 and 2.11 Chunking the task

8. Designing another game level featuring the building action at a higher granular-
ity degree and a trading action: With an initial in-field testing study completed,
we began to design the desert level that features the building action at a higher
granularity and complexity level (e.g., via adobe blocks rather than containers)
and a different scene/setting (i.e., a desert). We also added a training action (in
addition to the collection action) with the building materials. Considering the
2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 27

Fig. 2.12 An initial tutorial level

game world exploration issue, we designed natural boundaries such as a

sandstorm, a fence, and tall bluffs to the desert set to keep the players on the
construction site without wandering away (Figs. 2.13 and 2.14).
9. Refining the interface of the trading and building actions: We refined the user
input interface of the trading action to necessitate cost planning and calculation
by the player. In the original movement control of the building action, the item
movement followed the cursor, but items would easily overlap with each other
and cause a popcorn effect. To fix this issue, we changed the movement control
so that the player moved the objects along the two displayed axes that can be
changed by moving the camera around the object. Items actually moved on a
variable-sized grid. The actual location would be designated with a green out-
line when the player decided to place the item (Figs. 2.15 and 2.16).
10. Expanding the “desert” and “island” episodes with levels (tasks) development:
Both episodes were expanded with tasks (levels) development. As more levels
became available, an episode tab became more useful. The episode tab of the
menu gave a wide overview of the tasks for this landscape. The task tab gave
specific instructions for completing each level. As the levels and game grew more
complex and longer in duration, a way to save was necessary. Originally the
saves were all local. This forced the user to use the same computer from session
to session. Levels with variant difficulty options were given to the user. This dif-
ficulty selection was associated with variables and their dependencies in a task,
the number of problem-solving steps involved, and the complexity of the embed-
ded mathematical competency. To assist with problem-solving, an interactive
28 2 Chronicle of Designing a Game-Based Learning Platform

Figs. 2.13 and 2.14 Adding a “desert” level featuring building at a higher granularity level

problem-solving help section was added. Levels were also even more discretely
chunked with more frequent performance feedback (Figs. 2.17, 2.18, and 2.19).
11. Adding a “school” episode and more levels: A school setting was added as an
urban environment in which complicated building tasks, such as classroom and
stadium stands building, were developed (Figs. 2.20 and 2.21).
12. Creating cohesive UI: A new UI was created to be cohesive across the game
episodes. The new version opted to use the level chunks as smaller individual
levels, each with a single task to complete with a number of constraints (or
obstacles). This version also saw the immersion of adventure and building
2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 29

Figs. 2.15 and 2.16 Trading action and the item movement control

modes—the player does not need to shift between the two modes through a
toggle but has a “fly” mode that fully integrates ego- and exocentric perspec-
tives in the virtual game world. The interface of the collection action was
updated to require action-related mathematical reasoning. The end-of-level,
textual performance review was replaced with a visual, star system. The player
profiles were associated with game user names to help organize their gaming
logs (Figs. 2.22, 2.23, 2.24, and 2.25).
30 2 Chronicle of Designing a Game-Based Learning Platform

Fig. 2.17 “Desert” being a multi-level game episode

Fig. 2.18 Interactive problem-solving help

2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 31

Fig. 2.19 End-of-level textual performance feedback

Fig. 2.20 Adding a school episode

32 2 Chronicle of Designing a Game-Based Learning Platform

Fig. 2.21 Being a 3-episode and 13-level game

Fig. 2.22 A new cohesive UI

2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 33

Fig. 2.23 Intermediary interface of the collection action

Fig. 2.24 A visual, 3-star system as the end-of-level performance feedback

34 2 Chronicle of Designing a Game-Based Learning Platform

Fig. 2.25 Player profile associated with a game user name for log organization

Fig. 2.26 A cohesive game world: Island

13. Creating a cohesive game world and a new “farm” game episode: The level set
design was updated to be consistent across game episodes. More levels/tasks
were added throughout the game, including a “farm” episode, to better support
the learning and assessment of mathematical competencies specified (e.g., ratio
and proportion, composition and decomposition of shapes, and transformation
of geometric shapes, angle, perimeter, area, volume) while presenting a gradual
learning progression. At this point, the game had 19 levels in four episodes
(Figs. 2.26, 2.27, 2.28, 2.29, and 2.30).
2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 35

Fig. 2.27 A cohesive game world: Desert

Fig. 2.28 A cohesive game world: School

14. Updating the game reward with a badge system and the UI to foster cognitive
interactivity: The three-star system did not provide enough information to the
payers on their game-based mathematics problem-solving performance. To
promote reflective inquiry, we used a multi-badge system to provide more
detailed feedback on different aspects of the game-based problem-solving
36 2 Chronicle of Designing a Game-Based Learning Platform

Fig. 2.29 A new episode: Farm

Fig. 2.30 Representation of a 4-episode, 19-level game

performance (e.g., fluency, accuracy, and thriftiness). We also refined the user
interface again to increase cognitive interactivity. For example, in the collection
action, related object information was presented as a tooltip in tables or charts.
The log-in was moved to the server to allow online play on Chromebooks
(Figs. 2.31, 2.32, and 2.33).
2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 37

Fig. 2.31 A badge system

Fig. 2.32 Support for cognitive interactivity in the collection action

38 2 Chronicle of Designing a Game-Based Learning Platform

Fig. 2.33 Online login

Fig. 2.34 New training levels

15. New training levels and the movement interface: To better reduce the entry
learning curve and the technical functionality of object maneuvering in a 3D
space, we added more training levels and designed a new item manipulation
system. The new system provides unobstructed views on the item’s three-­
dimensional movement to allow easier manipulation and a higher level of
responsiveness. At this point, the game had 43 game levels distributed across
four game episodes (Figs. 2.34, 2.35, and 2.36).
2.3 Chronicle of Iterative Design, Testing, and Refinement of E-Rebuild 39

Fig. 2.35 New movement interface

Fig. 2.36 Representation of a 4-episode, 43-level game

In comparison with the iterative revisions in the gameplay and level development,
the game logs have remained relatively stable after the core game actions were
specified. This allows for comparison between the older logged play data and the
new ones, as well as the refining and validation of the statistical model for stealth
assessment. Where the game has gotten more complex, the logs have become ­simpler
during the development process to allow for quicker parsing of data and smaller
file sizes.
40 2 Chronicle of Designing a Game-Based Learning Platform

Fig. 2.37 Heat map analysis example

Originally, nearly everything was logged. The player’s position was recorded
approximately once per second to allow for a heat map analysis of their paths, as
you can see in Fig. 2.37. This required an xml block similar to:
<Auto Position>
<runningTime>2.009376</runningTime>
<Position>(-29.0, 0.8, 42.1)</Position>
</Auto Position>

The log files were flooded with the above entries, which caused problems with
file size and parsing, as well as reading them in a glance. In the next revision, we
removed the auto position recordings. Still, nearly every action was recorded. The
log files (see Fig. 2.38) were more manageable but still required parsing to get to the
variables we used in the Bayesian network. We then further trimmed the log struc-
ture to only include what we would use in the statistical model. The logs (see
Figs. 2.39 and 2.40) then varied by level but could be interpreted in a glance.

2.4 Researcher Positionality and Reflection Notes

As Van Manen (1990) pointed out, our preunderstandings predisposed us to inter-

pret the nature of the phenomenon in a phenomenological inquiry. As part of the
effort to bracket these predispositions, we presented a brief introduction of our roles
in the design team as well as a reflective description of individual perspectives and
perceived experience of the interdisciplinary design process for game-based
learning platform development. Our motives, domain-specific backgrounds, and
habits of and dispositions toward design certainly impact our reflective analysis of
the design process and the findings that emerged as they are part of the lens brought
to the project. However, it is important to understand that these preunderstandings
did not remain unbracketed during design and research. Additionally, the authors
applied peer debriefing and member checks in the inspection and rectification of the
data and interpretations.
2.4 Researcher Positionality and Reflection Notes 41

Fig. 2.38 Refined game

log example 1

Fig. 2.39 Refined game

log example 2

Fig. 2.40 Refined game

log example 3
42 2 Chronicle of Designing a Game-Based Learning Platform

2.4.1 Fengfeng Ke

My belief is that mathematics is a way of thinking, and learning is to develop powerful ways
of thinking.

Role and responsibility in the design team I am the principal designer of the
E-Rebuild project and responsible for leading all facets of the design and research
work, including the design, in-field testing, and refinement of E-Rebuild, the man-
agement of the project team’s communication and collaboration, as well as the data
collection and analysis. As an associate professor of educational psychology and
learning system designer at Florida State University, my design and research inter-
ests lie in digital game-based learning and innovative learning environments design.
Besides E-Rebuild, I have participated in multiple other research projects on
game- and simulation-based interactive learning systems.

Personal goal or interest for the E-Rebuild design As a former gamer, I prefer
games that involve puzzle solving and construction. As a teacher and a lifelong
learner, I believe learning is ultimately self-regulated, and teaching is a process of
instilling thinking or lifelong learning skills. These dispositions are naturally infused
into my vision for the E-Rebuild design. As a learning system designer, I think the
vital role of technologies for learning is not offering an alternative delivery media or
a motivation tool but facilitating a new way of understanding and organizing
­learning. My goal for designing E-Rebuild is hence not creating a single digital
learning tool but using it as a test bed to examine and refine our conjectures of math-
ematics and learning, to experiment with previous hypotheses governing the inte-
gration between learning and play, and to discern new tools and methods to designing
an active, meaningful learning experience.

Perceptions of interdisciplinary design of E-Rebuild Designing a learning sys-

tem is naturally an interdisciplinary process. During previously designed projects, I
have worked with content experts, educators, developers, and evaluators from differ-
ent fields. These works, however, were more multidisciplinary than interdisciplinary,
in that the design and development efforts of team members are frequently segmented
by time (e.g., the evaluation won’t start until the functional prototype is ready) or by
a clear-cut responsibility division. In spite of the occasional brainstorming or peer
debriefing opportunities, the overlapping or truly collaborative design efforts in those
projects are usually limited. E-Rebuild project, differently, is an extensive and longi-
tudinal project that requires and allows for a full understanding of the discipline-
specific subcultures, a gradual development of project-specific design culture, and
iterative testing of interdisciplinary perspectives and methods in the design artifact.
In spite of the numerous discussions of generic principles of interdisciplinary col-
laboration or professional development, I have found the interdisciplinary design of
E-Rebuild, by itself, is a learning by doing process. In our earlier project meetings, I
felt less comfortable or occasionally lost with the team communications; I wondered
whether and how those lengthy interdisciplinary communications would be a worthy
2.4 Researcher Positionality and Reflection Notes 43

investment to the quality of the final artifact, given a tight project schedule. All these
uncertainties, however, got addressed once the mutual understanding of the language,
stances, and concepts were achieved and when we gradually found our protocol to
channeling differing interests and expertise into the tangible working plan.

Perspectives of interdisciplinary design and research for educational game

development In spite of a plethora of research on game studies, designing an
­educational game, like art, is an utmost ill-structured problem. Intuitive design,
modifying a worked example, trial and error (or design experimentation), and work-
ing backwards (e.g., an analytical or evaluative review of exemplary games to
inform the current game design) are frequent strategies adopted by educational
game designers. Designers, educators, or evaluators may have an overarching goal
or vision about the desirable game-based learning experience, but it is difficult to
operationalize such an experience as tangible objectives. Even when objectives are
set, the problem space for achieving these game design objectives is murky. Through
conducting and reporting a reflective and phenomenological analysis of the core
facets and salient patterns emerged from a longitudinal design project, I hope our
work will help to illustrate and better define the problem space of interdisciplinary
education game design.

2.4.2 Valerie Shute

Measure twice and cut once!

Role and responsibility in the design team My role in the E-Rebuild project is
primarily related to all assessment-related aspects of the project. For a brief back-
ground, I am the Mack and Effie Campbell Tyner Endowed Professor of Education
in the Department of Educational Psychology and Learning Systems at Florida
State University. Before coming to FSU in 2007, I was a principal research scientist
at Educational Testing Service where I was involved with basic and applied research
projects related to assessment, cognitive diagnosis, and learning from advanced
instructional systems. My general research interests hover around the design, devel-
opment, and evaluation of advanced systems to support learning—particularly
related to the twenty-first-century competencies and STEM content. My current
research involves using games with stealth assessment to support learning—of cog-
nitive and noncognitive attributes. As the originator of the term and technologies
surrounding “stealth assessment,” I’m pleased to see it become more broadly
accepted and used in new research projects (including this current one).

Personal goal or interest for the E-Rebuild design When we began thinking and
talking about this project, prior to funding, our goals were to design a platform that
integrated architectural design and building with evolving and deepening mathemat-
ical understanding. Moreover, this was to be designed and developed in the context
44 2 Chronicle of Designing a Game-Based Learning Platform

of a game with which kids would want to engage and learn. For the past four years,
and across a lot of iterative pilot testing, our original goals are successfully being
met. With regard to the assessment part of the story, we have focused on refinements
to both the local and server-side game logs as a way to capture relevant gameplay
data to inform the math competency model (developed at the onset of the study).
Concurrently, we continue to explore the viability of other data mining approaches
to supplement the stealth assessment within E-Rebuild. Together, the more t­ op-­down
(theoretically driven) stealth assessment approach coupled with the more bottom-up
(empirically driven) data mining approach should provide more accurate estimates
of competency states than either alone. And having more accurate estimates—avail-
able at any time and at various grain sizes—will provide the basis for more targeted
and effective math learning supports in the game, including the automatic selection
of the next best task to provide to the learner based on his/her current needs.

Perceptions of interdisciplinary design of E-Rebuild Our weekly meetings of

the entire interdisciplinary team have been, overall, enjoyable and interesting.
Because everyone has their own areas of expertise, it is often a very educational
experience as well. Furthermore, when a topic is discussed, multiple perspectives
are available to look at an issue from all sides. Communication is respectful, even
when there are differing views on solving a particular problem.
We also have been able to integrate “lessons learned” from other currently funded
and directly related research. For example, in one of our current National Science
Foundation (NSF) cyber-learning grants using stealth assessment within a game
called Physics Playground, we needed to come up with some difficulty indices in
order to rate the various levels in the game relative to their difficulty. Originally, we
had derived indices related to particular aspects of the level (e.g., if there were a lot
of obstacles in the problem that made it harder, and so on). However, we soon real-
ized that additionally considering the nature of the physics content related to the
game level was necessary to make decisions about “level difficulty” overall. The
information gleaned from the cyber-learning project has thus been used in the current
E-Rebuild project. That is, we have developed an approach in E-Rebuild to catego-
rize all of the various game tasks based on a difficulty index that consists of both (1)
the complexity of the problem representation and solution processes in a given level
and (2) the mathematical competencies required in the solution, as well as prerequi-
site relationships between the competencies and sub-competencies. This in turn has
engendered the development of an adaptive game-based algorithm, which can select
appropriate learning tasks based on the real-time estimate of a learner’s competency
profile to support the development of focal math knowledge and skills.
Perspectives of interdisciplinary design and research for educational game devel-
opment In summary, working within an interdisciplinary design team has been a
great experience. As mentioned, because of the varying areas of expertise, I never
leave a meeting without learning something new (and often multiple new things).
The discussions of issues among the team members have resulted in an expansion
(and occasionally a refinement) of different areas given fresh eyes on a problem. What
2.4 Researcher Positionality and Reflection Notes 45

has been most enlightening to me is the direct relationship between architectural

design processes and math skills, at a general level. So in a sense, success in the game
provides a two-for-one—gaining architectural designing and building skills and asso-
ciated math knowledge and skills. Another eye-opener is what takes place when we
take our game into the schools. Invariably, the teachers and importantly the students
have a lot to share with regard to the functionality of the game. Both teachers and
students are very clear about what they want and need to succeed—as designers of
new levels (teachers using the platform) or as players of the game (students).

2.4.3 Kathleen M. Clark

A view from a non-gaming, non-assessment, mathematics teacher educator.

Role and responsibility in the design team I have just completed my 31st year in
mathematics education. I began teaching mathematics in 1987 in an inner-city
Miami high school. I taught mathematics in a variety of public schools in Florida
and Mississippi until 2001, when I was awarded an Einstein Distinguished Educator
Fellowship. During the 1-year fellowship, I worked in educational policy for the US
Senate, and though I did not return to my secondary teaching career afterward, I
decided to pursue my doctorate in Mathematics Education at the University of
Maryland. I have been at Florida State University (FSU) since August 2006. Early
in my FSU career, my research focused primarily on pre-service and in-service
mathematics teachers, with an emphasis on the role of history of mathematics in
mathematics teaching and learning. I currently spend the majority of my time con-
ducting research on the use of primary historical sources in the teaching of under-
graduate mathematics.
My role on the E-Rebuild team is to serve as the “mathematics education expert.”
In this role, I focus on providing support for the mathematical content both within
the game itself and within the various forms of assessments used in the project. The
content that informs the game and the assessments were—in the first development
of E-Rebuild—based upon the Common Core State Standards for Mathematics
(CCSS-M). However, as the game has developed, we included attention to the
Florida Mathematics Standards (FMS), and we do so because the game is currently
and primarily tested with Florida mathematics teachers and students, and we must
be cognizant of the minute differences between CCSS-M and FMS.
Personal goal or interest for the E-Rebuild design I was initially drawn to the
E-Rebuild project—particularly from the students’ perspective—because I often observe
limited appropriate use of multimedia tools (e.g., appropriate web tools, electronic or
Internet-based games, computer software, etc.) to promote middle grade students’ learn-
ing of mathematics. I was excited by the potential of E-Rebuild as a means by which
students would learn or, possibly practice, not only core mathematical concepts
46 2 Chronicle of Designing a Game-Based Learning Platform

(e.g., ratio and proportion skills, properties of geometric shapes and determining their
area, perimeter) but solve problems in a contextually rich and meaningfully engaging
game environment.
My experience with the E-Rebuild game itself has been one of healthy skepti-
cism. I have never been a gamer, so this may contribute to my skepticism, at least to
a small degree. That said, I find myself thinking about the numerous variables that
have the potential to impact the student learning experience as they play E-Rebuild.
For example, are the directions clear for the student in each game level? Is the game
visually and structurally appealing to young learners? What are the connections that
students make with regard to the mathematical concepts that they meet in the game?
What does the context and gameplay environment contribute to students’ engage-
ment with mathematical ideas? In my observations of student gameplay, I have
noted a high level of student frustration from the perspective of game design fea-
tures and not from the perspective of mathematical content. Thus, this prompts me
to think about how little expertise I have with regard to game-based learning—or
educational games in general—because I wonder about the amount of unproductive
struggle students experience with game features and how this may detract from the
role of productive struggle when solving the mathematical problems imbedded in
the game. And, due to my lack of expertise in “stealth assessment,” I do not under-
stand how the components of these more affective experiences with E-Rebuild may
factor in to whether students are learning as a result of their E-Rebuild play.
Perceptions of interdisciplinary design of E-Rebuild I have participated in a
number of interdisciplinary projects. I continue to be energized by the way in which
experts and practitioners from different disciplines and fields come together to cre-
ate, design, and test instructional tools (e.g., curriculum, materials, interventions) to
promote student learning. There are several meeting structures used within the
E-Rebuild project: Some are focused specifically on design of tasks, levels, and
episodes that comprise the game itself. Other meetings are focused on the inherent
mathematical concepts, skills, and problem-solving situations at the core of the
game, as well as the components of the architectural design elements that contribute
to the underlying mathematics. Regardless of the purpose or goal of any given
design meeting, I have found the discussions to be interesting, and they challenge
me to try to find ways in which I can learn as much as possible while also drawing
on my mathematics education background in order to contribute in optimal ways.
I believe the most challenging aspect of the design meetings has been navigating
the various communication styles within the group. We often assume we have
“taken-as-shared” knowledge, and since we do not, it often takes us several itera-
tions to operate with the same (or at least similar enough) knowledge. I have always
taken my own notes during the numerous meetings in which I have participated over
the years (though I often felt I was the only one—perhaps because I was the only
one who needed to!), and these have been helpful in being able to feel that I can
keep up with the various aspects of the E-Rebuild work. I think that consistent
use of agendas and assigned tasks for regular meetings would have been beneficial.
2.4 Researcher Positionality and Reflection Notes 47

For example, by scheduling and planning specific tasks and agenda items, I believe
a more equitably distributed range of ideas and foci would be addressed in the
meetings.
Perspectives of interdisciplinary design and research for educational game devel-
opment I am intrigued by the notion of what we can learn from the d­ evelopment and
the use of educational games in the learning of mathematics. I think there is much more
that we can know (and, consequently, contribute to the field) with respect to student
learning and E-Rebuild gameplay. However, as a qualitative researcher, I believe that
we must continue to conduct student and teacher interviews to better understand the
student experience with E-Rebuild. Additionally, the various teacher experiences are an
important part of this inquiry. Students have played E-Rebuild as part of the regular
mathematics classroom activity, in non-­mathematics classrooms, and within after
school and summer camp contexts. I believe it is wise to investigate the ways in which
students have played E-Rebuild in these different formal and informal contexts and
work to understand the assessment results with these different implementation lenses
in mind. I think there is great potential for the results of these investigations to assist us
in focusing on further game design and development to maximize the potential for such
a tool to improve middle grade students’ nonroutine problem-solving abilities.

2.4.4 Gordon Erlebacher

In my view, a perfect educational game should improve player skill (in an idealized situa-
tion) without the player being aware of the purpose of the game. Not unlike good advertis-
ing which influences the actions of a person without that person being aware of the ad.

Role and responsibility in the design team The E-Rebuild project team was
formed in 2011, and I participated from the onset. My role in this project is to ensure
that the team applies sound game design principles during the development of the
game and help manage implementation of its game mechanics. Prior to my arrival
at FSU, I was a researcher at NASA Langley and at the Institute for Computer
Applications in Science and Engineering. During that period, I conducted theoreti-
cal and computational research in the field of turbulence. In 2009, I created a first
course on Game Design at FSU, which attracts undergraduate and graduate students
over a wide range of science and nonscience departments. I have often felt that
properly designed games could help students achieve more of their potential and
mathematics is a natural subject matter given my background. I am currently Chair
of the Department of Scientific Computing at Florida State University, an interdis-
ciplinary department whose mission is to develop computer algorithms with appli-
cations in the applied sciences. Since 1996, I have been involved in computational
science, developing algorithms in the fields of visualization, GPU programming,
flow simulation, and more recently in computational neuroscience and deep learn-
ing. Over the past 30 years, I have been exposed to many frameworks, software,
computer languages, and technologies, which has helped me develop the skills to
interact with researchers in different fields.
48 2 Chronicle of Designing a Game-Based Learning Platform

Personal goal or interest for the E-Rebuild design In the early stages of the project,
I envisioned that we would eventually develop an exciting and engaging game that
middle-school students would want to play and as a result undergo a gradual improve-
ment of their mathematical skills. The game would adapt to the player, choosing sce-
narios and tasks commensurate with the player’s ability. This would be accomplished
through the collection of data (player movement, mouse clicks, imagery) that would
then feed into an assessment system that would measure the mathematical proficiency
of the player. Together with player history, new scenarios would be generated, adapted
to each player. Through a team effort, the game evolved somewhat differently. School
teachers would be responsible to creating the scenarios and tasks appropriate for their
students, and the players would be aware that the game was meant to improve their
mathematical skills. We have been refining the competency model and field-testing the
game to validate the model and ensure that the students are indeed improving through
gameplay. At this time, my goals are more ambitious. The difficulty of finding enough
participants to test the game and model leads to the idea of creating a synthetic player
based on machine learning principles with a “brain” that can adapt to gameplay.
Although this brain would be a poor facsimile of a human brain, my hope is that it
would learn in different ways and at different rates depending on the sequence of tasks
and scenes presented, and this would lead to insights into further game improvements.
At this stage, this remains just a vision for future research.

Perceptions of interdisciplinary design of E-Rebuild Early in the project, after

the project was funded, the whole project team met on a weekly basis to discuss the
nuts and bolts of the game, such as game mechanics and backdrops, along with
some initial tasks the player would have to complete. These meetings were fun but
challenging in the early days because of the different backgrounds and past experi-
ences of the participants, which ranged from architecture to educational games to
instructional systems to mathematics education to assessment and science computer
simulations. Our different past experiences led to rather different ideas on what
would go on in the mind of an eighth grader, which in turn would impact decisions
related to game development. Discipline vocabulary, which differed between disci-
plines, also impeded progress on occasion, and it behooved us to allow for differ-
ence of opinion which was not necessarily substantive but could be the result of a
different understanding of word usage. A recent case in point was confusion in the
use of the term “elevation view,” which for an architect refers to a side view but to
a game designer might mean view from above or a third-person view. Individual
members also had a unique vocabulary, which sometimes/often led to entire discus-
sions when different people used the same word in discussions, but with different
meanings. Once, the majority of a meeting was spent debating the meaning of the
words task, scene, objective, and how to use them to describe different components
of the game. Another early discussion revolved around the task of construction;
should objects be controlled via the keyboard or using the mouse? How would the
player respond to these different choices? How much help should the player be
given and in what form should it be made available. Should the help be in-game or
out-of-game? Should the game flow be interrupted or should we maintain player
References 49

immersion into the game. There was disagreement on all of these aspects, which
eventually were resolved. The cases that remained open were decided via testing by
the students. Overall, I enjoyed the meetings, gaining an appreciation of other fields
of study and how they related to ideas presented.

Perspectives of interdisciplinary design and research for educational game

development Interdisciplinary research and interdisciplinary design are concepts
that have been around and promoted for many years. From my personal experience,
the E-Rebuild project is probably the most satisfying interdisciplinary effort in
which I have participated. Very often, participants meet once a year or biannually
and share what they have accomplished since the previous meeting. On the other
hand, our whole-team meetings were weekly initially and, more recently, biweekly.
All ideas are shared and discussed, and progress made since the previous meeting is
presented. I have learned a lot from our interactions, not only in terms of game
design but also on the personal level. I have learned to appreciate some of the tools
used by architects, educators, and programmers to help others learn and enjoyed the
process through which different goals eventually merged into a game that pleased
the entire team. Eventually, it became clear that stated goals during our meetings
often reflected the background of the person. For example, our architect pushed for
a game that was much more architecturally oriented in terms of realism, while oth-
ers on the team were more interested in gameplay at the expense of realism.
Good teamwork produces a game that compromises between these different visions
without sacrificing the vision of the team leader. For large projects, a truly
interdisciplinary approach is often indispensable, particularly when a game touches
upon many different disciplines as it does in our case (construction, education,
mathematics, and simulation).

References

Almond, R. G., Mislevy, R. J., Steinberg, L., Yan, D., & Williamson, D. (2015). Bayesian networks
in educational assessment. New York: Springer.
Breuer, J., & Bente, G. (2010). Why so serious? On the relation of serious games and learning.
Journal for Computer Game Culture, 4(1), 7–24.
Butler, E., Andersen, E., Smith, A. M., Gulwani, S., & Popović, Z. (2015). Automatic game pro-
gression design through analysis of solution features. In April (Ed.), Proceedings of the 33rd
annual ACM conference on human factors in computing systems (pp. 2407–2416). New York:
ACM.
Dewey, J. (1929). The quest of certainty: A study of the relation of knowledge and action.
New York: G.P. Putnam’s Sons.
Eriksson, I., & Lindberg, V. (2016). Enriching ‘learning activity’ with ‘epistemic practices’–
enhancing students’ epistemic agency and authority. Nordic Journal of Studies in Educational
Policy, 2016(1) Retrieved from: https://www.tandfonline.com/doi/full/10.3402/nstep.v2.32432.
Giorgi, A. (2009). The descriptive phenomenological method in psychology: A modified Husserlian
approach. Pittsburgh, PA: Duquesne University Press.
50 2 Chronicle of Designing a Game-Based Learning Platform

Groenewald, T. (2004). A phenomenological research design illustrated. International Journal of

Qualitative Methods, 3(1), 42–55.
Hiebert, J., Carpenter, T. P., Fennema, E., Fuson, K., Human, P., Murray, H., et al. (1996).
Problem solving as a basis for reform in curriculum and instruction: The case of mathematics.
Educational Researcher, 25(4), 12–21.
Hunicke, R., LeBlanc, M., & Zubek, R. (2004, July). MDA: A formal approach to game design
and game research. In Proceedings of the AAAI Workshop on Challenges in Game AI (vol. 4(1),
p. 1722–1726) Palo Alto, CA: AAAI Press.
Mislevy, R. J., & Haertel, G. D. (2006). Implications of evidence-centered design for educational
testing. Educational Measurement: Issues and Practice, 25(4), 6–20.
Rodriguez, H. (2006). The playful and the serious: An approximation to Huizinga’s Homo Ludens.
Game Studies, 6(1). Retrieved from: http://gamestudies.org/0601/articles/rodriges.
Van Manen, M. (1990). Beyond assumptions: Shifting the limits of action research. Theory Into
Practice, 29(3), 152–157.
Weingart, P. (2000). Interdisciplinarity: The paradoxical discourse. In P. Weingart & N. Stehr
(Eds.), Practising interdisciplinarity (pp. 25–41). Toronto, ON: University of Toronto Press.
Chapter 3
Interdisciplinary Design Activities
and Patterns

Abstract In this chapter, we provide a reflective and analytical description of the

interdisciplinary design activities of E-Rebuild, identify driving design questions
and salient design patterns that capture and frame the essence of E-Rebuild develop-
ment, and discuss distilled meta-generalizations that help to decompose the inter-
disciplinary learning game design processes to inform future work related to design,
research, and deployment. The description of salient patterns of interdisciplinary
game design in this chapter provides a contextualized design narrative/account of
core design processes along with an analytical synthesis of core design pattern ele-
ments—a design problem statement with its context and specifics, the solution or
technique to solving the stated problem, and the pattern of transferring or scaling
this design solution or design move.

3.1 Introduction

There has been a growing recognition that the process of designing and developing a
computer-supported, highly interactive learning system in general and games or simu-
lations for learning in particular, is by its very nature an interdisciplinary and challeng-
ing task. Prior works of design science and serious game design highlight the importance
for learning and game designers to record, identify, and distribute not only detailed,
contextualized design narratives but also higher-level, distilled design knowledge such
as design patterns (Pratt, Winters, Cerulli, & Leemkuil, 2009; Winters & Mor, 2008).
A design pattern is a high-level specification for a solution methodology to a
design problem, by specifying the particulars of the problem and highlighting recur-
ring solutions that are field tested in real-world application development. In this
chapter, we provide a reflective and analytical description of the interdisciplinary
design activities of E-Rebuild, identify driving design questions and salient design
patterns that capture and frame the essence of E-Rebuild development, and discuss
distilled meta-generalizations that help to decompose the interdisciplinary learning
game design processes to inform future work related to design, research, and
deployment.

© Springer Nature Switzerland AG 2019 51

F. Ke et al., Interdisciplinary Design of Game-based Learning Platforms,
Advances in Game-Based Learning, https://doi.org/10.1007/978-3-030-04339-1_3
52 3 Interdisciplinary Design Activities and Patterns

The design meeting notes, sketches, aids, paper and electronic prototypes, as
well as conversation emails during the E-Rebuild project have all been archived.
Iterative and reflective interviews were also conducted among design team mem-
bers. Adopting the phenomenological qualitative research approach (Moustakas,
1994), we have conducted a thematic analysis to explore salient themes or patterns
that capture the nature and features of design choices and actions by the interdisci-
plinary design team. While coding salient themes emerging from the data, we have
also referred to the prior research in the field of design science when seeking the
boundary and meaning of a theme or pattern. The description of salient patterns of
interdisciplinary game design in this chapter provides a contextualized design nar-
rative/account of core design processes along with an analytical synthesis of core
design pattern elements—a design problem statement with its context and specif-
ics, the solution or technique (with its key structure or mechanism) to solving the
stated problem, and the pattern of transferring or scaling this design solution or
design move.
During the design of E-Rebuild, a variety of driving questions or problems initi-
ate the negotiation of the design plans and the coordination of the design knowledge
among interdisciplinary team members. It then takes interdisciplinary cases/exam-
ples exploration, infield testing, and iterative refinement for the team to settle on a
design solution. Moreover, design leadership during the decision-making takes on
an important role in the project management, especially within a constrained
timeline.

