Computers and Education Open 6 (2024) 100160

Contents lists available at ScienceDirect

Computers and Education Open

journal homepage: www.sciencedirect.com/journal/computers-and-education-open

Digital game-based language learning for vocabulary development

Mahjabin Chowdhury a, *, L．Quentin Dixon b, Li-Jen Kuo c, Jonan Phillip Donaldson d,
Zohreh Eslami e, Radhika Viruru f, Wen Luo g
a
Center for Teaching Excellence, Texas A&M University, Texas, United States
b
Independent Educational Researcher, Texas, United States
c
Associate Professor of Literacy Education, Texas A&M University, Texas, United States
d
Texas A&M University Center for Teaching Excellence, Texas, United States
e
Professor of ESL Education, Texas A&M University, Texas, United States
f
Clinical Professor, Teaching, Learning and Culture, Texas A&M University, Texas, United States
g
Professor and Associate Department Head, Educational Psychology, Texas A&M University, Texas, United States

A R T I C L E I N F O A B S T R A C T

Keywords: Current literature suggests that one of the impressive benefits of digital game-based language learning is its
Games potential to create engaging vocabulary learning experiences. However, this literature is dominated by game-
Elementary education play approaches, rather than game-making approaches. Drawing upon constructivist, motivation, dual-coding,
Collaborative learning
and constructionist theories, this study examined elementary English language learners’ vocabulary learning
Pedagogical issues
Distance education and online learning
experiences in a game-based learning context and investigated their experience as they built their own games
based on their understanding of different non-fiction texts. This mixed-methods case study involved novel use of
network analysis to examine student learning, investigate the relationships between various aspects of learner
experiences, and evaluate the strengths and weaknesses of the program. Findings indicated that a complex set of
interdependencies between enactments of theoretical principles including generativity, agency, tinkering,
intrinsic motivation, and contextualized learning lead to enjoyment and powerful language learning experiences.
Design principles for future digital game-based language learning include optimizing visual aspects, spending
more time on technical skill development, increasing learner agency, and focusing more on generative and joyful
learning experiences.

Introduction The purpose of this study is to evaluate the effectiveness of a digital

Digital game-based language learning (DGBLL) could afford English
language learners as game designers a motivating opportunity to learn
the necessary language skills, giving these students greater access to
learning and success in life [1,2]. English language learners are the
fastest-growing sector of the K-12 school population, and the enroll­
ments are increasing in most parts of the country [3]. These children can
be difficult to teach due to differences in language and culture [4].
Current literature shows that vocabulary is the most salient and prac­
ticed language skill in the context of DGBLL where researchers use
readily available commercial digital games [5,6]. However, instead of
using a digital game to teach specific skills, DGBLL can also be used to
promote learning through creating games [7]. The current study
employed DGBLL with students in charge of creating their own video
games based on non-fiction texts of which they had no prior knowledge.
game-based language learning (DGBLL) workshop, focusing on gains in
vocabulary and reading comprehension. It explores learners’ experi­
ences, analyzing the interplay between game-based learning, language
learning, and affective experiences. The study examines how these ex­
periences align with learning theories like constructionism and dual
coding theory. It aims to identify strengths, weaknesses, and leverage
points in a DGBLL workshop in which students designed their own
games, providing insights for future educational technology and lan­
guage pedagogy.
When discussing digital game-based language learning (DGBLL), it is
important to understand the differences among the terminologies such
as game-based learning, serious games, digital games, digital game-
based learning, gamification, and DGBLL. Game-based learning in­
volves learning through playing and/or making games including phys­
ical or non-digital games [8]. A digital game is any game played on a

* Corresponding author.
E-mail addresses: mahjabin@tamu.edu (M. Chowdhury), lijenkuo@tamu.edu (L.-J. Kuo).

