Received: 7 June 2022 Revised: 20 November 2023 Accepted: 16 December 2023

DOI: 10.1111/jcal.12938

ORIGINAL ARTICLE

Mathematical experience in game-based problem-solving

Fengfeng Ke 1 | Chih-Pu Dai 2 | Luke West 1

1
Department of Educational Psychology and
Learning Systems, Florida State University, Abstract
Tallahassee, Florida, USA
Background: Game-based learning can frame problem-solving as a sense-making
2
Department of Learning Design and
Technology, College of Education, University
experience with domain-specific tasks for school students. However, multiple chal-
of Hawaiʻi at M
anoa, Honolulu, Hawaii, USA lenges arise when trying to support learners in such a complex, problem-oriented

Correspondence
learning environment.
Fengfeng Ke, Department of Educational Objectives and Methods: With an architecture-themed mathematics learning game,
Psychology and Learning Systems, Florida
State University, 3205-F Stone Building,
we conducted two mixed-method studies to explore the impact and design of
Tallahassee, FL 32306-4453, USA. game-based mathematical experience on the math problem-solving performance of
Email: fke@fsu.edu
middle school students.
Funding information Results and Conclusions: The study findings suggested a positive impact of
National Science Foundation, Grant/Award
Number: 1720533 game-based math experience on math problem-solving for middle school students.
Problematization-oriented game-based math tasks with structuring features
enhanced students' reasoning with problems and channelled it to doing mathematics.
Takeaways: The current research findings support the initiative to frame learning as
a sense-making experience with domain-specific tasks and inform the design of
game-based mathematical experience and learning support.

KEYWORDS
conation, game-based learning, mathematics, problematize, structuring

1 | I N T RO DU CT I O N realistic mathematics education (Van den Heuvel-Panhuizen &

Drijvers, 2020).
Current epistemological and pedagogical perspectives of However, challenges arise when engaging learners in realistic
mathematical education emphasize realistic mathematical practices and mathematical (math) experiences driven by domain-specific tasks.
sense-making experiences that activate integrative acquisition and First, tackling a complex task while coordinating the acquisition or
application of multi-faceted mathematical proficiency (Grootenboer & application of domain-specific strategies and content knowledge is
Edwards-Groves, 2019; Schoenfeld, 2020). A salient proposition is that cognitively demanding for novice learners (Kalyuga & Singh, 2016;
by conforming learners' experience with mathematics to the way it Van Merriënboer & Sweller, 2005). Learners may focus on completing
is done in the real world, learners will get the opportunity to tasks rather than on learning and explanatory goals, leading to incon-
develop mathematical point of view with everyday tasks or prob- gruence between interim task performance and sustainable learning
lems, learn to coordinate constituent knowledge and skills for (Reiser, 2004). Successful problem-solving does not necessarily trans-
effective task performance, and achieve deeper understanding with late to better learning in the longer term, because students may only
mathematical objects or tools (Van Merriënboer & Kester, 2014). leverage intuitive and generic prior knowledge to solve a domain-
The focus on framing learning as a sense-making experience with realis- specific problem without understanding or exploring the underlying
tic, domain-specific tasks can find its roots in multiple instructional the- disciplinary content (Kapur, 2014, 2016). As such, it is important to
ories or learning approaches, such as project- or problem-based learning explore design strategies that assist learners in transitioning from intu-
(Blumenfeld et al., 1991; Hmelo-Silver, 2004), experiential learning the- itive or non-reflective work to reasoning with domain-specific frame-
ory (Kolb et al., 2014), case-based reasoning (Kolodner, 2014), and works during problem-solving.

J Comput Assist Learn. 2024;1–15. wileyonlinelibrary.com/journal/jcal © 2024 John Wiley & Sons Ltd. 1
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 KE ET AL.

The proposed solutions to assisting learners in accomplishing and mathematical experience in everyday classroom practice is challenging
learning from complex tasks evolve around the design of tasks, task- while barely studied (Kertil & Gurel, 2016).
focused interface, and task-related learning supports. They are aimed Prior research on learning by practicing like a professional, or learn-
to problematize a learning experience as challenges that invite learners ing to think and act like a professional (Shaffer, 2004; Wenger, 1999),
to attend to critical ideas and connections that might be otherwise sheds light on the design of meaningful experiences with mathematics. It
overlooked, and structure it to make it tractable for learners proposed that by undertaking simulated, domain-specific practices, such
(Reiser, 2004). The investigation of these design propositions for as architectural design and modelling, students can gain access to the
mathematical experiences, however, is lacking in the literature. ways of knowing, thinking, and valuing that are fundamentally grounded
Second, learners may lack conation—self-direction, planning, self- in these activities (Shaffer, 2004). Solving architectural design problems,
regulation, and persistence—with a complex, challenging problem when as an exemplary domain-specific practice, can convey a meaningful math-
it is not personally meaningful (Clark et al., 2006; Goldin, 2019). Com- ematical experience (Jurow, 2005; Shaffer, 2007). There is an intimate
plexity is mediated by both prior knowledge and conation a learner relationship between architecture and mathematics that provides a con-
brings to bear on a task; as complexity increases, a learner's conation genial setting for math experiencing (Taylor, 2009). Moreover, thinking
becomes a more significant predictor of learning (Clark et al., 2006; and learning are naturally situated in the general practices of design and
Pintrich & Schunk, 2002). When tackling a complex task, cognitive diffi- making (Harel & Papert, 1991). Prior research has reported that students
culties with negative emotions such as frustration and confusion could are able to formulate math conjectures and develop a better understand-
emerge. Striving to pursue a line of reasoning rather than resorting to ing of math concepts and procedures during and in consequence of
avoidance strategies or rote procedures, as a characterization of cona- design- or making-based problem-solving (e.g., Ke & Im, 2014; Healy &
tion, is requisite for learners to overcome cognitive and affective obsta- Kynigos, 2010; Kafai, 1995; Kolodner et al., 2003).
cles to problem-solving (Barnes, 2019). Research on how to foster However, a full-scaled domain-specific practice requires the
learners' conation in reasoning with a complex problem is still limited. acquisition and application of organized content knowledge, strategies
With an architecture-themed mathematics game, in this research we for representing and transforming problems, and metacognitive pro-
conducted two mixed-method studies to explore the impact and design of cesses that guide the selection of knowledge and strategies while
digital mathematical experience on the math problem-solving perfor- monitoring the performance (Abassian et al., 2020). The activity can
mance of middle school students. Specifically, we studied tasks and be overly challenging for novices (Silver, 2013).
support features that problematize and structure students' mathe- Two design conjectures were therefore proposed for creating
matical experience while fostering their conation in problem-solving. learner-appropriate math practices or experiences. Based on the per-
spectives of realistic mathematics and educational modelling (Blum &
Ferri, 2009; Freudenthal, 1991; Van den Heuvel-Panhuizen &
2 | T H E O R E T I C A L BA C K G R O U N D Drijvers, 2020), a streamlined and game-based simulation of the original
authentic tasks can frame an accommodating math experience, by
2.1 | Mathematical experience enabling guided discovery while getting students to adopt an active
stance toward problem-solving. Further, prior works on cognitive devel-
The vast body of mathematical practices and concepts has historically opment in an interactive and complex learning environment suggest that
emerged from and been linked with our everyday life; to certain integrating and balancing problematization and structure in a learning
extent everybody experiences with or does mathematics consciously experience helps to promote both improved problem-solving perfor-
(Davis et al., 2012). We learn mathematics by doing, especially doing mance and acquisition of domain-specific schemata (Kapur, 2011, 2016;
problems, along with thinking, contemplation, and conversation. A Mayer, 2004; Pea, 2004; Piaget, 1976; Reiser, 2004). Problematization
mathematical experience frequently starts with working with helps to prompt learners to inquire about critical ideas and connections
real-world problems (Abassian et al., 2020; Blum & Ferri, 2009; Kaiser related to the task-in-progress. Structuring, on the other hand, is aimed
et al., 2011; Zbiek & Conner, 2006), encompassing: (a) exploring and at reducing the problem space of the task-in-progress to increase the
identifying key patterns or variables of a potentially messy situation, likelihood of learners' effective action while recruiting and focusing their
(b) trying to explain and represent these structural characteristics of attention onto relevant task elements. These two design functions are
the given real-world situation using tools that evolve from a primitive the flipped side yet complementary to each other in supporting domain-
gadget (e.g., a drawing) to a mathematical model (e.g. an equation), specific problem-solving (Kapur, 2011, 2016; Reiser, 2004). Empirical
(c) working with a constructed math model to produce and validate a research investigating these design conjectures, however, is still lacking.
math solution that will address the current and future real-world prob-
lems with similar structural characteristics and/or (d) developing an
increased understanding, fluency, or motivation of using math tools, 2.2 | Game-based mathematical experience
procedures, and concepts during these mathematized problem-solving
processes. Such an experience of reasoning with and modelling real- Gaming can serve as a vehicle for play, imitation, and invention, three
world situations and problems using mathematics, however, is still human activities that are considered natural primers of problem-solving
deficient in mathematics education. A main reason is that creating and crucial for intellectual development (Piaget, 1976). A digital game-
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
KE ET AL. 3