3.2 Core Design Patterns and Activities

3.2.1  efining the Design Goal: A Trinity of Architectural

D
Simulation, Mathematical Learning, and Gaming

Problem statement During the initial design meetings of E-Rebuild, interdisci-

plinary team members debated and went back and forth to better understand
E-Rebuild as “an architectural simulation game for math learning.” Some notable
selections from the debate include statements such as: “Should it be mainly an
architectural building simulation that provides an architectural design education
while embedding mathematical conceptual development at the same time?
Should it focus on the representation of mathematical concepts and problems,
with architectural scenarios as a meaningful context for mathematical represen-
tation? Or should it prioritize the gameplay and fun elements situated in the prac-
tice of architectural and mathematical puzzle solving?” These debating topics, in
general, revolve around an innate design problem for learning game creation,
What is the creative stimulus or the ultimate inspiration for the creation of
E-Rebuild as a trinity of content learning, simulation, and gaming?
3.2 Core Design Patterns and Activities 53

Exploration of the design solution The aforementioned design problem roots in

the typical challenge of designing a learning game—to integrate learning into core
game elements while not violating or corrupting what is enjoyable about games
(Garris, Ahlers, & Driskell, 2002; Ke, 2016). Prior research argued that the prospect
of using a game to motivate learning lies in the assumption that this game will
involve learners in what is fundamentally engaging about the subject, via either a
dynamic representation or the simulation of the most “fun” (dynamic) component
of the academic domain (Ke, 2016). Yet it remains debatable as to whether and how
pursuing “fun” or gameplay as the preliminary design inspiration will drive the uni-
fied representation and simulation of content and content-relevant epistemic prac-
tice. For example, the following design notes documented a design goal motivated
primarily by the gameplay exploration:
We will start by building a “Block” level in which the player will select, drag, and drop
the “block” object to build a specific architecture form. The process is somewhat akin to a
3D version of block stacking or jigsaw game – the player must “fill up” an empty, 3D
structure using the building blocks (i.e., revive a building). Task assessment proceeds by
testing whether and how much the destined 3D architectural space has been occupied,
whether this constructed 3D structure will collapse due to a shaking landscape, etc.
The embedded math concept can be geometric shapes, area/perimeter, and
transformational geometry (rotation, reflection, and dilation), which can be integrated
with the organizational principles of architectural design (such as symmetry or balance,
hierarchy, and rhythm).

The notes portray a design inspiration that seeks a gameplay that can unify or align
with mathematical content and math-related architectural practices. Such a gaming-
driven design exploration, however, has been found by the content experts and
instructional designers in the team as incapable of capturing the depth and multifac-
eted nature of content learning. It prioritizes certain mathematical competencies or
levels of learning (e.g., composing and decomposing geometric forms for concep-
tual understanding) while ignoring the others (e.g., problem-solving that involves
area, volume, or unit rate computation).
As an alternative and refinement, the team shifted toward domain content and
architectural design as the dynamic starting points. In particular, we started by
developing graphical competency models to outline the structure or typology of
targeted mathematical competencies. Each node in the math competency model rep-
resents a performance objective and highlights learning actions. Exemplary mathe-
matical context problems for each node were also gathered. The practice of
composing child nodes or decomposing a parent node along with their supporting
mathematical problems has helped us conceptualize varied tasks or quests with cor-
responding actions within the game (Fig. 3.1).
Meanwhile, by referring to a pre-developed Architecture and Children Pilot
Curriculum by Dr. Anne Taylor (our architecture and design education expert in
the team), we drafted an Architecture Terms and Principles document that gathers
basic architectural concepts and skills that may relate to mathematical thinking.
This 24-page working document, illustrated in Figs. 3.2 and 3.3, provides
definitions and descriptions with rich visuals and real-life examples. Similar to
54 3 Interdisciplinary Design Activities and Patterns

Fig. 3.1 An example of graphical, competency models

the mathematical competency model, this manual also highlights the componen-
tial performances and actions when synthesizing the skills architects and design-
ers depend on. Given these features, this manual has served as a handy manual to
frame not only the task development but the game world and object design in
E-Rebuild.
The simultaneous exploration of the targeted subject domain and content-­relevant
epistemic practices has helped the project team better delineate the design problem
space—an externalized representation of the multifaceted design goal of E-Rebuild.
Given this clearly defined, action- or performance-themed design problem space,
the team then began anew and were productive in searching, selecting, and assem-
bling the necessary gameplay components (e.g., core actions, rules, and backdrop
missions) to further frame the design goal and the problem space from a “gaming”
perspective.
Design pattern summary Defining the design goal of the simulation-based learn-
ing game is the delineation of a multifaceted, interdisciplinary design problem
space. The framing of the problem space is initiated by a synchronized modeling of
domain-specific competencies and the cataloging of simulated epistemic practices.
Modeling and cataloging should highlight the fundamental performance expecta-
tions and supporting actions for content representation and practice simulation. The
competency-driven, action-themed design problem space will then drive the explo-
ration and selection of gameplay that captures and unifies content learning and
simulated epistemic practices.

3.2.2  etting Core Gameplay: Meaningful Mechanics that

S
Connect to Intellectual Learning and Artifact Design

Problem statement The subsequent design question that receives considerable

discussion takes the form, “What will a player do in the game, and why is that
engaging and intelligent?” In other terms, what and how will the core game
3.2 Core Design Patterns and Activities 55

TABLE OF CONTENTS

1. THE ORGANIZING PRINCIPLES OF DESIGN

2. GEOMETRIC FORMS IN ARCHITECTURE

3. ARCHITECTURAL VOCABULARY – STRUCTURES

4. SPATIAL RELATIONSHIPS

5. DESIGN AND VISUAL VOCABULARY

(a repetitive process for design)

6. LANDSCAPE DESIGN

7. SKILLS ARCHITECTS AND DESIGNERS NEED AND USE

2. GEOMETRIC FORMS IN ARCHITECTURE

Point (Zero D) - a precise location or place on a plane; usually represented by a dot.

Line (1D) - a geometrical object that is straight, infinitely long and infinitely thin.

Plane (2D) - a flat surface that is infinitely large and with zero thickness.

Volume (3D) - volume is the measure of the amount of space inside of a solid figure.

Circle - a2 -dimensional shape made by drawing a curve that is always the same distance from a
center.

Sphere - a3 -dimensional surface, all points of which are equidistant from a fixed point.

Figs. 3.2 and 3.3 Table of content and exemplary pages of the “Architectural Terms and
Principles” document
56 3 Interdisciplinary Design Activities and Patterns

Figs. 3.2 and 3.3 (continued)

mechanics of E-Rebuild connect intellectual content and architectural design to

motivate and guide math conceptualization and problem-solving? Prior research
suggests that learning occurs only when play experience connects to intellectual
content; yet the process that leads to the attainment of an integral and continuing
relationship between gameplay and the content to be learned remains murky
(Habgood & Ainsworth, 2011). Moreover, there had been some divergence within
the E-Rebuild team as to what share of (architectural) design knowledge relates to
the mathematical one to be embodied and activated by the gameplay actions and
rules.

Exploration of the design solution We have reviewed exemplary design cases in

relevant disciplines and explored discipline-specific inquiries or tasks when selecting
3.2 Core Design Patterns and Activities 57

the core game mechanics of E-Rebuild. The discipline-specific design cases and
ideas were typically presented by a discipline expert, reviewed and commented on
by the team at design meetings. Shared and prominent ideas emerging from these
reviews and discussions were then aggregated to be paper-prototyped as the initial
version of gameplay.
Treating architectural design as the epistemic practice of mathematical concep-
tualization and problem-solving, the team started by listing the key elements and
actions of architectural design and those of mathematical reasoning. The team also
searched for architectural design and construction cases that inspire the selection of
salient actions for intellectual gameplay, as illustrated by the following design notes
documented after several rounds of architectural design case discussion.
Can we simulate the “building block” and landscape-management design ideas
from the New Zealand earthquake rebuild project (https://www.psfk.com/2012/12/
shipping-container-mall.html)?
• Building blocks (tools: math tools, design tools, building tools, shape/texture).
• Building actions: Navigating/exploring, drop/drag/customize, watch/predict/
reflect, drawing and modeling (outside the game or using mixed reality)?
• Building tasks (reflecting a bottom-up approach, integrating Anne’s ideas and
content in the Architecture and Children Pilot Curriculum):
Architectural floor plan design—Bubble diagram and a related task: Designing
a fast food vegetarian restaurant (spaces only, without equipment) with pre-­
design exercise (such as decorating a shoe).
Structure in architecture (arch, triangle, asymmetrical vs balanced) and a related
task: Building a space frame structure for the roof of the picnic shelter with
pre-design exercise (e.g., toothpick puzzles).
Form in architecture or architecture in nature (may involve measurement, data
interpretation, scale, and proportion) and related tasks: Using organic struc-
tures in nature in a “fantasy” architectural structure for the “moon,” with
pre-design puzzle (scavenger hunt for basic geometric shapes and forms in
nature) or designing a “people’s” bridge across the canyon to build a jogging
and bike path (designing and modeling natural, beam, suspension, or arch
bridges).
Visual design, creative design.
Landscape, city planning.
• Building tasks (reflecting a top-down approach).
Building blocks and site planning—Using given building objects (e.g., shipping
containers) to complete the construction of a site (e.g., a village). This exer-
cise involves the organizational principles of design (e.g., Symmetry, Balance,
and Proportion) and a basic understanding of the geometric concepts of
shape and transformation (e.g., reflection, rotation, and dilation).
Reference: Architectural Ordering Principles (document is uploaded to the
project site)
58 3 Interdisciplinary Design Activities and Patterns

Architectural
system Math system Gameplay element
Primary elements: Geometry: Reviving an architecture demolished during
Point, line, plane, 1D to 2D to 3D earthquake—Jigsaw game: Recomposing an
volume shapes architectural structure via the scattered elements/pieces
Area, perimeter,
and volume
Properties of forms: Geometry: Scavenger hunt game: Collecting building items in
Shape, size, color, Shapes varied forms and properties
texture Area, perimeter,
Position, and volume
orientation, visual Proportion
inertia

As the above design note portrays, a real-life post-quake rebuild story, including
architectural building elements, actions, and tasks, have driven and inspired the
selection of applicable types of gameplay. In addition, a purposeful exploration of
the association between architectural and mathematical elements leads to the selec-
tion of gameplay that helps unify the math learning and architectural building
actions.
On the other hand, the learning system designers and game programmers in the
team were concerned that the performance of certain architectural building ele-
ments, such as creative and visual design or landscape and city planning, were not
only beyond the scope of the design goal but also hard for the digital gaming system
to gauge. There was also a concern that a variety of differing game mechanics
derived from the simulation and representation of varied architectural building tasks
and principles would lead to a lack of focus and consistency in gameplay and hence
diminish engagement. Instead, members of the team proposed that only one or two
gaming actions be deployed, as suggested in the following design note:
A player is given a set of materials and must create a structure. Performance is evaluated
in the game by the number of waste materials left over, time to completion, and structural
soundness of the structure. Peers can also rate the aesthetic beauty/originality of the
structure. There can be several constraints in E-Rebuild, such as 1) amount of materials,
2) strength of materials, 3) weight of materials, 4) variety of materials, 5) time, or 6)
building specifications. Let’s take constraints 1 and 4. In this case, assume that the player
is given 1000 cubic feet of oak wood and 100 square feet of glass. Based on these
constraints, the player must decide how to build a house.
This game design corresponds to a measurement core mechanic. A person must
create a blueprint of the structure that will be used to cut the materials to be used for the
structure. This blueprint specifies the dimensions of every element in the structure. A
player sketches a blueprint with a pencil and paper outside the game. With the blueprint
created, the player enters every unique component (e.g., material, quantity, dimensions) of
the structure into a menu within the game. An exemplary entry might take the form: wood
board, 50, 2 X 4. The material is always determined by the materials given in the problem
(wood, glass). Then the game automatically “cuts” the materials. I do not think the player
should cut material as this would get very tedious and repetitive. Once the materials are
cut to size, the player must assemble the materials using the mouse and the blueprint.
3.2 Core Design Patterns and Activities 59

In this design proposal, the core game mechanic is measurement, which consists of
the secondary mechanics of structure planning/sketching, material cutting, and
material composing. The core gameplay rules are design constraints such as effi-
ciency, thriftiness, and soundness. These ideas support simplicity and intrinsic con-
tent integration. Yet the practice of a major gaming action (e.g., blueprint sketching)
outside the game world may lead to the interruption of game flow. The menu input
for the cutting action led to a “design” action that felt like a mathematical drill that
was tedious.
By integrating the aforementioned two interdisciplinary perspectives, we decided
to focus on three core architectural design actions and constraints that identify and
unify the most dynamic processes of architectural building and mathematical rea-
soning and are in-game play. These core actions included (a) gathering materials
(including site surveying, material collection, and transformation), (b) crafting
(measurement + cutting), and (c) building (with given design constraints, evaluated
according to thriftiness and sturdiness). These core actions and rules defined the
first version of E-Rebuild, which was prototyped, user tested, and evaluated based
on the feasibility, playability, and learning-necessitating potential. After iterative
prototyping and user testing, the material crafting action was later replaced by a
material-trading action; an allocation (of space) action was added later.
Design pattern summary Gameplay exploration consists of identifying gaming
actions and rules that (a) discern and unify the most dynamic and engaging pro-
cesses of the performance systems to be simulated, (b) are feasible (e.g., being actu-
ated and assessed in the game), and (c) are streamlined and scalable (to compose
diverse tasks or scenarios). The gameplay exploration is driven by sharing, review-
ing, and aggregating interdisciplinary task examples and design cases. A purposeful
mapping of the association between applicable game mechanics and the design
cases/tasks proposed, along with the subsequent prototyping and user testing, will
then help to classify and consolidate the game mechanics.

3.2.3  esign Constraints for Rebuilding: Evaluating

D
and Rewarding Performance in the Game

Problem statement A design inquiry related to the exploration of gameplay

defines the evaluation of gaming success. Some team members argued that players
require flexibility with their design tools and goals, while others pointed out that
specifying design constraints and measurable rules is critical for computerized
assessment of a design artifact in the game. A variety of relevant questions arose
during the team discussion: What are the success criteria for post-quake rebuilding?
Restore (replicate) the pre-quake structures/models or withstand earthquakes of dif-
ferent magnitudes? Should the rebuilding task be accelerated given a forthcoming
earthquake, or self-paced with the option of testing the structure with an earthquake
60 3 Interdisciplinary Design Activities and Patterns

simulation? Do we want computerized or peer evaluation? What are the design

constraints and performance evaluation criteria?

Exploration of the design solution For the inquiry of rebuilding performance

evaluation, the architecture and scientific computing experts led the team to a
document research and case review of the earthquake architecture. This exercise
contributed a list of factors that mediate the challenging nature of a building task
(in terms of sturdiness of the structure against the earthquake vibration), includ-
ing the types of earthquake (and hence the necessity of designing diverse support
structures) and other associated design parameters (e.g., liquidation issue of the
terrain). After the initial prototyping and user-testing of the earthquake architec-
ture, we found that simulating all these factors was not only time-consuming but
also led to learning challenges far beyond the targeted competencies. For exam-
ple, creating an anti-­earthquake building would require a player to define the
center of mass of a composite structure and calculate friction. Evaluating the
collapsed objects to gauge the structure’s sturdiness is also tricky. For example, it
is difficult to define the degree of “collapsing.” What would be the consequence
if the structure only had one layer and hence would only rotate or move instead
of falling down given the earthquake vibration? Given all these design consider-
ations, the team decided to focus on structure/model rebuilding rather than build-
ing anti-quake structures.
In the initial design phase, the team could not persuade each other and hence
came to a compromised decision that we would enable both approaches of evaluat-
ing a gamer’s design performance: (1) computer evaluation based on the structure
or functionality of the architecture (e.g., whether the structure will collapse with
shaking, how much the site space is filled and used efficiently) and (2) evaluation by
peers on the organizing principles and/or art/beauty of the architecture (e.g., peer
critique, or design-protocol-based, portfolio-based evaluation), which can promote
reflection on action. Yet with a clearer definition of the design goal in the later phase
and after more in-depth discussion on the logistics and training required for imple-
menting peer evaluation with middle-school students, the team collectively decided
to drop the peer evaluation idea.
By synthesizing the reviewed cases and documents, the architect and learning
environment designers in the team co-proposed a list of the design parameters act-
ing as the performance evaluation standards for E-Rebuild gameplay. These param-
eters then drove the design of the game reward mechanism, as the following figure
illustrates (Fig. 3.4).
Design pattern summary The gameplay rules, which involve mainly the evalua-
tion of game performance and the game reward mechanism, are confined by the
success criteria of the simulated epistemic practice and the predefined design prob-
lem space. Research of the exemplary cases in the subject domain will contribute
evaluation rules that are aligned with the authentic performance success, while the
review of the predefined design goal and prototyping along with infield user testing
will help delimit goal-oriented and feasible rules.
3.2 Core Design Patterns and Activities 61

Fig. 3.4 Design parameter and game reward mechanism

3.2.4  esign Execution: Grounding Ideas in the Computer

D
Game System

Problem statement The execution or implementation of design propositions via a

computer game engine involves grounding and concretizing sketchy ideas related to
the user interface, tools, functions, interactive objects, landscape, logging structure,
narratives, and scaffolds/aids. The processes and end products that result from
developing these game system components are interdisciplinary in nature, convey-
ing and testing varied interpretations by the team members of game-based and
architectural design-oriented learning engagement. These salient design execution
processes in E-Rebuild evolved around a series of prompting questions or debatable
perspectives, such as the following:
• How will the player actually move materials around? What is the most intuitive
interface for rotating, stacking, and customizing (e.g., cutting and scaling)
objects? What exemplary interface mechanisms can act as a reference?
• Avoid text entry, if possible.
• Will “model sketching?” be handled from within or outside the game environ-
ment? Or should we make a set of model sketches (simulating real models from
Google Earth via Google SketchUp) available to the player and then perhaps
have the player construct model sketches at advanced levels?
• How will NPC characters be designed? For example, should we embed a “mini-
­me” avatar of the player to promote identity development and improve under-
standing of the proportion concept?
• Is it possible to embed an artificial intelligent agent who can scaffold game play
emotionally and cognitively?
Given the above questions, the teams’ perceptions varied and could be character-
ized as four discipline-related dispositions that prioritized, respectively, (a) intu-
itiveness in gaming interactions and hence an increased opportunity for game
62 3 Interdisciplinary Design Activities and Patterns

engagement, (b) motivated practice of mathematical reasoning and computation, (c)

increased opportunities for (architectural) design thinking, or (d) sufficiency and
quality of evidence collection to drive game-based learning assessment. These
diverse perspectives triggered lengthened, controversial discussions on the format
of a user interaction widget (e.g., numerical value entry, or multiple choice, versus
scrollbar), granularity or fidelity level in the simulation of core design actions (e.g.,
building, model sketching) in relation to the prospect of facilitating math or design
learning, the design of game world and objects, the proportion and framing of nar-
ratives in the game world, the approach of logging and assessing gaming actions,
and the approach of embedding in-game support (e.g., background help versus
active prompts).
Exploration of the design solution The uncertainty in the design execution of
E-Rebuild was resolved and settled via iterative design prototyping, expert review,
and infield user testing. Feasibility, learnability, and playability are three integral,
yet sometimes controversial, facets to be balanced when we gauged and settled on
game development details.
The interaction interface for core gaming actions (e.g., to build item maneu-
vering) was debated in relation to whether activating physics during the building
actions would be beneficial or disruptive to gameplay and learning engagement. On
the one hand, positioning, stacking, and rotating 3D physical objects are more chal-
lenging than maneuvering nonphysical ones. With the physics law activated, objects
can bump off each other, fall down, and crash the structure. Thus, building a struc-
ture becomes a carefully planned and precisely executed series of actions and
movements. Such an interface creates a learning curve and a reduced sense of
autonomy for game players, especially those not used to 3D gaming. On the other
hand, simulating the authenticity of the building process is a natural part of an
architectural simulation game. We also observed that by decreasing the intuitive-
ness of object maneuvering interaction, the game has made it compulsory for a
player to engage in planning their building movements (e.g., where to position each
item and how to rotate it precisely for stacking) and hence reinforced their spatial
and geometrical reasoning. Besides, adding a “snapping together” feature to the
stacking function made layering an object on top of another more efficient. In other
terms, the team compromised on a trade-off and purposeful integration between
learnability and playability during the gaming interface development.
Another example of compromise appeared in the decision to adopt a numerical
value entry, rather than the choice button or scrollbar, for user input. The human-­
computer interaction, and gaming literature in general, suggests that text entry may
interrupt the game flow. Yet our initial user testing of the choice button and scrollbar
for core gaming actions (such as training items and allocating resources) indicated
that they were associated with frequent exertion of guessing and random clicking,
and the lack of mindful mathematical computation. Learning system designers and
educational measurement experts then argued that prompting players to enter spe-
cific numerical values would not only increase the chance of mathematical thinking
but also provide more direct evidence for game-based learning assessment. This
3.2 Core Design Patterns and Activities 63

proposition was executed and infield tested. The team agreed that an intermediary
interface (e.g., text entry) helped enhance game-based learning engagement.
Tool development for the simulated design practice and mathematical problem-­
solving in E-Rebuild was another salient process experiencing iterative review and
refinement. For example, a cutting/scaling tool that “cut/scale items along the three
coordinate direction x, y, and/or z” was proposed as the following development note
outlined:
Scale and cutting tools: Using a drag-and-highlight tool to click on the vector points of
the original object (click the right area, then the area would glow to show the tolerance)
to define the x, y, z value of the target cube, or to define the radius of the target sphere, or
the other relative key values of the other geometry forms, to have it scaled or cut.

This tool aims to represent various mathematical concepts dynamically. Specifically,

the cutting interaction would assist the conceptual development in composing/
decomposing geometric forms. The scaling interaction would assist proportional
reasoning and help correct a misconception that division (multiplication) does not
always create a smaller (bigger) element. Yet when prototyped and infield tested, we
found that this cutting/scaling tool led to incoherent game play, due to (a) the diffi-
culty of spawning irregular shapes/figures (e.g., cutting a pyramid or a polyhedron),
(b) its conflict with the trading mechanic that aims to regulate what players can buy
(since they could buy a big item and cut/scale it), and (c) the potential of breaking
the thrifty rule (because scaling creates the potential to waste material). The team
debated as to whether the tool should be removed. It was considered a powerful tool
facilitating mathematical reasoning and learning, yet difficult to be fully executed
and not aligned with the core game mechanics. The final decision was that the tool
would be activated only for particular levels/tasks wherein item customization
would be aligned with other gaming actions and work as a reward of successful
game play to motivate purposeful geometrical reasoning.
The “highlighter” was another tool streamlined during the development phase.
This tool, originally proposed to enable quick object “positioning” and integration
of architectural design/planning, was built based on a concept of grid view by refer-
ring to the editing interface of Portal 2 Puzzle Maker. The following development
note is a brief outline of object positioning action using a highlighter tool:
(The player) defines the construction area using highlighter tool; the defined construction
area will show the grid view (with the unit of grid to be defined by the player).
(Click to) select the highlighted shipping container (or any constructional element/
object).
(Click to) select the grid in which the container (constructional object) will be positioned;
the object will be positioned or placed onto the grid.
Adjust the object’s position: move, rotate, and elevate up and down (with highlighted,
dotted 3D grid viewing showing).

During prototyping, the development team found it difficult to program an active

3D grid view for the gameplay mode using Unity 3D. In addition, the introduction
of a user-defined grid-view into the gameplay led to more difficult computerized
gaming assessment.
64 3 Interdisciplinary Design Activities and Patterns

Coding game objects and developing the log structure are salient, interactive
processes that frame the 3D game world and in-game learning/performance assess-
ment. When coding major game objects, the team co-explored and outlined the
types, key properties, and associated basic functions of game objects based on (a)
an architectural building scenario script that describes and visualizes the exemplary
landscape of E-Rebuild (see Fig. 3.5) and (b) the targeted mathematical concepts
outlined in the aforementioned competency model. A constructional item object, for
example, can be outlined in the features of forms (e.g., cuboid, cylinder, triangular
prism, along with the basic functions of transforming, cutting, and scaling), physics
(e.g., friction, bounciness, density, and joint strength), materials (e.g., stone, wood,
brick), mass (e.g., solid or hollow), size (e.g., height, width, depth, radius, base,
angle), position, and occupancy (livable or not, empty to not). The key properties
and functions of each object class were then refined during the game programming
phase to better align with the interface of the Unity 3D development system.
Based on the defined object classes, the game programming sub-team then
drafted the game log structure that would drive the capturing of game performance
and provide the gameplay data to be mined for game-based learning assessment.
This log structure (see Fig. 3.6) was then reviewed and refined by the educational
measurement experts in the team to ensure that the actions and states logged were
in line with the task model, which was developed alongside the competency model,

Fig. 3.5 An Exemplary scenario script document of E-Rebuild. Historic Rural Southwest
Scenario: Adobe making yard with stacks of some adobes and mud pile, Basic landscape and not
many trees, Kiva with ladder, Mountains in the background, Orno ovens in yards, Plaza with rec-
tilinear or square adobe houses around it, Water of some kind…old-fashioned well and pumps,
White church with graveyard, Sheep herd. Suggested Uses of Elements for Design: Adobe bricks
(pick a size and this will be a great math problem), Mud for mortar, Sun for drying adobes
3.2 Core Design Patterns and Activities 65

Fig. 3.6 An example of game log structure draft

66 3 Interdisciplinary Design Activities and Patterns

and to provide sufficient high-quality data for a stealth assessment of the targeted
mathematical competency development.
Assessment of game play success was a major execution challenge that con-
fined the breadth of the previously defined game task evaluation and reward mecha-
nism (see Fig. 3.4). Detecting and scoring the player’s architectural artifact in the
game world was challenging mainly due to the conflicting demands of creativity
and feasibility. For example, a fully computerized evaluation of a 3D free-form
object based on the full set of architectural design and visual principles (e.g., sym-
metry, hierarchy, rhythm, transformation) involves developing and coding an
extremely extensive (if not interminable) list of measurable/observable features in
order to concretize the aesthetic representation, leading to endless programming
and debugging work of the task assessment. To confine the evaluation space, clear
and measurable design parameters were further delineated, prototyped, user tested,
and iteratively refined. The occupancy, position, size, and shape (or form), in rela-
tion to the limits of material, time, and living needs, were settled by the team as the
core design parameters for the evaluation of a built architectural structure. The
sub-team of game programming and the scientific computing experts then led the
discussion and the crafting of the specific formula for computing/scoring a build-
ing task (see Fig. 3.7).

Fig. 3.7 A white-board sketch of the scoring of an architectural artifact

3.2 Core Design Patterns and Activities 67

Design pattern summary Design execution is a critical phase during which the
team reviews and gauges the integration and balance of feasibility, learnability, and
playability of each and every design proposition for the game mechanic, in-game
learning, and assessment. Mainly via iterative prototyping in game engine, interdis-
ciplinary expert review, and infield user testing, the team selects design features that
integrate learning and play given limited time and resources and better plan the
technical details that frame the interactive game world, user interactions, and the
interaction performance logging and scoring.

3.2.5 Structure and Navigation of Game Levels

Problem statement Designing the structure of game levels to create a responsive,

game-based learning path is another critical design inquiry for the E-Rebuild team.
For example, what is the appropriate difficulty level for the initial game task/level?
How should a task, level, and multilevel episode be structured? Should the naviga-
tion across the game levels be linear or nonlinear? When addressing these questions,
one could resort to multiple approaches of game structuring, including the segmen-
tation of an in-game adventure story, a gradual introduction of game mechanics, a
scaffold of mathematical learning and problem-solving, and/or an accumulation of
performance data to drive the training and test of an assessment model (e.g., a
Bayesian network). One might also argue for learner-selected, nonlinear navigation
to support learner autonomy versus a computer-controlled, sequenced navigation to
guide gaming and learning skill development.

Exploration of the design solution The team had tried to integrate a post-
quake-­themed narrative, introduction of basic game rules and actions, and initia-
tion of mathematical conceptual learning when designing the initial game level,
as portrayed by the following design note:
The protagonist is trapped in a hole and needs to get out. All that is available are some
wooden blocks. The protagonist then needs to cross a river and all that is available are
some branches. (Initial problem/task used to train the actions of moving, rotating, and
stacking items.)
The protagonist comes across some victims who need a home built before a heavy rain
comes in. (Second problem—for mathematical problem-solving)
The victims can include four types of being, human adult, human child, animal adult,
and animal child, or animal above and under a specified weight. A human adult has a
standard requirement of space or cubic space (area or volume); the other three beings
will require different proportions of standard (cubic) space.
The site or the safe plane available for the home construction is limited (e.g., only 300
square feet, in certain irregular shape).
The protagonist comes across some victims who need a dam built before a river gets too
high and floods the village (Another problem for a future level, potentially).
The protagonist comes across some villagers who complained about their homes or
shelters as lacking light or in wrong direction toward wind, in comparison against their
neighbors (so in need of transformation such as rotation, reflection, or translation); some
68 3 Interdisciplinary Design Activities and Patterns

Fig. 3.8 A design sketch of the initial game level

others complained about the size of their homes, wanting them to be scaled up (other
problems for future levels, potentially).