https://doi.org/10.1016/j.caeo.2024.100160
Received 27 February 2023; Received in revised form 17 December 2023; Accepted 17 January 2024
Available online 20 January 2024
2666-5573/Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
M. Chowdhury et al.
digital device such as a computer, tablet, or mobile device. Digital games
are goal-oriented, challenging, promote interactions, and are governed
by a set of rules and feedback [9]. Serious games are designed to fore­
ground specific learning objectives, often pushing game aspects such as
play, learner agency, and enjoyment into the background [10,11], and
sometimes seeking to trick students into doing things they do not wish to
do through “chocolate covered broccoli” approaches [12]. According to
Prensky [13], Digital game-based learning combines education and
games, uses a digital platform as the medium, and uses game elements to
teach a specific skill or achieve a specific learning outcome. It takes the
core content and objectives and makes the learning experience fun for
many students [14,15]. Unlike digital game-based learning, in DGBLL,
game elements specifically focus on teaching a language skill or on
reaching a specific language learning outcome. DGBLL can create an
environment where education is mostly learner-centered, provide op­
portunities for socialization when well-organized, and can awaken the
will to win and competitive desire in some social contexts [16,17]. In
contrast to game-based learning, gamification is the application of game
mechanics in a non-game context to promote desired behavior and
learning outcomes [18]. Game-based learning integrates games with the
learning process to teach a specific skill or achieve a learning aim [15,
19]. For this study, the focus will be only on digital game-based lan­
guage learning (DGBLL).
Background literature
Words are the building blocks of a language [20], and vocabulary
knowledge is a major factor underlying the achievement of reading
comprehension and successful language learning for all learners [21].
Learners develop vocabulary knowledge through systematic learning of
vocabulary– targeted vocabulary instruction along with discussion and
opportunities to learn the words in context [22] as well as through
incidental exposure to new words [23], which involves authentic tasks
such as read-aloud, repeated exposures, and multimodal interactions to
make word-form and word-meaning connections [24]. In the past
decade, researchers have found that technology can improve learners’
vocabulary development, and reading comprehension (e.g., [25,26])
and increase student motivation [27].
Different affordances of DGBLL, such as student agency, goal-
oriented learning and feedback, opportunities for collaboration, and
have the potential to increase English language learners’ vocabulary
knowledge [28,29]. DGBLL may be beneficial for developing early lit­
eracy [30]. DGBLL may also lead to successful reading comprehension
development because DGBLL aligns semiotic interrelationships among
multiple sign systems (e.g., words, images, actions, and artifacts) [31].
By creating a game, students create a shareable artifact that permits
them to convey their personal understanding of the concept in support of
their reading comprehension [32]. Despite earlier research indicating
promising results for DGBLL with vocabulary development, improving
reading comprehension via DGBLL at the elementary level has not yet
been studied using learning designs grounded in theory.
Theoretical framework
One of the core principles in the learning sciences is that learning is
complex and messy [33,34], and therefore multiple learning theories are
necessary for robust investigation of learner experiences in DGBLL
contexts. This study is based on constructivist theory, constructionist
theory, and motivation theory because these theories are dominant
theories used in game-based learning studies, and dual coding theory
because this is relevant to vocabulary development [22]. Constructivist
learning theory is grounded in the work of Piaget [35], Vygotsky [36],
and Bruner [37], who defined learning as the active construction of
knowledge and meaning. Vygotsky [36] and Bruner [37] expanded
constructivist theory from an individual cognitive process [35] to
address the constructive interaction between and across individual
2
Computers and Education Open 6 (2024) 100160
minds and their social context including peers, more knowledgeable
others, tools, languages, and history. Constructivist theory not only sees
learning as a generative process, but also views knowledge as constructed
as opposed to discovered [38].
Constructionist learning theory expands the constructivist perspec­
tive on learning [39] by placing more emphasis on the art of learning
and on the significance of learning through making [40]. By construct­
ing artifacts, activities, processes, and other observables that reflect and
embody the meaning being constructed, a positive feedback loop
emerges: the embodied construction contributes to the construction of
knowledge, which contributes to the construction of the artifacts, and so
on [41]. Learning is inextricable from context and is the result of in­
teractions between learners and features of their environments, such as
other learners, tools, resources, language, and social structures [36].
Constructionist theory provides a set of principles for the design of
learning environments and experiences. Learning is most powerful when
learners make things of their own design. The constructed artifacts