based learning environment can simulate a constrained aspect of the (Reiser, 2004; Schmidt & Bjork, 1992). The literature of game-based learn-
real-world domain-specific practices, by instantiating the central objects ing suggested mainly two design approaches to channelling generic
and relations of a domain into a concrete or semi-concrete form for problem-solving into experiencing mathematics in gameplay. One focuses
people to explore and learn (Edwards, 1991; Hoyles & Noss, 1992). on integrating in the game world external knowledge representations that
Through interactive storytelling, it can motivate players to actively do the learner will interact with and transform (Barzilai & Blau, 2014;
and value math ideas in the context of meaningful activities. Holbert & Wilensky, 2014). For example, written task narratives, pictorial
Notably, practices of gaming and mathematics share multiple com- or symbolic inscriptions on game props, outfits of game characters, and
monness in that (a) both focus on seeking patterns on the basis of discov- interactive game objects have acted as multiple external representations
ery and evidence, (b) both involve the acts of speculation and proving of core concepts (e.g., ratio, fraction and proportion). Students get to pro-
during discovery and (c) for both, learning proceeds through construction cess, manoeuvre, or transform these knowledge representations during
more than absorption (Ke, 2016; National Research Council, 2011; gameplay (Author, 2016, 2020; Habgood & Ainsworth, 2011).
Schoenfeld, 2020). Prior research reported that digital game-based learn- Another approach is to design game-based tasks that prioritize
ing has positive cognitive and motivational effects on the development of math experiencing, by problematizing learners' undertakings to make
multi-stranded math proficiency: understanding, problem-solving, and them grapple with, thus increasing the perceived utility of, domain-
positive disposition (Ke, 2008; Ke & Clark, 2020; Brezovszky et al., 2019; specific strategies and concepts (Barzilai & Blau, 2014; Ertmer &
Tokac et al., 2019; Wang et al., 2018). In these studies, game players dis- Glazewski, 2019). Problematization is a key feature of game-based
covered, experimented with, and learned from contextualized math learning: making the most foundational elements of subject matter
problem-solving to fulfil a combination of self-set and assigned goals. ‘problematic’ to stimulate inquiry and curiosity, thus intrinsically inte-
Nevertheless, game-based math experiences required self-regulation grating disciplinary knowledge construction in gameplay (Author, 2016;
in sense-making as well as reflective inquiry. Prior research reported that Engle & Conant, 2002; Hiebert et al., 1997; Reiser, 2004). Problematiza-
not all students were self-regulated learners or capable of sense-making— tion will trigger a higher-level of cognitive engagement with the typi-
attending to and connecting events and cues in the environment to cally ignored, domain-specific parts of real-world situations
anticipate their trajectories and act effectively—during game-based (Puntambekar & Hubscher, 2005; Reiser, 2004). Prior research reported
problem-solving (Ke & Clark, 2020; Chang et al., 2012; Wouters & Van that problematized learning tasks can afford students opportunities to
Oostendorp, 2013). Engaged gameplay may not guarantee domain- engage in pattern discovery and experimentation that are germane for
specific experiencing, especially when players lack the motive, knowledge, learning, thus promoting learning in the longer term despite an initial
or skills to perform domain-specific problem-solving (Author, 2016; Garris task performance dip (Kapur, 2016; Schmidt & Bjork, 1992).
et al., 2002). The design and effectiveness of game-based math experi- On the other hand, problematizing the subject matter may pre-
ence on math problem-solving and learning warrants further research. sent more challenges and degrees of freedom than students can han-
dle, thus making domain-specific experience largely unattainable
(Kirschner et al., 2006; Nachtigall et al., 2020). Adding structuring fea-
2.3 | Challenges and design propositions of game- tures that guide and explicate the connection between gameplay and
based math experience domain-specific practice is therefore desirable (Puntambekar &
Hubscher, 2005; Reiser, 2004). Structuring features in or outside of a
Translating a real-life problem-solving scenario to a math practice is not game, such as aids with task decomposition, worked examples, pre-
a naturalistic evolvement (Abassian et al., 2020; Silver, 2013). Prior configured concept maps, glossaries, and multimedia tutorials, have
research reported that learners may only leverage intuitive or domain- been applied to guide or prompt students on problem representation
generic cognitive strategies rather than developing mathematical point and solution modelling (Shute et al., 2020; Wouters & Van
of view to tackle a novel problem (Author, 2016; Kapur, 2014, 2016), or Oostendorp, 2013). Structured or guided experiential and problem-
they may fall back into their prior conceptions instead of exploring and based learning has been found to outperform unguided learning or
discovering new patterns or hypotheses, thus failing to acquire disciplin- direct instruction, though it is also reported to decrease learner initia-
ary framework or knowledge from the problem-solving experience tive and make gameplay less autonomous (Charsky & Ressler, 2011;
(Nebel et al., 2016). In a game-based learning environment, the motiva- Kirschner et al., 2006; Mayer, 2004). How we should structure a pro-
tion to win the game may also have a deleterious effect on learning blematized domain-specific experience for learning remains an open
from specific tasks (Israel-Fishelson & Hershkovitz, 2020). question.

2.3.1 | External knowledge representation, 2.3.2 | Supporting conation in game-based math

problematization and structuring experience

Designing game-based math experience, therefore, needs to attain a bal- Actively exploring and engaging with the environment is how children
ance between tapping into learners' domain-generic strategies and requir- learn during gameplay. Game-based problem-solving, in particular,
ing them to develop and work within disciplinary frameworks emphasizes exploring and discovering the patterns governing one's
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4 KE ET AL.

interaction with the game tasks and objects (Author, 2016; Brezovszky and examining an architecture-themed math game (called eRebuild) with
et al., 2019; Wang et al., 2018). It will be a self-regulated activity by a middle school students. Specifically, the studies addressed the following
motivated agent, which involves novel cognitive operations in the face questions:
of an open-ended, discovery environment with distractions (Clark
et al., 2006; Pintrich & Schunk, 2002). 1. What is the impact of game-based experiencing with math tasks
Prior research has frequently reported inappropriate ‘off task’ on students' math problem-solving performance?
behaviours, chanciness, insufficient attention to reflection or reevalua- 2. How do students interact with game-based math tasks and
tion, and reduced mental effort portrayed by the learners with game- learning-support features, in relation to their task performance
based domain-specific tasks (e.g., Abdul Jabbar & Felicia, 2015; Baker and math experience?
et al., 2010; Ke et al., 2019; Rowe et al., 2009). A principal cause of
these issues is the lack of conation of learners with problem-solving
when they are overwhelmed by a complex and dynamic environment 3.1 | A game-based learning environment for math
(Clark et al., 2006; Reiser, 2004). A related manifestation is functional experiencing
fixedness: players may fall back into their previous entertainment gam-
ing strategies or pre-existing ideas even when the current game-based In this study, we used eRebuild as a testbed for investigating game-
learning task contradicts them, thus failing to create and test extensive based experiencing with mathematics. Developed using Unity, eRebuild
new hypotheses for an increased understanding (Nebel et al., 2016). is a 3D architecture game aimed at engaging middle school students in
A recommended strategy to reinforce learners' conation to apply math contextual problem-solving—discovering and using math tools to
intellectual energy diligently and productively to a specific math task over conceptualize and solve math-related problems in a real-life situation
time is to emphasize the task's meaningfulness, by creating its perceived (Van den Heuvel-Panhuizen & Drijvers, 2020), with a focus on ratios
relevance (Reitan & Wolfson, 2000). Specifically, task goals that are spe- and proportional reasoning, geometric figures (de)composition and
cific and congruent with learners' prior experience, knowledge, or self- properties calculation, as well as expressions and equations. Differing
identified aspirations will help them to direct and focus attention and from word problems used in traditional math instruction, eRebuild math
effort on these goals (Huitt & Cain, 2005). Learners are also found to tasks are represented in a multimodal form and their parameters are dis-
perform better and persist longer when both task goal and performance tributed as interactive objects in the game world.
are acknowledged rather than anonymous (Seijts et al., 1997). Therefore, For example, in a ‘Row House’ task the player needs to build
in this study we proposed to design and examine a game-supported a multi-room shelter replicating a pre-earthquake row house using a
math experience that aims to leverage students' daily-life experience minimum number of shipping containers, which will accommodate a
with a series of enactive tasks to empower self-direction, regulation, and batch of families that differ in their living-space needs. This contextual
sustaining of their mental effort. Specifically, these game-based, enactive problem is spatially-configured in that it is represented by multimodal
math tasks comprise: (a) problematized quests that tap into students' information elements which are distributed in the game world, such as:
prior knowledge or life experience as well as their general interest in the opening task narrative, interactive 3D game objects (e.g., shipping
making or building, while stimulating domain-specific inquiries into the containers of different sizes, families in varied structures, and the house
unknown; and (b) explicit game-action rules (or incentives) to structure model), pictorial 2D diagrams (e.g., the floor plan), in-game tools (e.g., a
the students' mental efforts during gameplay. ruler that measures distance and angle), visual and syntactic gameplay
In summary, engaging learners in a sense-making, mathematical feedback, and the landscape designating the construction site. Given
experience with realistic, domain-specific tasks is proposed in prior such kind of multimodal and spatially configured contextual problem,
research and rooted in multiple instructional and learning theories. the player is challenged to investigate, identify, and coordinate the
Simulating domain-specific tasks (e.g., architectural design problems) embedded formal math notations and task-relevant parameters in these
in a digital game-based learning environment can support and frame a information elements to represent the problem mathematically and
meaningful mathematical experience in everyday classroom practice. arrive at a solution.
Nevertheless, empirical research investigating the design and impact Specifically, core problem-solving actions entail the practice of
of game-based math experience that channels generic problem- comprehending and modelling with multiple math representations.
solving into experiencing mathematics during gameplay is still limited. For example, a ‘building’ action will enact the practice of representing
Exploring task and support features that problematize and structure prisms using geometric nets (e.g., folding panels into a shelter, compos-
learners' mathematical experience while fostering their conation in ing a stadium stair-bench using cuboids), and that of investigating and
solving a problem mathematically is warranted. calculating geometric properties (e.g., area, perimeter, and volume in
painting a basketball court, fencing a farm, or filling a fish pond). When
manoeuvring these interactive mathematical representations in the
3 | MATERIALS AND METHODS 3D game world, the players also need to coordinate them with
the math notations embedded in the task narrative and mouse-over
We conducted two mixed-method studies to explore the design and tooltips to fully identify and map the structure of task-related parame-
impact of game-based mathematical (math) experience by implementing ters. They are expected to engage in sense-making—acting upon the
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
KE ET AL. 5