Based on the above design note, we then developed a mock-up of the initial game
level that encompassed the prototypes of the game world (e.g., earthquake simula-
tion, blocks and containers as the building items, and victim characters), basic gam-
ing actions (e.g., collecting and maneuvering items to build, allocating victims to
the built structure), constraints/rules (e.g., time, materials, design parameters, and
living space requirements), user interface (e.g., a help panel, feedback, a measure-
ment tool), and evaluation rules of the structure built (e.g., checking whether the
size/location/direction replicates the pre-quake model to 90%). Specific mathemati-
cal context problems embedded within the game levels were drafted by the mathe-
matical education expert in the team based on a collection of state and common core
test items for middle-school students. An exemplary draft/sketch of a game-based
math problem along with the design notes are provided below (Figs. 3.8 and 3.9):
Building a shelter via shipping containers to accommodate the victims saved:
16x10 + 8π square meters open space for potential construction, 67 people (40 children,
20 adults, 7 pets), 16 small containers (6 x 3 x3), and 8 big containers (12 x3 x3)
requiring 5 meters higher than the base for living space
Minimum space for pet, child, and adult: 1:2:4
Maximum space for pet, child, and adult (meaning numbers bigger than this maximum
will not influence living being’s happiness any more): 2:4:8

The mock-up was iteratively refined through user testing with a small group of
middle-school students in a local school’s after-school program. It then served as
the initial archetype for the development of additional game levels.
3.2 Core Design Patterns and Activities 69

Fig. 3.9 Computerized prototyping of the first E-Rebuild level

Structuring and navigation To plan and explore the structure and navigation of
game levels, the team referred to instructional design strategies (particularly elabo-
ration theory, Reigeluth, 1992), prior work on flow experience in gaming (e.g.,
Csikszentmihályi, 1990; Kiili, 2005), and discussions of evidence accumulation
across tasks in the evidence-centered design approach (Mislevy, Almond, & Lukas,
2003). A conceptual framework on how to sequence and navigate levels based on
the embedded math competencies, the highlighted architectural design scenarios,
70 3 Interdisciplinary Design Activities and Patterns

Fig. 3.10 Game level sequencing and its integration of math competencies and design scenarios

and the difficulty index of tasks, was then drafted (see Fig. 3.10). We then proposed
that the player could proceed across levels by (a) exploring tasks of the same archi-
tectural design scenario and same subset of math topics but varied difficulty levels,
and then advancing to a different scenario with different subsets of math topics (as
portrayed by the blue line in Fig. 3.10) or (b) exploring varied architectural design
scenarios along with different sets of math topics at a similar difficulty level and
then advancing to more challenging tasks (as portrayed by the red line in Fig. 3.10).
Path a or b can be set by the computer as a fixed, linear sequence for all players.
Alternatively, the game could allow the players to self-choose a path, to skip certain
levels, or even to explore any levels without a sequence/path.
The team was indecisive as to which path to employ. Path a appeared to be more
aligned with E-Rebuild level development sequence and hence was executed first in
our initial infield feasibility studies with 66 6th–7th graders in their science or math
classes. The infield observation indicated that players differed in their gameplay
progress and naturally shared/compared their gaming experiences. While some
remained stuck in Scenario 1, others had proceeded to Scenario 2 or even Scenario
3. The former group demonstrated obvious frustration when their peers showed off
the new game landscapes and items. Besides, not all students managed to complete
all scenarios within the study sessions and hence process all embedded mathemati-
cal concepts, which made it challenging to accumulate sufficient evidence or data
across tasks for the game-based math competency/learning assessment. This also
reduced the potential to provide an equivalent access to game-based mathematical
learning content. Given these infield findings, we then switched to Path b in the later
infield implementation studies and found that the aforementioned issues resolved
themselves. Another infield finding from our design experiments with the game level
3.3 Constructive Conflict, Group Synergy, and Leadership 71

structuring and navigation was that middle-school students were not adaptive to the
open-ended task structure and frequently requested help with chunking and sequenc-
ing steps of the design problem-solving in a game level (Ke et al., 2017). Even
though some E-Rebuild team members had argued for the sense of autonomy
afforded by a nonlinear navigation with the player choice, our experimental findings,
consistent with prior research (Kim, Almond, & Shute, 2016), implied that nonlinear
gameplay did not produce better game engagement or task performance but could
negatively impact the evidence accumulation for game-based learning assessment.
Therefore, a linear gaming structure was adopted.
Design pattern summary In a learning game that aims to incorporate content
learning, authentic problem-solving, and stealth learning assessment without inter-
rupting the game flow, the structure and navigation of game levels should be
designed to scaffold play, learning, and assessment simultaneously. Specifically,
the initial game levels should train the player on core game mechanics and frame
the backdrop mission or theme narrative that embed, represent, and contextualize
domain-specific concepts or problem-solving. Game structuring, encompassing the
structuring of targeted competencies, scenarios, and tasks variant in difficulty,
should scaffold conceptual or skill development as well as evidence accumulation
for learning assessment. Game-based navigation or learning path should be selected
and dynamically adjusted based on learners’ response and performance during the
infield testing.

3.3 Constructive Conflict, Group Synergy, and Leadership

Aside from the aforementioned salient design processes and patterns, constructive
conflicts, group synergy, and leadership were three emerging, prominent themes
found to advance segmented multidisciplinary perspectives and skills toward coher-
ent interdisciplinary design. They developed given purposeful preparations that
took some effort. These preparations involved (a) confronting and discussing issues
of scope and definitions to frame a common design goal or space, (b) using a com-
mon set of design aids to identify linking points between disciplinary propositions
and activate group synergy in the solution exploration, and (c) building participative
and transformational leadership to reduce deconstructive conflict and create effi-
cient decision-making.

3.3.1 Framing a Common Design Goal and Space

Team members of different disciplines initially proposed interpretations of the

design goal or the scope of design space for E-Rebuild, as implied by the following
meeting note that recorded team members’ perspectives on the most unique feature
of E-Rebuild.
72 3 Interdisciplinary Design Activities and Patterns

• Mini-me (embodied via a 1.5 meter middle-school student figure) performing

various architectural design/building activities.
• Integrating math history into the gameplay, for example:
An AI agent (history agent) has time-traveling power and can present not only
the views of the site/architectures before an earthquake but also the history of
ancient construction tools (e.g., perpendicular lines in the coordinate plane)
or practices (e.g., proportion reasoning when measurement units are not stan-
dardized or existing).
• 3D object maneuvering and transformation
• Development and validation of game-based stealth assessment of learning.
The above perspectives were presented by the scholars from the disciplines of
architecture, mathematics education, scientific computing, and educational mea-
surement, respectively. Such a diverse range of design priority propositions were
found to expand the scope and create a segmented agenda when framing the design
space. Thus, our initial design meetings mainly focused on consolidating design
goal interpretations.
Moreover, an interdisciplinary design team has different customs of classifying or
naming design concepts, which made the design discussion and brainstorming at the
initial design meetings especially challenging. For example, scholars of learning sys-
tems design, architecture, and scientific computing all had variant usages for terms.
Some examples include “blueprint” versus “floor plan” or “site plan,” “God’s view”
versus “plan view” or “elevation view,” and game “level, episode, or chapter,” which
led to an initial discussion on the design of gaming perspective and the game world
projection confusing and clumsy. Documents explicitly defining architectural vocab-
ulary, disciplinary concepts, and gamer terms were hence drafted to assist interdisci-
plinary design communication, as the following example illustrates (Table 3.1).

3.3.2 Using Common Design Aids

Exemplary design cases in architectural and learning games were used as the
“seeds” or the design aids to help the team identify linking points between disciplin-
ary propositions and activate group synergy in the solution exploration. The follow-
ing project meeting note highlighted the usage of design examples as the starting
point for the team to identify the linking points between disciplinary propositions
and activate group synergy in the solution exploration:
9/28
We have decided that before next meeting, we will further explore gaming tasks and core
mechanics of E-Rebuild by reviewing the Bridge! Construction game, Minecraft, and
Google SketchUp. The following is a suggested list of design questions to be considered/
addressed during the review.
What are the major game mechanics? Is it given resources and design constraints,
build or compose architecture?
3.3 Constructive Conflict, Group Synergy, and Leadership 73

Table 3.1 An exemplary sketch specifying gaming and scientific computing vocabulary
Building items (gaming) Objects (scientific computing/coding)
Forms Forms: cuboid, cylinder, sphere, triangular prism,
cone, pyramid
Transformation, cutting, scaling
Materials (e.g., stone, wood, brick, and Physics: friction, bounciness, density, joint strength
steel, glass?)
Solid/hollow Mass (Can hollow objects be cut? Center of mass?)
Size Height, width, depth, radius, base, angle
Positioning
Livable space Occupancy

What is the potential of mixing survival mode (speeded construction task) and creative
mode (simulation mode in which one can test the construction product with the
earthquake tool)?
Game actions: Moving, stacking, cutting, and scaling? Model sketching within the
game?
Rule: Limit of constructional items as a constraint of design? Reward the level pass
with more inventory items?
Will a building block (e.g., a cube or shipping container) be given by the system? How
and why should the system give the player shipping containers? Any story element?
What game rewarding system?
How will the tools be given to players, one at each game level, or once for all at
beginning?
How will players learn to play?
We have decided that each teammate should present a design sketch or a written
document to address the above questions and then have all thoughts synthesized during
next meeting.

3.3.3 Building Participative and Transformational Leadership

As the above design note illustrates, the E-Rebuild team resorted to participative
design practice that values the input of every team member and holds everyone
responsible for design brainstorming. A prerequisite of such a design style was the
lengthy discussion and congregation of each and every perspective in the project
meetings, extending the duration of the group meeting. Frequently, a common syn-
thesis could not be reached, and the meaningful negotiation was abruptly stopped
when a meeting adjourned. To increase the efficiency of decision-making and ensure
the design discussion would be thorough, we divided the whole team into groups
based on major design responsibilities, with each group led by a disciplinary expert.
We also separated the design meetings into two sections, the whole-project meeting
and the sub-team design meeting. At the former ones, we focused on the big picture,
sharing the reports of every sub-team and conducting high levels of communication/
decision-making to accomplish design goals. Smaller and detail tasks were dele-
gated to each sub-team who would hold separate sub-team design meetings to fully
74 3 Interdisciplinary Design Activities and Patterns

discuss and explore the solutions to discipline-specific design inquiries. These

solutions would be brought back to the whole-project meetings for review and
critique. In general, we observed that such a mixture of participative and transfor-
mational leadership (Burns, 1998) aided in efficient and constructive design deci-
sion-making in an interdisciplinary team.

References

Burns, J. M. (1998). Transactional and transforming leadership. In G. R. Hickman (Ed.), Leading

organizations: Perspectives for a new era (pp. 133–134). Thousand Oaks, CA: SAGE.
Csikszentmihályi, M. (1990). Flow: The psychology of optimal experience. New York: Harper &
Row.
Garris, R., Ahlers, R., & Driskell, J. E. (2002). Games, motivation, and learning: A research and
practice model. Simulation & Gaming, 33(4), 441.
Habgood, M. P. J., & Ainsworth, S. E. (2011). Motivating children to learn effectively: Exploring
the value of intrinsic integration in educational games. The Journal of the Learning Sciences,
20(2), 169–206.
Ke, F. (2016). Designing and integrating purposeful learning in game play: A systematic review.
Educational Technology, Research and Development, 64(2), 219–244.
Ke, F., Shute, V., Clark, K., Erlebacher, G., Smith, D., Fazian, P., Lee, S., & Xu, X. (2017, April).
Math learning through game-based architectural design and building. Paper presented at 2017
American Educational Research Association Annual Meeting, San Antonio, TX.
Kiili, K. (2005). Digital game-based learning: Towards an experiential gaming model. The Internet
and Higher Education, 8(1), 13–24.
Kim, Y. J., Almond, R. G., & Shute, V. J. (2016). Applying evidence-centered design for the devel-
opment of game-based assessments in physics playground. International Journal of Testing,
16(2), 142–163.
Mislevy, R. J., Almond, R. G., & Lukas, J. F. (2003). A brief introduction to evidence-centered
design. ETS Research Report Series, 2003(1). Retrieved from https://www.ets.org/Media/
Research/pdf/RR-03-16.pdf
Moustakas, C. (1994). Phenomenological research methods. Thousand Oaks, CA: Sage.
Pratt, D. D., Winters, N., Cerulli, M., & Leemkuil, H. (2009). A patterns approach to connecting
the design and deployment of mathematical games and simulations. In Technology-enhanced
learning (pp. 215–232). Dordrecht, Netherlands: Springer.
Reigeluth, C. (1992). Elaborating the elaboration theory. Educational Technology Research &
Development, 40(3), 80–86.
Winters, N., & Mor, Y. (2008). IDR: A participatory methodology for interdisciplinary design in
technology enhanced learning. Computers & Education, 50(2), 579–600.
Chapter 4
Design of Gameplay for Learning

Abstract A common skepticism about educational games is that learning and play
are frequently not well integrated—the skill or content to be used and learned lacks
a semantic or meaningful relation with the fantasy and challenge elements and can
be easily swapped without influencing gameplay. In this chapter, we describe and
analyze design challenges associated with the core components of gameplay—
game mechanics and the narrative scheme as it relates to learning—and review the
gameplay design propositions and infield test findings of E-Rebuild. Via a retro-
spective investigation of design features and strategies in terms of learnability and
playability (i.e., capability of activating knowledge-based cognitive performance
without interrupting gameplay), this chapter aims to report and discuss how domain-­
specific learning is integrated in, and activated by, core game actions, rules, game
objects, and the game world design.

4.1 Introduction

Gaming is an organized play structured by a set of rules and an obstacle-tackling

goal (Klopfer, Osterweil, & Salen, 2009; Schell, 2014; Suits, 1978). Learning is an
innate element of gaming, in that players interact with a game to learn rules and play
strategies and then adapt and improve play skills to make progress in the game
(Lindley & Sennersten, 2008). Yet learning in games varies in the degree to which
the acquired skills and strategies emphasize “domain-specific knowledge” (Ke,
2016; Tricot & Sweller, 2014). The more higher-degree domain-specific knowledge
is needed, the more careful design effort is required for the attainment of an integral
relationship between gaming and learning (Habgood & Ainsworth, 2011; Kafai,
1995; Malone & Lepper, 1987; Torbeyns, Lehtinen, & Elen, 2015). A common
skepticism about educational games (and, hence, a relevant design issue) is that
learning and play are frequently not well integrated—the skill or content to be used
and learned lacks a semantic or meaningful relation with the fantasy and challenge
elements and can be easily swapped without influencing gameplay. As a result,
players may be distracted by the play part and not achieve the learning goals, or they

© Springer Nature Switzerland AG 2019 75

F. Ke et al., Interdisciplinary Design of Game-based Learning Platforms,
Advances in Game-Based Learning, https://doi.org/10.1007/978-3-030-04339-1_4
76 4 Design of Gameplay for Learning

may become disengaged because learning elements may corrupt what is enjoyable
about games (Garris, Ahlers, & Driskell, 2002; Miller, Lehman, & Koedinger,
1999). For example, in E-Rebuild we could design a “trading” challenge that
requires the player to complete a screen of math questions to obtain the corresponding
number of game tokens and then use the tokens earned to purchase items needed.
Such a game mechanics (i.e., the question-answering point system) could be reused
for another challenge (e.g., earning time credits to buy more time for a “building”
action) only with the content or math topics of the questions swapped as needed. An
extrinsic integration of content in gameplay, as the example illustrates, appears to be
a rapid and all-purpose gameplay design strategy. But it is, as prior research suggests,
deficient in reinforcing game-based engagement and won’t likely promote desirable
active and deep learning, especially for learners who lack the competency of and
positive disposition toward the subject matter (Ke, Xie, & Xie, 2016; Richards,
Stebbins, & Moellering, 2013).
In this chapter, we describe and analyze design challenges associated with the
core components of gameplay—game mechanics and the narrative scheme as it
relates to learning—and review the gameplay design propositions and infield test
findings of E-Rebuild. Via a retrospective investigation of design features and strate-
gies in terms of learnability and playability (i.e., capability of activating knowledge-­
based cognitive performance without interrupting gameplay), this chapter aims to
report and discuss how domain-specific learning is integrated in and activated by
core game actions, rules, game objects, and the game world design.

4.2  onceptualization of Game Mechanics, Narrative,

C
and Learning Integration

Endorsing the perspectives of the previous game studies, we have designed and
examined the construct of gameplay in two layers: the “ludus” or game mechanics
layer that involves rules and actions, and the narrative layer that comprises the setting
(or scenario), backdrop mission, and game objects (Ang, 2006; Frasca, 1999). It is
commonly believed that gameplay lies at the meaningful interplay between the two
layers, although whether game design is more the design of experience (Salen &
Zimmerman, 2004) or a narrative architecture (Jenkins, 2002) is still inconclusive.
According to Järvinen (2008) and Sicart (2008), the term game mechanics refers
to an activity structure consisting of rules and play actions. Rules are designed to
determine the conduct and standard for both play behaviors and the winning/losing
state. These rules lead to the creation of player strategies and actions with which the
player can interact with game elements to “influence the game state towards the
attainment of a goal” (Järvinen, 2008, p. 255). The game mechanics hence is “a
compound activity composed of a suite of actions” that players, abiding by the
rules, recurrently perform and directly apply to achieve the goal state (Salen &
Zimmerman, 2004, p. 316; Sicart, 2008).
4.3 Learning-Play Integration in E-Rebuild 77

Not all games tell a story, and hence the narrative layer is not a defining feature
of games (Ke, 2016). But many games do have narrative aspirations, or at least tap
into the player’s memory of previous narrative experiences (Jenkins, 2002).
According to Jenkins (2002), narrative can be integrated into a game as a broadly
defined goal, such as a backdrop plot or mission, (b) a localized incident or plot
developed in game level(s), and/or (c) an open-ended game world that allows play-
ers to define their own stories via authoring- or construction-based play. Notably,
a game can create an immersive narrative experience and convey storytelling via
the “spatiality” of the game world, in which the narrative element is infused into a
space that a player navigates through and interacts with. In E-Rebuild, we have
explored the aforementioned three approaches of the narrative layer during the
design process and in particular focused on examining the feasibility and effec-
tiveness of representing a mathematical story problem via the spatial narrative of
the game world.
In a recent literature review of domain-specific content learning via gameplay
(Ke, 2016), the typical processes of integrating learning into game actions are
described as representation, simulation, and contextualization. Specifically, repre-
sentation involves designing game objects as external representations of the tar-
geted concepts and interactions with game objects as the conceptual exploration or
application processes. Simulation is achieved via designing the game world as a
scientific problem or system and hence the game actions as an iterative process of
problem-solving and discovery learning. Contextualization, conveyed via a back-
drop mission, a plot, game characters, and/or meaningful scenarios, is then employed
to increase the pertinence and fascination of content representation and problem-­
solving simulation for players.

4.3 Learning-Play Integration in E-Rebuild

The design of E-Rebuild has been inspired by previous sandbox or architecture

game examples (e.g., Minecraft, SimCity), 3D modeling tools (e.g., Google
SketchUp), and the previous architectural design education programs for children
(e.g., Architecture and Children Pilot Curriculum; Taylor, 2009). Different from
and extending these design stimuli, E-Rebuild aligns and combines construction-/
authoring-themed gameplay, accessible 3D design, and interaction interface, as well
as architectural design missions and setting, to facilitate mathematical learning and
problem-solving for middle-school students. Designing core game mechanics that
manage to embody such an interdisciplinary design goal is critical and challenging.
What should be the nature or foundational “act” of the interdisciplinary, learning-
constructive gameplay? What corresponding game actions and rules are found to
reinforce the game’s learnability and playability? How will backdrop and virtual
narrative environment contextualize or “legitimize” the target learning and play
actions?
78 4 Design of Gameplay for Learning

The aforementioned questions underlay our design inquiry on the core game
mechanics of E-Rebuild. We conducted a retrospective and thematic analysis with
the design notes, meeting records, design talk among teammates at meetings, and
the observation notes of players’ actions and reactions when test-playing the proto-
typed E-Rebuild game levels. The analyses focused on delineating salient design
propositions and beliefs governing the core game mechanics and narrative scheme,
refined perceptions and design moves during infield testing, and the identified, func-
tional design heuristics that promote in-game learning actions without interrupting
the flow of play. These design research findings, with support of citations and exam-
ples, are presented below in a succession of themes that collectively constitute the
thinking process and core sectors of the game mechanics and narrative design
process.

4.3.1 Setting Genre and Nature of the Core Gameplay

There were two intertwining threads of design effort in the initial inquiry of core
gameplay design in E-Rebuild—searching the genre of building-themed gameplay
and investigating the nature of building-relevant mathematical experience. The for-
mer thread of design effort highlighted the element of play by addressing the fol-
lowing questions: “What will be building-themed play? What is the core action of
play? What will be the major obstacle and goal for such a play act?” The latter
thread of design focused on the learning experience by defining the core compo-
nents of a salient or desirable mathematical experience or thinking process. The
intent was to seek the salient and unified actions existing between architectural
design and mathematical knowledge application and gauge how and why these
interdisciplinary actions could be challenging while personally meaningful for stu-
dent players to be “fun.”
Genre or type of building-themed gameplay Why is building fun? was the head-
line title of the first design meeting memo of the E-Rebuild project, exemplifying
the predominant and early design effort that was contributed to exploring the poten-
tial genres of serious play enabled by the processes of planning, designing, and
constructing buildings and other physical structures. Such an exploration was acti-
vated via (a) an inspection of toys, games, simulations, and other hypermedia exam-
ples highlighting the theme of building, especially survival rebuilding activities, and
(b) a reflective design conversation among the design team members in relation to
our own experiences and preferences of building-themed gameplay.
The design team members of E-Rebuild have, individually and collectively,
sought, reviewed, and test-played a collection of architectural construction- or
building-oriented toys, games, simulations, and other hypermedia applications. The
assorted prior examples of building-themed serious play demonstrated various
modes of play, such as block stacking puzzle (e.g., https://play.google.com/store/
apps/details?id=com.ketchapp.stack), block building in the real or virtual world
4.3 Learning-Play Integration in E-Rebuild 79

(e.g., Magna-Tiles, Minecraft), adventure (e.g., Raw Danger!, Camelot Unchained),

strategy game (e.g., http://www.stopdisastersgame.org/en/home.html), and physics
simulation (e.g., Bridge Constructor, Bridge Construction Simulator). These exam-
ples of building-themed gameplay could evolve around a designer-defined game
goal or the preset criteria for game success, such as height and balance in block
stacking, sturdiness of the construction, survival, and/or people and property protec-
tion against a disaster. It may also depict an open-ended, self-determined goal of
play, such as creative play via block building in a sandbox game. Core player or
user actions include block stacking, (information or object) collection, and decision-­
making, with real-time physics simulation activated or deactivated during those
actions. Reflecting a focus on an adventure or a construction mode of play, the core
gaming action adopts a plan view or an elevation view, whether in first- or third-
person perspective. Ultimately, the obstacles to be resolved to achieve the play suc-
cess are typically the challenge posed by the properties of the building materials in
relation to the construction goal (e.g., the weight, shape, structure, variety, and/or
scarcity of construction materials in a block stacking puzzle or a bridge construction
game), time limitation, and insufficiency of prerequisite knowledge or skills of a
construction task (e.g., engineering concepts in Minecraft; physics in Bridge
Constructor; science, economics, and global issues in Stop Disasters; or generic
problem-solving skill in Raw Danger!).
The design case review and test-play notes of the design team indicate variability
among the team members in terms of the play preference, discipline knowledge, and
hence the thought processes for design. For example, regarding two exemplary con-
struction- or building-themed games, a game designer commented, “I think what we
can learn from these two games are the game actions and the potential design con-
straints we can have: Given a landscape or terrain, given the potential risk of future
hazards and limited budget and materials, where and what kind of post-­quake archi-
tectures should we invest in and build?” The comment carried the game design
language (e.g., “game actions” and “constraints”) and a designer perspective gov-
erning the common core game mechanics in the game examples. Yet two team
members, a mathematician and an educational measurement specialist, made differ-
ent annotations (provided below). These annotations highlighted the player percep-
tion, indicating different player preferences.
https://eduweb.com/portfolio/bridgetoclassroom/engineeringfull.html
http://www.stopdisastersgame.org/en/home.html
Cool games. I just looked them over quickly, and they seem interesting!
I played the bridge game and it was boring. The other earthquake game was a bit more
engaging, but not by much. We have a chance to build a more compelling game for this
niche!

The second commenter then expressed her preference of adventure games (e.g.,
Raw Danger!) that got concurred by another team member who was also a fan of
sandbox games like Minecraft.
Indeed, play preferences tended to mediate how individual team members framed
the role of building or constructing in the play experience. Some perceived the task
of construction or building by itself as a fun process, thus framing the core ­gameplay
80 4 Design of Gameplay for Learning

as the simulation of building actions and replicating architectural design constraints

and specification as the obstacles and goal state. Others considered building mainly
a backdrop mission or a meaningful context that legitimizes the gameplay of cre-
ative problem-solving and adventure. Another individual characteristic that medi-
ated the thought processes for building-themed gameplay design is disciplinary
knowledge or epistemology. The architecture educator, for example, insistently
emphasized the basic principles of design (e.g., balance in visual distribution,
rhythm, unity) and creativity as core goal states or success evaluation criteria. Yet
game designers and programmers in the team preferred measurable objectives (e.g.,
time, thriftiness in material usage, structural soundness, and building specifications
of a structure) and expressed concern as to whether and how some principles and
elements of design can be defined and specified quantitatively for computerized in-
game evaluation. Certain team members also worried that giving players unlimited
materials and seeing what they create would eliminate constraints, thus reducing the
challenge and hence fun in the game.
After multiple rounds of reflective and active conversations with the design cases
and teammates on the type of building-themed gameplay, we decided that our proto-
type of the core game mechanics would try to integrate different design perspectives,
which would be infield tested by the learners and then refined based on the testing
data and learners’ feedback. Our initial prototype of building-themed gameplay,
therefore, encompassed the following elements: (a) shifting between a first-­person,
adventure mode (with collection, site measurement, and space allocation as sector-
specific game actions) and a third-person, building mode (with trading and building
as core game actions, Fig. 4.1); (b) simulating physics (e.g., earthquake waves) to
test the structural soundness of a building artifact; (c) enabling both plan and eleva-
tion view in the building mode (Fig. 4.2); (d) depicting building obstacles as a variety
of landscapes, time, and material limit; (e) defining/evaluating the goal stated mainly
via time, materials usage, structural soundness (e.g., not collapsing given variant
earthquake waves), and preset building specifications (e.g., size, location, and direc-
tion aligned with a pre-quake house model); and (f) letting peers and potentially the
teacher rate the aesthetic design elements and originality of the structure outside the
game during post-game debriefing. This prototype later got iteratively reviewed,
user-tested, and refined considerably, as described in a later section.
Building-relevant mathematical experience Another concurrent thread of design
inquiry on core gameplay in E-Rebuild was to investigate the salient, common fea-
tures of both a mathematical experience and an architectural design/construction
experience, or What makes a good building task or activity that creates engaging
and effective mathematical experiences for participants? Exploring building-rele-
vant mathematical experience is critical for an intrinsic integration of learning and
building in gameplay. The exploration process evolved through the interviewing of
disciplinary experts, the literature and design case review, and an inspection of
exemplary design education curricula (e.g., Architecture and Children Pilot
Curriculum; Taylor, 2009) that promote the dynamic architectural and mathematical
experiences for children.
4.3 Learning-Play Integration in E-Rebuild 81

Fig. 4.1 Adventure and building modes of play

The design discussions on the joint features of architectural and mathematical

experiences involved the analysis of core actions, basic resources or materials, fre-
quently used tools, requisite knowledge and skills, and a set of archetypal practices
typifying the aforementioned elements. For instance, the following game design note
carried initial discussions on the potential conjoint architecture-math experience
elements.
82 4 Design of Gameplay for Learning

Fig. 4.2 Plan and elevation views

Tools
• Measuring/layout tools: rule, square, protractor, etc.
• Sketching tools: brush.
• Construction tools: e.g., cutting, transportation, elevation tools, etc.
Basic materials
• Building unit: line, plane, form, or prefabricated block.
• Shape: cube, arch, cylinder, etc.
• Texture: stone, brick, etc.
Actions
• Navigate and explore (i.e., measure, collect data, and analyze data).
• Draw or sketch via extended devices and apps—mixed reality.
• Build: select, drag, drop, and customize.
• Test (e.g., with simulated earthquake waves) and predict, decide, or reflect.
• Collect, trade, and/or manage resources (e.g., time, space, material cost).
Exemplary Practices
• Develop architectural plans and a model for a “mouse” house: Bubble dia-
gram –> floor plan (a plan view) –> elevation drawing (2D drawings that show
the outside walls of a building, e.g., the front side with door) –> perspective
drawing (a house that is three-dimensional looking) –> model construction.
• Related task (p. 8 in Taylor, 2009): Designing a vegetarian fast-food restaurant
(eating and preparation areas, delivery space storage, restrooms, parking,
etc.—spaces only, not equipment).
• Predesign exercise or puzzle: decorate a shoe
• Structure in architecture: arch, triangle, or asymmetrical vs. balanced.
• Related task: The picnic shelters must use a space frame structure for their roofs.
Each shelter will have four picnic tables, and the space frame will rest on at least
four columns.
4.3 Learning-Play Integration in E-Rebuild 83

• Predesign exercise or puzzle: Toothpick puzzles

• Form in architecture, involving measurement, data interpretation, scale, and
proportion.
• Related task: Use organic structures found in nature in a “fantasy” architectural
structure for the “moon.”
• Predesign exercise or puzzle: Scavenger hunt for basic geometric shapes and
forms in nature (e.g., crystal, plants’ spirals, spider’s net) (p. 31).
• Related task: Design a “people’s” bridge across the canyon to build a jogging
and bike path. This bridge should be as beautiful as the landscape that surrounds
it and should appear to gracefully and sturdily cross the canyon (p. 35).
• Predesign puzzle: Designing and modeling natural bridges—a fallen tree as a
beam bridge, intertwining vines as suspension chains or bridge, and a rockfall
and later the water eroded some rocks away forming a simple arch over the
stream (i.e., the arch bridge) (p. 33)
• Visual design: Super wall graphics using concrete block or block grid pattern
enlargement method (scale, proportion, measurement, introduction of area and
perimeter) (p. 38).
• Related task: Drawing and painting a supergraphics for your village’s school,
which will represent your community.
• Predesign exercise or puzzle: Painting an existing graphic projected onto a brick
wall. Mixing color and texture
• Creative, imaginative design: My favorite place.
• Related task: Design and model an “ideal classroom” using a scale, and then
construct it.
• Similarities between body and building systems (membrane-skin, mechanical
systems-muscular system, heating/cooling-circulatory system, electrical system-
nervous system, structural system-skeleton, waste disposal system-digestive sys-
tem, ventilation-respiratory system, mechanical devices that lift and push
things-muscular system).
• Related task: Transformer “You are architecture.”
• Landscape (p. 45)—collaboratively make a section drawing from a topographic
map of a site, schematic drawing of individual designs, and create models and
the final working architecture.
• Related task: Redesign a creative play park for the surrounding community and
the school.
• City planning—collaboratively plan, model, and build transportation, residen-
tial areas, business district, recreation and open spaces, and the capitol complex
(p. 50) (data collection and interpretation, in addition to all prior knowledge).
• Predesign exercise or puzzle: mapping and touring.
As illustrated in the above design notes, our initial exploration of building-rele-
vant mathematical experience has pinpointed a short list of core actions—site analy-
sis (surveying), sketching, building (design) experimentation, and resource
management. Yet there remains much ambiguity and disagreement within the team
discussions: Will the experience or actions highlight design/sketching (e.g., drawing
84 4 Design of Gameplay for Learning

and visual pattern design) or building/modeling (e.g., the composition and

decomposition of the forms and shapes) or both? What is the desired the granularity
level of the building action? What are the various forms assumed by the building
units? What and how will the tools be integrated into the game setting?
The design discussion on the role of design/sketching versus building/model-
ing in the gameplay experience was actually a comparative analysis of the two
prospective game mechanics in enabling and legitimizing targeted mathematical
practices. In particular, we discussed whether and via what enabling acts or moves
the two mechanics would embody the multimodal conceptualization and applica-
tion of the targeted math knowledge (e.g., angle measure, area, and volume of
geometric figures). Via expert interviews and exemplary task performance obser-
vation along with task analyses, we found that the game mechanics of building or
modeling tends to optimally align with the targeted learning actions (Table 4.1).
We then wondered about the applicable user input interface for each action and
the feasibility of in-­game tracking and assessing of design or building actions.
Design, especially sketching, as the game mechanics would entail high-precision
touch screen interaction (e.g., tablets or laptops with stylus); the complexity
involved in architectural drawing and creative design also presented challenges to
real-time, in-game behavior tracking and design artifact evaluation. As such,
gameplay in E-Rebuild portrayed more building/modeling actions than design/
sketching actions.
The granularity level in the basic unit to be maneuvered during the building/
modeling actions would vary. Specifically, using a “block” approach, the game
could provide prepackaged structures (e.g., a shipping container to be remodeled as
a house) as building blocks to be composed, decomposed, or modified to create vari-
ant structures. Alternatively, a player could build using the basic architectural forms
(e.g., bricks, doors, windows) to compose higher-level elements (e.g., walls, floors,
and roofs), then a house or bridge, then a neighborhood, and so forth. During design
discussions, the team perceived that both approaches have advantages. The design
hypotheses were that the previous approach would enable a novice player to
­construct a complicated structure efficiently while focusing his effort on planning
before building. The latter, though demanding a higher level of design knowledge

Table 4.1 Actions of learning and building

Learning actions Building actions
Identification and representation (e.g., of 2D/3D Collection and trading (of construction
geometric figures) elements)
Measurement and calculation (e.g., of length, angle, Construction site survey and plan
circumference, area, and volume)
Creation Building—positioning, rotating,
stacking, joining, copy-pasting
Analysis Allocation of space, time, and other
limited resources
Evaluation Design evaluation: fulfilling needs,
efficiency in cost and time
4.3 Learning-Play Integration in E-Rebuild 85

and technical fluency, would situate building units in a variety of forms in line with
the system of geometric shapes to be learned. Hence both approaches were used and
integrated across game episodes and tasks in E-Rebuild.
To implement building-relevant gameplay, the user input interface is another
critical design element. The interface should enable core game-based learning
actions while keeping players in control. In other words, it should facilitate action-
based mathematical conceptualization or knowledge application beyond affording
simplicity and efficiency. To explore appropriate user input interfaces for core game
actions, we designed multiple papers, toys, and computerized prototypes, iteratively
tested them with expert and novice gameplayers, and evaluated the interaction expe-
riences to select the interface and interaction tools that are not only user-friendly but
also conducive to learning. Detailed discussions of the user input interface or tool
design are provided in a later section.