mirror the construction of meaning occurring in the minds of the
learners. The artifacts are tangible objects-to-think-with, or tools of
embodied cognition [42]. Learner agency is the key. When learners have
autonomy and authority over the goals, processes, roles, and nature of
the artifacts, those artifacts take on personal significance [43]. Addi­
tionally, learners should be situated as designers on design teams who
engage in the negotiation of goals, roles, procedures, tools, and mean­
ings. To be meaningful, artifacts must also be authentic. If learners
create artifacts that they know will only be seen just by teachers, the
construction of artifacts is akin to drill-and-practice activities. Learners
need to know that the artifacts they are creating will have a real-world
impact on authentic audiences [44]. Resnick and Rosenbaum [45] argue
for focused tinkering, which they define as “a playful, experimental,
iterative style of engagement, in which makers are continually reas­
sessing their goals, exploring new paths, and imagining new possibil­
ities” (p. 164). Situating learners as designers fosters a more focused
form of tinkering which facilitates the development of skills in framing
and reflection-in-action. Focused tinkering provides a means of
balancing the product-driven activity of artifact construction with the
joy and freedom of exploration.
In addition, constructionism views learning as both an internal and
external process [46]. The internal process includes the active process
where the students construct knowledge. The external process includes
making artifacts that they can share with an authentic audience [44].
The constructionist approach can aid language learning because con­
structing knowledge by making artifacts may help the learners build
their cognitive knowledge and language skills.
Motivation theory, particularly self-determination theory [47], fo­
cuses on the factors that engage learners in the learning process and help
to explain their learning outcomes, either intrinsically or extrinsically.
Learning is most engaging when intrinsic motivation is optimized
through a combination of learner autonomy, self-efficacy in one’s
competence, and a sense of relatedness with others [47]. In the context
of DGBLL, structural playable elements such as rules, goals, objectives,
outcomes, feedback, conflict, competition, challenge, opposition, rep­
resentation, and narrative are designed to inspire learners’ extrinsic
motivation [13]. In contrast, intrinsic motivation relates to the learner’s
interests and inner mental fulfillment, which can be interpreted as the
enjoyment one experiences, the learning that the game offers, or the
feeling of accomplishment one perceives (Figueroa, 2015).
In dual coding theory (DCT), cognition involves the processing of
information via two different subsystems, a verbal system that is directly
composed of language, and a nonverbal system that processes nonverbal
objects and events. Hence, knowledge–and the learning processes
through which knowledge is constructed–consists of the use of verbal
and nonverbal systems together (Sadoski, 2005; Paivio & Clark, 2006).
In the context of DGBLL, digital games can involve learners in learning
new words by connecting word forms and meanings successfully using
audio, video, pictures, graphics, and the like, in their virtual world
M. Chowdhury et al.
(Sadoski, 2005).
Scratch programming
One of the promising tools for DGBLL is Scratch. Scratch (https
://scratch.mit.edu/) is a programming platform developed by the
Massachusetts Institute of Technology Media Lab that includes a
development environment and a website where the community can host
their projects, play games, reuse other programs, and share ideas or
suggestions with others [48]. In this free-to-use block-based program­
ming language, students can create their own interactive stories, digital
games, and animation programming. The code pieces are displayed as
puzzle pieces that only fit together if they are syntactically correct [49].
In general, students who are exposed to Scratch report positive learning
experiences [50].
Even though researchers have found Scratch to be helpful for ac­
commodating different learning preferences and projects [48], the
benefits of Scratch have not yet been adequately investigated in terms of
DGBLL, vocabulary development, and reading comprehension. In a
small-scale intervention study, Sourani and Ihmaid (2019) found that
elementary students using Scratch programming performed better than
the control group who received traditional instruction for vocabulary
development. However, the researchers did not mention how long the
treatment lasted or whether the treatment had any impact on students’
language learning. Similarly, Burke and Kafai [51] conducted a writing
workshop in which students had to create stories using Scratch instead
of with pencil and paper. Both students and researchers stated that the
overall experience had been very motivating and positive not only for
improving digital literacy, but also for improving the learners’ writing
skill development through digital storytelling. However, the critical is­
sues of generativity and learner agency were not addressed and no in­
formation about how the students chose new vocabulary within their
story retellings was provided. Therefore, more research is needed link­
ing the DGBLL and generative platforms, such as Scratch, in terms of
vocabulary development and reading comprehension.
The present study
The present research aimed to address the limitations outlined
above, including DGBLL and generative platforms, as well as to expand