game-based math task and understanding how the task states (con- gender and prior gaming experience data were collected before the
veyed by the dynamic multimodal representations in the game world) study. The gaming participants' in-game task performance were
react to their actions, thus developing task-related conceptions or logged and archived. They were also surveyed on their enjoyment,
strategies that correspond coherently to the math notations. perceived challenge, and perceived learning from gameplay via three
The aforementioned math representations are contextualized in 5-point Likert items respectively.
architecture-themed design problems in eRebuild. Design or creation, Participants in the gaming group were oriented with the user
as the core mechanics of eRebuild, is innately meaningful for children interface and basic game mechanics of eRebuild by going through its
and will leverage their prior experience with sandbox games or com- orientation levels (see Figure 1). The game's backdrop mission (post-
munity participation, thus making game-based math experience mean- disaster community rebuilds) was presented via in-game task narra-
ingful. Nevertheless, solving a spatially configured math problem in a tives along with a verbal gameplay introduction. Participants then
dynamic game would be still novel and cognitively demanding for played eRebuild individually on Chromebooks for 45 min at a math
school students, requiring their ample conation in mathematically class, twice a week for 5 weeks. Teachers acted as facilitators provid-
planning, inspecting, and refining problem-solving actions. We ing encouragement and answering individuals' game-based math
conjectured that by problematizing math content through questions as needed during the gaming sessions. Students in the con-
architecture-themed inquiries while structuring the inquiry process trol group performed their conventional math class activities led by
using the in-game learner supports, eRebuild will support mathemati- their teachers, including completing math worksheets and practice
cal experiencing to activate and focus students' cognitive effort in quizzes on the similar math topics.
representing and solving the problems with mathematics. To deepen our understanding on the design and impact of game-
based math experience, in Study 2 we re-examined the research ques-
tions with eRebuild added with two structuring-oriented support fea-
3.2 | Two mixed-method studies tures aimed at motivating and structuring students' engagement in
mathematical task-solving via explicit incentive and planning aid. The
We conducted two mixed-method studies to investigate the above first feature—badges for math-related gameplay—was designed to
conjectures on the impact and nature of game-based math explicate and reward mathematical experiencing/practicing during
experiencing when students interacted with game-based math tasks task-solving (see Figures 2 and 3). For example, badges of ‘Thrifty
and learning supports. In Study 1, we used an experimental control- Shopper’ and ‘Minimalist’ awarded the math practice of ‘attending to
group pre-test post-test design to investigate the impact of using eRe- precision’ during the ‘trading’ and ‘building’ tasks. Badges of ‘Spa-
build, in comparison with a business-as-usual math class condition, on tially Aware’, ‘Built to Scale’, and ‘Efficient Mover’ encouraged the
the math problem-solving performance of middle school students. practice of mathematical reasoning in investigating and preplanning
Thirty six 8th graders, with approximately 16.7% of them not playing an optimal solution for the ‘building’ and ‘allocation’ tasks. A design
or playing digital games occasionally (i.e., a few times a month) and conjecture on these math-experiencing badges was that they would
50% being girls, were recruited from a rural public school in the stealthily direct students to self-regulate and mathematize their cogni-
Southeastern United States. They were randomly and individually tive efforts and strategies for problem-solving. The second structuring
assigned to the gaming (n = 18) and the control groups (n = 18). Both feature, a sub-goal checklist, was intended to assist students in
groups were facilitated by the same math teacher, and received a decomposing or appraising sub-goals of a multi-step complex task.
math problem-solving test before and after the study. Participants' Presented with the task narrative, the feature was expected to

F I G U R E 1 An orientation/training
level: manoeuvring 3D objects.
13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
KE ET AL.

F I G U R E 3 Support features: Sub-

An example of task

goals and badges.

FIGURE 2
narrative.
6
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
KE ET AL. 7