4.3.2  esigning Learning-Play Integration in Game Actions

D
and Rules

During the past four project years, we have iteratively designed, tested, and refined
more than 30 prototypes of E-Rebuild. Learnability and playability are two key
objectives driving the iterative design tests and refinements. During iterative testing,
we have observed a continuum of gameplay engagement framed by the designed
game actions and rules—from play engagement (i.e., game-based play irrelevant to
the designed tasks), task engagement, and cognitive engagement to content engage-
ment, being incremental in demonstrating commitment to mathematical thinking,
knowledge application, and problem-solving. After iterative refinement and elimi-
nation, the resulting game actions and rules have the following core feature or func-
tionality: necessitating cognitive and content engagement within gameplay actions
and rewarding mathematical thinking as the most warranted gameplay strategy.
Cognitive and content engagement in game actions As discussed above, the
design conjecture of gameplay in E-Rebuild was that players would purposefully
engage in architectural building types of game actions, solve problems framed by
each game action using strategy and knowledge, and consequently experience a vari-
ety of action-based mathematical thinking processes. These thinking processes, cor-
responding to game actions (see Table 4.1), can involve analyzing the underlying
numerical, logical, and structural essentials of each problem, representing and apply-
ing conceptual knowledge to the problem, and then engaging in reasoning and finally
proof. As such, the gameplayer would develop the targeted knowledge and skills to
perform mathematical problem-solving or engage in mathematical thinking. Game
actions in E-Rebuild essentially necessitate both cognitive and content engagement.
Cognitive engagement in the game-based setting refers to engaging in generic cogni-
tive endeavors and making cognitive investment in tasks with purposiveness and
strategy use (Ge & Ifenthaler, 2017; Ke et al., 2016). Content engagement refers to
86 4 Design of Gameplay for Learning

engagement in which players also establish positive value systems and get motivated
in mathematical content (i.e., concepts, relations, formulas, and theorems) process-
ing, representation, and application (Ke et al., 2016).
The user-testing findings with our initial prototypes of E-Rebuild game actions,
however, indicated that many players by default lacked a disposition of looking at
an architectural building or design problem in a mathematical way, or looking for a
logical explanation, which is a key prerequisite of mathematical thinking. Instead,
they could seek task-irrelevant gameplay, such as exploring the game world without
purpose or creating their own play experiences with the game objects (e.g., using
the measurement/marking tool to sketch on the ground rather than to survey the
construction site and imagine the construction site as an arena and maneuvering 3D
building materials as armaments or vehicles). They might tackle game actions and
the associated problems in a mindless way, such as guessing with trial and error.
They would avoid math-relevant planning or strategic thinking, lacking the acts of
pre-estimating the construction cost during a trading action or predetermining the
structural properties of a construction in a building action. They would also shy
away from calculating the exact amount of resources available and needed during an
allocation action while resorting to testing and refining rough estimates, randomly
or with a content-generic strategy. They would spend limited time reflecting on or
reasoning with the basis and consequence of major gameplay inputs or the connec-
tion between them, thus lacking the practice of inductive or deductive reasoning.
More critically, we observed that student players generally lacked initiative in read-
ing during gaming, showing reluctance to peruse the mathematical problem or con-
tent information in a task narrative or the in-game help panel.
In summary, we found that cognitive and content engagement by players are not
promised even when the core game actions embody or operationalize the salient
elements of architectural and mathematical problem-solving. An earlier user-testing
study indicated that the percentage of cognitive engagement of players in all game-
play actions was 41% and that of content engagement was only 29%. To improve
the efficacy of E-Rebuild game actions in requiring cognitive and especially content
engagement, we experimented with the following design strategies for both game-
play actions and rules: (a) prioritizing “cognitively active” gameplay inputs and
behaviors that extrapolate the deed of representing or employing mathematical
knowledge and skills and curtail the chance of guessing and other mindless strate-
gies; (b) rewarding mathematical thinking (planning, calculation, and reasoning), or
domain-specific gameplay, as the most justified gameplay strategy by players; and
(c) presenting in-game mathematical information via visuals, action feedback, or
properties of interactive game objects, besides background narratives.
Prioritizing “cognitively active” gameplay inputs and conducts that extrapolate
the deed of “doing” mathematics and curtail the chance of guessing or other mindless
strategies is a major design strategy. An underlying index for our selection and refine-
ment of a core game action is whether and to what extent the action will n­ ecessitate
cognitively active gameplay inputs. Specifically, we have prioritized the “producing”
type of gameplay input (such as generating or constructing a numeric or quantitatively
4.3 Learning-Play Integration in E-Rebuild 87

accurate solution) to “choice-making” type of input (e.g., inspecting a given set of

options from which to choose a solution). We found the latter expedited and increased
the usage of trial-and-error strategy by players. We also tried to prioritize the game
action of which the associated challenge (or problem) requires the actual deed of
exploiting the resources of mathematical knowledge and skills, rather than just heu-
ristics strategies or general reasoning abilities, during problem-solving (Schoenfeld,
1985; Stacey, 2006). These design decisions are inferences extrapolated from the
iterative infield testing of game action archetypes, as the following example shows.
The learning objective of the “collection” game action, for example, is to enable
the action of identifying and representing the numerical and structural properties of
everyday objects and phenomena (e.g., shapes and measurements of construction
materials, structure of a family with the ratio of adult to child in their minimum liv-
ing space needs). Our original design assumption was that players would proac-
tively perform a variety of relevant learning acts (as depicted by an initial design
note below) enabled by and during the collection action, with the help of a measure-
ment tool, a goal statement for the collection action (e.g., using containers to con-
struct a shelter), and visual/verbal cues (e.g., a family cartoon signifying the
composition of family member, a mathematical statement about the unit space
need).
Exploration, including measuring and appraising the landscape, is part of material gather-
ing. We could integrate site analysis and terrain judging (e.g., on what part of the site is
higher or has solid ground for building?).
Decision-making: What and how much materials to gather, given design needs and the
limited time and capability (limited storage space or limited tool resource). How can we
softly push a player to perform accurate calculations when gathering materials in the game?
Transformation: If the geometry forms or elements exist somewhere, simply gathering
them is a lucky act. The problem is that if one needs to use a complicated component, he or
she will need to comprehend and use the complex geometry or geometry transformation.
How can one decompose and recompose diverse geometry components?

The infield observation, however, indicated that most players treated the collection
action literally as the “treasure hunting” gameplay. They focused on searching for
objects dispersed in the game world, with few purposefully identifying, checking,
or measuring the dimensions and other structural properties of the objects collected.
They did not fully process the goal statement or attend to the cues on the structural
or numerical properties of game objects, demonstrating the intention to treat this
information as trivial to their “at-hand” game objective—gathering all objects to be
gathered.
These observations suggested (a) a misalignment between the action-specific
learning and game objectives in that the instant or by-default solution to the action-­
specific challenge is learning-irrelevant (e.g., navigating the game world to obtain
objects), (b) participants’ lack of motivation to tackle a game action and challenge
via a mathematical strategy, and (c) participants’ lack of a planning perspective to
anticipate the relation between individual pieces of information or individual moves
to an overall plan, and hence falling short of identifying and analyzing the mathe-
matical properties of individual objects (e.g., a shipping container or a family) to
88 4 Design of Gameplay for Learning

fulfill future game actions and objectives (e.g., allocating the limited space afforded
by a certain number of containers to a set of families). The consequential design
insight is that our target middle-school students typically lack attitudes, behaviors,
or skills expected to perform purposeful, cognitively effortful, and mathematical
gameplay. Motivating participants’ purposiveness and proactivity in cognitive and
content engagement is important for the game action design.
To foster the cognitive and content engagement associated with the collection
action, we tried prioritizing the investigation of the numerical and structural essen-
tials of each item to be collected, rather than item-hunting in the virtual world, as the
primary action-specific challenge. For example, the player had to measure and com-
pute the living area of a shipping container and the living space needs of a family
before the container and the family could be gathered. We tried adding the “produc-
ing” type of user input that requires the actual deed of exploiting the resources of
mathematical knowledge and skills, such as prompting the player to accurately com-
pute and inscribe the numerical values of the numerical and structural properties of
the collected items. We also experimented with the strategies of externalizing,
chunking, and presenting the sub-goals of the multistep game task, in an attempt to
cue the player on the integration of the at-hand moves and individual information
pieces to the future game actions and objectives. These design strategies have evi-
dently improved the players’ time spent on and the frequency of enactments of
mathematical information processing and reasoning processes, as evidenced by both
the game-logged task performance and gaming behavior analysis (Ke et al., 2017).
Rewarding domain-specific gameplay (i.e., mathematical planning, calculation,
and reasoning) as the most effective gameplay strategy for players is critical for
motivating game-based mathematical practice. As discussed in the literature of
game-based learning (Hamlen, 2012) and observed in our own design experiments,
the goals and values of individual uses are not necessarily allied with the designers’
expectations or conjectures. A hypothesis of learning games is that gameplay should
be both result- and process-driven—achieving game objectives and the processes of
problem-solving to achieve the objectives are both important and enjoyable. Yet
students frequently prioritize the result state (e.g., “passing this level”) over differ-
ent challenges, sometimes to an extent that they would try to bypass difficult tasks
instead of working through them, adopt certain “cheating” strategies, or give up
when a task was difficult. Fostering a positive disposition and the grit to tackle a
game challenge via a mathematical strategy is therefore a critical goal and condition
of game-based learning. In the following subsection, we describe three salient
themes that emerged from our design research and governing game action-specific
constraints and rewards that helped to motivate E-Rebuild participants’ attitudes
and behavior change in terms of mathematical thinking and practices.
From fluky psychology to mathematical reasoning During interviewing, we found
that participants’ pre-study gaming experiences appeared to foster a fluky psychol-
ogy or belief in randomness during game-based problem-solving. An exemplary
participant quote is, “This (E-Rebuild) is different from Minecraft. (In) Minecraft
4.3 Learning-Play Integration in E-Rebuild 89

you do whatever you want. This is picky.” Fluky psychology was observed in the
following gaming behaviors:

(In a “trading” game action) Bob1 repeated the same guessing and trial-and-error strategies
as before. He read the task description in less than 5 seconds and did not attend to the item
price bulletin inside the store. He entered random numbers (for the bid of a construction
item) into the ordering area, executing a trial-and-error strategy. He tried five times and
realized that it was not an easy guess after repetitive failures, before finally returning to the
virtual store to read (process) the price bulletin.
Bryan continued the similar trial-and-error play with the trading action. His neighbor com-
mented, “This is Bryan! You don’t care.” Bryan laughed and started to refine his play strat-
egy. In the ordering area, he placed 10 as the number of bricks to be ordered, thought for a
minute, and then put 45 as the total cost (which is an accurate answer, reflecting his com-
prehension and application of the concepts of unit price and percentage embedded in price
bulletin—“12% discount for ordering 6+ bricks”). He successfully obtained 10 bricks via
mathematical reasoning. Yet when he proceeded to purchase the next construction item, he
resumed guessing plus trial and error, by entering an estimated initial value and then
increasing it by 5 post each failed trial. He spent around 3 minutes guessing, got stuck,
appeared frustrated, and still showed no intention to review the price bulletin.

In the above examples, guessing coupled with trial and error was the primary play
strategy. The participants appeared to dodge mathematical reasoning and calcula-
tion during the trading action, until being prompted or when they realized that math-
ematical practices were indispensable or more efficient for problem-solving.
In some other cases, participants were engaged in logical reasoning without
applying mathematical knowledge or accurate calculation, as the following exam-
ples demonstrate.
Andy placed one of the smallest and one of the largest families into the first container, while
inspecting the family inventory chart during the allocation action. For the second and third
containers, he continued the same strategy – matching the smallest family with the largest
one in the to-be-assigned list and placed the pair to each container. He commented, “I think
I get it. So I still have four more families and three more containers.” He finished allocating
each of the remained largest family to each of the remained two containers to pass the level,
“Oh, Yay! Oh, Yay! I beat it! Oh, Yay!” He was very excited, waving and clapping the hands.
George started by assigning the biggest family to each container. He did not process the
statements on the unit and ratio governing family members’ living space needs in the task
description. Neither did he perform any mathematical calculations. He tried to place two
second-largest families into another container, failed to do so, and then replaced one with a
smaller family. For the next container, he tried adding the second-largest family, and then
the smallest family. He repeated the same strategy for the fifth container. He went to the last
container and remarked, “I think I am going to do it this time!” He managed to add the last
family to the last container, and got a Congratulations screen, “YES!” His neighbor asked,
“How did you do it? How many (containers) did you put?” He answered, “Put the big one
(container) on the bottom, and the fairly big one on the top.” He did not realize that the two
containers he referred to actually have the same living areas.

As shown above, participants were involved in generic logical reasoning about

the space allocation, such as evaluation, information mining, planning, testing,
and solution refining. Yet these reasoning processes did not involve ratio- and
area-­related mathematical calculations.

1
All participants’ names cited in this paper are pseudonyms.
90 4 Design of Gameplay for Learning

Notably, there was a pattern of failure-driven learning among participants:

recurrent failures resulting from fluky psychology, guessing, or generic reasoning
would trigger the acts of mathematical knowledge application and calculation. The
game behavioral analysis data indicated that the percentage of explicit content
engagement in the participants’ gaming actions usually doubled from the earlier
gaming sessions to the later ones.
Prioritizing problem-solving efficiency There was an observed association between
the perceived value of mathematical solutions and participants’ engagement in
endorsing those solutions for game problems. In other words, when the evaluation
of game performance prioritized efficiency, it motivated the players’ exploration of
mathematical solutions. Such a pattern is illustrated by the following example.

Rather than estimating or calculating the amount of bricks and other building materials
beforehand, Sandy was ordering those items as needed during the construction process. She
would order four bricks initially, stacked them as a row along the foundation, and then went
back to order another four bricks.

What Sandy did was repeated by multiple other participants. Game conversations
among participants indicated two main reasons underlying the act of repetitively
ordering a small set of items: (1) avoiding the act of calculating the amount of
needed items based on the unit size of an item and the targeted living area of the
structure to be built and (2) avoiding the processing and application of percentages
(e.g., 6% discount for ordering 6+ bricks each time) for the cost calculation. These
participants apparently deemed preplanning for building (that frames mathematical
thinking and calculation) as effortful and unnecessary when tackling a game
challenge.
The dodging of mathematical planning and reasoning, prominently, was curbed
by the introduction of action-specific rules that prioritize problem-solving effi-
ciency—the “time limit” and “material credit” as the reward or evaluation criteria of
a building action.
Sandy managed to complete the four walls surrounding the foundation along with the door
and two windows. She went back to the virtual store to order more bricks to build the roof,
but got a pop-up window notifying that she had run out of the material credit. The second
time when she tried the task, she started to order 10 bricks each time and computed the cost
with the discount rate included during the trading action. She was almost done with build-
ing, yet failed the level again for exceeding the time limit. The third time she tried the task,
she read the price bulletin board carefully, used a paper sheet to preplan and calculate the
number of bricks used for each side of the wall around the foundation, pre-ordered the
doors and two windows, and entered an accurate value for the cost of a batch of bricks in
the ordering area. This time Sandy managed to pass the game level. And while she
­complained to her neighbor how challenging the task was, she instantly went on to play the
next game level/task.

As Sandy’s case shows, only when mathematical planning, calculation, and reason-
ing are warranted for gameplay and perceived as high-paying strategies to tackle
game challenges would they be implemented. Game rules that prioritize efficiency
and preplanning in the action-specific evaluation would frame the associated
4.3 Learning-Play Integration in E-Rebuild 91

mathematical solutions as the most anticipated strategy and hence help to foster
participants’ situational interest, commitment, and potentially a positive disposition
toward mathematical problem-solving.
Reward versus punishment with trade-off between ruling and inclusiveness When
experimenting with game action-specific rules, we found that setting a time limit
with a standard threshold is challenging, because individual learners differ in both
play fluency (e.g., of maneuvering 3D objects or comprehending the gameplay
rules) and math problem-solving competency. A standard time limit tended to pun-
ish learners who do not have much gaming experience or are not confident about
game-based math problem-solving. Frequently, participants complained that they
were just “one brick away from the success” when they were timed out. Although
failure-driven learning appeared to promote cognitive and content engagement in
gameplay as discussed above, iterative failures due to the time limit appeared to
discourage participants who needed scaffolding or a lenient environment for alter-
ative types of game-based learning. Some participants, for example, were observed
working on the game problems offline using paper sheets, exchanging strategies
with peers, or asking for help from a facilitator or teacher amidst gameplay. These
actions, though potentially extending the on-task time, are part of game-based learn-
ing and valued cognitive or affective support for the participants.
A similar issue is the level of “forgiveness” toward the minor defect in the player’s
action-specific performance. When evaluating the players’ design solution against the
preset evaluation parameters (e.g., thrifty in material cost and time, reproducing the
shape, size, position, and functionality of the target structure and the resource usage
rate), setting the necessary degree of accuracy relates to the gauge of an optimal chal-
lenge for learners. For example, in a shelter-building task, the game evaluates whether
a built structure is adequately secured against a sandstorm, by testing whether any
single particle would blow through the walls/roof to touch the foundation. The size of
the particle can be up- or downscaled to test the strength of the structure. We found
that a low level of “forgiveness” tended to punish the participants with a lower level
of fluency in maneuvering 3D objects (e.g., bricks) but comparable in the action-
related math competency than others. In other words, these participants could pro-
duce the output that is accurate in its mathematical properties but defective in its
design features (e.g., having a half-inch seam in between two bricks on the roof) due
to an inept performance of brick stacking. Executing a standard level of accuracy
across all design evaluation parameters would hence create a conflict between content
and technical competencies for individual learners. Moreover, we also found in cer-
tain cases, a high-fidelity evaluation of certain design parameters (e.g., the sturdiness
of the structure or whether it will collapse given a seismic wave) would involve the
evaluation and hence the exploitation of the resources of physics knowledge (e.g.,
force, friction, elasticity, mass, and joint) that is beyond the project scope and the
expected knowledge base of the middle-­school users.
Consequently, we have refined the rules on time and material cost as rewarding
rules rather than the pass-or-fail punishment. Specifically, badges of different criterion
categories (i.e., time and material credit) are awarded to players who outperformed the
92 4 Design of Gameplay for Learning

threshold value in each category. A relatively lenient, minimum requirement of design

evaluation is set to encourage learner agency in the initial game levels, with the level
of accuracy gradually scaled up in more advanced game levels when the players are
expected to develop an appropriate level of fluency with gameplay in addition to
game-based math problem-solving. The design parameters that leverage an out-of-
scope knowledge base were detached from the ruling base for game action evaluation.
The infield testing of the refined game action rules on participants’ game-based prob-
lem-solving and learning has supported our design decisions (Ke et al., 2017).

4.3.3 Designing Intermediary User Interface

Learning in E-Rebuild is to proactively interact with, interpret, and coordinate mul-

timodal elements of a building-oriented mathematical phenomenon or problem, by
selecting, studying, and maneuvering these elements to map out their relationship
and finally configure the problem solution. The interface that enables players’
active and meaningful interactions with the problem elements (or game objects) is
hence a main facet of our design exploration and has been iteratively developed,
tested, and refined.
An important user interface in E-Rebuild is the controls for viewing and navigat-
ing the 3D game world. To enable multiple perspectives and alternative modes of
inspecting game objects (or game-based math problem elements) and hence foster
representational flexibility in game-based math learning, we had to leverage both
keyboard and the mouse for the control of navigation: using arrow or WASD keys
for moving, right click and drag to turn/rotate the direction/angle of viewing and
moving, and the scroll (mouse) wheel to zoom in and out. These navigation con-
trols, we found, were novel to users not familiar with 3D gaming or virtual world
environments. Middle-school players used to touch pads faced a learning curve in
controlling the keyboard and mouse simultaneously, though they all managed to
master and use these controls naturally after playing through five to six orientation
levels. The comments like “change your (viewing) angle,” “turn the position of your
camera,” or “move up and to the other side” were frequently heard during the play-
ers’ game talk, when they tried to mentor each other on how to inspect or enter an
object or structure in the game. These comments reflect their effort and cognizance
in interpreting, tracking, and positioning objects or ­benchmarks in this 3D virtual
world. Those user controls, though not very intuitive, activated a conscious practic-
ing and reasoning process of the participants in inscribing spatial relationships
among 3D structures or objects while shifting the perspective and field of view. That
is, they may lack technical efficiency but present pedagogical or learning affordance
for the learning game.
Such a pattern governing the integration and trade-off between technical and
pedagogical usability (i.e., being intuitive versus cognizant) is also observed for the
design of user controls for the building actions, such as moving, rotating, stacking,
joining, and copying/pasting building items. Different from other 3D sandbox
4.3 Learning-Play Integration in E-Rebuild 93

games, object maneuvering and building in E-Rebuild must follow the basic laws of
physics to reflect the “rebuild” theme; thus the objects will collide or rebound with
other objects. Moving, rotating, and stacking physical objects in the xyz coordinate
axis system to constitute a structure are indeed challenging for middle-school stu-
dents. Participants in the initial gaming sessions were frequently found either strug-
gling in selecting the appropriate coordinate axis for moving or rotating an object,
stressed out when having to frequently shift the viewing angle to move and position
the object accurately to the target spot in a 3D space, or frustrated when trying to
stack the object over others yet accidently colliding objects. But there was an obvi-
ous increasing trend in participants’ fluency and accuracy levels in maneuvering the
3D objects during gameplay. The gaming behavior analysis indicated that the aver-
age error rate (the percentage of failed or unintended trials in all observed rotation
acts) of the user control with object rotation was 0.40, and the error rates with other
building actions (e.g., moving and stacking) were 0.15 to 0.16. The high error rate
associated with object rotation control, on one hand, could be due to the low techni-
cal usability of the interface—it needs time for the player to execute the keyboard
shortcuts (“Tab” for shifting among the three axis coordinates, “,” and “.” for clock-
wise or anticlockwise rotation). On the other hand, it showed that the middle-­school
students were in need of training in the 3D rotation task for the spatial skills devel-
opment. We found that the intermediary interface actually had necessitated a con-
scious performance of preplanning and extrapolating the specific axis and direction
to rotate or move an object before actually trying them out, since instinctive guess-
ing would involve significantly more clicks and hence more effort for the player.
The infield studies have demonstrated that E-Rebuild participants, in comparison
with the control group students, demonstrated improved mental rotation task perfor-
mance (Ke et al., 2017).
To improve the technical and pedagogical usability of the intermediary interface
of building, we added a joining and a copying/pasting function. These two functions
enable the players to join neighboring items into a panel or a structural subsystem,
copy/paste it to create multiple panels or subsystems, and then compose them into a
complicated structure. This “panel” approach, in comparison with the “brick-by-­
brick” approach, is not only more efficient but also fosters the performance of
design planning before building. For example, when building a multiroom adobe
house or a stadium bench, the players would find it difficult to adopt a “brick-by-­
brick” approach. When they resorted to the “panel” approach, they got to analyze
the structural subsystems of the house or the bench (e.g., foundation, walls, and
roof; support structure, seating, and railings), calculate the configuration of the sub-
systems (e.g., number of bricks, size, and shape of each subsystem), and gauge the
number of bricks needed in the inventory to copy/paste a panel or substructure. In
other words, the players have been engaged in composing and decomposing three-­
dimensional shapes and building-related mathematical calculation, when using
joining and copying functions for building. It should also be noted that both func-
tions involve only left clicks (to select the neighboring items to be joined) that are
assisted with visual cues (with the selected items highlighted) and are found high in
technical usability (with a low 0.07 error rate average).
94 4 Design of Gameplay for Learning

In certain cases, we had to prioritize the learning affordance of a user interface to

its technical usability. In the allocation action, for example, we found that the left
click, button control (e.g., “+” and “-” buttons) for adding or removing an item to an
enclosed space (e.g., allocating families to the shelter or fishes to a pond) was signifi-
cantly and positively correlated with the guessing and trial-and-error gaming strate-
gies. Instead, the mechanism of drag and drop the items by unit from a personal item
inventory to the target space tended to reduce the behaviors of wild guessing. It is
because the drag-and-drop act is more dynamically linked with the visualized conse-
quence (e.g., a four-member family being dragged/dropped to physically occupy part
of the space), thus promoting participants’ awareness of the relationship between
occupants and the space to better comprehend the math problem to be solved.

4.4 Game World Design

In our design and research of E-Rebuild, we have focused on particularizing a back-

drop mission via the spatial narrative, or environmental storytelling, of the game
world in which the narrative element is infused into a space that a player explores
and interacts with. The game world presents multimodal, contextualized representa-
tions of a mathematical problem via a series of background scenarios or landscapes,
interactive game objects and structures, game characters, and landmarks. It helps to
(a) legitimize and motivate architectural design-based mathematical problem-solv-
ing activities and (b) embody a multimodal problem space that portrays dynamic
beginning, intermediate, and goal states in consequence of the player’s problem
exploration and solving actions.

4.4.1  ame World as the Scenario for Contextualized

G
Problem-Solving

The overarching mission in E-Rebuild is to design and rebuild a disaster-damaged

space to fulfill the design parameters and varied residential needs specified by vari-
ous architectural scenarios. These architectural scenarios were presented via diverse
landscapes (or environment settings) along with associated “rebuild” themes, such
as a post-quake island with a shelter to be built, a desert with adobe houses threat-
ened by the sandstorm, an urban school site with portable classrooms and a stadium
to be restored, or a rural farm with livestock shelters to be constructed against preda-
tors. These environment settings, with the focal points (or landmarks) and other
objects (e.g., building materials, tools, and other structures and props) on the set,
were designed and developed by the architect and game designer collaboratively.
The architect in the team first designed the layout and visual content of a series of
environmental settings using Google SketchUp (as Figs. 4.3, 4.4, and 4.5 exempli-
fied). The game designers then simulated and developed these settings in Unity 3D.
4.4 Game World Design 95

Fig. 4.3 Desert and historical rural pueblo

Fig. 4.4 Rural farm

When developing environmental settings, we tried to reproduce a similar level of

fidelity and visual aesthetic appeal to the architect-designed sets and landscapes. Yet
when we infield tested the prototype with the students at the local school district, the
high-fidelity game worlds loaded very slowly using the school network and con-
flicted with the older version Chromebooks used prevalently in the school class-
rooms. To better design the game as a learning tool compatible with the existing IT
96 4 Design of Gameplay for Learning

Fig. 4.5 Urban school

infrastructure of the school setting, we scaled down the level of fidelity as well as the
scope of the map and adjusted the visual presentation as a conversion between the
original design and the typical sandbox game world (e.g., Minecraft). We found that
a lower level of fidelity in the game world design did not reduce the engagement
level of the participants. Actually, the downscaled map and game space helped focus
the players’ attention on the at-hand task rather than on exploring the virtual world.
For example, a heat map analysis with the players’ navigation behaviors in the origi-
nal game world demonstrated that the most frequently visited or explored parts of
the set were not the construction site or the focus point but boundary spots (e.g.,
underwater ditches or mountains) that were irrelevant to the problem or l­earning
task. Such an off-task game space exploration behaviors dropped by 50% in the
simplified game world version.
In addition to the implicit storytelling—background scenario presentation—via
the game space, we used a task panel to present a description of the game environ-
ment story associated with a game episode (comprising multiple levels with the
same set/landscape) and game level-specific objectives. We tried to make the task
description concise, dialogic, and chunked because the infield testing indicated that
middle-school players tended to spend minimal time reading texts on screen during
gameplay.

4.4.2 Game World as the Problem Space

A theoretical conjecture for E-Rebuild design is that meaningful interactions with

multiple forms of external representations will convert an individual’s mental repre-
sentation for mathematical problem-solving to the sense-making process in which
References 97

relevant cognitive structures are experienced richly, multi-encoded, and integrated

flexibly. Given this conjecture, the game world will act as the externalized problem
space that animatedly portrays the beginning, intermediate, and outcome states of
problem-solving, via interactive game objects and dynamic visualizations (e.g.,
visual cueing, feedback, and altered states of objects) in response to players’ inter-
actions with these objects.
In E-Rebuild, each math problem is represented in multimodal forms in the game
space. These multimodal external representations include (1) a verbal/syntactic task
description, including mathematical and design vocabulary and syntax; (2) pictorial
presentations, including interactive 3D game objects (e.g., varied types of construc-
tion items and the target structure), 2D diagrams (e.g., a floor plan), and a spatial
configuration of the landscape designating the location, size, and direction of the
structure to be built; (3) formal mathematical notations embedded in the task
description, inscribed onto game objects, or presented via cursor-on-target cues;
and (4) concrete stimuli, such as a ruler that measures distance and angles.
Given these multiple external representations in the game world, students must
actively investigate, transform, and integrate them into a coherent, internal problem
representation to arrive at the solution. In E-Rebuild, core game actions, such as site
survey, item collection, trading, building, and allocation, are designed to necessitate
active encoding of the external mathematical representations. The gaming objects in
the game space, in addition to the game actions and rules, collectively frame and
foster game-based mathematical problem representation and solving processes,
ranging from processing and selecting information to selecting, testing, and refining
mathematics-related problem-solving strategies. In summary, students’ internal
mathematical representations can be altered through each interaction with the game
world and interactive game objects and linked automatically and continuously to
one another to facilitate mental model development.