the scope of existing research in several directions. The present study
included: (1) the development of a game-based literacy curriculum
focusing on developing both lower-level and higher-level processes that
are aligned with the Common Core Standards, and (2) the imple­
mentation of a game-based literacy workshop, using concept-based in­
struction so participants can actively use targeted language skills for
game-based activities. The research questions were:
1. Were there gains in terms of vocabulary and reading comprehension
performance from the pre- to post-test?
2. What was the learners’ experience in the program?
a. What were the relationships among game-based learning experi­
ences, language learning experiences, and affective experiences?
b. What were the leverage points in networks of relationships
regarding aspects of learner experiences that can be used to design
DGBLL experiences in the future?
c. How were learner experiences related to the principles of learning
theory in which the program was grounded?
d. How do learner experiences indicate how different aspects of the
constructionist theory are related to different aspects of dual coding
theory and/or motivation theory?
1. What were the strengths and weaknesses of the program, based on
learners’ experiences?
3
Computers and Education Open 6 (2024) 100160
a. How were the strengths and weaknesses of the DGBLL program
related?
b. What were the leverage points which could be used to address
weaknesses?
This paper is organized as follows: Section 2 outlines the methodo­
logical approaches, including participants, data collection, and analyt­
ical procedures. Section 3 presents the findings of the study, with a focus
on the relationships between various interdependent aspects of the ex­
periences of the learners, as well as the impact on vocabulary develop­
ment in a DGBLL environment. Throughout Section 3, we discuss the
findings, drawing on constructivist, motivation, dual-coding, and
constructionist theories to interpret the results. We propose design
principles for future DGBLL endeavors. In Section 4 we conclude by
summarizing the main contributions of the study, discussing limitations,
and suggesting directions for future DGBLL research. By addressing
these areas, we aim to elucidate the potentials and challenges of digital
game-based language learning for vocabulary development, particularly
from a game-making perspective.
Method
This was a convergent mixed methods case study using paired sam­
ple t-test, qualitative data coding, and network analysis. Students
participated in a game-based language learning workshop that aimed to
improve their vocabulary and reading comprehension through devel­
oping digital games. The DGBLL workshop consisted of eight 50-minute
online sessions over eight weeks. Before the workshop and again after
the workshop, the thirteen students were assessed for reading compre­
hension and vocabulary development. After the conclusion of the pro­
gram, we interviewed the participants. We are situating our work in a
constructivist and expansive paradigm in which generalizability is not a
goal [52,53]. Rather, the goal is contextual understanding in order for
readers to interpret theory-grounded principles in light of their own
contexts. As constructivist researchers with expertise in game-based
learning, we seek to understand how students experience DGBLL and
refer the reader to the literature (see previous section) for research
regarding the efficacy of game-based learning.
Research context and sample
A DGBLL workshop was designed on constructionist principles: stu­
dents were immersed in the constructionist learning environment for
incidental vocabulary development; they constructed their own game in
Scratch based on the assigned nonfiction text to encourage embodied
cognition [42]. The non-fiction texts were five-paragraph informative
pieces about the water cycle, the lifecycle of frogs, African elephants,
and school uniforms. In addition, the students were situated as game
designers, and the final artifacts (digital games) were personally
meaningful to the students because they have learners’ agency resulting
in the embodiment of the knowledge that learners constructed [43]. The
students also shared their final artifacts with their peers and instructors
for feedback, providing them with authentic audiences [44]. During this
whole process, students went back and forth between their non-fiction
texts and their games multiple times to remember the facts to embed
in their games, to add details to their games, to visualize the facts in the
games, or to connect the games with what they read in the non-fiction
topic, which could lead to incidental learning of vocabulary [54]. The
workshop was divided into four 2-session units, with a different
non-fiction text for each unit where students created new games for each
unit. This iterative process gave the students ample opportunities for
focused tinkering [45], where they reassessed their game-making stra­
tegies and information from the non-fiction text to design their game.
Thirteen elementary school age (9–13 years old; 7 male, 5 female)
learners participated in this study. Participants were second-generation
Bangladeshi American English language learner students. All of the
M. Chowdhury et al.
students speak Bengali at home. Parents reported that these students had
beginner to intermediate-level proficiency in Bengali and were strug­
gling readers in English and had low English vocabulary knowledge. All
participants had computer and internet access at home. None of the