provide an advanced organizer and a gauge for planning and monitor- such as ‘Can you describe what you did (with this level or task)?’
ing one's problem-solving moves, thus reducing the cognitive demand ‘How do you feel?’ and ‘What did you learn?’
of solving a spatially configured contextual problem. It was hypothe-
sized that the two support features would promote participants' in-
game math task performance and the post-gaming math problem- 3.2.2 | Statistical analyses
solving test performance.
We used control-group pre-test–post-test design in Study 2 to In each study, we conducted ANCOVA tests to examine the impact of
investigate the impact of using the revised game-based experience plat- game-based math experience, in comparison with a control condition,
form (with the added structuring supports), in comparison with a on participants' math problem-solving test performance. The game-
business-as-usual math class condition, on the math problem-solving task performance data were delineated from the participants' game
performance of middle school students. Both quantitative and qualita- logs for a descriptive statistical analysis. Descriptive statistics and
tive data were collected from 65 middle school students who were exploratory association analyses were conducted to examine the
recruited from a urban public school in the Southeastern United States potential association between the game design features, the partici-
and randomly assigned to the gaming (n = 39) and the control groups pants' in-game task performance, and their post-gaming math
(n = 26). Approximately 35.4% of the participants were non- or casual- problem-solving test performance. All statistical analyses were per-
gamers, and 40% being girls. The study procedure and the data collec- formed using SPSS v.25.0.
tion process were similar to those of Study 1. We conducted qualitative thematic analysis with the observation
notes and interview responses to delineate salient themes governing
players' game-based problem-solving processes in relation to the in-
3.2.1 | Instruments and qualitative data collection game task and support features. These qualitative themes defined the
nature of commonly observed and unique behavioural patterns, along
In both studies, math problem-solving performance was measured with the critical properties of each salient pattern (e.g., actions, objects
externally via a math problem-solving test before and instantly after engaged with, and relevance to math problem-solving). We conducted
the intervention. The test items and the corresponding scoring scale preliminary open coding independently, before having iterative group
were developed based on sample items from the state's math stan- and debriefing discussions to congregate codes and resolve the disagree-
dard tests, validated by a panel of math educators and measurement ments. After reaching 100% interrater agreement on the thematic codes,
experts, and infield tested in prior research (Ke, 2019). The test evalu- we then collaboratively classified and defined the relations among
ates middle-school students' application of the targeted math knowl- themes to delineate convergent and deviating patterns emerging from a
edge (e.g., ratio/proportion reasoning, area/volume calculation) for comparative analysis with the data across participants and studies. These
math problem-solving. The test contains two equivalent forms (for the patterns were interpreted and reported with the support of observation
pre- and post-tests), with each form containing 10 multi-step, math notes and interview quotes. The qualitative analysis results supplemen-
word problems. Each item is 3-point scale (with the full test score ted and explained the statistical analysis findings. They helped to illus-
being 30) and scored by two trained graders independently based on trate how math experiences would come about during participants'
a preset grading rubric (see Appendix). The inter-rater reliability was interactions with game-based math tasks and support features.
0.97. The two graders then met and discussed on the discrepancy to
achieve 100% agreement, and the final scores were used in the data
analyses. The average Cronbach's alpha of the pre- and post-tests is 4 | RE SU LT S OF ST U D Y 1
0.61 and 0.69 respectively in Studies 1–2. Participants' game-based
math task performance, including the frequency and time stamp of 4.1 | Impact of game-based math experience
successful and failed task trials, the frequency, accuracy, and duration
of each game act performed, and the user inputs with interactive scaf- Descriptive statistics of the study groups' pre- and post-test perfor-
folds, was logged automatically and analysed as a measure of the in- mance are presented in Table 1. An ANCOVA test was conducted to
game math problem-solving performance. examine the effect of the study grouping on the post-intervention
We did onsite observation of all study sessions and interviewed test performance, with the pre-test as the covariate. The testing of
gaming participants in focus groups during and at the end of weekly homogeneity of variances for the ANCOVA assumption was satisfied.
study sessions to examine their experiences and perspectives with The ANCOVA analysis result indicated that the gaming group
game-based math experiencing and problem-solving. Both observa- (M = 11.29, SD = 5.58, n = 14) significantly outperformed the con-
tion and interviewing were semi-structured. The observation focused trol group (M = 7.14, SD = 3.65, n = 11) in the post-intervention
on participants' gaming behaviours and reactions when tackling game- math problem-solving test with a large effect size, F(1, 22) = 4.33,
based math tasks and sub-tasks, along with their talk-alouds or con- p < 0.05, partial η2 = 0.17. This finding suggested a positive impact of
versations during gameplay. The interview questions were open- game-based math experiencing, in comparison with a business-
ended and aimed at delineating participants' perceptions of their as-usual control condition, on the math problem-solving test perfor-
experiences, strategies undertaken, and potential insights developed, mance of the study participants.
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 KE ET AL.

TABLE 1 Descriptive statistics of Study 1. (a) whether they were goal-driven—navigating and interacting with

Math test Pre-test M (SD), n Post-test M (SD), n a the game objects in line with a given objective in the task narrative;
(b) how timely and carefully they attended to and processed their sur-
Control group 10.63 (3.58), 18 7.14 (3.65), 11
roundings in the game world; and (c) to what extent they appraised
Gaming group 9.22 (3.04), 18 11.29 (5.58), 14
and refined in-game acts in response to the environmental feedback.
a
Certain participants failed to attend the post-test, resulting in attrition. As depicted in Figure 2, a task narrative outlined the overarching
mission and salient task rules. Over 57% of the gaming participants
portrayed purposefulness in seeking and identifying task-relevant
A paired t-test with the gaming participants' pre- and post-test objects and localities distributed in the game world after reading the
performance did not indicate a significant result, though there was a task narrative. The others (43%), however, was found clicking around
numerical increase with a small effect size, t(13) = 1.05, p > 0.05, the screen randomly or engaging in playful moves (e.g., checking back-
Cohen's d = 0.28. Correspondingly, the gaming participants self- ground props and characters). These ‘wandering-around’ players
reported a lower level of self-perceived learning (2.57 out of 5, by showed passivity or difficulty with processing a context-rich task, fre-
average) than that of enjoyment (4 out of 5) with gameplay, and a low quently requesting peers or their teacher to explain task goals ver-
level of perceived challenge (2.14 out of 5). It should be noted, how- bally, ‘What am I supposed to do?’
ever, that the relatively small sample size and the lack of statistical Over 50% of the gaming participants failed to identify or inspect
power for the paired t-test may have failed to detect the effect of spatially distributed task parameters. They rarely fixed their mice over
gaming participation. game objects, thus failing to delineate the object-associated informa-
The gaming participants showed much variability in their game- tion. Less than 50% of the participants voluntarily used the in-game
based math task performance, with the success rate ranging from mini-map to gauge the location or orientation of the target construc-
21.4% to 60% (M = 46.7%, SD = 10.7%). The gaming participants, by tion. They appeared to rely on coarse estimates and visual feedback
average, attempted 24.5 (SD = 11.8) game-based math tasks (e.g., randomly assigning resident characters to an enclosed 3D space
(or levels) and completed 12.5 (SD = 5.6) tasks over the study ses- to test their spatial compatibility), rather than processing and comput-
sions. The correlation analysis between the number of game levels ing with math notations associated with the game objects or charac-
tried and the gaming participants' post-gaming math test performance ters (e.g., the math properties of a block, the resident space need).
indicated a positive association, r(14) = 0.45, p = 0.10 (significant for Less than 30% of the gaming participants portrayed volitional acts of
one-tailed test). discerning the in-game task parameters (e.g., using an in-game ruler to
measure the geometric properties of a task-relevant object).
A cross-participant comparative analysis with the game logs indi-
4.2 | Learner interactions with game-based math cated that among gaming participants who steadily progressed
tasks and environment through the game levels, approximately 50% failed to instantly refine
their gaming actions in response to the environmental feedback
The qualitative analysis with the onsite observation, game log, and (e.g., visual or written feedback upon an inappropriate act), especially
participants' interview data indicated two thematic patterns on partic- when a failed attempt involved multiple potential contingencies. Take
ipants' experiencing of the game-based math tasks. First, participants a ‘panelized construction’ task (i.e., composing prefabricated building
were found actively interacting with distributed game objects, but components into a structure) for example. The task encompasses mul-
they lack demonstrated effort in structuring or appraising the tiple sub-goals, including computing the size and amount of building
interactions, thus lacking a full noticing of environmental feedback to components needed for the construction, sequencing ‘join’ and ‘copy’
interpret how the surroundings react to their actions during gameplay. moves, and calculating the construction efficiency in terms of mate-
Second, participants portrayed fixedness in their problem-solving actions: rials and moves used. Approximately two-thirds of the gaming partici-
they were inept in generating or testing alternative problem-solving pants were found repeating an inefficient or erroneous construction
strategies when the environmental feedback contradicted their preex- sequence, such as initializing the construction without acquiring suffi-
isting ideas. Once a student resorted to trial-and-error and chanciness cient building materials. Instead of processing the post-action in-game
during a task trial, it appeared difficult for him/her to shift to the act feedback to trouble-shoot and refine problem-solving actions, they
of systematic testing and estimation. attributed the failure to a game bug (‘The join (or copy) tool does not
work!’) or sought instructions from others, ‘How many times did I
copy?’ or ‘Do I have enough bricks? Should I buy more?’
4.2.1 | Variability in structuring interactions with Correspondingly, participants differed in terms of the logged
game-based math environment efficiency and accuracy in task-solving, such as the number of trial-
and-error associated with each core game action (e.g., trading, con-
Consistent with the quantitative finding on the learner variability in struction, or allocation), duration of time spent for each action, and the
game-based math task performance, participants demonstrated beha- precision of the construction (e.g., accuracy in the orientation, position,
vioural differences in interacting with these tasks, with respect to and size of a building). Prominently, the participants who portrayed an
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
KE ET AL. 9

above-average level of efficiency and accuracy in the logged task per- TABLE 2 Descriptive statistics of Study 2.
formance, indicative of a higher level of action-outcome contingency Math test Pre-test M (SD), na Post-test M (SD), na
appraisal, acquired a numerically higher gain score in their post-
Control group 4.50 (5.05), 25 4.68 (3.13), 25
gaming math problem-solving test than the others.
Gaming group 5.10 (2.66), 39 7.23 (2.71), 35
a
Certain participants failed to attend the pre-test and/or the post-test,
resulting in attrition.
4.2.2 | Fixedness with game-based problem-solving