References

Ang, C. S. (2006). Rules, gameplay, and narratives in video games. Simulation & Gaming, 37(3),
306–325.
Garris, R., Ahlers, R., & Driskell, J. E. (2002). Games, motivation, and learning: A research and
practice model. Simulation & Gaming, 33(4), 441–467.
Frasca, Gonzalo (1999). Ludology meets narratology: similitude and differences between (video)
games and narrative. Retrieved from http://www.ludology.org/articles/ludology.htm
Ge, X., & Ifenthaler, D. (2017). Designing engaging educational games and assessing engagement
in game-based learning. In R. Zheng & M. K. Gardner (Eds.), Handbook of research on serious
games for educational applications (pp. 255–272). Hershey, PA: IGI Global.
Habgood, M. P. J., & Ainsworth, S. E. (2011). Motivating children to learn effectively: Exploring
the value of intrinsic integration in educational games. The Journal of the Learning Sciences,
20(2), 169–206.
Hamlen, K. R. (2012). Academic dishonesty and video game play: Is new media use changing
conceptions of cheating? Computers & Education, 59(4), 1145–1152.
Järvinen, A. (2008). Games without Frontiers: Theories and Methods for Game Studies and
Design. Tampere, Finland: Tampere University Press.
98 4 Design of Gameplay for Learning

Jenkins, H. (2002). Game design as narrative architecture. In P. Harrington & N. Frup-Waldrop

(Eds.), First person. Cambridge, MA: MIT Press Retrieved from: http://www.anabiosispress.
org/VM606/1stPerson_hjenkins.pdf
Kafai, Y. B. (1995). Minds in play: Computer game design as a context for children's learning.
New York: Routledge.
Ke, F. (2016). Designing and integrating purposeful learning in game play: A systematic review.
Educational Technology Research and Development, 64(2), 219–244.
Ke, F., Xie, K., & Xie, Y. (2016). Game-based learning engagement: A theory-and data-driven
exploration. British Journal of Educational Technology, 47(6), 1183–1201.
Ke, F., Xu, X., Lee, S., Moon, J., Dai, Z., Pan, Y., Shute, V., Clark, K., & Erlebacher, G. (pre-
sented 2017, April). Math learning through game-based architectural design and building.
Presentation at 2017 American Educational Research Association Annual Meeting, American
Educational Research Association, San Antonio, TX.
Klopfer, E., Osterweil, S., & Salen, K. (2009). Moving learning games forward: Obstacles, oppor-
tunities, & openness. Boston: The Education Arcade Retrieved from http://education.mit.edu/
papers/MovingLearningGamesForward_EdArcade.pdf
Lindley, C. A., & Sennersten, C. C. (2008). Game play schemas: From player analysis to adaptive
game mechanics, 216784. International Journal of Computer Games Technology Retrieved
from: https://doi.org/10.1155/2008/216784
Malone, T. W., & Lepper, M. R. (1987). Making learning fun: A taxonomy of intrinsic motivations
for learning. Aptitude, learning, and instruction, 3(1987), 223–253.
Miller, C. S., Lehman, J. F., & Koedinger, K. R. (1999). Goals and learning in microworlds.
Cognitive Science, 23(3), 305–336.
Richards, J., Stebbins, L., & Moellering, K. (2013). In J. Richards, L. Stebbins, & K. Moellering
(Eds.), Games for a digital age: K-12 market map and investment analysis. New York: The
Joan Ganz Cooney Center at Sesame Workshop.
Salen, K., & Zimmerman, E. (2004). Rules of play: Game design fundamentals. Cambridge, MA:
MIT press.
Schell, J. (2014). The art of game design: A book of lenses. Boca Raton, FL: CRC Press.
Schoenfeld, A. H. (1985). Metacognitive and epistemological issues in mathematical understand-
ing. Teaching and learning mathematical problem solving: Multiple research perspectives,
89(4), 361–380.
Sicart, M. (2008). Defining game mechanics. Game Studies, 8(2), 1–14.
Stacey, K. (2006). What is mathematical thinking and why is it important. Progress report of the
APEC project: collaborative studies on innovations for teaching and learning mathematics in
different cultures (II)—Lesson study focusing on mathematical thinking.
Suits, B. H. (1978). The grasshopper: Games, life and utopia. Toronto, ON/Buffalo, NY: University
of Tronto Press.
Taylor, A. (2009). Linking architecture and education: Sustainable design for learning environ-
ments. Albuquerque, NM: UNM Press.
Torbeyns, J., Lehtinen, E., & Elen, J. (2015). Describing and studying domain-specific serious
games. Cham, Switzerland: Springer.
Tricot, A., & Sweller, J. (2014). Domain-specific knowledge and why teaching generic skills does
not work. Educational Psychology Review, 26(2), 265–283.
Chapter 5
Interweaving Task Design and In-Game
Measurement

Abstract There are two important design issues related to game-based learning
(GBL) in school settings: (a) the intrinsic integration of content-related tasks in
gameplay and (b) the real-time capture and analysis of in-game performance data.
In this chapter, we describe an integrative design approach that is aimed to inter-
weave game-based task design with in-game assessment of learning. Extending
other GBL projects in which the mechanism of data mining for assessment was
created after game development, in E-Rebuild we have designed the evidence-­
centered, data-driven assessment during the course of game design. Design-based
research findings on emergent core design processes and functional conjectures on
the approaches of task generation and evidence accumulation are discussed, with
support of infield observations on the implementation feasibility and outcomes of
various design assumptions.

5.1 Introduction

In E-Rebuild, we examined two important design issues related to game-based

learning (GBL) in school settings: (a) the intrinsic integration of content-related
tasks in gameplay and (b) the real-time capture and analysis of in-game perfor-
mance data. In spite of the plethora of research on GBL, the design descriptions or
operational conjectures on how to develop and integrate domain-specific learning
tasks in gameplay are scarce (Habgood & Ainsworth, 2011; Ke & Shute, 2015). In
addition, learning in games has generally been assessed via external tests in a post
hoc manner (Shute & Ke, 2012). Adopting an educational data mining approach,
real-time assessment of learning based on the dynamic performance of players
could better capture process- and performance-oriented evidence on competency
development (Shute, Ke, & Wang, 2017). This has the potential to drive the design
of personalized or adaptive game-based learning while not being intrusive to dis-
tract players’ gameplay or state of flow.

© Springer Nature Switzerland AG 2019 99

F. Ke et al., Interdisciplinary Design of Game-based Learning Platforms,
Advances in Game-Based Learning, https://doi.org/10.1007/978-3-030-04339-1_5
100 5 Interweaving Task Design and In-Game Measurement

In this chapter, we review an integrative design approach that intends to

interweave game-based task design with in-game assessment of learning. Different
from other GBL projects in which the mechanism of data mining for assessment
was created after game development, in E-Rebuild we designed the evidence-cen-
tered, data-driven assessment during the course of game design. Design-based
research findings on emergent core design processes and functional conjectures on
the approaches of task generation and evidence accumulation are discussed, with
support of infield observations on the implementation feasibility and outcomes of
various design assumptions.

5.2 Conceptual and Design Perspectives

5.2.1 Content Modeling for Game Task Design

Games in general can be defined as organized play that is structured by a set of

rules and an obstacle-tackling goal (Klopfer, Osterweil, & Salen, 2009; Schell,
2014). A long-held concern about using computer games for learning is that stu-
dents may be distracted by the play part, thus not achieving the learning goals
(Miller, Lehman, & Koedinger, 1999). It is argued that the extent to which the
content is intrinsic to the game tasks and associated actions, rules, and goals will
influence the game’s learning effectiveness (Richards, Stebbins, & Moellering,
2013). Nevertheless, empirical and theoretical research examining the design of an
intrinsic integration between learning and game tasks is still limited and inconclu-
sive (Habgood & Ainsworth, 2011).
A recent review on the design of intrinsic integration of domain-specific GBL
(Ke, 2016) argued that it is critical for learning game designers to clarify what, how,
where, and when learning and content will be embedded and activated by gameplay
actions, rules, and game world or narrative design. Successful integration of learn-
ing in games is regarded as a process rooted in an exploration of content or learning
elements that are fundamentally engaging within an academic discipline and repre-
sent meaningful interactions between the player and the epistemic frames—salient
activities, understanding, and value—of the target subject matter (Klopfer et al.,
2009; Shaffer, 2006). These engaging learning elements should provide players
with legitimacy in content engagement (Barab, Gresalfi, & Ingram-­Goble, 2010).
Therefore, meaningful game task design should be driven by content modeling, a
process of defining the ontology of the targeted subject domain with a mapping-
building-block notion. During the content modeling process, componential con-
cepts and skills (or competencies) along with their associated resources, properties,
and examples (e.g., conceptual representations or skill demonstrations), as well as
the relationships between these componential competencies, are specified and pre-
sented via cognitive mapping.
5.2 Conceptual and Design Perspectives 101

5.2.2 Integrative Design of Task and Assessment

From the perspectives of construct-centered assessment (Messick, 1994) and

evidence-­centered design for educational testing (Mislevy & Haertel, 2006; Almond,
Mislevy, Steinberg, Yan, & Williamson, 2015), task construction operationalizes the
underlying cognitive theory and the proficiency model of the targeted competencies
and hence is a core part of the assessment design. Task construction follows the
operational definition of targeted competencies and the analysis of behaviors that
reveal the competencies. Thus the selection or construction of relevant tasks that are
intended to elicit those behaviors is based on evidence, and interacting with such
tasks serves as both the source and evidence of knowledge and skill development.
The literature of learning and assessment highlights the following two salient
claims for the integrative design of specific tasks and the overall assessment: (a) the
core of authentic tasks is to simulate the representative challenges and criteria of the
real-world application of target competencies (Wiggins, 1990) and (b) the perfor-
mance of complex skills (e.g., those of mathematics and science) is moderated by
contexts, and hence it is necessary to construct cross-contextual tasks with varied
difficulty and implementation contexts to effectively measure skill development
(Resnick & Resnick, 1992).

5.2.3 I n-Game Integration of Active Learning and Authentic

Assessment

Games and simulations can be instrumental in the aforementioned task and assess-
ment design processes because they can (a) present a broad range of complex sce-
narios and tasks to model the functional context of the real world, (b) enable
interactions with authentic tasks and contexts, and (c) achieve a comprehensive,
cross-contextual assessment of the complex skill. For example, the conventional
approach to designing math word problems typically “attempts to strip the problem
context of all irrelevancies, retaining only the task information needed to engage the
focal knowledge and skills involved in task processing” (Messick, 1994, p. 18). Yet
task-based skill practice and assessment are subject to the moderation of context
variables. Game-based challenges, presenting controllable and customizable crite-
rion situations to simulate varied authentic task contexts, can afford the assessment
of math development that ranges between situated knowledge and abstract, decon-
textualized understanding (Pratt & Noss, 2002).

5.2.3.1 Data-Driven, Game-Based Assessment

Process-oriented data mining and learning analytics methods, such as Bayesian net-
works, social networks or structural analysis, visual or graphical analysis of event
paths, and sequential analysis of time series, can capture the complex and open-­ended
102 5 Interweaving Task Design and In-Game Measurement

learning trajectories in a game setting. Prior research has suggested that educational
data mining and learning analytics can and should be used together to exploit game-
based performance data to inform on students’ on-task or off-task behaviors, compe-
tency development related to the targeted subject matter, and hence the effectiveness
and design of learning games. Four recent projects (Dede, 2012; Levy, 2014; Shaffer
et al., 2009; & Shute & Ventura, 2013) have exemplified the potential and applicabil-
ity of game-based assessment via educational data mining. These projects all adopted
a data-intensive, evidence-based approach by collecting, measuring, analyzing, and
reporting dynamic data about learner performance and contexts (e.g., online log data)
to understand and optimize learning and the environments in which it occurs. Multiple
methods of data analyses (e.g., quantitative psychometric modeling, network or
structural analysis, and path analysis) and visualization (e.g., algorithms, models,
network graphs, or spatial and chronical maps) were used. On the other hand, in some
previous projects, data-driven assessment design tends to be a post hoc justification
or evaluation for the game development.

5.3 Design-Based Inquiry

Adopting the design-based research approach (Sandoval & Bell, 2004), we explored,
iteratively tested, and refined the core elements and design features of the game task
and assessment of learning through multiple mixed-method case studies. The suc-
cessive iterations and testing played a role similar to that of systematic variation in
a controlled experiment (Cobb, Confrey, Lehrer, & Schauble, 2003). Specifically,
the design-based investigations focused on extracting design heuristics that would
enable an integrative design of task and in-game learning assessment, by addressing
the following questions:
• What are the core processes that define the integrative design of game tasks and
assessment of learning?
• What are the functional conjectures on the design strategies and features of the
GBL tasks and assessment derived from the iterative experimental findings?
Extensive, longitudinal data sets were collected during the course of our design
experiments. These data sets involved qualitative and quantitative resources, includ-
ing participatory observations of the project team’s design meetings, screen- and
video-recording of participants’ gameplay actions/reactions, game activity logs,
participants’ gameplay think-aloud transcriptions, interview data, and results from
math knowledge tests. One hundred and twenty middle-school students participated
in the E-Rebuild learning program and test-played the prototypes of the GBL assess-
ment system that was iteratively and systematically refined based on the successive
design-based research findings.
We conducted retrospective analyses (Cobb et al., 2003) with the longitudinal,
design experiment data to generate situated accounts of design conjectures (Sandoval
& Bell, 2004) and to specify the functional contexts of those conjectures. Specifically,
we conducted thematic analysis using the design meeting records, coded the
5.4 Core Design Sectors for In-Game Learning and Assessment 103

gameplay recordings to delineate the types and properties of major gameplay

actions and reactions via behavior analysis (Ke et al., 2017), and then extracted the
descriptive data (e.g., attempts made, levels passed, and time spent) from students’
gameplay activity logs. Students’ gameplay activity logs, as well as their pre- and
post-gaming math knowledge test results, were also used to calibrate and refine the
game-based learning evidence and assessment models to validate the game-based
assessment mechanism.

5.4  ore Design Sectors for In-Game Learning

C
and Assessment

The process of interweaving game-based task and assessment development in

E-Rebuild includes five core design sectors: (a) designing templates of tasks based
on the selected gaming (i.e., architectural design) actions that necessitate, as well as
the scenarios (i.e., overarching design missions) that frame, the practice of the tar-
geted competencies; (b) setting the gameplay rule and reward system that scaffolds
mathematical practices; (c) developing and organizing tasks based on the compe-
tency, evidence, and task models; (d) drafting the game logs, including the specifi-
cation of in-game performance observable variables, along with the assessment
statistical model (e.g., a Bayesian inference network); and (e) testing, refining, and
validating the feasibility and effectiveness of the tasks and assessment models. The
ordering of these core design sectors refers to their operational sequence in our
game design process. However, they frequently interact with each other, thus yield-
ing a holistic system rather than a collection of segmented processes.

5.4.1  esigning Task Templates: Action-Oriented, Scenario-­

D
Based, and Math Practice-Necessitating
5.4.1.1  ction-Oriented, Contextualized, and Multimodal Mathematical
A
Problems

Our design-based investigation indicated that participants’ gaming actions (i.e., the
core part of gameplay and main behavioral unit to be tracked in gameplay perfor-
mance) underlie the nature of the math content (e.g., qualitative understanding,
numerical calculation, and expressions and equations) and hence the targeted GBL
actions (e.g., identification, procedure execution, and problem-solving). Via itera-
tive expert review and user testing of various architectural design actions (as
described in Chap. 4), we have settled on a set of architectural design-based actions:
(a) collecting and customizing construction items, (b) site surveying and building,
(c) material trading, and (d) space and resource allocation. After composing these
learning-constructive game actions and incorporating relevant backdrop design sce-
narios, we then developed an assortment of game task archetypes.
104 5 Interweaving Task Design and In-Game Measurement

In E-Rebuild, each task template was designed to depict an archetype of

contextualized math story problems. These problems aim to motivate learners to
investigate and coordinate alternative representations of the constituent math
parameters, variables, and relations to solve a realistic design problem. The archi-
tecture-themed scenarios served to contextualize the final construction of tasks
and the development of task narratives. For example, a task in E-Rebuild asks the
player to rebuild a multiroom structure by referring to a pre-earthquake house
model and using a minimum number of shipping containers. The player then must
assign different family configurations to the shelter, with the ratio of an adult’s
required living space to a child’s required living space being x to y (e.g., 2:1). A
multimodal representation of the problem is conveyed through (a) a verbal/syntac-
tic task narrative; (b) pictorial presentations, including 3D game objects (e.g.,
shipping containers of different sizes and the house model), 2D diagrams (e.g., the
floor plan and a material price board in the form of a table of equivalent ratios),
and a spatial configuration of the landscape designating the location, size, and
direction of the shelter to be built; (c) formal math notations, such as ratios embed-
ded in the task narrative, numerals inscribed onto game objects, and operations
presented via gameplay feedback and the help panel; and (d) concrete stimuli, such
as cutting and scaling tools and a ruler.
Given such a multimodal math story problem, the player needs to actively inves-
tigate and transform math representations embedded in the game world and then
select and integrate the information into a coherent problem representation to
arrive at a solution. In E-Rebuild, core game actions necessitate active encoding
and coordination of external math representations. For instance, a single “build-
ing” game action involves the acts of planning (e.g., site survey or measurement,
floor or landscape planning via object positioning), composing and decomposing
(prisms, such as joining and stacking cuboids), covering (the space, such as paint-
ing an area), surrounding (a structure, such as fencing a yard), and filling (the cav-
ity, such as filling a rectangular prism with unit cubes or spheres). These actions
will activate participants’ interactions with the verbal, graphical, and numerical
representations of a geometrical figure and facilitate a coordination among these
mental representations.

5.4.1.2 Core Elements and Properties of a Task Template

Task archetypes can be classified based on core gameplay actions, such as collect-
ing, building, trading, and allocating. Certain task types, such as building, could
also encompass sub-tasks of planning, composing and decomposing, covering, sur-
rounding, and filling. Each basic E-Rebuild task template integrates a list of key
structural elements. These structural elements include (a) the core/driving action
and its enabling act(s); (b) action-appropriate goal state of the problem and the
criteria conditions that define the satisfactory goal state (e.g., rebuilding a pre-
quake house model with a minimum number of items); (c) start state of the prob-
lem, including unsatisfied criteria condition(s) in comparison with the goal state;
5.4 Core Design Sectors for In-Game Learning and Assessment 105

and (d) obstacles to be resolved or tackled to achieve the satisfactory goal state.
Obstacles will consist of action-specific constraints (e.g., limited space for an allo-
cation action or limited materials for a building action), sub-goal hierarchy, objects
to be maneuvered (e.g., actable objects), and/or those to be explored (e.g., proper-
ties of the structure to be built and construction materials).
Prior research on problem structures (e.g., Kaller, Unterrainer, Rahm, &
Halsband, 2004) and mathematical problem-solving (e.g., Schoenfeld, 2014) has
shown that the difficulty of a task archetype can be predicted by the requisite math
knowledge and skills needed to solve the task, as well as the structural features of
the task, like the number of constraints—where more constraints mean that more
variables need to be considered during problem solution. Another task feature which
can influence its difficulty concerns the ambiguity of goal hierarchy (i.e., the obscu-
rity of goal priorities in the assembly of sub-goal states). Finally, the existence of
suboptimal, alternative solutions in a given task renders it more difficult to solve.
Overall, the prerequisite math knowledge in a given task archetype reflects the
required level of competency in the content domain, while the task’s structural fea-
tures affect math problem-solving performance, including problem interpretation,
problem-solving strategy selection (or planning), execution, and monitoring.

5.4.1.3 Composing Tasks for Game Level Development

E-Rebuild is a multi-episode game. Each game episode consists of multiple levels.

A game level can be developed as either an architectural design problem with mul-
tiple instantiations of task archetypes or as a sector of an overarching inquiry (i.e.,
episode) which carries only one instantiation of a task archetype. We have experi-
mented with these two options of task composition for game level development and
found that the former was challenging for novice student players who lacked skills
related to problem interpretation, sub-goal generation, and analysis. In addition, a
typical duration of a class session (and hence the in-class gaming session) was only
45 min, which made it difficult for learners to fully practice and complete a compre-
hensive inquiry-based game level. To bolster engagement in the game and learning,
we found it necessary to simplify the task composition by further decomposing a
task within a game level.
Our initial game level prototype (referred to as “row house”) illustrates our
design decisions governing task composition. Figure 5.1 depicts the game level
which serves as the backdrop (or design inquiry) for the player who has to build a
shelter for survivors using shipping containers. The design inquiry consists of three
types of tasks—collecting (survivors and building materials), building (a shelter),
and allocating (survivors to the shelter built). During collection, the player is
expected to calculate the space needed to accommodate each type of surviving fam-
ily (i.e., the various adult-child ratios in the family composition and associated
space requirements), measure the living space (area) afforded by each container,
and then figure out the number of containers needed before collecting both survi-
vors and containers within a limited time frame. During building, the player is
106 5 Interweaving Task Design and In-Game Measurement

Fig. 5.1 Composition of tasks in a game level

expected to review the location, three-dimensional shape, and size of the pre-quake
structure and replicate it using collected shipping containers. During allocation, the
player then has to allocate the various families into the limited number of container
rooms in the shelter. For each type of task, multiple constraints are applied, such as
the variety and number of types of families and containers to collect, the time limit,
the design criteria, and the space limits per room compared to the space needed by
the different families. Additional constraints may be added to further increase dif-
ficulty, such as minimizing the transportation cost of collecting a container.
In the earliest version of “row house,” all component tasks were presented as a
holistic inquiry without in-game feedback or cueing on the composition of sub-­
tasks. During its infield testing, we observed that middle-school students typically
lacked the awareness or skill to identify and plan sub-goals for the inquiry. They
were quick in figuring out how to collect families and containers yet failed to pur-
posefully check and calculate the space needed per family or use the family collec-
tion results to plan the container collection act. Hence they would randomly collect
containers and then either not replicate the row house or build the structure but not
allocate all of the families given the wrong number and/or type of containers col-
lected. To scaffold their sub-goal generation, we decomposed the broader inquiry
5.4 Core Design Sectors for In-Game Learning and Assessment 107

into segments and had the game check the player’s collection performance before
allowing him to proceed to the next step/task (as Fig. 5.1 illustrates). A consequence
of such a linear segmentation, however, was that certain players adopted the generic
strategy of guessing, along with trial-and-error actions to pass the collection sector,
rather than processing and applying the task-related math information. Segmentation
also caused players to inadequately develop an overarching interpretation of the
problem. For instance, they failed to recall the living space needs of the different
families (i.e., information identified in the collection sector) during the allocation
sector. We additionally experimented with task-equivalent game level development,
in which each component task represented a separate level. The component tasks
comprised a partial problem to solve.
Now, in the game, after the player completes a sequence of related task levels, he
then must tackle a complete problem (boss level) which is an all-inclusive instantia-
tion of the sum of component tasks. Correspondingly, we created two levels of game
chunking—level (task-equivalent) as the child or lower-level units and episode
(multitask-equivalent) as the parent or higher-level units. Such a mixture of task
composition and decomposition to particular actions needed, as the field test data
indicated, was well received by student players and associated with an increased
level of task engagement and math problem-solving performance.
Chunked versus holistic task We experimented and conducted a qualitative com-
parative analysis with the users’ participation behaviors in between the holistic task
with which a player had to explore and chunk the componential problem-solving
phases by himself and the chunked task series in which each problem was purpose-
fully chunked into multiple phases (or levels) that players will proceed sequentially
with. When interacting with the sequenced phases (e.g., material collection/trading,
structure building, and space allocation) in a design task, participants followed a
framed problem-solving path to solve and connect each part of the puzzle. The com-
parative analysis of participants’ game-based problem-solving processes and perfor-
mance in the two task structures indicated that an explicitly chunked task structure,
by offering partial representation of a complex problem, fostered task engagement.
Participants with the holistic task structure demonstrated obviously more off-task
behaviors than those with the chunked task structure. For example, participants of
the former were frequently found wandering around the game world, casually play-
ing with a game object (e.g., using the measurement tool to draw line sketches
instead of site surveying), and reporting feeling stuck or lost during gaming.
On the other hand, an explicitly chunked task structure presented less opportuni-
ties for participants to perform failure-driven, reflective learning. For example, with
a holistic task, participants were frequently observed recalculating the size of each
container, the total living space needed by the families, and hence the number of
containers needed, when they failed to allocate all families. Yet with the chunked
task structure, such a failure-driven mathematical practice occurred less frequently
because at the end of the building phase, the game would evaluate the shelter built
and alert the player if containers used were not enough. In other words, the game
had done a critical part of the problem representation and solution for the player.
108 5 Interweaving Task Design and In-Game Measurement

Connection between tasks For multilevel episodes in E-Rebuild, we explored

alternative ways of constructing and relating game tasks or levels. For example, in
an “Island” game episode, every level presented a task following the same problem
template, while the specifics of core problem parameters (e.g., the size of a con-
tainer, the number and type of families, the adult-child ratio in space needed) varied
to create incremental difficulty in mathematical problem-solving. In a “Desert” epi-
sode, a sequence of levels represented a collection of tasks governed by an over-
arching architectural design scene. Across the levels there was an incremental
inclusion of new parameters (e.g., new building and site design criteria) in the prob-
lem to create incremental complexity in problem-solving. The gaming behavior
analysis and infield observation did not reveal any obvious difference between the
two task or level arrangements in relation to players’ task engagement or game-­
based problem-solving performance.
Instead, we observed that the degree of granularity required by a building task
(e.g., building a structure using prefabricated shipping containers versus using 2-by-­
2-meter blocks) in different game episodes moderated players’ performance in 3D
object maneuvering. It was reported and observed that the acts of rotating and stack-
ing 3D objects in a higher granularity level were more time-consuming than that in
a lower granularity level. The challenge of manipulating 3D objects during building,
interestingly, appeared to foster the behaviors of preplanning for building, such as
predicting the rotation and location of a block before placing and stacking it onto
another and gauging the minimum space needed or the most efficient way of build-
ing. These behaviors were associated with more observed events of mathematical
reasoning and calculation.

5.4.2  etting Gameplay Rules to Motivate Content-Related

S
Task Performance

Prior analyses of the video- and screen-recorded gaming behaviors of the partici-
pants suggested that intentionality and mindfulness were two prerequisite facets of
game-based engagement. Intentionality refers to the ability to create and maintain
goal-directed gaming behaviors. Mindfulness refers to the level of reflective think-
ing (e.g., diagnosing the reason of a failed game action) and degree of planning
(e.g., comprehending the purpose of a future game action) during gameplay. The
analyses with the gaming behaviors and game logs indicated a positive association
between the presence of intentionality and mindfulness in participants’ gaming
behaviors and game level completion (Ke et al., 2017). The two task engagement
patterns appeared to mediate the players’ processing and application of in-game
mathematical information. Yet even when primary game actions and tasks were pur-
posefully designed to stimulate mathematical problem-solving, players did not nec-
essarily demonstrate intentionality or mindfulness in their gaming moves. Guessing
and trial-and-error behaviors were two strategies prioritized by novice players for
game-based problem-solving. More critically, their use of trial-and-error and
5.4 Core Design Sectors for In-Game Learning and Assessment 109

guessing lacked goal direction, mindful planning, reflection, or mathematical

thinking. It appeared that novice players, especially those who were motivated to
play rather than to learn, were reluctant to accept the circumstance that gameplay in
E-Rebuild is an intentional and reasoning process; they tended to avoid much cogni-
tive and mathematical engagement. Therefore, a critical challenge of task design in
E-Rebuild as a learning game was to motivate players’ intentionality, mindfulness
(or cognitive engagement), and hence processing and application of task-related
mathematical knowledge (or content engagement) during game-based problem-­
solving. A theoretical perspective that helps to shed the light on this design chal-
lenge is what Schoenfeld (2014) described about the three challenges faced by
students in mathematical problem-solving:
In some cases, “much of the mathematical knowledge that the students had at their disposal,
and that they should have been able to use, went unused in problem solving. This was not
because they had forgotten it (a matter of resources) or because they ran out of time to use it
(a matter of control), but because they did not perceive their mathematical knowledge as being
useful to them, and consequently did not call upon it (a matter of belief systems).” (p. 13)

Students’ problem-solving performance is not simply the product of what the students
know; it is also a function of their perceptions of that knowledge, derived from their experi-
ences with mathematics. That is, their beliefs about mathematics – consciously held or
not – establish the psychological context within which they do mathematics. (p. 14)

In general, Schoenfeld (2014) argued that mathematical problem-solving involves

(a) problem-solvers’ operational understanding of the mathematical information (or
resources at disposal); (b) techniques (heuristics) problem-solvers have or lack for
making progress on unfamiliar problems; (c) the way (control) they use, or fail to
use, the information at their disposal; and (d) their mathematical world view (belief
system) that determines the ways that the knowledge in the first three categories is
used. The lack of positive mathematical belief systems actually underlies the obser-
vation in our study that players were not willing or had the awareness to access their
mathematical knowledge to solve the game tasks.
Based on the infield observations and the aforementioned importance of players’
positive perceptions of mathematical information during problem-solving, we spec-
ulated that setting gameplay rules that highlight mathematical knowledge applica-
tion as the most efficient, if not the only, strategy for problem-solving should
reinforce mindful and content-related performance in the game. Hence we designed
a set of task constraints and reward rules that aim to curtail absent-minded, casual
gaming actions and enhance players’ awareness of the connections between each
math skill-based game action and the reward of gaming progress or credit. For
example, during the allocation task, players intermittently added families to random
containers and used a trial-and-error approach to reshuffle them (families to con-
tainers) until the task was solved. Consequently, a material and time credit deduc-
tion/penalty was then added for each allocation-reshuffling move. Once the deducted
credits hit some threshold, the player has to restart the task/level. With this supple-
mentary gameplay rule for the allocation task, the observed guessing and trial-and-
error gaming acts were reduced.
110 5 Interweaving Task Design and In-Game Measurement

Similarly, in a trading task, the player is supposed to order building items in bulk
to gain a discount (in an increasing percentage depending on the amount ordered).
Yet players frequently found ways to circumvent the percentage calculation and
only order items in part, not bulk. They also did not preplan the number of items to
be traded before a building action; instead they would buy items as needed during
building. To motivate mathematical information processing and calculation, we
enforced a “transaction fee” for each trade made, reduced the default material credit,
and increased the distance (and hence transportation time) between the store and
construction site. We also added a badge system (in addition to the game level pass/
fail) that explicitly reviews and rewards the degree of accuracy and efficiency in a
player’s game-based problem-solving actions. All of these gameplay rules, as the
design-based research findings indicated (Ke et al., 2017), have subsequently moti-
vated and increased the frequency of observed game-based cognitive and content
engagement acts, reduced the number of failed game level attempts, and promoted
game-based task performance.