learners had prior experience with the game-designing platform Scratch.
The workshop was designed to increase the participants’ vocabulary
knowledge and reading comprehension skills in non-fiction texts. Par­
ticipants were recruited from mailing lists of Bangladesh student asso­
ciations at a major US research university. Parents received a study
invitation and the informed parental consent form via Email.
We designed the curriculum where in each session students read a
non-fiction text, learned new vocabulary, and designed games based on
their reading using Scratch. The instructors read the non-fiction text to
the students, discussed difficult or new words when necessary, and
demonstrated how to design a game. Note that after the demonstration,
students in pairs were encouraged to use their creativity to design
games, with the expectation that each game would be unique. This
direct instruction regarding new words may lead to successful vocabu­
lary knowledge development (Lesaux et al., 2014). During game-design
sessions, the students worked in groups. The sessions were conducted
online, and the participants remained at their respective homes. They
participated in both video and audio chat via Zoom.
The 50-minute online sessions included interactive guided learning.
All interactions were verbal. Sessions included game-designing activities
and discussions. The students were grouped by the facilitator during the
game design sessions. Each group worked in Zoom breakout rooms.
Even though students brainstormed together, the games were designed
individually. The facilitator rotated in the group breakout rooms to
observe and give feedback. To further facilitate and encourage collab­
orative learning, between sessions the groups worked in Google docu­
ments and in the Scratch platform asynchronously.
Instruments and qualitative data collection
Measures for this study included game-design journals, pre and post
vocabulary, and reading comprehension tests, and semi-structured in­
terviews which are described below. According to Maxwell [55], data
triangulation is a strategy for validity testing. Triangulation refers to the
collection of information of various types from different sources to gain
a better understanding of the explanations and interpretations of the
collected data [56]. This research study used multiple data sources to
collect data to gain an understanding of learners’ experiences and
strategies as they created their digital games. Data collection took place
over eight weeks of the game design workshop.
Since most of the participants were 2nd graders, the reading
comprehension tests were aligned with the 420L-650 L Lexile band of
2nd graders [57]. The design ensured that students encountered texts of
increasing complexity as they moved through each grade level. Pre- and
post- vocabulary tests consisted of 10 questions to measure vocabulary
growth. Fifty percent of the questions were the same between the pre-
and post-tests selected to be academic vocabulary common to 2nd grade
texts, and the other fifty percent were different questions grounded in
the contextual learning during the DGBLL workshop (see Appendix A
and [58–60]). The vocabulary tests included 15 targeted vocabulary
included in the intervention regarding the non-fiction concepts (e.g.,
author’s purpose, context clues, using diagrams to clarify a text).
Non-standardized assessments were selected for pretesting and post-
testing as more appropriate for this study. Standardized assessments are
designed to measure growth over a year or more. Therefore, gains after
an eight-week workshop are unlikely to be captured by standardized
tests. This study focused on examining the acquisition of specific vo­
cabulary words and non-fiction concepts introduced in the DGBLL
workshop, which would not be reflected in standardized tests. To
minimize the concern regarding the reliability of the tests, the reliability
coefficients for the pre- and post-tests using Cronbach’s alpha were
calculated. The reliability coefficients for the vocabulary development
4
Computers and Education Open 6 (2024) 100160
and reading comprehension for pretests were 0.68 and 0.63, considered
acceptable for research [61]. The reliability coefficients for the vocab­
ulary development and reading comprehension for post tests are 0.72
and 0.83, considered highly reliable [61].
Each student received a game design journal in each lesson in Google
slides where they could note new vocabulary, facts from reading, or
brainstorm their game-design ideas. Khalili [7] suggested that game
design journals could be helpful for students to think through their ideas,
collect their data, and observe the progression of their ideas. The jour­
nals were primarily used for brainstorming and note taking (see
Appendix B).
At the end of the workshop, students participated in semi-structured
interviews [55] conducted in Zoom to understand the overall experience
of the learners, which were transcribed verbatim. The interview ques­
tions were constructed to elicit rich data regarding their experiences in
the workshop in relation to the theoretical framework described in the
previous section (see Appendix B and C).
Data analysis
In this convergent mixed-methods research design, both qualitative
and quantitative data were collected essentially simultaneously, and
both data sets were analyzed and compared to understand the research
problem [62,63]. Data were analyzed using both quantitative and
qualitative research methods. Because the number of participants was