Another emerging pattern governing participants' game-based math predisposition and fixedness to ‘casual’ gameplay in the initial
experiencing is an indication of fixedness in how they framed and gaming-sessions, then gradually shifted to context-processing and
solved a game-based math problem. During the pre-gaming survey, strategic gameplay through practice.
approximately 67% of participants self-reported playing digital games
daily or weekly. Sandbox (e.g., Minecraft) and simple mobile
games that highlight free-and-casual play with self-set goals were 5 | RE SU LT S OF ST U D Y 2
reported as the most frequent type. These participants' predisposition
to undertaking casual-play was observed in their interactions with 5.1 | Impact of the game-based math experience
eRebuild. Random trial-and-error with chanciness was the most fre-
quently observed gameplay strategy among the participants. It Descriptive statistics of the pre- and post-test results of the two study
appeared that they fell back into the prior gaming habit and assumed groups are presented in Table 2. Both groups showed a low level of
a mental state of ‘least effort + luck test’ with game-based problem- prior math competency. An ANCOVA test was conducted to examine
solving. Upon a failed task-trial, these participants would retry the the study group difference in the post-intervention math test perfor-
task with the similar sequence or the same pattern of moves. mance, with the pre-test performance as the covariate. The testing of
The game logs showed that it took an average of 5.4 failed attempts homogeneity of variances for the ANCOVA assumption was satisfied.
per task for the participants to refine their in-game moves or The ANCOVA analysis result was significant: the gaming group
strategies. (M = 7.23, SD = 2.71, n = 35) outperformed the control
Notably, the portrayal of fixedness in the participants' game logs group (M = 4.68, SD = 3.13, n = 25) with a large effect size, F(1, 57)
and observed behaviours declined when the gaming sessions pro- = 15.91, p < 0.001, partial η2 = 0.22.
ceeded. It seemed that through practice, the participants started to The paired sample t-test also indicated a significant improvement
develop proactivity and mindfulness with contextual problem-solving: with a large effect size of the gaming participants in the math
they portrayed increasing acts of inspecting distributed task-relevant problem-solving test performance from the pre- to the post-test, t
objects, attending to environmental feedback, and adjusting their in- (34) = 5.00, p < 0.001, Cohen's d = 0.85. The gaming participants
game moves accordingly. This observation was echoed by the partici- reported moderate level of self-perceived learning from gameplay and
pants' comments on what they learned from playing eRebuild, such perceived challenge (both being 2.83 out of 5) and a medium-high
as: ‘how to be more precise when placing and measuring,’ ‘how to level of enjoyment (3.58 out of 5). The study findings, overall, indi-
place or allocate objects and how to compose them in a certain way,’ cated a positive impact and an advantage of game-based math
‘the shapes and sensitivity,’ and ‘spending credits wisely.’ Neverthe- experiencing, in comparison with a business-as-usual control condi-
less, approximately 10% of the participants also showed fatigue or tion, on the math problem-solving test performance of middle school
stress with the cognition-inducing game tasks; they frequently students.
requested short breaks during gameplay. A multiple linear regression analysis indicated a collective significant
effect of the pre-test math test performance, the game levels com-
pleted, and the number of math-gameplay badges acquired in predicting
4.3 | Summary the gain score of the math problem-solving test, F(3, 30) = 4.31,
p = 0.01, R2 = 0.30. The individual predictors were then examined fur-
The Study 1 findings indicated that participants of game-based math ther, with pre-gaming test performance being a significant predictor in
experiencing in the math classroom outperformed those in a business- the model, t = 3.11, p < 0.01. Notably, a lower pre-gaming math test
as-usual control condition in a post-intervention math problem- predicted a higher gain score in the post-gaming math test performance.
solving test. However, the effect size of the pre-test–post-test The results with the variables of the game task performance (i.e., the
improvement was small. Qualitative findings suggested that not all number of game tasks completed and the number of math-gameplay
gaming participants purposefully planned or appraised their interac- badges acquired) as predictors were not statistically significant, but
tions with the spatially distributed task parameters. These participants there was a mild trend toward significance for the number of badges
differed in whether they noticed the in-game mathematical situation, acquired as a predictor: t = 1.63, p = 0.1, 95% CI [ 0.04, 0.39]. The
or how efficiently and accurately they refined their problem-solving number of math-gameplay badges acquired tended to predict the gam-
moves in response to environmental feedback. They demonstrated ing participants' gain scores in the math problem-solving test
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
10 KE ET AL.

performance in a positive way. In agreement, a bivariate Pearson's cor- trading moves drop from 6.35 times per task-trial in Study 1 to 0.16
relation coefficient also indicated a significant and positive association times in Study 2. The participants were found vigilant in: measuring
between the number of badges acquired and the gain score of the math the dimensions of 3D game objects using the in-game ruler; appraising
problem-solving test, r(32) = 0.43, p = 0.02. the locality, shape, and size of building components in relation to the
Gaming participants of Study 2 shared a similar level of game construction site and the structure; and attempting alternative ways
tasks attempted (M = 24.5) as that of Study 1 gaming participants of composing prefabricated elements into the structure. These obser-
(M = 25.5). Despite a lower level of math competency demonstrated vations indicated that Study 2 participants got to investigate the geo-
in the pre-test, Study 2 participants achieved a higher game task suc- metric properties and interrelationships of game-based math objects,
cess rate (the number of tasks completed divided by the number of which helped to explain why the average completion rate of the build-
tasks tried) in Study 2 (M = 0.61, SD = 0.12, n = 35) than that in ing action by the Study 2 participants increased to 46% from 25% in
Study 1 (M = 0.53, SD = 0.11, n = 17) with a medium effect size: t Study 1.
(50) = 2.14, p < 0.05, Cohen's d = 0.69.

5.2.2 | Prevalence of qualitative math conceptions

5.2 | Learner interactions with game-based math
tasks and environment Action-driven, qualitative math conceptions were prevalent in gaming
participants' problem-solving strategies. For example, when tackling
Salient behavioural patterns governing how students approached the ‘space allocation’ task as depicted in Figure 4, multiple partici-
math tasks, given the added structuring-oriented support features in pants were found using a generic strategy that featured the process
the game-based math environment, were delineated from the qualita- of ‘arrangement’ instead of ‘allocation.’ Without computing the living
tive data and game log analyses. These patterns indicated that Study space needed for different types of residents or the space afforded by
2 participants did purposefully appraise their in-game moves in rela- each structure, they were found seeking an optimal strategy for
tion to the task sub-goals, and attend to precision via acts of mathe- arranging and fitting a combination of residents (e.g., a bigger-size
matical computing and measuring. Qualitative math conceptions— family plus a smaller one) into a given structure (e.g., a container-
informal, context-based descriptions or pre-symbolic representations converted house). A validated arrangement strategy was then
of mathematical concepts or situations—surfaced and prevailed during repeated or refined for the next structure, until all residents were
their game-based experiencing with math tasks. arranged. Such an ‘arrangement’ strategy appeared to be a mixture of
numerical knowledge and domain-generic reasoning, while lacking a
quantitative math model of the problem. It depicted a conception of
5.2.1 | Purposefulness and mathematical precision ratio (e.g., the adult-to-child ratio in terms of the space needed per
in problem-solving moves person) as a qualitative comparison (e.g., ‘bigger and smaller’) rather
than a quantity that specifies the comparative relationship (e.g., how
The ‘wandering-around’ causal-play behaviours were rarely observed much bigger or smaller). Similarly, the notion of livable floor area sur-
with the Study 2 gaming participants. Over 86% of their in-game faced as a qualitative concept—the plane that afforded (or ‘fitting in’)
moves targeted task-relevant game objects, portrayed in a sequence a limited occupancy.
aligned with the provided sub-goal checklist. These in-game moves
were coupled with observed self-checking with the on-screen, sub-
goal completion status during moves; they featured iterative refine- 5.3 | Summary
ments with the angle, distance, and shape of the target structure
before clicking on the level-completion button. Confirming these The Study 2 findings further supported a positive impact of game-
observed efforts to goal-alignment and precision for in-game moves, based math experiencing, in comparison with the business-as-usual
the game logs showed that the Study 2 participants spent more time control condition, on the post-intervention math problem-solving test
on each task trial (387 s per trial across all gaming participants, performance of middle school students. There was a significant pre-
vs. 289 s in Study 1). In consequence, they showed a higher comple- test–post-test improvement with a large effect size for the game-
tion or accuracy rate with core game actions, at an average of 46% based math-experiencing participants. The number of badges acquired
(vs. 25% in Study 1) for the ‘building’ game action and 73% (vs. 59% in the in-game math tasks was significantly, positive correlated with
in Study 1) for the ‘allocation’ action. These findings helped to explain the gain score of the math problem-solving test. The gaming partici-
a higher task success rate in Study 2. pants scored a higher in-game math-task success rate than Study
The gaming participants' problem-solving moves also portrayed a 1 peers, which was corroborated by the qualitative finding on the
stronger presence of mathematical computation and measurement. presence of purposefulness and mathematical precision in partici-
For example, we observed significantly reduced guessing behaviour pants' in-game moves. These findings suggested that adding two
and increased performance of arithmetic calculation by the gaming structuring-oriented in-game supports—badges for mathematical
participants with the ‘trading’ action, making the average erroneous moves and a sub-goal check list—appeared to enhance participants'
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
KE ET AL. 11