5.4.3  enerating and Organizing Tasks with the Assessment

G
Models

A retrospective analysis of the observational data and design artifacts from the
E-Rebuild design meetings suggested that game task template generation and task
instantiation are confined and substantiated by the development of competency, evi-
dence, and task models for in-game assessment of learning. The sequential record
of design events indicated a concurrent, interactive association among the afore-
mentioned task and assessment design processes.
In E-Rebuild, we adopted the evidence-centered assessment design approach
(Mislevy, Almond, & Lukas, 2003) to construct game-based assessment.
Specifically, we started by defining the claims to be made about participants’
math competencies (i.e., competency modeling), establishing what actions or ele-
ments of gameplay constitute valid evidence of the claim (i.e., evidence model-
ing), and determining the nature and form of game tasks that will elicit that
evidence (task modeling) (Shute & Ke, 2012; Shute et al., 2017). Although assess-
ment design flows from competency to task modeling, in practice it is more itera-
tive. Diagnosis flows in the opposite direction. That is, the learners’ performance
(recorded by game logs) during a game level/task will provide the evidence or
data (e.g., logged scores of observable variables) that are passed on to the compe-
tency model, which in turn updates the claims (e.g., probabilities) about relevant
competencies. Based on the competency claims made, E-Rebuild, in the long
term, can dynamically present adaptive help or personalized game tasks/levels to
the player.
The competency model, as a framework for defining the targeted skills and
knowledge requirement in the game-based learning system, is a collection (or hierarchy)
5.4 Core Design Sectors for In-Game Learning and Assessment 111

of competencies that jointly define the ultimate achievement, successful ­performance,

or proficiency. Mapping the task templates with the competency model helps to
ensure game-based learning and assessment of the targeted competencies and skills
(see Fig. 5.2). The evidence model, presented as a Q-matrix (Fig. 5.3), specifies
which elements (or observed performance measures) of major game task archetypes
will demonstrate the practice and hence elicit the evidence of competency facet(s)
and how they will be combined to afford the learning and assessment of the targeted
competencies. The task model, presented as another Q-matrix (Fig. 5.4), states
which specific game level/task will provide the evidence for which competency
facet(s). The evidence and task models help the design team to (a) estimate whether,
when, and what tasks generated will accumulate enough evidence and (b) gauge the
difficulty, discrimination quality, and hence sequencing of generated/instantiated
tasks across game levels and episodes.

5.4.4  esigning Game Log with Observables to Record Game-­

D
Based Learning Evidence

During the course of E-Rebuild design, we designed the game logs to track game-­
based performance measures intended to provide evidence/data to propagate and
validate the statistical model (e.g., Bayesian network) for game-based assessment.
Game logs are XML files that are created at the end of gameplay. Since the goal is
to assess a participant’s competency from how the participant plays the game,

Fig. 5.2 An exemplary design document depicting competency model and game-based task tem-
plate design (Shute et al., 2017)
 Reason with ratio and proportional reasoning
112

Compare ratios
with whole
Task Name ObsName number Recognize a ratio Recognize a ratio Recognize a ratio
measurement relationship relationship between relationship between Represent a ratio Represent a ratio Represent a ratio Calculate the unit Recognize a percent of
using tables of between 2 quantities 2 quantities in 2 quantities in relationship via relationship via relationship via rate (a/b) associated a quantity as rate per
equivalent ratios in numerical form verbal form symbolic form numerical form verbal form symbolic form with a ratio (a : b) 100
timeToCompletion 0 1 1 1 1 0 1 1 0
Material Credit 0 0 0 0 0 0 0 1 0

Allocation Task scratchpad

editing(math related) 0 0 0 0 1 0 0 1 0
assignment
operation 0 0 0 1 0 0 1 1 0
# of trades 1 1 1 0 1 0 0 1 1
5

scratchpad
editing(math related) 0 0 0 0 1 0 0 1 0
Trading Task percentage lost in
trade avg 1 1 1 0 1 0 0 1 1
cut (for resourcing) 0 0 0 0 0 0 0 0 0

scale (for resourcing) 0 0 0 0 0 0 0 0 0

structure size 0 0 1 0 0 0 1 0 1
structure location 0 0 0 0 0 0 0 0 0
structure direction 0 0 0 0 0 0 0 0 0
Building Task # copy/paste failed 0 0 0 0 0 0 0 0 0

scratchpad
editing(math related) 0 0 0 0 1 0 0 1 0
ruler record 0 0 1 0 0 0 1 0 0
timeToCompletion 1 1 1 1 1 0 1 1 1
Game Task Material Credit 1 1 1 1 1 0 1 1 1
Happiness Credit 0 1 1 1 1 0 1 1 1

Fig. 5.3 Part of an evidence model (Q-matrix) example (Shute et al., 2017). Facets of the focus competency are listed in columns, and the indicators are listed
in rows
Interweaving Task Design and In-Game Measurement
 Comprehend a (De)compose Find surface Compute the
ratio Solve problems quadrilaterals Compute the areas of 3D volume of right Compute area
relationship via involving finding Calculate the and polygons area and figures using rectangular and
numerical, the whole, given unit rate (a/b) into (right) perimeter of nets of prisms (V = l w circumference of
verbal, and a part or a associated with triangles and triangle and rectangles and h, V = b h) a circle using
Episode Level symbolic forms percent a ratio (a : b) rectangles rectangle (right) triangles formulas
Island Collect Container 1 0 0 0 0 1 0 0 0
Island Collect Family 1 1 0 1 0 0 0 0 0
Island Build Training 0 0 0 0 0 0 0 0
Island Fill Training 1 0 0 0 1 0 0 0
Island Build 1 1 0 1 1 1 0 0 0
Island Fill 1 1 0 1 0 1 0 0 0
Island Collect Container 2 0 0 0 0 1 0 0 0
Island Collect Family 2 1 0 1 0 0 0 0 0
Island Build 2 1 0 1 1 1 0 0 0
Island Fill 2 1 0 1 0 1 0 0 0
Island Collect Container 3 0 0 0 0 1 0 0 0
Island Collect Family 3 1 0 1 0 0 0 0 0
Island Build 3 1 0 1 1 1 0 0 0
Island Fill 3 1 0 1 0 1 0 0 0
Desert Copy Training 1 0 1 1 1 1 0 0 0
Desert Copy Training 2 0 1 1 1 1 0 0 0
Desert Placement Training 1 0 0 0 0 0 0 0 0
Desert Build 1 1 1 1 1 0 1 0 0
Desert Angle Build 1 1 1 1 1 0 1 0 0
Desert Location Build 1 1 1 1 1 0 1 0 0
5.4 Core Design Sectors for In-Game Learning and Assessment

Desert Fill 1 1 0 1 0 1 0 0 0
School Place 1 1 1 1 0 1 0 0 0
School Place 2 0 0 0 0 0 0 0 1
School Fill1 1 0 1 0 1 0 0 0
School Paint 1 0 1 1 0 0 0 0 1
School Stadium 1 0 1 1 1 0 0 0 0
School Stadium 2 0 1 1 1 0 0 0 0
School Paint 2 0 1 1 1 0 0 0 0
Farm Angle 1 0 1 1 0 0 0 0 0
Farm Perimeter 1 0 1 1 0 1 0 0 0
Farm Area 1 0 1 1 0 1 0 0 0
Farm Volume 1 0 1 1 0 0 0 1 0
16 14 25 11 17 3 1 2
113

Fig. 5.4 Part of a task model (Q-matrix) example

114 5 Interweaving Task Design and In-Game Measurement

Table 5.1 Sets of observables for the given game levels

Exemplary level Observables
21ContainerCollect Time, NumWrong, MaterialCredits
26FamilyPlacement Time, NumWrong, AssignmentComplete
MaterialCredits
SchoolAssignment01 Time, AssignmentComplete, NumAssignments, NumFailedAssignments,
Num-FamilyCollected LevelComplete
IslandBuild02 Time, NumBlocks, NumTrades, Total-lost, MaterialCredits, distance,
size, angle BuildingComplete, LevelComplete
All levels Angle, AssignmentComplete, building-
Complete, distance, LevelComplete, MaterialCredits, NumAssignments,
NumBlocks, NumFailedAssignments NumFamilyCollected, NumTrades,
NumWrong, size, time, TotalLost

information that capture evidence of such competencies are logged in the XML file.
The following is the content of an example XML file for user “abc” for the game
level “SchoolPlacement02”:
<?xml version="1.0" encoding="utf-8"?>
<root>
<Name>abc</Name>
<Level>SchoolPlacement02</Level>
<Time>40.7275429</Time>
<NumBlocks>0</NumBlocks>
<NumTrades>0</NumTrades>
<TotalLost>0</TotalLost>
<MaterialCredits>10000</MaterialCredits>
<LevelComplete>true</LevelComplete>
</root>

The XML file has elements with tags like root, Name, Level, and Time which are
enclosed within “<” and “>.” For example, <MaterialCredits>10,000</
MaterialCredits> is a MaterialCredits element with content 1000. The elements
other than root, Name, and Level are the observables for the game level
SchoolPlacement02.
E-Rebuild prototype has 34 game levels and each level has its own set of observ-
ables. Table 5.1 lists a few game levels along with their observables. The last entry
of Table 5.1 is the union of the observables from all the game levels. As it is evident
from Table 5.1, each game level logs only a subset of the total observables.

5.4.5 Testing and Refining Task and Assessment Development

We have iteratively tested and refined the task templates, including game actions
and their interfaces, along with the instantiated tasks via iterative design experi-
ments. To examine the affordance of game tasks in fostering math problem-solving
5.4 Core Design Sectors for In-Game Learning and Assessment 115

skills and explore the nature of problem-solving processes, we collected longitudinal,

gameplay data of participants in every gaming session. Data were collected via
infield observations of students’ gameplay, informal interviews, game logs, as well
as video- and screen-captured gameplay behaviors and think-aloud protocols. We
conducted a qualitative thematic analysis with observation notes and gameplay
recordings to extract salient themes on participants’ gaming actions and their game-­
based problem-solving processes. These qualitative themes defined the nature of
commonly observed and unique player actions, along with the critical properties of
each action (e.g., objects engaged, purpose, and relevance to mathematical problem-­
solving). We then classified them into categorical themes depicting participants’
cognitive and affective engagement states and their major gaming actions.
Using the categorical themes and BORIS (an event logging software), we per-
formed pattern matching and systematically coded the recordings of a representa-
tive sample of participants’ gameplay actions and reactions, using 30 second time
intervals as the primary coding unit. The qualitative thematic and systematic behav-
ioral coding results informed about the nature of game-based mathematical
­problem-­solving processes in the context of variant types of game actions and tasks
and illustrated the relative effectiveness of each type of game actions, tasks, and
gaming interfaces. Based on the analyses results, specific game actions and tasks
were then refined and tested iteratively across design experiments.
Training and Testing of the Game-Based Learning Assessment Model Using a
Bayesian Network We have trained and tested the game-based math assessment
using a Bayesian network via Netica. Here are some reasons why we chose Bayesian
network for in-game learning assessment in this project:
• It is flexible in allowing domain experts to encode the domain knowledge by
defining nodes (i.e., competency and observable variables) and edges (i.e., the
relationships among the variables) in a directed acyclic graph.
• It handles missing data. This feature is a requirement in our case because each
game level provides only a subset of the observables.
Specifically, we trained, refined, and substantiated the Bayesian network using
both users’ gameplay data and expert review (Ke & Shute, 2015). Users’ game-
play data, as well as external pre- and post-gaming math knowledge test results,
were collected during iterative design experiments in E-Rebuild. We consolidated
the data stored in multiple XML game logs along with the corresponding pretest
scores into a single CSV (comma-separated values) format file. Each entry or row
in the output CSV file corresponds to a particular user and her game log for a par-
ticular game level played at a certain time. The consolidated data file contains
values of different types: integers, strings (categories), and floats. Since we have
selected Netica’s Bayesian network with discrete nodes for the game-based assess-
ment of learning, all raw values were converted to categorical values. The category
thresholds were determined from the training data and reviewed/refined by the
educational measurement, math education, and game design experts in the project
team (Fig. 5.5).
116 5 Interweaving Task Design and In-Game Measurement

Fig. 5.5 A snapshot of a Bayesian network model

We used Netica to create the Bayesian network described above and conducted
experiments on the preprocessed categorical data set. The goal was to learn a model
based on the training set about the relationships between observables across all
gameplay and the users’ competencies. Later, the learned model can take as input a
set of observables from the gameplay of an unknown user and predict competencies
of the user.
We fed the training data set into the Bayesian network via the Netica case file
format. We chose the expectation-maximization algorithm to train the model. To
establish initial values for the conditional probability tables in our Bayesian net-
work model, we considered each level played as an individual case. The initial val-
ues for the student’s mathematic ability are taken from their external math test
results. The network is trained using these values along with the log data. The net-
work is then tested against the posttest data.
After the training was completed, we continued with the testing process. We first
created a Netica control file which controls the output we want from Netica. We
then fed the testing data set from Netica to populate nodes in the Bayesian networks
to obtain the predictions for the participant’s mathematical competencies. An asso-
ciation analysis between the in-game assessment results (i.e., Bayesian network
predictions for individual learners) and the external post-gaming math test results
was conducted to validate the Bayesian network model. The correlation analysis
was conducted to examine the consistency between the predicted result of the cur-
rent Bayesian network model (i.e., low, medium, or high in the targeted compe-
tency) and the categorized posttest performance of middle-school gaming
participants. The analysis result indicated a significant association, r = 0.40,
p = 0.02. Error rate was used as another metric to evaluate the trained Bayesian
network model. The error rate measures the overall accuracy of predictions. The
error rate was 0.44 when we validate the model trained on Pa with Pb. This shows
that the model trained on a certain population can be used to make predictions on an
entirely different population.
References 117

5.5 Summary and Future Investigation

By examining and discussing the task design and assessment modeling processes of
E-Rebuild, we were able to explore how the two core game design elements func-
tion together to develop a game-based learning system. Such a system can poten-
tially serve as both a learning and an assessment tool. The five core design sectors
illustrate the design heuristics of the learning-play integration during game task and
level development; data mining-based assessment models for the problem-oriented,
task-centered learning; and the non-interruptive assessment mechanism in a playful
learning environment. They also support the innate association among content/
domain modeling, task template design and instantiation, and assessment model
development.
Multiple design issues remain unaddressed and warrant more design-based
investigation. The main issues are:
1. The existing game tasks are developed by the project team and cannot afford the
activation and assessment of varied sets of math competencies; salient parame-
ters of learning-centered game tasks need to be extracted to inform the genera-
tive method to enable cost-effective, scalable, and participatory game task
development by teachers and other direct users.
2. The mechanism that enables the real-time collection and analysis of the logged
gameplay data needs further development, which will involve the design of an
application that dynamically extracts the data from the game logs into the
Bayesian network or other applicable statistical models.
3. Incoherency still exists in designing game as a learning tool versus as an assess-
ment tool, in that the sequencing of tasks for learning focuses on scaffolding via
a content and difficulty progression and iterative practices for deep learning,
whereas the arrangement of tasks for assessment prioritizes selecting tasks that
discriminate most and presenting them to collect and accumulate evidence.

References

Almond, R. G., Mislevy, R. J., Steinberg, L., Yan, D., & Williamson, D. (2015). Bayesian networks
in educational assessment. New York: Springer.
Barab, S. A., Gresalfi, M., & Ingram-Goble, A. (2010). Transformational play: Using games to
position person, content, and context. Educational Researcher, 39(7), 525–536.
Cobb, P., Confrey, J., Lehrer, R., & Schauble, L. (2003). Design experiments in educational
research. Educational Researcher, 32(1), 9–13.
Dede, C. (2012, May). Interweaving assessments into immersive authentic simulations: Design
strategies for diagnostic and instructional insights. In Invitational research symposium on tech-
nology enhanced assessments.
Habgood, M. P. J., & Ainsworth, S. E. (2011). Motivating children to learn effectively: Exploring
the value of intrinsic integration in educational games. The Journal of the Learning Sciences,
20(2), 169–206.
118 5 Interweaving Task Design and In-Game Measurement

Kaller, C. P., Unterrainer, J. M., Rahm, B., & Halsband, U. (2004). The impact of problem
structure on planning: Insights from the tower of London task. Cognitive Brain Research,
20(3), 462–472.
Ke, F. (2016). Designing and integrating purposeful learning in game play: A systematic review.
Educational Technology Research and Development, 64(2), 219–244.
Ke, F., & Shute, V. J. (2015). Design of game-based stealth assessment and learning support. In
C. Loh, Y. Sheng, & D. Ifenthaler (Eds.), Serious games analytics (pp. 301–318). New York:
Springer.
Ke, F., Shute, V., Clark, K., Erlebacher, G., Smith, D., Fazian, P., Lee, S., & Xu, X. (2017). Math
learning through game-based architectural design and building. Paper presented at 2017
American Educational Research Association Annual Meeting, San Antonio, TX.
Klopfer, E., Osterweil, S., & Salen, K. (2009). Moving learning games forward: Obstacles, oppor-
tunities, & openness. Boston: The Education Arcade. Retrieved 28 April 2011 from http://
education.mit.edu/papers/MovingLearningGamesForward_EdArcade.pdf.
Levy, R. (2014). Dynamic Bayesian network modeling of game based diagnostic assessments.
CRESST Report 837. National Center for Research on Evaluation, Standards, and Student
Testing (CRESST).
Messick, S. (1994). The interplay of evidence and consequences in the validation of performance
assessments. Educational researcher, 23(2), 13–23.
Miller, C. S., Lehman, J. F., & Koedinger, K. R. (1999). Goals and learning in microworlds.
Cognitive Science, 23(3), 305–336.
Mislevy, R. J., Almond, R. G., & Lukas, J. F. (2003). A brief introduction to evidence-centered
design. ETS Research Report Series, 2003(1), i–29.
Mislevy, R. J., & Haertel, G. D. (2006). Implications of evidence-centered Design for Educational
Testing. Educational Measurement: Issues and Practice, 25(4), 6–20.
Pratt, D., & Noss, R. (2002). The microevolution of mathematical knowledge: The case of random-
ness. The Journal of the Learning Sciences, 11(4), 453–488.
Resnick, L. B., & Resnick, D. P. (1992). Assessing the thinking curriculum: New tools for educa-
tional reform. In Changing assessments (pp. 37–75). Dordrecht: Springer Netherlands.
Richards, J., Stebbins, L., & Moellering, K. (2013). Games for a digital age: K-12 market map and
investment analysis. New York, NY: Joan Ganz Cooney Center at Sesame Workshop.
Sandoval, W. A., & Bell, P. (2004). Design-based research methods for studying learning in con-
text: Introduction. Educational Psychologist, 39(4), 199–201.
Schell, J. (2014). The art of game design: A book of lenses. Boca Raton: CRC Press.
Schoenfeld, A. H. (2014). Mathematical problem solving. Orlando, FL: Academic Press.
Shaffer, D. W. (2006). Epistemic frames for epistemic games. Computers & Education, 46(3),
223–234.
Shaffer, D. W., Hatfield, D., Svarovsky, G. N., Nash, P., Nulty, A., Bagley, E., et al. (2009).
Epistemic network analysis: A prototype for 21st-century assessment of learning. International
Journal of Learning and Media, 1(2), 33–53.
Shute, V. J., & Ventura, M. (2013). Measuring and supporting learning in games: Stealth assess-
ment. Cambridge, MA: The MIT Press.
Shute, V. J., & Ke, F. (2012). Games, learning, and assessment. In D. Ifenthaler, D. Eseryel, &
X. Ge (Eds.), Assessment in game-based learning (pp. 43–58). New York, NY: Springer.
Shute, V., Ke, F., & Wang, L. (2017). Assessment and adaptation in games. In P. Wouters & H. van
Oostendorp (Eds.), Instructional techniques to facilitate learning and motivation of serious
games (pp. 59–78). Chem, Switzerland: Springer.
Wiggins, G. (1990). The case for authentic assessment. Washington, DC: ERIC Digest.
Chapter 6
Designing Dynamic Support
for Game-Based Learning

Abstract The role of support for game-based learning cannot be overemphasized.

It remains inconclusive as to what, when, and how support for learning should be
designed and implemented to foster learners’ extended engagement, in-game per-
formance, and game-based disciplinary knowledge learning and transfer. In this
chapter, we review prevalent support features in digital games, prior theoretical and
empirical research on scaffolding and support in game-based learning, and the sup-
port design conjectures deemed effective. We then share our observations of the
obstacles that learners experienced in game-based learning processes when using
E-Rebuild, describe the corresponding learning support strategies and features, and
report the findings from the iterative testing and refinement of these support fea-
tures. Propositions for future research and the design of support for game-based
learning are discussed in relation to the current project findings and prior research.

6.1 Introduction

The role of support for learning in game-based learning contexts cannot be overem-
phasized (Wouters & Van Oostendorp, 2013). A closely related and frequently
examined construct of support for learning is scaffolding. Though the definition and
scope of scaffolding is still inconclusive in the literature, scaffolding originally
referred to situations “in which the learner gets assistance or support to perform a
task beyond his or her own reach if pursued independently when unassisted” (Pea,
2004, p. 430; Wood, Bruner, & Ross, 1976). An adaptive level of support and the
fading of the support are implied as intrinsic components of the scaffolding process.
In this chapter, we describe our design and research of support for game-based
learning in a broader sense that includes both scaffolding with fading (or support as
needed) and general learner support without fading. We will use the terms of sup-
port and scaffolding interchangeably to include all support features in game-based
learning, whether adaptive or general, internal or external.
Support for learning in gaming often involves two alternative avenues: internal
scaffolding as the innate design feature of the game (or tool-/material-mediated

© Springer Nature Switzerland AG 2019 119

F. Ke et al., Interdisciplinary Design of Game-based Learning Platforms,
Advances in Game-Based Learning, https://doi.org/10.1007/978-3-030-04339-1_6
120 6 Designing Dynamic Support for Game-Based Learning

support) and external scaffolding as part of the game-based learning experience or

activity (or human agent-mediated support). Research on the desirable implementa-
tion context and the corresponding relative effectiveness of the internal tool-­
mediated learning support or the external learning support with human agency is
lacking. It remains inconclusive as to what, when, and how support for learning
should be designed and implemented to foster learners’ extended engagement, in-­
game performance, and game-based disciplinary knowledge learning and transfer.
In this chapter, we review prevalent support features in digital games, prior theoreti-
cal and empirical research on scaffolding and support in game-based learning, and
the support design conjectures deemed effective. We then share our observations of
the obstacles that learners experienced in game-based learning processes when
using E-Rebuild, describe the corresponding learning support strategies and fea-
tures, and report the findings from the iterative testing and refinement of these sup-
port features. Propositions for future research and the design of support for
game-based learning are discussed in relation to the current project findings and
prior research.

6.2  rior Design and Research on Support in Game-Based

P
Learning

6.2.1 Support in Digital Games

Support features found in digital games are not necessarily designed for content
learning purposes but focus on teaching novice players the flow of a game or the
game mechanics (i.e., basic actions, rules, and controls) while providing an optimal
level of challenge to support game engagement or flow (Chen, 2007). A common
design is a non-interactive in-game tutorial that provides a static introduction of
fundamentals, though they are frequently ignored by players who generally have
short attention spans and are not looking to digest masses of information. To aid or
motivate the tutorial processing, the designer can reward the action of attending to
the tutorial with achievement points. An engaging and active substitute of the in-­
game tutorial is to design an assortment of training levels in which the player starts
in a “noob cave” (or fail-proof) game level and then experiences or learns key game
controls or actions one by one via and during gameplay. Rather than providing
explicit instructions, certain games (e.g., Plants vs Zombies) try to use universal,
basic rules of gameplay or a fairly standardized user interface, assuming the player
has some familiarity with them and do not have to relearn everything. They also
encourage gameplay as active self-experimentation.
To create an optimal challenge, a frequent design strategy is to progressively
increase challenge or the complexity level by increasing the variables to be man-
aged in a problem scenario (Bos, 2001). In the game SimCity, for example, the
designer associates access to new game mechanics with the progression curve and
6.2 Prior Design and Research on Support in Game-Based Learning 121

gradually increases the number of changeable task parameters or rules through levels.
Another exemplary support feature is an in-game “technology tree” (e.g., in the
game of Civilization) that limits the players’ access to advanced items or tools until
they are able to master the usage of the basic mechanics.
Apart from internal game support, game community mechanisms, such as “let’s
play” gaming videos on YouTube and community support via forums, conferences,
and blogs, have provided both learning and socioemotional support by peers and
developers. Prior research (e.g., Ho & Huang, 2009; Steinkuehler, 2006) on online
game communities has reported that online communities for gamers have reinforced
not only participatory culture but also transactional and social learning processes.
Because the game designers’ focus is on preventing frustration and reinforcing
enjoyment more than the effectiveness of learning embedded domain-specific
knowledge, the scaffolding or support approaches in digital games may have the
potential to act as learning barriers (Sun, Wang, & Chan, 2011). For example, instant
feedback and demonstration scaffolding that aim to reduce frustration levels can
reduce the potential for experimentation and failure-driven learning, create overreli-
ance on system prompts without internalizing knowledge acquired from interac-
tions with the game system, and hence undermine learning if overused. The
mechanism that supports a sense of playfulness could also be in conflict with the
development of purposiveness and the effort contributed to “problematizing” the
game challenges with respect to academic disciplines or subjects (Reiser, 2004).

6.2.2 Support in Game-Based Learning

Games are a complex learning environment in which players—especially novices—

can be overwhelmed by the abundance of information to be processed, the dynamics
of the game world, and the complexity of the task and consequently deterred from
activities that foster learning (Ke, 2016; Wouters & Van Oostendorp, 2013).
Moreover, the dynamic and instant linking between a player’s action and changes in
the game world may lead to trial-and-error learning in which knowledge remains
intuitive without being explicated or conceptualized (Leemkuil & de Jong, 2011).
Hence it is warranted to design supports in game-based learning that facilitate not
only the selection and processing of relevant information but also an active organi-
zation and conceptualization of intuitive insights for substantive knowledge
development.
During the past decade, increased research has examined the design features that
promote game-based learning, especially learning or instructional support features.
Designing supports or scaffolds to fit the game-based task properties as well as the
associated control interfaces is a critical design issue (Soloway, Guzdial, & Hay,
1994). Though empirical studies examining particular game-based learning support
features have increased in the past decade, research on a systematic framework of
game-based learning support design is still lacking. The literature has, however,
122 6 Designing Dynamic Support for Game-Based Learning

shed light on a design framework for computer-assisted learning support in general.

Wood, Bruner, and Ross (1976) initially specified scaffolding as the processes of
(attention) recruitment, reduction in degrees of freedom for the problem at hand,
direction maintenance, marking critical features, frustration control, and demon-
stration (or modeling). Building on such a perspective, Pea (2004) and Reiser (2004)
observed that scaffolding serves three main major purposes: (a) channeling and
focusing (or task structuring), such as reducing the degrees of freedom for the task
at hand by providing constraints that increase the likelihood of a learner’s effective
action and recruiting and focusing learners’ attention by marking relevant task fea-
tures in what is otherwise a complex stimulus field; (b) modeling, such as demon-
strating a solution to the task, with or without human agency (e.g., modeling by
software features or a socially interactive other); and (c) content problematizing—
tools or features that can shape a learner’s performance and understanding of the
task in terms of key disciplinary content and strategies while problematizing the
disciplinary content with the task at hand. Quintana and his colleagues (2004) then
tried defining a scaffolding design framework for computer-assisted scientific learn-
ing environments via a theoretical review and a synthetic analysis of exemplary
courseware. They classified a list of design heuristics and techniques of scaffolding
based on the basic problem-solving processes to facilitate, such as sensemaking (or
problem and content representation), process management (or task structuring along
with modeling), and articulation and reflection (for content problematizing). These
earlier frameworks of scaffolding, along with the previous empirical studies on spe-
cific game-based learning support features, have all informed our design of learning
support features in E-Rebuild.
Wouters and Van Oostendorp (2013) conducted a meta-analytic review of 29
studies on the effects of instructional support in game-based learning. They reported
that players benefit from instructional support for knowledge (d = 0.33, I < 0.001),
skills (d = 0.62, I < 0.001), and in-game performance (d = 0.19, p < 0.001). Support
features that facilitate learners in attending to and selecting relevant information
(d = 0.46, p < 0.001) are more effective than the ones that stimulate organizing and
integrating new information (d = 0.14, p < 0.01). Specifically, modeling (or explica-
tion of the problem-solving procedure), multimodality, personalization, and feed-
back are effective techniques to support learners in selecting relevant information.
Their analysis indicated that the most effective instructional support for information
organization and integration is reflection (d = 0.29), in which learners are explicitly
asked to think about their actions or decisions. Instructional support that less explic-
itly stimulates the organization and integration of information, such as narrative
elements and collaboration, seems to be less effective (d = 0.11 and d = 0.14, respec-
tively). They also reported that the effectiveness of instructional support in game-­
based learning was significant for elementary school and college/university students
(d = 0.19, p < 0.05; d = 0.41, p < 0.001, respectively), but not significant for middle
or high school students.
The recent work of Kao, Chiang, and Sun (2017) on scaffolding in game-
based learning environments reported that marking critical features facilitated
6.2 Prior Design and Research on Support in Game-Based Learning 123

conceptual knowledge acquisition better than demonstration, and the two scaf-
folding types facilitated different dimensions (sensitivity and flexibility) of
design creativity. Johnson and Mayer (2010) reported that people who played the
circuit game learned faster and were better able to transfer to new problems if
they received in-­game guidance in the form of non-interruptive self-explanation
(i.e., selecting a reason for each action) or feedback (i.e., being shown the reason
for the correct action). They concluded that the educational impact of a game can
be substantially improved by incorporating support features aimed at guiding the
learner’s cognitive processing during playing. Their conclusion was supported
and extended by a later study by O’Neil et al. (2014) who found that self-expla-
nation prompts aimed at helping game-based learners make connections between
game variables and disciplinary concepts (or content problematizing) are espe-
cially effective.
Other studies have similarly examined external supports for game-based learn-
ing. For example, Barzilai and Blau (2014) found that conceptual external scaf-
folds provided before gameplay led to better problem-solving (in comparison
with the play-only condition or external scaffolds provided after gameplay) but
lowered perceived learning. Presenting external scaffolds may have conceptual-
ized learners’ understandings of the game by connecting them to disciplinary
knowledge, but its effectiveness is moderated by the timing of the scaffolding.
Tsai et al. (2013) examined both teacher-initiated content presentation before
gameplay and student-­controlled in-game question prompts as game-based learn-
ing scaffolds. They reported that scaffolding before and during gameplay pro-
moted knowledge test performance better than scaffolding only during gameplay
and non-scaffolding conditions. Promisingly, the embedding of background con-
tent objects or question prompts did not influence students’ perceptions of their
gaming experiences. Chen and Law (2016) examined and compared two types of
external supports for game-­based learning: nonadaptive question prompts pro-
vided after gameplay (called hard scaffolds) and peer discussions during collab-
orative gameplay (called soft or dynamic scaffolds). They found a significant
positive effect of both external supports on learning performance but a negative
impact on motivation (competence, autonomy, and interest). They also found that
after-gameplay question prompts can reinforce the effectiveness of collaboration
for game-based learning.
In summary, the previous study findings suggest that the content of the scaffolds,
as well as the timing of their provision, should be carefully designed according to
the game features to achieve specific instructional purposes. Supporting the argu-
ments of Wood et al. (1976) and Pea (2004), in-game learning scaffolds that high-
light channeling, focusing, and content problematizing while being non-interruptive
are found effective. External scaffolds, such as collaboration and question prompts
aimed to explicate knowledge development, can be effective, but their effectiveness
will be moderated by the timing of the scaffolds.
124 6 Designing Dynamic Support for Game-Based Learning