small, analytical differentiation by gender was not possible. To answer
research question 1, the pre- and post-workshop vocabulary and reading
comprehension assessments were analyzed quantitatively using paired
t-test statistical analysis in STATA to determine how the participation in
the study related to the learners’ vocabulary and reading comprehension
performance from pre- to post-test. For qualitative data coding, re­
searchers used the MAXQDA Analytics Pro-qualitative analysis software
to understand the learner’s experience. Learner’s experience includes
the affective, interpersonal, intrapersonal, and cognitive aspects of how
the learners experienced the DGBLL workshop. Each sentence was coded
according to multiple codes in the codebook when appropriate. All the
qualitative data were coded in vivo for emergent codes [64] except for
categories based on the theories in which this study is grounded where
we used a priori coding [65]. The initial codebook included a category
for theory with sub-categories related to constructivist theory, motiva­
tion theory, constructionist theory, and dual coding theory, as well as
categories for the game-based learning experience, affective experience,
language learning experience in terms of vocabulary development and
reading comprehension, and research design strengths and challenges as
perceived by the learners.
Next, we conducted axial coding, which is a process of analyzing
codes in terms of relationships among codes [64]. Axial coding was
conducted to relate the data together to reveal relationships between
codes, which was organized according to categories and subcategories
[66]. Then we cleaned the codebook, deleting the emergent codes that
occurred only once since they cannot be used in a pattern [67]. We also
merged codes if they were two versions of the same idea to avoid
redundancy [66]. If a code occurred too many times, we split the code
into multiple codes to provide a greater depth of understanding [66]
resulting in 57 codes.
We negotiated the codes, categories, and subcategories with an
expert regularly to minimize coder bias. However, there might still be a
concern about whether a single data coder can “interpret the same data
and record exactly the same value for the same variable each time these
data are collected?” ([68], p. 277). To respond to this concern, we
conducted an intra-rater reliability test [69] where 20 % of the data was
coded twice by the first author. Then, we ran an intra-rater reliability
test in MAXQDA [66] to determine Kappa values. Using the scale for
strength of agreement developed by Landis and Koch (1977), we found
the Kappa value to be 0.64, indicating substantial reliability [69].
To answer research questions 2 and 3, we conducted semantic
M. Chowdhury et al.
network analysis [34]. Semantic network analysis is the use of network
analytic techniques on multiple paired associations based on their lin­
guistic connections and semantic proximity within the cognitive schema
[70]. Taken together, these associations represent the meaning inherent
in the data [71].
We calculated the code co-occurrence correlations for all pairs of
codes within each category, as well as pairs across multiple categories.
The unit of analysis [72] is the network representing the relationships
among codes found from the coding of all qualitative data, and the unit
of observation [73] is a dyad consisting of two nodes (codes) and their
relationship [74,75]. In this study, each code consisted of a sentence
[70].
Correlations were calculated using Pearson Product-Moment Corre­
lation [76] in MAXQDA Analytics Pro-to identify significant correlations
between each code and other codes within each coding category. These
correlations were exported as symmetrical correlation matrices in MS
Excel format constructed at the p<0.05, p<0.01, and p<0.001 confi­
dence levels that were statistically significant. We used UCINET/Net­
Draw network analysis software to assist with axial coding through the
analysis of relationships among categories [64] via semantic network
maps [71,70]. These maps visually expressed the multiple networked
relationships in the correlated data, and then each map was analyzed to
determine which patterns were meaningful from a theoretical perspec­
tive [64]. We also used the Girvan-Newman algorithm to apply cluster
analysis [77]. Cluster analysis indicated the property of two nodes that
were both neighbors of the same third node having a heightened prob­
ability of also being neighbors of one another. Cluster analysis allowed
us to identify patterns of aspects of learner experiences that were more
closely related to each other than to other aspects.
To answer the second research question, clustered semantic network
maps of learner experiences in terms of affect, game-based learning, and
language learning were used to identify how various aspects of learners’
experiences were related. After that, betweenness measures of centrality
were used to identify leverage points – that is, aspects of the learning
experience that would have the greatest impact on other aspects if these
aspects were to be modified. Betweenness measures of centrality
represent the importance of a node in connecting other nodes with each
other in a network [78]. These measures of centrality identified leverage
points, which indicated the places to intervene in a network map where
a relatively small change in one part of a system can lead to relatively big
changes in the whole system [79,80]. Leverage points could be used to
design more powerful DGBLL experiences in the future.