F I G U R E 4 A game-based math task

featuring the ‘allocation’ action.

planning of and self-regulated efforts to math task- and precision- adding structuring-oriented in-game supports, can promote math
oriented in-game moves, thus promoting their experiencing and per- problem-solving and learning as a conative, sense-making experi-
formance of in-game and post-intervention math problem-solving. ence (Barzilai & Blau, 2014; Holbert & Wilensky, 2014).
Nevertheless, the participants were apt to leverage qualitative math
conceptions and action-oriented conjectures when reasoning with a
game-based math task. This observation helped to explain a modest 6.2 | Cultivating game-based math experience
level of self-reported learning by the gaming participants who may
lack reflection on how the qualitative insights connect with formal Findings across the two studies suggested that game-based math
math notations. experiencing—tackling spatially configured math contextual problems in
a dynamic game world—is a mindful and self-regulated activity that
needs to be cultivated. Previously reported challenges to game-based
6 | DISCUSSION AND CONCLUSIONS domain-specific experience, such as students' lack of self-direction,
insufficient mental effort with domain-specific problem-solving, chanci-
6.1 | Promise of the game-based math experience ness, and functional fixedness (e.g., Abdul Jabbar & Felicia, 2015;
Author, 2016, 2019; Baker et al., 2010; Nebel et al., 2016; Rowe
The findings of Studies 1 and 2 suggest the feasibility and promise et al., 2009), were observed in Study 1.
of game-based math experience in promoting math contextual In Study 2, gaming participants have portrayed purposefulness
problem-solving for middle school students. The study findings indi- and mathematical precision in their in-game problem-solving moves,
cate that game-based math experiencing in math classroom, in com- which corresponds with the improved in-game math-task success rate
parison with the business-as-usual control condition, promotes and a larger effect size in the post-intervention math problem-solving
students' post-intervention math problem-solving test performance. test-performance improvement. These findings provide preliminary
There is a trend of the pre-test–post-test performance improvement evidence for the benefit of structuring-oriented supports—subgoaling
for the gaming participants, from a numerical difference with a small and cognition-inducing incentive (badges)—in cultivating the partici-
effect size in Study 1 to a statistical difference with a large effect pants' conation for mathematical problem-solving and experiencing.
size in Study 2. Students with a lower-level math prior knowledge On the other hand, these support features appear to fall short of
have achieved a higher gain score in the post-gaming math test per- advancing participants' qualitative math insights to formal math nota-
formance. The number of game badges acquired, indicative of the tions when interacting with math contextual problems. Alternative
game-based math task performance, is significantly and positively scaffolding strategies that enhance the connection and coevolution
correlated with the gain score of the math problem-solving test. This between informal and formal math knowledge during problem-
finding supports the design proposition that purposively-designed oriented math experiencing are warranted.
in-game incentives, by acknowledging mathematical in-game moves
or problem-solving strategies, should help to structure students'
mental efforts and foster their conation with math experiencing dur- 6.3 | Implications
ing gameplay (Puntambekar & Hubscher, 2005; Reiser, 2004; Seijts
et al., 1997). The study findings, overall, support the conjecture that The current research findings suggest that by simulating and structur-
a game-based multimodal math environment, through problematiz- ing architecture-themed math contextual problems, a game-based
ing math experience as architecture-themed design problems while learning environment can challenge and improve the problem-solving
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
12 KE ET AL.

practice of school students who are used to conventional math word that better channel and upscale the evolved qualitative conceptions to the
problems, and make them proactively collect, process, and coordinate construction and awareness of formal mathematical models for problem-
problem parameters (or information elements) represented in different solving. Alternative support strategies, such as collaborative gameplay,
modalities and distributed in an interactive environment. It is observed teacher-facilitated debriefing, and reflective demo of expert play, can be
that students will learn to be goal-driven in planning problem-solving examined for game-based math experiencing in a future study.
moves, noticing environmental feedback, and purposefully appraising The lack of statistical power for the pairwise t-test in Study
and refining each move in response to the environmental feedback. 1 may have failed to detect the impact of gaming participation. Future
Teachers could integrate and facilitate game-based math experiencing research should build on and replicate the current studies in alterna-
as part of classroom learning activities. The game-based learning envi- tive implementation settings with a larger sample size. The Cronbach's
ronment may also act as a congenial mind tool that trains students on alpha values of the current math problem-solving test, though accept-
basic causal reasoning—context processing and perceiving contingency able, should be improved in future research by increasing the number
between action and outcome. of test items while adjusting the items' difficulty levels.
Students' experiencing of problem-oriented math masks, as the
current research illustrated, differs and is contingent on their conation AUTHOR CONTRIBU TIONS
with respect to the tasks. This finding is in alignment with the perspec- Fengfeng Ke: Conceptualization; data curation; formal analysis; funding
tive of Johnson et al. (2017) that there is incongruence between the acquisition; investigation; methodology; project administration; resources;
student perspective and the designer's original intent with a learning software; supervision; validation; visualization; writing – original draft;
task. Our findings support the previous report of game-based learning writing – review and editing. Chih-Pu Dai: Formal analysis; investigation;
that the motivation to complete the game (rather than learning from visualization; writing – review and editing. Luke West: Formal analysis;
the game) has a deleterious effect on acquiring and practicing disciplin- investigation; visualization; writing – review and editing.
ary frameworks in game-based problem-solving (Israel-Fishelson &
Hershkovitz, 2020; Nebel et al., 2016). A salient implication for creating FUNDING INF ORMATI ON
game-based math experience, therefore, is to nurture and enhance stu- This material is based upon work supported by the National Science
dents' conation in developing and endorsing mathematical point of Foundation under grant #1720533.
view when tackling a novel task or problem.
Students may constantly engage in a cost-utility analysis before CONFLIC T OF INTER E ST STATEMENT
selecting or endorsing the domain-specific tool, strategy, or frame- The authors declare no conflicts of interest.
work when interacting with the digital math tasks and environment. In
the current research, the cognition inducing and rewarding incentives, PE ER RE VIEW
such as badges that reward mathematical precision, help to escalate The peer review history for this article is available at https://www.
the perceived utility of mathematical strategies for problem-solving. webofscience.com/api/gateway/wos/peer-review/10.1111/jcal.
The sub-goal check list, by chunking and explicating the multi-step 12938.
problem-solving process, can reduce the cognitive cost for planning
and inspecting problem-solving moves and that for exploring the DATA AVAILABILITY STAT EMEN T
underlying structure of a math contextual problem. Moreover, The data generated and analysed in the current research are not pub-
the two support features are integrated into the game mechanics and licly available due to privacy/ethical restrictions, though the support-
the game world respectively, making the participants' support usage a ing data is available from the corresponding author on reasonable
natural part of game-based math experiencing. A resultant proposition request and following the university IRB policy.
for math experience design, therefore, is to craft and follow a cost–
benefit principle in designing both problematized learning tasks and IRB APPROVAL
structuring-oriented supports in relation to students' conation. This The current research was approved by the Florida State University
design proposition echoes the argument of Ainley et al. (2006) that Institutional Review Board (IRB) committees.
we should clearly define the purpose and utility of a task or domain-
specific problem-solving strategy from the perspective of the student OR CID
rather than the task designer. Fengfeng Ke https://orcid.org/0000-0003-4203-1203

RE FE RE NCE S
6.3.1 | Limitations and future research Abassian, A., Safi, F., Bush, S., & Bostic, J. (2020). Five different perspec-
tives on mathematical modeling in mathematics education. Investiga-
tions in Mathematics Learning, 12(1), 53–65. https://doi.org/10.1080/
Nevertheless, qualitative mathematical conceptions are still dominant in
19477503.2019.1595360
the study participants' game-based math problem-solving strategies; their Abdul Jabbar, A. I., & Felicia, P. (2015). Gameplay engagement and learning in
self-perceived learning is relatively low. Future research should further game-based learning: A systematic review. Review of Educational Research,
explore the design and support features of game-based math experiencing 85(4), 740–779. https://doi.org/10.3102/0034654315577210
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
KE ET AL. 13