6.3  esign Conjectures and Observed Challenges in Game-­

D
Based Learning

The design of game-based learning in E-Rebuild, as described in the previous chap-

ters, assumes that core game actions will necessitate transactional and problem-­
based learning in which players interact with dynamically represented elements of
a building-themed math problem, select and coordinate related information embod-
ied in game objects, plan and experiment with the problem-solving actions in an
interactive game world, and comprehend the task (with its variables and relationship
structure) in terms of key disciplinary content and strategies, thus engaging in math-
ematical conceptualization and problem-solving. These design conjectures imply
multiple expectations on the game-based learning processes—attentiveness, purpo-
siveness, and need for cognition – of a problem-based or discovery learner, an
acquired motive or sense of legitimacy toward “content problematizing,” as well as
a prerequisite resource base for performing entry-level tasks (e.g., the basic math
content knowledge and skills for building-themed problem-solving).
Yet heterogeneity in the school student group and the complex nature of game-­
based problem-solving induce observed participation and outcome variability in
which not all participants respond to game-based learning tasks in an expected or
favorable way. In particular, we found that many students lack training in compre-
hending and representing the mathematical relationships in a multistep context
problem, need help to connect and structure distributed task information embodied
in game objects and game actions, and do not automatically engage in the cognitive
or affective processes of game-based learning. Some of them also lack entry-level
conceptual knowledge or general understanding of design problem-solving.
In the initial prototype of E-Rebuild, we tried presenting a multistep, building-­
themed math problem in its lifelike form via a single game level. A mission descrip-
tion (or design problem statement) in the opening task panel worked as an
overarching task narrative. Designed as an open-ended learning environment, the
game world was a simulated natural landscape in which building items (or materi-
als), a potential construction site, and the structure to be built (with its criterion
properties) were situated but not explicitly denoted. Relevant cueing was presented
within the opening task narrative (e.g., “those shipping containers should be of
help”) or as learner-controlled, displayable visual aids (e.g., a “historical view” por-
trays the structure to be rebuilt in the pre-quake landscape, via a 2D slideshow as
well as a 3D emulation of the pre-quake structure and landscape). These design
efforts were intended to highlight aspects of mystery and puzzle in gameplay and
game-based, realistic math problem-solving.
Frequently, however, participants were not fully processing or digesting a written
task narrative at the beginning of gameplay, even when the narrative was concise
and evocative. They appeared to treat it as an affective plot opening rather than a
mental riddle containing the critical task message and hence lacked attentiveness to
delineate the embedded problem goal and variable information. The open-ended
game world also reminded some players of commercial building games that they
6.3 Design Conjectures and Observed Challenges in Game-­Based Learning 125

played (e.g., Minecraft), who then intuitively transferred the related gaming actions
or strategies (e.g., exploring the game world freely and building items at one’s own
will and standard) to the current game. As observed, they were often wandering
through the 3D landscape, clicking around, and picking up collectible objects ran-
domly. They then assembled the collected items casually, without an understanding
of the building goal/problem and the related design criteria. They became puzzled
upon receiving the level-failed message and asked, “What are we supposed to do?”
It was only after multiple failed trials would they realize that their old gaming habits
did not work and hence show more attentiveness in reading the task narrative and
other in-game cues. Other learning habits or preferences, such as the need for
explicit, step-by-step procedural guidance along with the lack of appreciation of
independent puzzle solving (or a low need for cognition), made them easily frus-
trated from game bottlenecks and frequently asking for instant help from peers or a
facilitator, “What should I do now?” Instead of looking to maximize “hard fun,”
these players wanted to pass the level and end the problem-solving experience as
soon as possible.
We also observed that players tended to lack persistent effort in task structuring
(i.e., mapping sub-goals and planning steps beforehand). For example, we expected
that a player would proceed with a logical process of site surveying, planning the
structure to be built (in terms of the size, shape, and position), collecting/trading
items needed, building as designed, and allocating inhabitants to the compartments
of the structure. Yet most players were involved in intuitive, trial-and-error game
actions that lacked system thinking: building instantly with only items at hand while
searching more items as needed, building a random structure, finding it impossible
to allocate all inhabitants, rearranging/rebuilding the structure, reallocation (poten-
tially with repeated mistakes), and so on. This unsystematic problem-­solving pro-
cess, to some extent, generated failure-driven, reflective understanding about the
presence of a mathematical relationship among task variables. It was, however, non-
mathematical and thus made it difficult for the players to complete sufficient levels
within a class session. To assist these intuitive-thinking players, we provided a short
list of sub-tasks (or marking the task structure) on the task panel. This design feature
did significantly improve the players’ in-game performance (e.g., the number of
levels completed within a session) but reduced the opportunity for the participants
to practice representation for multistep problem-solving (Ke et al., 2017). This find-
ing is consistent with prior research which found the big challenge for designing
game-based learning support is to find a balance between supportive tool availabil-
ity and encouraging learners to engage in discovery learning and accept some level
of frustration from game bottlenecks (Sun et al., 2011; Yelland & Masters, 2007).
E-Rebuild learners also experienced a learning curve in coordinating information
distributed across objects and actions in the 3D virtual world. Situating a mathemat-
ical problem in a 3D world, representing related problem variables as interactive
objects or action feedbacks that are distributed across game space and time, and
making one actively search and connect (rather than being provided with) the
embedded information were generally novel practices to middle-school students. In
the initial E-Rebuild gaming sessions, we commonly observed that participants
126 6 Designing Dynamic Support for Game-Based Learning

lacked purposiveness in storing and using information obtained during the previous
game object or action to guide their next move. Even when sub-tasks were marked
explicitly, participants tended to treat them as separate assignments rather than part
of a holistic problem. Based on think-aloud protocols and interviews, participants
generally lacked a solid understanding of the systematic nature of mathematics—
the need to work on a higher logical plane in problem-solving situations—and
understanding which and why mathematical relationships and ideas are useful in a
particular context for problem-solving.

6.4 Pros and Cons of Learning Support Design Strategies

Driven by both the literature on game-based learning support and the observed
learning challenges in E-Rebuild, we experimented with multiple learning support
design strategies: (a) channeling learners’ attention toward the final target task by
displaying sub-tasks in structure and constraining the space exploration to maintain
directedness of the learner’s activity toward task achievement; (b) presenting/
sequencing game tasks (or levels) with gradually increased degrees of freedom, by
reducing the numbers of variables in the task structure involved in the initial prob-
lem scenarios; (c) modeling or scaffolding the representation of mathematical rela-
tionships in a task with interactive step-by-step prompts, an in-game scratch pad,
multimodal cues, and feedback; and (d) externally supporting content problematiz-
ing by encouraging learning-constructive game talk among peers and external help
as needed from a facilitator or teacher. The pros and cons of these learning support
design strategies, with the illustration of our design experiment findings, are
described in the following section.

6.4.1 Structuring and Sequencing of Game Tasks

To assist players in mapping and systematically planning problem-solving steps, we

have experimented with alternative ways of chunking and structuring a game task.
One method involves chunking and sequencing sub-goals (or componential steps)
of a composite task, while another involves chunking and then gradually increasing
the set of influencing variables of the problem/task.
Sub-Goal Chunking and Structuring In our earlier prototype, we tried the first
chunking/structuring strategy by evaluating each sub-goal as a componential task,
sequencing them linearly as a succession of tasks, and showing only the “instant”
sub-goals with a task completion progress bar (Fig. 6.1a). The next sub-goal or task
(e.g., allocation) was locked until the previous one (e.g., building) was completed.
This design arrangement highlighted a “step-by-step” guidance on task structuring.
Such an arrangement has obviously fostered in-game task performance and on-task
6.4 Pros and Cons of Learning Support Design Strategies 127

Fig. 6.1 (a-b) Marking and sequencing of sub-goals

time, in comparison with the initial design that presented a multistep task as it is.
Yet we also observed that the explicit and linear sub-goal structuring reduced the
opportunity for the participants to proactively reflect on and refine their under-
standing of mathematical relationships underlying the task structuring. For exam-
ple, in the episode depicted in Fig. 6.1, the building problem (using multiple
containers to build a shelter resembling the pre-quake home and allocating all sur-
viving families to containers) includes two puzzles: (1) the number of shipping
containers needed and (2) the way the structure should be composed. When inter-
acting with the initial version, participants were frequently found transferring
between the two steps of “rebuilding” and “sheltering all families” to self-check
and refine the answer to the first puzzle. Yet under the linear and chunked task
structure, both puzzles had to be solved at the building step, and the participants
lost meaningful action feedback (e.g., whether all families get sheltered). Hence
more participants started trial-and-­error gameplay: building with all containers, let-
ting the computer or game check the built structure, and reducing one container at
each new trial until they passed this step.
In a later prototype, we presented all sub-goals of the composite game task as a
to-do list (Fig. 6.1b), with which the player can check off each completed sub-task
at a nonlinear sequence. All sub-goals or componential tasks are parallel and not
prerequisite to each other. The flexible sub-goal list helped to channel the players’
attention while allowing them to navigate and convert between sub-goals (or tasks).
128 6 Designing Dynamic Support for Game-Based Learning

As the gaming behavior analysis indicated, the nonlinear sub-goal chunking, in

comparison with the linear one, was associated with more enactments of content
learning engagement (e.g., information processing, calculation, knowledge applica-
tion) and generic cognitive engagement (e.g., planning, evaluation, refining).
However, providing the players with a specified sub-goal list has reduced the neces-
sity for learners to proactively construct and experiment with personally meaningful
representations of a multistep math context problem. They thus missed the opportu-
nity to practice problem identification with task structuring, a critical element for
realistic math problem-solving.
Variables Chunking and Escalation To enable an exploration or experience of
the systematic nature of mathematical relationships in a problem while scaffold task
structuring, we tried examining the number or structure of problem variables as an
additional salient entity for game level organization and sequencing. Specifically,
we designed and sequenced game tasks with a gradually increasing number of influ-
encing variables (or degrees of freedom) involved in the underlying mathematical
relationships. In a classroom building level of the School episode (Fig. 6.2a–b), for
example, the problem involves multiple variables: unit classroom/container space,
unit space need of each student of each subject, the total space/area needed, and the
positioning (and distance) of classrooms. The logged failed attempts, the frequency
of help requested, and the observed frustration level of participants were all higher
than expected in the original version (Fig. 6.2a). We therefore chunked the original
level into an easier version (Fig. 6.2b) where the positioning/distance variable was
temporarily removed. In the next level, a more composite version of the task was
then presented with the full set of the variables. The revision increased participants’
task performance and content-engaging actions.

Fig. 6.2 (a, b) Chunking and sequencing by task variables

6.4 Pros and Cons of Learning Support Design Strategies 129

Fig. 6.3 The plan view of the school set design

Constraining Game Space and Highlighting Information Points of Interest We

additionally found that constraining the game space helped to reduce the task-­
irrelevant locale exploration acts and channel participants’ attention toward prob-
lem information processing and selection. Take the aforementioned School episode
as an example; the earlier version presented a full replication of an urban school
setting (Fig. 6.3), designed by the architect consultant on the team. Proceeding from
the previous Desert episode to this lifelike School episode, participants (at a local
school district in Florida) all demonstrated a high level of excitement with the rich-
ness and attractiveness of the 3D game world. They became immersed in the envi-
ronmental story portrayed by the game space: navigating and checking every
structure and background object of the set and asking the facilitator whether they
could drive a school bus or how to enter a locked townhouse. They spent more time
exploring the task-irrelevant game space than fully processing the task narrative and
related information objects. Due to this observation and the need to downsize the
130 6 Designing Dynamic Support for Game-Based Learning

game file to run on the local school infrastructure, we reduced the magnitude of the
school set and background structures/objects. To better focus the players’ attention
on the salient problem variables (or the embodied game objects), we purposefully
made the information point of interest—the object that is critical for the problem
identification—as the starting point of gameplay at each level. For instance, in a
school stadium building level (Fig. 6.4), participants had to analyze the structure of
the stadium’s bench seating and find a way to build it in an efficient way. The prob-
lem requires the player to decompose and compose geometric shapes (e.g., a trap-
ezoidal prism as the base and a rectangular prism as the seat) and join and copy/
paste small cuboids of standard or varied dimensions (e.g., bricks) to construct the
base and seat of the stadium bench. This task, as observed, was novel to our middle-­
school participants who struggled in analyzing the composition of the structure,
usually spending more than 30 minutes in a single attempt or trial, and frequently
requesting hands-on help with the building process. Specifically, they did not fully
inspect the bench to notice its hollow rear. To correct this misapprehension, we set
an isometric view of the bench’s rear (instead of the elevation view of its front side,
as in the earlier version) as the starting point of the level. We also colored the base
and seat section of the bench differently to highlight its geometric composition. All
of these design revisions with the game space have clearly reduced the off-task time
and actions by the participants.

Fig. 6.4 The school stadium level—starting point

6.4 Pros and Cons of Learning Support Design Strategies 131

Fig. 6.5 (a, b) Scaffolds of mathematical relationship representation in iconic and symbolic
formats

6.4.2 I n-Game Scaffolding of Mathematical Relationship

Representation

To scaffold players’ comprehension and representation of the mathematical rela-

tionships in a game task, we tried adding step-by-step, interactive prompts, pre-
sented in both pictorial (or iconic) and symbolic formats, in the Help panel (see
Fig. 6.5). An empirical investigation of this scaffolding feature (Lee, 2016; Lee &
Ke, 2016) found a significant and positive impact of using this representation scaf-
fold, especially in symbolic form, on learners’ conceptual and procedural mathe-
matics knowledge development via gameplay. Yet the infield observations indicated
that players would not necessarily attend to or process the mathematical scaffolds
unless they were associated with explicit game rewards.
Tina got stuck on the numerical expression and calculation of a family’s space
need and started to use a trial-and-error strategy with the prompt—entering random
values to the text area until she guessed the right one. The scaffolding window pre-
sented instant feedback on every entry (e.g., reddening the wrong entry without
proceeding to the next screen), aiding the trial-and-error practice. Proceeding to the
next prompt, Tina kept guessing on a related math variable—the total space needed.
It appeared that she did not understand or remember the ratio statement and the vari-
able of family space processed/calculated during the previous step. Her neighbor
asked, “How did you figure out what (number) goes here (a variable in the equa-
tion)?” She answered, “I was just typing numbers.”
As this observation shows, certain participants failed to associate the interactive
scaffolds on numerical expressions with the game task at hand and deemed these
132 6 Designing Dynamic Support for Game-Based Learning

scaffolds as boring hindrances during gameplay. In contrast, other participants,

driven by multiple failed trials of the game task, were cognitively engaged in inter-
acting with the scaffolds.
Jeremy got stuck in the family allocation task and failed multiple times. He
started to check the help panel. It appeared that this was his first time reading the
scaffolding section seriously. He completed the first three ratio prompts smoothly.
He entered correct variable values in the equation for calculating the space needed
for a 1-adult-2-child family, but did not calculate the total correctly until the third
trial, indicating a lack of competency in interpreting and calculating the mathemati-
cal expression. He got stuck with the equation for a 2-adult-1-child family in the
next step; he entered wrong values for the space needs of each adult and child, lack-
ing a solid understanding of the math concept of unit. He went back to the previous
step and reviewed the equation of the 1-adult-2-child family. After fully compre-
hending that mathematical expression, he then reworked on the equation for the
2-adult-1-child family. This time he entered correct values for both the unit and total
space needs.
In the above example, the participant had developed a better understanding of the
unit concept and the mathematical equation representing the space needs of varied
families. As exemplified, in-game scaffolds assisted participants’ encoding of the
math problem using mathematical expressions or notations, but only when they
were processed mindfully. Therefore, it is important to design the in-game scaffold
of mathematical representation in a way that the players will perceive the scaffold
as a sensible investment of cognitive effort and an essential support for their game-
play or game-based problem-solving process.
In-Game Scratch Pad In E-Rebuild, we designed an in-game scratch pad (Fig. 6.6)
as a virtual calculator and a virtual worksheet with which the player can write, plan,
and record their problem-solving steps and calculation processes. Specifically,
when conducting a site survey using the measurement tool, the player could conve-
niently select and record measurements to the scratch pad for quick numerical cal-
culations. The infield testing indicated that players frequently used it as a calculator
and a notepad for measurement recording, yet few used it to scribe or plan problem-
solving steps. A critical reason for this finding is that electronically typing mathe-
matical expressions and equations is effortful for middle-school students. In
addition, the players could not map or encode the mathematical relationships via
freehand diagrams or visuals, on the scratch pad. Hence instead of using the virtual
scratch pad, the players preferred to use the offline paper worksheet to aid their
problem-solving and calculation processes. However, using an offline worksheet
could be interruptive to the players’ gameplay flow and not quite compatible with
the desk space that was already packed with the computer, keyboard, and mouse.
Consequently, participants tended to either work on a small sheet of paper in
between the equipment or choose to bypass the problem-solving scribing or map-
ping processes. Creating a user-friendly, efficient, and multimodal in-game scribing
tool is a continuing design challenge in E-Rebuild.
6.4 Pros and Cons of Learning Support Design Strategies 133

Fig. 6.6 An in-game

scratch pad

Multimodal Cues and Feedback To recruit and maintain the players’ attention on
the salient objects and features in the task structure and facilitate the reflection and
purposive refining of problem-solving actions, we designed in-game scaffolds also
as tooltips, action-specific feedback, in-game tools, and end-of-level badges. These
in-game scaffolds are multimodal, in line with the game mechanics and game world,
and presented in response to a game act or at the end of a game level.
In the original prototype of E-Rebuild, we tried putting all salient problem infor-
mation, including core variables and their features, into the task narrative being
portrayed on the Episode panel when one entered a game level and ever-present as
a background information object. However, we found that on average, participants
spent less than 1 minute reading the task narrative, with some simply skipping it. We
then tried presenting salient information as a tooltip associated with each game
object or element. The information appears when an anchor was positioned over the
related object, as shown in Fig. 6.7. As observed, interacting with diverse informa-
tion objects in the game world had involved participants in an in situ and active
investigation of the related problem variables and helped them to comprehend the
problem through hands-on discoveries.
During the design experiments, we found that action-specific cues, in compari-
son with ever-present cues, are better received and processed. For example, in the
trading action, we used to present items’ prices via a background bulletin board that
replicates the format of a math ratio table (Fig. 6.8a). We observed that middle-­
school participants frequently ignored the board in the store house. They would
134 6 Designing Dynamic Support for Game-Based Learning

Fig. 6.7 Salient variable

information presented as a
tooltip

make a random estimate of an item’s price, make an unreasonably high offer,

purchase certain items, run out of resource credits during the task, and thus needed
to retry the task. Only after multiple failed attempts in the trading action would they
be willing to carefully read the price bulletin board. In response, we added a tooltip
presentation of building items’ prices in the inventory, which has reduced trial-and-­
error processes and promoted numerical calculation behaviors by participants dur-
ing the training action.
A similar pattern was also observed with the action-specific feedback. For the
allocation action in E-Rebuild, we designed an item-allocating panel (Fig. 6.8b).
During the infield testing, participants complained that the panel was blocking the
game world, and they could not see whether and what item (e.g., a specific type of
family) was added or not. Correspondingly, they were found randomly clicking on
the adding or reducing buttons without purposive planning or mathematical reason-
ing. As a result, we changed the allocating interface to enable action-specific, imme-
diate visual feedback: a player can drag items from the inventory to the target space,
with the items either appearing in the space (if applicable) or reverting back to the
inventory (in a spaced-out situation). We also added a reddened alert at the bottom
of the screen cueing them about the reason for each unfitting occasion. The provi-
sion of action-specific, intuitive feedback has again fostered participants’ mathe-
matical reasoning and calculation behaviors during the game action.
To activate students’ reflection on their actions, we tried to present written, infor-
mative feedback in a game performance summary at the end of each game level. But
again, few participants carefully read the summary; most only skimmed through it
before proceeding to the next game level. Although participants did attend to the
three game credits presented on the left bottom corner of the screen, many of them
did not fully comprehend the relationship between each credit and their specific
gaming behaviors. To encourage action-based reflective learning and mathematical
6.4 Pros and Cons of Learning Support Design Strategies 135

Fig. 6.8 (a, b) Price bulletin board for the training action and the item-allocating panel

reasoning during gameplay, we have instead designed and presented a badge system
on the post-level summary (Fig. 6.9). Each badge is designed and presented to
encourage a player to improve the efficiency and accuracy in their game-based
problem-solving strategies and solutions and hence motivate a purposive usage of
mathematical reasoning and knowledge in the future game levels.
2D/3D Visuals for Alternative Processing and Comprehension An example dem-
onstrating the benefit of multimodal in-game scaffolds for problem representation
136 6 Designing Dynamic Support for Game-Based Learning

Fig. 6.9 Post-level game

performance badges

Fig. 6.10 2D floorplan slideshow and 3D historical view of the structure

was the pairing between a 2D visual sketch and a 3D model of the target architec-
tural structure in E-Rebuild (Fig. 6.10). The former was presented as a slideshow,
while the latter was presented as a historical view of the structure and construction
site with which the player could actively inspect and survey. When interviewed,
participants generally reported that they referred to both forms of the visual-spatial
representation during gameplay, which was confirmed by the infield observation of
their gaming behaviors. During the building action, participants frequently shifted
between reviewing the static, visual depiction (with numerical marking) of the tar-
get structure and exploring/measuring the structure model situated in the simulated
construction site. This pattern confirmed our speculation that such a multimodal
representation scaffold enabled learners to comprehend static visuals while interact-
ing with a maneuverable object or model, thus sufficing both visual and kinesthetic
ways of information processing.
6.5 Conclusive Design Insights on Game-Based Learning Support 137

In a recent study on the patterns of scaffolding usage of participants of E-Rebuild

(Dai, Ke, & Pan, 2017), we found that the help panel was the most frequently used
in-game scaffold and was used for the longest time, followed by tooltips, prompts,
and summary screen. However, the time spent on these in-game scaffolds was gen-
erally short, ranging from 6.85 to 44.65 s per 45-min gaming session. The frequency
of using in-game scaffolds has a positive correlation (r = 0.20, p < 0.05) with the
frequency of learning engagement.

6.4.3 External Scaffolding

In addition to the in-game scaffolding, we also encouraged learning-constructive

game talk among peers and external help as needed from a facilitator during partici-
pants’ gaming sessions. Even though E-Rebuild was developed as a single-player
game, participants voluntarily partnered up with their neighbors during gameplay,
updating or showing off their gaming progress with each other, sharing tips or strat-
egies, or venting frustration. Though the game talk was not fully mathematics rele-
vant, it acted as an external support for game task engagement. Via the game talk,
student participants frequently sought help and affective support from each other.
They also relied on the external scaffolding offered by a trained facilitator. The
facilitator’s scaffolding in the E-Rebuild program included question-answering on
the game mechanics, modeling of the 3D object maneuvering procedure, cueing on
the information to be processed or the potential way to unblocking a bottleneck, and
prompting for action-based reflective thinking. Notably, we found that external
scaffolding by the facilitator was especially beneficial for the participants who were
underachieving in mathematics. There was a positive association between the facili-
tator scaffolding usage and these students’ post-gaming math knowledge test per-
formance. Yet these students, in comparison with their peers, were significantly
more passive in seeking external help. We thus requested the facilitators to provide
proactive and recurrent scaffolding to the underachieving students, though it was
difficult in cases when multiple students required help.
Our study on the scaffolding usage in E-Rebuild (Dai, Ke, & Pan, 2017) indicated
that student participants generally sought a similar number of human facilitator scaf-
folds and in-game scaffolds, but they spent more time with human facilitators. In addi-
tion, participants received much more scaffolding, in both frequency and time spent,
from adult facilitators than from their peers. There was a significant correlation between
facilitator scaffolding and learner’s engagement behaviors (r = 0.26, p < 0.05).

6.5  onclusive Design Insights on Game-Based Learning

C
Support

Our design and research findings in E-Rebuild have highlighted and confirmed a
salient claim in the literature on computer-assisted learning support—only when a
support is context-aware and intrinsically integrated in the technical environment
138 6 Designing Dynamic Support for Game-Based Learning

will the support be effective. In game-based learning, a support should not only be
context- or environment-coherent but also action specific. Game-based learners
have a short attention span for textual information processing in a dynamic game
world and are occupied with in-game problem-solving. Consequently, they tend to
interact with learning supports when they are game action- and object-associated,
perceived as a legitimate investment (e.g., a worthy solution to block repetitive fail-
ures), and presented in a multimodal and non-interruptive way (i.e., not constituting
extra information-processing load).
Confirming the literature on game-based learning support (Wouters & Van
Oostendorp, 2013; Mayer & Johnson, 2010), the current project findings suggest that
learning supports that scaffold mathematical relationships (or task structuring) are
beneficial to learners’ in-game task performance and learning outcomes. Specifically,
chunking and sequencing sub-goals as well as gradually increasing componential
variables in the game tasks help to channel the players’ attention and effort toward
the overarching task goal and a systematic problem-solving process. However, it
remains a design challenge to create an in-game task-structuring tool that enables
mathematical diagraming or freehand scribing or a tool that bridges action- and visu-
alization-based qualitative insights to mathematical expressions of the variables with
their analytical relationships in an architectural building problem.
In the E-Rebuild project, we have found that internal and external scaffolding
can complement each other to create a coherent learning support system. Proactive
and individualized external support particularly assists underachieving students. Yet
individualized learning support via a trained facilitator is demanding in the school
settings. A potential solution and future research topic of the E-Rebuild project is to
provide in-game adaptive learning support that is driven by the real-time assessment
of the cognitive and affective states of the learners during gameplay. For example,
the game data-driven assessment of a learner’s math competency and affective states
can drive an adaptive selection and sequencing of game levels that supports an opti-
mal learning path and level of challenge. Within each game level, the real-time
tracking of the learner’s in-game task performance and engagement state will then
trigger multimodal support features that are adaptive in the content (e.g., associated
with the action at hand or the object being interacted with), timing (e.g., when a
threshold value of failed trials or the bottleneck state is reached), and format (e.g.,
modeling or prompting for a partial solution). These adaptive learning support strat-
egies can augment the current in-game scaffolding features and balance learner
autonomy in selecting the preferred support mechanism.

References

Bos, N. (2001). What do game designers know about scaffolding? Borrowing SimCity design
principles for education. Ann Arbor, MI: University of Michigan and CILT, PlaySpace Project,
Retrieved from http://concepts.concord.org/projects/PlaySpace/application/pdf/Learning%20
from%20SIMCITY.pdf.
References 139

Barzilai, S., & Blau, I. (2014). Scaffolding game-based learning: Impact on learning achievements,
perceived learning, and game experiences. Computers & Education, 70, 65–79.
Chen, C. H., & Law, V. (2016). Scaffolding individual and collaborative game-based learning in
learning performance and intrinsic motivation. Computers in Human Behavior, 55, 1201–1212.
Chen, J. (2007). Flow in games (and everything else). Communications of the ACM, 50(4), 31–34.
Dai, Z., Ke, F., & Pan, Y. (presented 2017, July). How can we help? Examining learning behavior
patterns in a math game. Presentation at 2017 Digital Games Research Association Conference,
Digital Games Research Association, Melbourne: Australia.
Ho, S. H., & Huang, C. H. (2009). Exploring success factors of video game communities in hier-
archical linear modeling: The perspectives of members and leaders. Computers in Human
Behavior, 25(3), 761–769.
Johnson, C. I., & Mayer, R. E. (2010). Applying the self-explanation principle to multimedia
learning in a computer-based game-like environment. Computers in Human Behavior, 26(6),
1246–1252.
Kao, G. Y. M., Chiang, C. H., & Sun, C. T. (2017). Customizing scaffolds for game-based learn-
ing in physics: Impacts on knowledge acquisition and game design creativity. Computers &
Education, 113, 294–312.
Ke, F. (2016). Designing and integrating purposeful learning in game play: A systematic review.
Educational Technology, Research and Development, 64(2), 219–244.
Ke, F., Shute, V., Clark, K., Erlebacher, G., Smith, D., Fazian, P., Lee, S., & Xu, X. (2017, April).
Math learning through game-based architectural design and building. Paper presented at 2017
American Educational Research Association Annual Meeting, San Antonio, TX.
Lee, S., & Ke, F. (presented 2016, April). The effects of representation format in problem repre-
sentation on qualitative understanding and quantitative proficiency in a learning game con-
text. Presentation at American Educational Research Association 2016 Annual Convention,
American Educational Research Association, Washington, DC.
Lee, S. (2016). The effects of representation format in problem representation on qualitative
understanding and quantitative proficiency in a learning game context. (Doctoral dissertation,
The Florida State University).
Leemkuil, H., & de Jong, T. (2011). Instructional support in games. In S. Tobias & J. D. Fletcher
(Eds.), Computer games and instruction (pp. 353–369). Charlotte, NC, US: IAP Information
Age Publishing.
Mayer, R. E., & Johnson, C. I. (2010). Adding instructional features that promote learning in a
game-like environment. Journal of Educational Computing Research, 42(3), 241–265.
O’Neil, H. F., Chung, G. K., Kerr, D., Vendlinski, T. P., Buschang, R. E., & Mayer, R. E. (2014).
Adding self-explanation prompts to an educational computer game. Computers in Human
Behavior, 30, 23–28.
Pea, R. D. (2004). The social and technological dimensions of scaffolding and related theoretical
concepts for learning, education, and human activity. The Journal of the Learning Sciences,
13(3), 423–451.
Quintana, C., Reiser, B. J., Davis, E. A., Krajcik, J., Fretz, E., Duncan, R. G., et al. (2004). A scaf-
folding design framework for software to support science inquiry. The Journal of the Learning
Sciences, 13(3), 337–386.
Reiser, B. J. (2004). Scaffolding complex learning: The mechanisms of structuring and problema-
tizing student work. The Journal of the Learning Sciences, 13(3), 273–304.
Soloway, E., Guzdial, M., & Hay, K. E. (1994). Learner-centered design: The challenge for HCI in
the 21st century. Interactions, 1(2), 36–48.
Steinkuehler, C. A. (2006). Why game (culture) studies now? Games and Culture, 1(1), 97–102.
Sun, C. T., Wang, D. Y., & Chan, H. L. (2011). How digital scaffolds in games direct problem-­
solving behaviors. Computers & Education, 57(3), 2118–2125.
Tsai, F. H., Kinzer, C., Hung, K. H., Chen, C. L. A., & Hsu, I. Y. (2013). The importance and use of
targeted content knowledge with scaffolding aid in educational simulation games. Interactive
Learning Environments, 21(2), 116–128.
140 6 Designing Dynamic Support for Game-Based Learning

Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in problem solving. Journal of
Child Psychology and Psychiatry, 17(2), 89–100.
Wouters, P., & Van Oostendorp, H. (2013). A meta-analytic review of the role of instructional sup-
port in game-based learning. Computers & Education, 60(1), 412–425.
Yelland, N., & Masters, J. (2007). Rethinking scaffolding in the information age. Computers &
Education, 48(3), 362–382.
Chapter 7
An Evolving Design Framework
for Game-Based Learning Platforms

Abstract In this concluding chapter, we discuss an evolving, experiential design

framework for game-based learning platforms, by synthesizing the salient design
problem-solving events, experiences, and solution-exploration findings reported in the
previous chapters. It is not aimed to be prescriptive or exhaustive, but acts as a starting
point for specifying the structuring and important concepts of the interdisciplinary
design of game-based learning. Results of our phenomenological inquiry assert that
solving a game design problem is mainly about problem structuring or transforming
an indeterministic problem space to partially limited. The structuring of the interdisci-
plinary design of game-based learning platforms consists of three norms: (a) transfor-
mation of the design goals into functional specifications of the design artifact, (b)
coevolution of the problem and solution spaces as well as exploration and syntheses
of partial solution of the subproblems, and (c) use of a common symbol (or representa-
tion) system to communicate and focus information while augmenting collective
memory and processing.

7.1 Introduction

Design is innately an ill-structured problem and a phenomenon difficult to model.