Finally, to answer the third research question, clustered semantic
network maps of strengths and weaknesses of the design were used to
identify the relationships between design strengths and weaknesses.
Betweenness centrality measures were calculated and node (code) sizes
indicated the level of betweenness. Nodes with higher betweenness
centrality measures helped to identify leverage points by which partic­
ular strengths can be used to address weaknesses. Clustered semantic
network maps of student experiences in relation to elements of the
theories in which the program was grounded were used to determine the
relationships between elements within each theory and elements across
theories. Betweenness measures in these semantic network maps were
used to identify leverage points in each theory which could provide
insights to strengthen other theories.
Results and discussion
Moving from the discussion of our methodological considerations,
we now turn our attention to the results and their implications. In this
section, we delve into an exploration of the data, beginning with the
outcomes from our qualitative data analysis, and weave a narrative of
the students’ experiences, vocabulary development, and the emergent
interdependencies in the digital game-based language learning context.
5
Computers and Education Open 6 (2024) 100160
Results from quantitative data analysis
To answer research question 1, we identified the null hypothesis as
that there would be no difference in vocabulary performance scores
between the pre and post vocabulary test, whereas the hypothesis was
that there would be higher vocabulary performance scores in the post
tests compared to the pre-tests. The Shapiro-Wilks test showed that the
pre and posttest vocabulary scores were normally distributed with a p-
value of 0.18 and 0.16 (p>0.05). We conducted a paired sample t-test on
the pre and pre and post vocabulary test scores. The post mean score
(3.35) was higher than the pre mean score 2.08, indicating that the
students gained in their post vocabulary scores compared to their pre
vocabulary scores. The mean difference was 1.28 with a standard de­
viation of 1.08. The obtained t-value was 4.25 and the p-value was 0.001
(p<0.05). Therefore, the null hypothesis was rejected and the research
hypothesis was accepted: there was a statistically significant gain be­
tween the two mean vocabulary scores of 1.28 (95 % CI, 0.62 to 1.93), t
(12) = 4.25, p<0.05.
We also wanted to see whether there was any significant difference in
the reading comprehension score from pre to post test. Shapiro-Wilks
test found that the pre-test was normally distributed with a p-value of
0.98 (p>0.05) for the pretest and 0.99 for the post test. We ran the paired
sample t-test on the pre and post reading comprehension mean scores.
The post mean score (3.28) was higher than the pre mean score (2.24),
suggesting that students had a gain in their post reading comprehension
scores compared to their pre reading comprehension scores. The mean
difference was 1.03 with a standard deviation of 0.81. The obtained t-
value was 4.58 and the p-value was 0.0006 (p<0.05). Therefore, the
null hypothesis was rejected and the research hypothesis was accepted:
there was a statistically significant gain between the two mean reading
comprehension scores of 1.03 (95 % CI, 0.54 to 1.53), t(12) = 4.58,
p<0.05.
To find out about the gains in students’ performance from the pre to
post test, we also looked at the understanding level of the students based
on the students’ open-ended responses to the test items. The responses
were scored based on the categorical variables of “complete under­
standing,” “almost understanding,” “incomplete/vague understanding,”
and “misunderstanding.” In the case of vocabulary pre and post tests, it
was found that there was an increase in “complete understanding” (28.2
% to 65 %), and a decrease in “almost understanding” (65.5 % to 7.5 %),
“incomplete/vague understanding” (58.3 % to 15 %), and “misunder­
standing” (40 % to 6.67 %). In the case of reading comprehension pre
and post tests, it was found that there was an increase in “complete
understanding” (22 % to 78 %), and a decrease in “almost understand­
ing” (85.5 % to 14.4 %), “incomplete/vague understanding” (69.4 % to
13.9 %) and “misunderstanding” (50 % to 0 %). Looking at the mean
scores of the repeated test items in the target vocabulary tests, we found
that there was a gain in the mean scores from the pre to post test (Fig. 1).
All mean score differences for the repeated vocabulary items were
statistically significant (Fig. 1). In the reading comprehension test cat­
egories, there were gains in the mean scores from the pre to post tests
(Fig. 2).
The overall mean scores differences for the comprehension cate­
gories in the reading comprehension pre and posttest (Fig. 2) were sta­
tistically significant.
Discussion of results from quantitative data analysis
There were statistically significant gains from pre to post tests in both
vocabulary development and reading comprehension. Cronbach’s alpha
coefficients indicated that the assessment items had medium to high
reliability. The students’ responses in the pre- and post-tests in both the
vocabulary and reading comprehension showed that there was an in­
crease in “complete understanding” and a decrease in “almost under­
standing,” “incomplete/vague understanding,” and
“misunderstanding.” There was a gain in the mean scores of the repeated
test items in the pre and post vocabulary and reading comprehension
M. Chowdhury et al. Computers and Education Open 6 (2024) 100160