Ainley, J., Pratt, D., & Hansen, A. (2006). Connecting engagement and Habgood, M. J., & Ainsworth, S. E. (2011). Motivating children to learn
focus in pedagogic task design. British Educational Research Journal, effectively: Exploring the value of intrinsic integration in educational
32(1), 23–38. https://doi.org/10.1080/01411920500401971 games. The Journal of the Learning Sciences, 20(2), 169–206. https://
Baker, R. S., D'Mello, S. K., Rodrigo, M. M. T., & Graesser, A. C. (2010). Bet- doi.org/10.1080/10508406.2010.508029
ter to be frustrated than bored: The incidence, persistence, and impact Harel, I. E., & Papert, S. E. (1991). Constructionism. Ablex Publishing.
of learners' cognitive–affective states during interactions with three Healy, L., & Kynigos, C. (2010). Charting the microworld territory over
different computer-based learning environments. International Journal time: Design and construction in mathematics education. ZDM Mathe-
of Human-Computer Studies, 68(4), 223–241. https://doi.org/10.1016/ matics Education, 42(1), 63–76. https://doi.org/10.1007/s11858-009-
j.ijhcs.2009.12.003 0193-5
Barnes, A. (2019). Perseverance in mathematical reasoning: The role of Hiebert, J., Carpenter, T. P., Fennema, E., Fuson, K., Human, P., Murray, H.,
children's conative focus in the productive interplay between cogni- Olivier, A., & Wearne, D. (1997). Making mathematics problematic: A
tion and affect. Research in Mathematics Education, 21(3), 271–294. rejoinder to Prawat and Smith. Educational Researcher, 26(2), 24–26.
https://doi.org/10.1080/14794802.2019.1590229 https://doi.org/10.3102/0013189X026002024
Barzilai, S., & Blau, I. (2014). Scaffolding game-based learning: Impact on Hmelo-Silver, C. E. (2004). Problem-based learning: What and how do stu-
learning achievements, perceived learning, and game experiences. dents learn? Educational Psychology Review, 16(3), 235–266. https://
Computers & Education, 70, 65–79. https://doi.org/10.1016/j. doi.org/10.1023/B:EDPR.0000034022.16470.f3
compedu.2013.08.003 Holbert, N. R., & Wilensky, U. (2014). Constructible authentic repre-
Blum, W., & Ferri, R. B. (2009). Mathematical modelling: Can it be taught sentations: Designing video games that enable players to utilize
and learnt? Journal of Mathematical Modelling and Application, 1(1), knowledge developed in-game to reason about science. Technol-
45–58. ogy, Knowledge and Learning, 19(1–2), 53–79. https://doi.org/10.
Blumenfeld, P. C., Soloway, E., Marx, R. W., Krajcik, J. S., Guzdial, M., & 1007/s10758-014-9214-8
Palincsar, A. (1991). Motivating project-based learning: Sustaining the Hoyles, C., & Noss, R. (1992). A pedagogy for mathematical microworlds.
doing, supporting the learning. Educational Psychologist, 26(3–4), 369– Educational Studies in Mathematics, 23(1), 31–57. https://doi.org/10.
398. https://doi.org/10.1080/00461520.1991.9653139 1007/BF00302313
Brezovszky, B., McMullen, J., Veermans, K., Hannula-Sormunen, M. M., Huitt, W., & Cain, S. (2005). An overview of the conative domain. Educa-
Rodríguez-Aflecht, G., Pongsakdi, N., Laakkonen, E., & Lehtinen, E. tional psychology interactive. http://www.edpsycinteractive.org/
(2019). Effects of a mathematics game-based learning environment on brilstar/chapters/conative.pdf
primary school students' adaptive number knowledge. Computers & Edu- Israel-Fishelson, R., & Hershkovitz, A. (2020). Persistence in a game-based
cation, 128, 63–74. https://doi.org/10.1016/j.compedu.2018.09.011 learning environment: The case of elementary school students learning
Chang, K. E., Wu, L. J., Weng, S. E., & Sung, Y. T. (2012). Embedding game- computational thinking. Journal of Educational Computing Research,
based problem-solving phase into problem-posing system for mathe- 58(5), 891–918. https://doi.org/10.1177/0735633119887187
matics learning. Computers & Education, 58(2), 775–786. https://doi. Johnson, H. L., Coles, A., & Clarke, D. (2017). Mathematical tasks and the
org/10.1016/j.compedu.2011.10.002 student: Navigating “tensions of intentions” between designers,
Charsky, D., & Ressler, W. (2011). “Games are made for fun”: Lessons on teachers, and students. ZDM Mathematics Education, 49, 813–822.
the effects of concept maps in the classroom use of computer games. https://doi.org/10.1007/s11858-017-0894-0
Computers & Education, 56(3), 604–615. https://doi.org/10.1016/j. Jurow, A. S. (2005). Shifting engagements in figured worlds: Middle school
compedu.2010.10.001 mathematics students' participation in an architectural design project.
Clark, R. E., Howard, K., & Early, S. (2006). Motivational challenges experi- The Journal of the Learning Sciences, 14(1), 35–67. https://doi.org/10.
enced in highly complex learning environments. In J. Elen & R. E. Clark 1207/s15327809jls1401_3
(Eds.), Handling complexity in learning environments: Research and theory Kafai, Y. B. (1995). Minds in play: Computer game design as a context for chil-
(pp. 27–43). G.B. Elsevier Science Ltd. dren's learning. Routledge.
Davis, P. J., Hersh, R., & Marchisotto, E. A. (2012). The mathematical experi- Kaiser, G., Schwarz, B., & Buchholtz, N. (2011). Authentic modelling prob-
ence, study edition. Birkhäuser Boston. lems in mathematics education. In G. Kaiser, W. Blum, R. Borromeo
Edwards, L. D. (1991). Children's learning in a computer microworld for Ferri, & G. Stillman (Eds.), Trends in teaching and learning of mathemati-
transformation geometry. Journal for Research in Mathematics Education, cal modelling, ICTMA 14 (Vol. 1, pp. 591–601). Springer Science & Busi-
22(2), 122–137. https://doi.org/10.5951/jresematheduc.22.2.0122 ness Media.
Engle, R. A., & Conant, F. R. (2002). Guiding principles for fostering pro- Kalyuga, S., & Singh, A. M. (2016). Rethinking the boundaries of cognitive
ductive disciplinary engagement: Explaining an emergent argument in load theory in complex learning. Educational Psychology Review, 28(4),
a community of learners classroom. Cognition and Instruction, 20(4), 831–852. https://doi.org/10.1007/s10648-015-9352-0
399–483. https://doi.org/10.1207/S1532690XCI2004_1 Kapur, M. (2011). A further study of productive failure in mathematical
Ertmer, P. A., & Glazewski, K. D. (2019). Scaffolding in PBL environments: problem solving: Unpacking the design components. Instructional Sci-
Structuring and problematizing relevant task features. In M. Moallem, ence, 39(4), 561–579. https://doi.org/10.1007/s11251-010-9144-3
W. Hung, & N. Dabbagh (Eds.), The Wiley handbook of problem-based Kapur, M. (2014). Comparing learning from productive failure and vicari-
learning (pp. 321–342). John Wiley & Sons, Inc. ous failure. Journal of the Learning Sciences, 23(4), 651–677. https://
Freudenthal, H. (1991). Revisiting mathematics education. Kluwer Academic doi.org/10.1080/10508406.2013.819000
Publishers. Kapur, M. (2016). Examining productive failure, productive success, unpro-
Garris, R., Ahlers, R., & Driskell, J. E. (2002). Games, motivation, and learn- ductive failure, and unproductive success in learning. Educational Psy-
ing: A research and practice model. Simulation & Gaming, 33(4), 441– chologist, 51(2), 289–299. https://doi.org/10.1080/00461520.2016.
467. https://doi.org/10.1177/1046878102238607 1155457
Goldin, G. A. (2019). Exploring a conative perspective on mathematical Ke, F. (2008). Computer games application within alternative classroom
engagement. In S. Chamberlin & B. Sriraman (Eds.), Affect in mathemat- goal structures: Cognitive, metacognitive, and affective evaluation.
ical modeling (pp. 111–129). Springer. Educational Technology Research and Development, 56(5-6), 539–556.
Grootenboer, P., & Edwards-Groves, C. (2019). Learning mathematics as https://doi.org/10.1007/s11423-008-9086-5
being stirred into mathematical practices: An alternative perspective Ke, F. (2016). Designing and integrating purposeful learning in game play:
on identity formation. ZDM Mathematics Education, 51(3), 433–444. A systematic review. Educational Technology Research and Develop-
https://doi.org/10.1007/s11858-018-01017-5 ment, 64(2), 219–244. https://doi.org/10.1007/s11423-015-9418-1
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
14 KE ET AL.