Design studies on ill-structured design problem-solving, such as those in the fields
of architecture and engineering, were mainly aimed to discern a generic design
model that is versatile in describing the invariant nature of design problem-solving
across the fields (e.g., Brown & Chandrasekaran, 2014; Dorst & Cross, 2001).
These generic design models shed light on our examination and understanding of
the game design process. But they come short of suggesting tangible solutions or
explaining applied reasoning processes for educational game design.
In this concluding chapter, we discern and discuss an evolving, experiential
design framework for game-based learning platforms, by synthesizing the salient
design problem-solving events, experiences, and solution-exploration findings
reported in the previous chapters. This framework is aimed to extend and enrich
previous generic design models with discipline-specific and functional design

© Springer Nature Switzerland AG 2019 141

F. Ke et al., Interdisciplinary Design of Game-based Learning Platforms,
Advances in Game-Based Learning, https://doi.org/10.1007/978-3-030-04339-1_7
142 7 An Evolving Design Framework for Game-Based Learning Platforms

knowledge emerged from our phenomenological inquiry. It is not aimed to be

prescriptive or exhaustive but acts as a starting point for specifying the structuring
process and important concepts of the interdisciplinary design of game-based
learning.

7.2 Structuring of the Game Design Spaces

Game design is characterized by an open-ended or unspecified goal state and an

extensive, indeterminate design problem space—in which the transition states (or
paths) between the design input and output (or goal) states can be unlimited and
their connections are unknown. Results of our phenomenological inquiry assert that
game design problem-solving is mainly about problem structuring (Newell &
Simon, 1972; Simon, 1973) or transforming an indeterministic problem space to a
partially limited one. The structuring of the interdisciplinary design of game-based
learning platforms consists of three norms: (a) transformation of the design goals
into functional specifications of the design artifact; (b) coevolution of the problem
and solution spaces, including defining the space constraints, decomposing the
problem into interconnected components, and exploration and syntheses of partial
solution of the subproblems; and (c) use of a common symbol (or representation)
system to communicate and focus information while augmenting collective mem-
ory and processing.

7.2.1  ransformation of Interdisciplinary Design Goals

T
into Functional Specifications

To formulate the design problem, we have tried to explicate, synthesize, and trans-
form the design goals into functional specifications of the design product. Because
game design is the experience design, the problem formulation investigation is
hence driven by a basic question: What learning-gaming experience will we design?
Specifically, a design team has to address the following sequence of questions to
specify the mechanics nature of the game-based learning experience:
• What composes a salient problem-solving or epistemic practice in the targeted
disciplines (e.g., mathematics and architecture)?
• What desirable aesthetic experiences—emotional and cognitive responses
evoked in the player—are associated with such a practice?
• What game objectives, actions, rules or constraints, and environmental storytell-
ing will promote the relevant disciplinary practices and aesthetic experiences?
• Is designing such an experience applicable considering the technical and prag-
matic constraints (e.g., time, budget, resources at hand, and the future implemen-
tation setting)?
7.2 Structuring of the Game Design Spaces 143

The questions above have helped to bring to bear interdisciplinary team

members’ personal and disciplinary knowledge on the design problem formula-
tion and channel their attention and discussion on the core game mechanics and
game world design.

7.2.2 Coevolution of the Design Problem and Solution Spaces

Prior design research observes that creative design is “a matter of developing

and refining together both the formulation of a problem and ideas for a solution,
with constant iteration of analysis, synthesis and evaluation processes between
the two notional design ‘spaces’—problem space and solution space” (Dorst &
Cross, 2001, p. 434). Such a design observation was confirmed by our design
inquiry findings. Our design chronicle indicates that we start by exploring the
problem space and identifying a partial structure (e.g., what is the core game-
play?). The partial problem structure then helps us to identify a partial structure
of the solution space (e.g., a “building” action along with the 3D movement
interface). We examine and use this partial structure to generate more initial
ideas for the form of a design concept (such as objectives, potential challenges,
and rules associated with a building action), and so extend and develop another
partial structure of the solution space (e.g., actions of allocation and collection).
These solution space evolutions transfer back to extend the problem space iden-
tification (e.g., how will a game task assemble and legitimize game actions in
meaningful context?) and, then again, trigger the exploration of the solution
space (such as the task design and chunking and the game world development).
Ultimately, we create a system of matching problem-­solution pairs that are also
interconnected with each other.
It should be noted that the evolution of the system of matching problem-solu-
tion pairs for game design can be inexhaustible. To define constraints or a func-
tional boundary to the design space, the interdisciplinary team needs to frequently
review and refine the aforementioned functional specifications of the design arti-
fact, and prioritize the design structures within the negotiated, goal-driven
boundary.
A foundational structuring process of the game design space is to decompose
(Alexander, 1964) the problem into subproblems that are not strictly encapsulated,
with an ongoing monitoring process that looks for interconnections across them. In
E-Rebuild, for example, the problem-solution space exploration is decomposed into
the following interconnected facets:
• Designing game actions and rule sets that necessitate and legitimize the compe-
tency performance
• Designing an intermediary user interface to enable and motivate both functional
and cognitive interactivity
144 7 An Evolving Design Framework for Game-Based Learning Platforms

• Designing and structuring game tasks, including contextual scenarios (the game
world or setting), to exemplify actions in a legitimate (or meaningful) context, to
scaffold learning progression, and to capture/accumulate performance evidence
defined by the competency/evidence models
• Designing the game logging structure along with the development of the statisti-
cal model for stealth assessment (in this case, Bayesian network)
• Designing game-based learning support to enhance theoretical thinking and
subject-­problematizing opportunities
In the later stage of the design space structuring, the sub-solutions for the above
subproblems are then aligned and adjusted coherently to create a synthetic design
solution for the whole game design problem. For example, in the later design phase
of E-Rebuild, game task development and sequencing, game logging structure
design, assessment model refinement and validation, and learning support design
have merged with each other so that the iterative design and refinement are pro-
ceeded with them all rather than a single facet.

7.2.3  onstruction and Use of a Common Symbol (or

C
Representation) System

Using a common symbol system for design concepts among the interdisciplinary
design team is imperative for a longitudinal, complicated design problem-solving
process. During the E-Rebuild project, we found that a substantial amount of time
during the earlier design meetings was spent in decoding, comparing, and coordi-
nating differing symbol or representation systems of the design ideas, more than
that for the design idea generation. Due to the lack of a common symbol or repre-
sentation system of the design knowledge among the interdisciplinary teammates,
the collaborative review and refining of the design spaces become even murkier. The
use of a common “pattern” language, as we discovered, helps to communicate, fil-
ter, and focus information and augment collective memory and processing. The
construction of such a design language system should actually occur as the com-
mencement of the design problem structuring process.

7.3  ore Design Concepts for Game-Based Learning

C
Platforms

Several core design concepts have emerged as the most dynamic, composite, or
impelling constructs in the design spaces of game design. They embody the unique
features and challenges that characterize game-based learning platforms and deserve
further discussions and a deliberate examination.
7.3 Core Design Concepts for Game-Based Learning Platforms 145

7.3.1 Intrinsic Knowledge Integration

Prior research of game-based learning has examined and discussed the importance of
integrating knowledge as an internal or intrinsic part of gameplay, such as the design
exploration of endogenous fantasy (Habgood, Ainsworth, & Benford, 2005; Malone,
1981) and a recent synthesis of major approaches of integrating purposeful learning in
gameplay (Ke, 2016). Actually, the nature of the knowledge is as important as the way it
is integrated in or structured/enabled by gameplay. But the former receives much less
design or research attention. Few game studies examine the nature of game-based knowl-
edge to justify why its development warrants gameplay. That is, an implicit assumption
of “embedding” or “integrating” knowledge in gameplay is that the knowledge is infor-
mation or static data to be represented or conveyed by multisensory game worlds.
In this project, we explore the design of game-based learning as more an experience
design, with the aim to convey knowledge in interactions rather than static data. The
assumption is that problem-based gameplay enables not only conceptual and proce-
dural math knowledge but also structural understanding—understanding of the princi-
ples and conceptual relations among mathematical variables and procedures in a
contextual math problem. We also hypothesized that learners will discern new problem-
solving methods from the experience, after being provided expressive ways to confront
with and test the previously developed methods (e.g., trial and error) and realizing that
they are restricted. These theoretical and design conjectures about learning in gameplay
have surely influenced our design exploration. It is important, but beyond the scope of
this book, to examine how the nature of knowledge (e.g., general schema versus
domain-specific knowledge; sign-mediated theoretical thinking versus empirical think-
ing) will mediate the design approach of game-based learning platforms.

7.3.2 Meaningful Play

Although games have been frequently cited as a motivation tool for learning, design
research that examines the construct and design of game-based learning motivation
is still lacking (Star, Chen, & Dede, 2015). It is important for the game designers to
explicate why and how different learners will find a game-based, intentional learn-
ing experience interesting (due to intrinsic motivation) or important (due to well-­
internalized extrinsic motivation). For example, the game-based learning designers
need to resolve the conflict between the gameplay that requests considerable sense-­
making or cognitive interactivity and learners who may lack need for cognition—an
inclination toward effortful cognitive activities.
Educational game design is to create meaningful, interactive, and challenging
worlds involving the user as the conductor of his own intellectual development. In
relation to this stance are three design concepts imperative for describing and design-
ing game-based learning motivation: autonomy with environmental regulation,
meaning in interactivity, and optimal challenge.
146 7 An Evolving Design Framework for Game-Based Learning Platforms

Autonomy balanced with environmental regulation Learner autonomy (e.g.,

user choices of the game video/audio setting, speed, and the play mode or path) is a
common theme in the discussion of game engagement and usability. But we discov-
ered that too much freedom in a game-based learning space is associated with off-­
task behaviors and the lack of purposiveness and mindfulness for knowledge
application and acquisition during gaming. Environmental regulations of learner
autonomy, such as rules and incentive structures that prioritize domain-specific
methods and tools as well as a game world with a natural boundary, are imperative
for game-based cognitive and learning engagement.

Meaning in interactivity Salen and Zimmerman (2005) discussed meaningful

play as an emotional and psychological experience in a game that emerges from the
relationship between player action and system outcome, when such a relationship is
both discernable and integrated into the larger context of the game. This proposition
is adopted and confirmed by our inquiry that the user interface should communicate
the result of the game action to the player in a perceivable way while demonstrating
the significance of the game action on the play experience both immediately and at
a later point in the game. On the other hand, we found that the mechanism of creat-
ing meaning in cognitive interactivity (or psychological and intellectual participa-
tion between a participant and a system) is much more complicated than that for
functional interactivity (or technical usability). For example, players may fail to
discern the advantages of using domain-specific methods and tools or associate
their significance with the larger context in the game. Design principles or strategies
aimed to make cognitive interactivity discernable and integrated in game mechanics
and settings warrant deliberation.
In the E-Rebuild, we have tried reifying values of cognitive interactivity into
gameplay rules and incentive structures (e.g., a game badge system rewarding the
efficiency and thriftiness of using mathematical methods or solutions). We have also
tried leveling the entry barriers in cognitive interactivity, by chunking and sequen-
tially embedding domain-specific challenges in game levels. We design the game
world and the fantasy environment as a context that justifies cognitive interactions.
Specifically, discovering and interacting with objects in the game world (e.g., the
actions of site surveying and item collection) are intertwined with intellectual par-
ticipation (e.g., coding and calculating the mathematical properties of the site and
items). Infield testing and observing the interactivity between players and the coded
game system have enabled us to analyze the association between the game task/
level features and the creation or reduction of the undesirable shortcut solutions
(solutions without domain-specific thinking). In general, the game tasks should be
designed to provide the player expressive ways to confront, test, and compare
domain-specific methods or tools with generic ones.
Optimal challenge is the most popular design concept or standard in the previous
game studies, whereas it remains the trickiest design challenge due to the high het-
erogeneity in relation to prior game skills, domain competencies, as well as the gen-
eral learning ability of the players. It is difficult to gauge the presence of optimal
7.3 Core Design Concepts for Game-Based Learning Platforms 147

challenge until the actual responses of diverse players toward the tasks are iteratively
observed. In the E-Rebuild project, we have iteratively retuned the usability of the
interaction controls and training levels and emphasized a sequential embedding of
novel challenges and mechanics in game levels that presents a positive slope of the
learning curve.
On the other hand, an assumption of game-based knowledge development is
to reinforce a scientific attitude, such as the persistence toward frustration from
bottleneck as well as independent problem-solving. Design efforts to increasing
the players’ game engagement level by chunking and reducing the complexity of
a game task, or the arrangements of external help from peers or experts, can
reduce the chance for players to practice the scientific attitudes. It remains hazy
as to when frustration is positive for scientific attitude development and through
what observable behavior tracking we can gauge such a state adaptively for dif-
ferent learners.

7.3.3 Learning and Assessment Integration

An implication of our design inquiry of E-Rebuild is to design and use games as

both a learning and an assessment tool to enable gradual learning progression along
with evidence accumulation for assessment in the game setting. An integral design
and development of the tasks that scaffold learning while discerning competency
status is achieved via evidence-centered learning task design and data-driven perfor-
mance tracking. Specifically, we need to construct a learning trajectory (across
game tasks) that maps (a) the interrelationships among the sub-competencies (and
their observables), (b) the complexity level of each and every learning task (e.g.,
based on the required domain competencies, the internal complexity, amount of
information, and disclosure of information in a game-based contextual math prob-
lem), and (c) the correspondence between the former two facets.
The actual process of interweaving game-based task and assessment develop-
ment in E-Rebuild includes five interactive design sectors: (a) designing templates
of tasks based on the selected gaming actions that necessitate, as well as the sce-
narios that frame, the practice of the targeted competencies; (b) setting the game-
play rule and reward system that scaffolds mathematical practices; (c) developing
and organizing tasks based on the competency, evidence, and task models; (d) draft-
ing the game logs, including the specification of in-game performance observable
variables, along with the assessment statistical model (e.g., a Bayesian inference
network); and (e) testing, refining, and validating the feasibility and effectiveness of
the tasks and assessment models. The ordering of these core design sectors refers to
their operational sequence in our game design process.
More design investigations on the interaction and alignment between learning
task and assessment mechanism design are warranted. In particular, it remains
inconclusive as to whether the pacing and mode of learning progression is fully
coherent with those for performance evidence accumulation. In other words, the
148 7 An Evolving Design Framework for Game-Based Learning Platforms

learning trajectory may not be fully overlapped with an assessment trajectory. The
rule that decides the selection of a sequential learning task after a failed trial can
differ from that of selecting a task for better evidence collection and diagnosis.

7.4  ther Analytic Generalizations on Designing Game-­

O
Based Learning Platforms

A few other analytic generalizations (Yin, 2003) have emerged from our phenome-
nological inquiry of the E-Rebuild design and can jointly contribute to the future
development of a design theory for game-based learning platforms. These emergent
generalizations, as outlined below, define functional and unique properties of the
mechanic, interface, and player support in an educational game:
• Game actions should act as the foundation entity for the integration of learning
and play in game mechanics development. Defining the basic actions that
embody the essential content-rich practice helps to explicate the nature of learn-
ing and gameplay. Derived from the action specification will be the description
of its structure (e.g., its objective and execution rule), potential obstacle (e.g.,
challenge or constraint), and meaningful context (e.g., the backdrop setting or an
environmental story). Consequently, core gameplay for learning is defined.
• A bicentric mode (e.g., fly-through mode) of play improves game engagement as
well as the game’s functional and cognitive interactivity. It features a naturalistic
coordination between egocentric and exocentric perspective in the interactions
(Dede, 2009). Compared with either elevation or orthographic view, a bicentric
perspective facilitates multimodal representations, both grounded and abstract,
for the user’s construction of visual-spatial schemata in encoding game objects
and worlds, thus assisting them in investigating the game-based, contextualized
math problem.
• An intermediary interface is the extension of an intuitive control to promote
action-related, content-rich theoretical thinking. It is characterized by an active
prompting that necessitates an overt presentation of purposive, strategic, and
domain-specific cognitive engagements during the user input. An intermediary
interface should enhance cognitive interactivity while maintaining adequate
usability (or functional interactivity).
• A mixture of granularity levels of the simulated interactions (e.g., building with
coarse-grained versus fine-grained units) enables the emphasis of different
domain-specific competencies. The finer granularity level an interaction has, the
higher procedural complexity it presents, and the more engagement in strategic
planning (e.g., in selecting the most efficient procedure) it will motivate.
• Dynamic learning support should be interactive, and intrinsically integrated into
the game mechanics, the interaction interface, and/or the game scenario. It
should clearly convey its significance on the immediate game action and the
later play experience to the game players. The agency of learning supports can
7.5 Future Design and Research of the Game-Based Learning Platform 149

be compound: Learner initiated, system controlled, and human agent presented.

There is an interaction effect between the presentation formats (symbolic or
iconic) of the in-game supports and the task complexity in promoting game-
based conceptual understanding and procedural skills. We also found that disad-
vantaged learners tend to be passive in requesting help and hence should be
provided with more system- and agent-initiated learning supports.

7.5  uture Design and Research of the Game-Based

F
Learning Platform

7.5.1  o-construction of a Library of Design Cases or

C
Chronicles

Game design is a vastly ill-structured problem. A generic, versatile game design

framework can foster a generic understanding of the design process and constructs,
but will come short of guiding a specific design project. Moreover, the learning
purpose and the requirement of cognitive interactivity of an educational game make
the design of knowledge-conveying gameplay significantly different from the design
of entertaining gameplay. Differing stances and cultures in relation to domain-­
specific knowledge and practices will mediate the design goal and problem formu-
lation and make the design problem space even more indeterminate.
An ecological solution to the aforementioned challenges, as this book illustrates,
is for members of the game design community to purposefully record and analyze
their design explorations and problem-solving experiences, and share and archive
their retrospective analysis results along with the design chronicle via a design case
library or database. It appears that game designers have already developed the cul-
ture of blogging about their game design processes. Building on such a culture,
gathering and archiving differing design cases while constructing a common sym-
bol or representation system of the design concepts and knowledge should be appli-
cable and vastly helpful to the community of game-based learning designers.

7.5.2  esign Solution Generalization and Participatory Level

D
Design

As the prior game design research and this project demonstrate, the design of game-­
based learning platforms requires a longitudinal, iterative design, testing, and refine-
ment process. At the same time, using a game-based learning platform in school
typically requires that the platform should be scalable to enable a continuous devel-
opment or localized customization of game tasks based on a dynamic competency
model and the varying needs of diverse school or class settings. The challenges of
150 7 An Evolving Design Framework for Game-Based Learning Platforms

using games for learning, therefore, are to find generative methods that allow game
tasks to be created cost-effectively and enable users’ (e.g., teachers’) customization
of game levels. Even through automatic game design or level development has been
a prominent theme in game design research (e.g., Butler, Andersen, Smith, Gulwani,
& Popović, 2015; Nelson & Mateas, 2007; Togelius & Schmidhuber, 2008), the
exploration of automatic design solution generalization in educational game setting
is particularly limited.
Analyzing the basic design features and patterns of successful design solutions
(e.g., engaging and effective game tasks or levels) is the starting point for auto-
matic educational game design or level development (Butler et al., 2015). The
evidence-­centered task development in the E-Rebuild project has integrated the
idea of using a set of semi-structured task templates that are defined by a set of key
task parameters. For the future, we will experiment with a parameter-based level
development approach (Sorenson & Pasquier, 2010) through which teachers can
customize the values of a set of game parameters and task variables to define the
basic features of a game level (or task) and generate game levels dynamically.

7.5.3  daptive Learning Support Based on Real-Time

A
Assessment

Another potential extension of the current integrative design of learning and data-­
driven assessment is to use real-time diagnostic assessment to drive adaptive and
unobtrusive in-game learning support. Engaged gameplay may not guarantee suc-
cessful math learning for all players, especially for those who lack the habits and
skills associated with critical and reflective thinking. Adaptive scaffolds for action-­
based, metacognitive learning should be designed via unobtrusive in-game learning
support, such as game task selection, regulation and reflection of content-specific
gameplay, intrinsic learning prompts, and the balance between regulation and
autonomy (Chang, Wu, Weng, & Sung, 2012; Wouters & Van Oostendorp, 2013).
By leveraging the stealth assessment mechanism, one can develop and experiment
with adaptive in-game learning support that consists of (1) a customizable trajectory
of gameplay, such as an optimum selection of game tasks, enabled by the user control
of the parameter-based level development and (2) an adaptive and optimum level,
form, and timing of scaffolding that integrate user initiation (autonomy) and computer
regulation. Learners’ interaction with in-game learning support features can also be
captured as the input evidence to substantiate the developed Bayesian networks.

References

Alexander, C. (1964). Notes on the synthesis of form. Cambridge, MA: Harvard University Press.
Brown, D. C., & Chandrasekaran, B. (2014). Design problem solving: Knowledge structures and
control strategies. San Mateo, CA: Morgan Kaufmann.
References 151

Butler, E., Andersen, E., Smith, A. M., Gulwani, S., & Popović, Z. (2015, April). Automatic game
progression design through analysis of solution features. In Proceedings of the 33rd annual
ACM conference on human factors in computing systems (pp. 2407–2416). New York: ACM.
Chang, K. E., Wu, L. J., Weng, S. E., & Sung, Y. T. (2012). Embedding game-based problem-­
solving phase into problem-posing system for mathematics learning. Computers & Education,
58(2), 775–786.
Dede, C. (2009). Immersive interfaces for engagement and learning. Science, 323(5910), 66–69.
Dorst, K., & Cross, N. (2001). Creativity in the design process: Co-evolution of problem–solution.
Design Studies, 22(5), 425–437.
Habgood, M. J., Ainsworth, S. E., & Benford, S. (2005). Endogenous fantasy and learning in digi-
tal games. Simulation & Gaming, 36(4), 483–498.
Ke, F. (2016). Designing and integrating purposeful learning in game play: A systematic review.
Educational Technology Research and Development, 64(2), 219–244.
Malone, T. W. (1981). Toward a theory of intrinsically motivating instruction. Cognitive Science,
5(4), 333–369.
Nelson, M. J., & Mateas, M. (2007, September). Towards automated game design. In Congress of
the Italian Association for Artificial Intelligence (pp. 626–637). Berlin/Heidelberg, Germany:
Springer.
Newell, A., & Simon, H. A. (1972). Human problem solving. Englewood Cliffs, NJ: Prentice-Hall.
Salen, K., & Zimmerman, E. (2005). Game design and meaningful play. In J. Raessens &
J. Goldstein (Eds.), Handbook of computer game studies (pp. 59–79). Cambridge, MA: MIT
Press.
Simon, H. A. (1973). The structure of ill structured problems. Artificial Intelligence, 4(3–4),
181–201.
Sorenson, N., & Pasquier, P. (2010, April). Towards a generic framework for automated video
game level creation. In European conference on the applications of evolutionary computation
(pp. 131–140). Berlin/Heidelberg, Germany: Springer.
Star, J. R., Chen, J., & Dede, C. (2015). Applying motivation theory to the design of game-based
learning environments. In Describing and studying domain-specific serious games (pp. 83–91).
Cham, Switzerland: Springer.
Togelius, J., & Schmidhuber, J. (2008, December). An experiment in automatic game design. In
Computational intelligence and games, 2008. CIG'08. IEEE symposium (pp. 111–118). https://
doi.org/10.1109/CIG.2008.5035629. IEEE Xplore.
Wouters, P., & Van Oostendorp, H. (2013). A meta-analytic review of the role of instructional sup-
port in game-based learning. Computers & Education, 60(1), 412–425.
Yin, R. (2003). Case study research (3rd ed.). Thousand Oaks, CA: Sage.
Index

A Construct-centered assessment, 101

Amory’s game object model, 6 Content problematizing, 122, 124
Anti-earthquake building, 60 Core game action, 18, 62
Architectural artifact, 66 Cross-cultural communication, 9
Architectural Terms and Principles document,
53, 55
Architecture-themed scenarios, 104 D
Desert and historical rural pueblo, 95
Design aids, 71–73
B Design-based investigation, 103
Bayesian network model, 116 Design-based research, 7
Bicentric mode, 148 approach, 102
Brick-by-brick approach, 93 findings, 100
Building modes of play, 81 Design chronicle, 20, 143
Building-themed gameplay, 78 Design execution, 61, 62, 67
Design goal and space, 71–72
Designing gameplay, 18
C Designing games, 9
Chromebooks, 95 Design parameter and game reward
Chunked vs. holistic task, 107 mechanism, 61
Coding game mechanics, 18 Design principles/strategies, 146
Coding game objects, 64 Desirable situation, 2
Cognitive engagement, 85 3D game world, 92, 129
Cognitive interactivity, 5 Digital games, 120
Collaborative design, 17 Downscaled map and game space, 96
interdisciplinary, 17 3D rotation task, 93
students, authentic and constructive 3D sandbox games, 92–93
learning, 17 3D virtual world, 92, 125
Comma-separated values, 115 Dynamic learning support, 148
Commercial off-the-shelf (COTS) games, 2
Common Core State Standards for
Mathematics (CCSS-M), 45 E
Common symbol system, 144 Educational game, 19
Competency models, 54, 110 Educational game design, 4, 9
Consequential design, 88 approaches and features, 6

© Springer Nature Switzerland AG 2019 153

F. Ke et al., Interdisciplinary Design of Game-based Learning Platforms,
Advances in Game-Based Learning, https://doi.org/10.1007/978-3-030-04339-1
154 Index

Educational researcher perspective, 6 experience, 142

E-Rebuild creation, 20 platform, 18
E-Rebuild design, 42, 96 support, 19
architectural design and building, 43 system, 19
data mining approach, 44 tasks, 19
designing an educational game, 43 Game-based math learning assessment, 115
elevation view, 48 Game-based problem-solving performance,
interdisciplinary design, 48 35–36
interdisciplinary process, 42 Game community mechanisms, 121
interdisciplinary projects, 46 Game conversations, 90
learner, 42 Game data-driven assessment, 138
principal designer, 42 Game design, 9, 16, 142, 149
role and responsibility, 43, 47 Game goal, 79
teamwork, 49 Game level prototype, 105
E-Rebuild game, 46 Game levels
E-Rebuild game actions, 86 embedded mathematical concepts, 70
E-Rebuild game levels, 78 experimental findings, 71
E-Rebuild gameplay, 47, 60 exploration, 67
E-Rebuild gaming sessions, 125 game-based math problem, 68
E-Rebuild learning program, 102 learning game, 71
E-Rebuild participants, 93 problem statement, 67
E-Rebuild project, 42, 78, 147 structuring and navigation, 69
E-Rebuild prototype, 114 Game level sequencing, 70
E-Rebuild team, 45, 67 Game log structure draft, 65
Evidence-centered assessment design Gameplay, 4, 20, 60
approach, 110 brick-by-brick approach, 93
Evidence-centered design approach, 19 building-relevant gameplay, 85
Evidence-centered task development, 150 building-relevant mathematical experience,
Exemplary design cases, 72 80, 83
Exemplary design document, 111 building-themed gameplay, 79, 80
Experiential gaming model, 6 cognitive and content engagement, 86
External scaffolding, 123, 137 components, 76, 78
content engagement, 85
core game mechanics, 77
F 3D building materials, 86
Failure-driven learning, 90 design conjecture, 85
First E-Rebuild level, 69 design effort, 78
Florida Mathematics Standards (FMS), 45 design hypotheses, 84
Florida State University, 42, 45 design research findings, 78
Fluky psychology, 88 design team members, 78
Forgiveness, 91 disciplinary knowledge/epistemology, 80
Formal math notations, 104 3D modeling tools, 77
Functional interactivity, 5 E-Rebuild, 77
game mechanics, 76
game studies, 76
G granularity level, 84
Game actions, 85, 148 iterative testing, 85
Game-based knowledge development, 147 learning constructive, 85
Game-based learning, 8, 9, 18, 19, 88, 99, literature review, 77
120–124 mechanics, 76
assessment, 64 narrative aspirations, 77
contexts, 119 open-ended game, 77
engagement, 63 play preferences, 79
environment, 19 player perception, 79
Index 155

rules, 76 core game mechanic, 59

skill/content, 75 description, 52
team members, 80 design execution, 62
user-testing findings, 86 design inquiry, 59
Gameplay and learning, 2 design inspiration, 53
Gameplay exploration, 59 design pattern, 51, 54
Game programmers, 58 design problem roots, 53
Game programming, 66 design solution, 60
Games and simulations, 101 discipline-specific design, 57
Games for learning, 8 E-Rebuild development, 51
epistemic practice, 3 gaming-driven design exploration, 53
experience design, 5 graphical competency models, 53
game design, 5 math learning, 52
game-based learning, 2–3 mathematical context problems, 53
gameplay, 1 model sketching, 61
knowledge, 2 modeling and cataloging, 54
knowledge acquisition, 3 team members, 52
learning and play, 1 technique, 52
MDA, 5 Interdisciplinary design team, 72
play and learning, 2 Interdisciplinary educational game
player’s perspective, 4 design, 16
scientific attitude, 3 Interdisciplinary research and interdisciplinary
teaching and learning context, 2 design, 49
Game space, 129 Intermediary interface, 148
Game task design, 100 Intermediary user interface, 92–94
Game World design Internal scaffolding, 119
architectural scenarios, 94 Interweaving game-based task, 103
environmental settings, 95 Intrinsic knowledge integration, 145
environmental storytelling, 94 Iterative player testing, 20
formal mathematical notations, 97
pictorial presentations, 97
Gaming, 75 K
Gaming behavior analysis, 93, 108, 128 Kiili’s model, 6
Gaming behaviors, 108
General reasoning abilities, 87
Generic design models, 141 L
Google SketchUp, 94 Learnability and playability, 85
Guessing and trial-and-error behaviors, 108 Learner autonomy, 146
Learning and assessment integration,
147–148
H Learning system designers, 58
Highlighter tool, 63
Human-computer interaction, 62
M
Mathematical equation, 132
I Mathematical knowledge and skills, 88
In-game scratch pad, 23, 132 Mathematical planning and reasoning, 90
Initial game level, 68 Mathematical problem-solving, 109
Instructional support, 122 Mathematical reasoning, 59
Interactivity, 5, 9, 18, 19, 35, 36, 143 Mathematical reasoning and calculation, 89
Interdisciplinary collaborative design, 17 Meaningful play, 5, 145–147
Interdisciplinary design activities Mechanics, dynamics, and aesthetics
architectural concepts and skills, 53 (MDA), 4
architectural design case, 57 Metacognitive learning, 150
156 Index

Mindfulness, 108 R
Multimodal cues and feedback, 133 Reflective inquiry, 3
Multimodal in-game scaffolds, 135 Retrospective analysis, 110
Multimodal math story problem, 104 Row house, 106
Multisensory rendering environment, 1 Rural farm, 95

N S
National Science Foundation (NSF), 44 Scaffold players, 131
Nongame educational settings, 6 Scaffolding, 123, 131, 138
School set design, 129
School stadium building level, 130
O School stadium level, 130
Open-ended game world, 124 Self-determination motivation theory, 7
Stealth assessment, 46
Stealth assessment mechanism, 150
P
Parameter-based level development
approach, 150 T
Phenomenological examination, 10 Task archetypes, 104
Phenomenological perspective, 15 Task-based skill practice, 101
Physics Playground, 44 Task construction, 101
Pictorial presentations, 104 Task model, 111
Plan and elevation views, 82 Trading and building actions, 27
Post-level game performance Transaction fee, 110
badges, 136 Treasure hunting, 87
Price bulletin board, 135
Problematizing mathematics, 3
Problem-solving efficiency, 90 U
Problem-solving performance, 108 Urban school, 96
Problem-solving skills, 114–115
Process-oriented data, 101
V
Variables chunking and escalation, 128
Q Verbal/syntactic task description, 97
Q-matrix, 111, 113 Video- and screen-recorded gaming
Qualitative data, 16 behaviors, 108