Ke, F. (2019). Mathematical problem solving and learning in an architec- Craig & D. Dicheva (Eds.), Proceedings of 14th international conference
ture-themed epistemic game. Educational Technology Research and on artificial intelligence in education (pp. 99–106). Springer.
Development, 67(5), 1085–1104. https://doi.org/10.1007/s11423- Schmidt, R. A., & Bjork, R. A. (1992). New conceptualizations of practice:
018-09643-2 Common principles in three paradigms suggest new concepts for train-
Ke, F., & Clark, K. M. (2020). Game-based multimodal representations and ing. Psychological Science, 3(4), 207–218. https://doi.org/10.1111/j.
mathematical problem solving. International Journal of Science and 1467-9280.1992.tb00029.x
Mathematics Education, 18(1), 103–122. https://doi.org/10.1007/ Schoenfeld, A. H. (2020). Mathematical practices, in theory and practice.
s10763-018-9938-3 ZDM Mathematics Education, 52, 1163–1175. https://doi.org/10.
Ke, F., & Im, T. (2014). A case study on collective cognition and operation 1007/s11858-020-01162-w
in team-based computer game design by middle-school children. Inter- Seijts, G. H., Meertens, R. M., & Kok, G. (1997). The effects of task impor-
national Journal of Technology and Design Education, 24(2), 187–201. tance and publicness on the relation between goal difficulty and per-
https://doi.org/10.1007/s10798-013-9248-6 formance. Canadian Journal of Behavioural Science, 29(1), 54–62.
Ke, F., Shute, V., Clark, K. M., & Erlebacher, G. (2019). Interdisciplinary https://doi.org/10.1037/0008-400X.29.1.54
design of game-based learning platforms. Springer. https://doi.org/10. Shaffer, D. W. (2004). Pedagogical praxis: The professions as models for
1007/978-3-030-04339-1 postindustrial education. Teachers College Record, 106(7), 1401–1421.
Kertil, M., & Gurel, C. (2016). Mathematical modeling: Abridge to STEM Shaffer, D. W. (2007). Learning in design. In R. A. Lesh, E. Hamilton, & J. J.
education. International Journal of Education in Mathematics, Science Kaput (Eds.), Foundations for the future in mathematics education
and Technology, 4(1), 44–55. (pp. 99–126). Routledge.
Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance Shute, V. J., Smith, G., Kuba, R., Dai, C. P., Rahimi, S., Liu, Z., & Almond, R.
during instruction does not work: An analysis of the failure of con- (2020). The design, development, and testing of learning supports for
structivist, discovery, problem-based, experiential, and inquiry-based the physics playground game. International Journal of Artificial Intelligence
teaching. Educational Psychologist, 41(2), 75–86. Education, 31, 357–379. https://doi.org/10.1007/s40593-020-00196-1
Kolb, D. A., Boyatzis, R. E., & Mainemelis, C. (2014). Experiential learning Silver, E. A. (Ed.). (2013). Teaching and learning mathematical problem solv-
theory: Previous research and new directions. In R. J. Sternberg & L. F. ing: Multiple research perspectives. Routledge.
Zhang (Eds.), Perspectives on thinking, learning, and cognitive styles Taylor, A. (2009). Linking architecture and education: Sustainable design for
(pp. 227–248). Routledge. learning environments. UNM Press.
Kolodner, J. (2014). Case-based reasoning. Morgan Kaufmann. Tokac, U., Novak, E., & Thompson, C. G. (2019). Effects of game-based
Kolodner, J. L., Camp, P. J., Crismond, D., Fasse, B., Gray, J., Holbrook, J., learning on students' mathematics achievement: A meta-analysis. Jour-
Puntambekar, S., & Ryan, M. (2003). Problem-based learning meets nal of Computer Assisted Learning, 35(3), 407–420. https://doi.org/10.
case-based reasoning in the middle-school science classroom: Putting 1111/jcal.12347
learning by design(tm) into practice. The Journal of the Learning Sciences, Van den Heuvel-Panhuizen, M., & Drijvers, P. (2020). Realistic mathematics
12(4), 495–547. https://doi.org/10.1207/S15327809JLS1204_2 education. In Lerman, S. (Ed.), Encyclopedia of mathematics education
Mayer, R. E. (2004). Should there be a three-strikes rule against pure dis- (pp. 713–717). Springer. https://doi.org/10.1007/978-3-030-15789-
covery learning? American Psychologist, 59(1), 14–19. https://doi.org/ 0_170
10.1037/0003-066X.59.1.14 Van Merriënboer, J. J., & Sweller, J. (2005). Cognitive load theory and com-
Nachtigall, V., Serova, K., & Rummel, N. (2020). When failure fails to be plex learning: Recent developments and future directions. Educational
productive: Probing the effectiveness of productive failure for learning Psychology Review, 17(2), 147–177. https://doi.org/10.1007/s10648-
beyond STEM domains. Instructional Science, 48, 651–657. https:// 005-3951-0
doi.org/10.1007/s11251-020-09525-2 Van Merriënboer, J. J. G., & Kester, L. (2014). The four-component instruc-
National Research Council. (2011). Successful K-12 STEM education: Identi- tional design model: Multimedia principles in environments for com-
fying effective approaches in science, technology, engineering, and mathe- plex learning. In R. E. Mayer (Ed.), Cambridge handbooks in psychology.
matics. The National Academies Press. The Cambridge handbook of multimedia learning (pp. 104–148). Cam-
Nebel, S., Schneider, S., & Rey, G. D. (2016). Mining learning and crafting bridge University Press.
scientific experiments: A literature review on the use of Minecraft in Wang, S. Y., Chang, S. C., Hwang, G. J., & Chen, P. Y. (2018). A
education and research. Educational Technology & Society, 19(2), microworld-based role-playing game development approach to engag-
355–366. ing students in interactive, enjoyable, and effective mathematics learn-
Pea, R. D. (2004). The social and technological dimensions of scaffolding ing. Interactive Learning Environments, 26(3), 411–423. https://doi.org/
and related theoretical concepts for learning, education, and human 10.1080/10494820.2017.1337038
activity. The Journal of the Learning Sciences, 13(3), 423–451. Wenger, E. (1999). Communities of practice: Learning, meaning, and identity.
Piaget, J. (1976). Piaget's theory. In B. Inhelder, H. H. Chipman, & C. Cambridge University Press.
Zwingmann (Eds.), Piaget and his school (pp. 11–23). Springer. Wouters, P., & Van Oostendorp, H. (2013). A meta-analytic review of the
Pintrich, P. R., & Schunk, D. H. (2002). Motivation in education: Theory, role of instructional support in game-based learning. Computers & Educa-
research, and applications. Prentice Hall. tion, 60(1), 412–425. https://doi.org/10.1016/j.compedu.2012.07.018
Puntambekar, S., & Hubscher, R. (2005). Tools for scaffolding students in a Zbiek, R. M., & Conner, A. (2006). Beyond motivation: Exploring mathe-
complex learning environment: What have we gained and what have matical modeling as a context for deepening students' understandings
we missed? Educational Psychologist, 40(1), 1–12. https://doi.org/10. of curricular mathematics. Educational Studies in Mathematics, 63(1),
1207/s15326985ep4001_1 89–112. https://doi.org/10.1007/s10649-005-9002-4
Reiser, B. J. (2004). Scaffolding complex learning: The mechanisms of
structuring and problematizing student work. The Journal of the Learn-
ing Sciences, 13(3), 273–304. How to cite this article: Ke, F., Dai, C.-P., & West, L. (2024).
Reitan, R. M., & Wolfson, D. (2000). Conation: A neglected aspect of neu- Mathematical experience in game-based problem-solving.
ropsychological functioning. Archives of Clinical Neuropsychology, 15(5),
Journal of Computer Assisted Learning, 1–15. https://doi.org/
443–453. https://doi.org/10.1093/arclin/15.5.443
Rowe, J. P., McQuiggan, S. W., Robison, J. L., & Lester, J. C. (2009). Off- 10.1111/jcal.12938
task behavior in narrative-centered learning environments. In S.
 13652729, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jcal.12938 by Bbw Hochschule, Wiley Online Library on [06/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
KE ET AL. 15

APPENDIX Grading scale

An example of math proplem test item 3 pts: 7.94 or 8 gallons are needed because (30  18) +
Direction (9  9  3.14) = 540 + 254.34 = 794.34 = 794 square feet.

You should show your work on the page itself, and you may complete • 2 pts: when the amount of paint needed for either the rectangle
the items in any order that you wish. or two semicircular ends is calculated, 5.4 gallons or 2.54
gallons.
• 1 pts: when the area of either the rectangle or two semicircular
Question # ends is calculated, 540 square feet or 254.34 (or 254)
square feet.
A pool contractor needs to paint a floor. The floor is a rectangle with two
semicircular ends as shown in the figure. One gallon of paint can cover
100 square feet. How many gallons of paint will the contractor need to
use to paint the floor? Express your answer to the nearest gallons.