Fig. 1. Pre and Post Vocabulary Items (Repeated target vocabulary).

Fig. 2. Pre and Post Comprehension Categories.

tests. Since there was no control group, these gains cannot be attributed
solely to the DGBLL workshop, but several aspects of the workshop such
as making a digital game [42], learners’ agency [43], authentic audience
[44], tinkering [45], or intrinsic motivation [47] may have contributed
to this gain.

Findings from network analysis of qualitative data

Because learning is complex, we used the complex systems analytical

approach of network analysis. We analyzed the qualitative data using
semantic network analysis to explore the learners’ experiences and the
strengths and weaknesses of the DGBLL workshop. Through cluster
analysis in the network map of correlated learner experiences reflected
in their game-design journals and semi-structured interviews, we iden­
tified groups of codes, and using betweenness centrality measures we
identified leverage points.
Findings from network analysis for research question 2
To answer research question 2a, a three-cluster semantic network
map (Q = 0.23) was generated using the categories “game-based
learning experience” (GBL), “language learning experience” (LL), and
“affective experience” (AFF) using the Girvan-Newman algorithm [77]
in the analysis software. The network map was created using the cor­
relation matrix at the p<0.001 confidence level (Fig. 3).
The reading-application cluster (circles cluster in Fig. 3) includes the
significantly correlated (as indicated by lines connecting nodes) LL ex­
periences codes of enjoyed reading/learning about nonfiction, positive,
going back to text, learned new facts, and context clues, related to students’
Fig. 3. GBL vs LL vs AFF with Leverage Points.
AFF experience code of confident. This cluster also includes GBL expe­
rience codes of enjoyed coding and translate information into an interactive
game. The fun-coding cluster (triangles cluster in Fig. 3) includes the
AFF codes of fun and easy, as well as a GBL code of improve coding
knowledge and wants to practice more. The game-making-learning cluster
(squares cluster in Fig. 3) includes two GBL experiences of making a game
and learned facts about the non-fiction topic, and an affective experi­
ence–motivated to learn more about the topic.
To answer question 2b, we calculated the betweenness measures of
centrality of nodes within the categories of GBL, LL, and AFF to see
which nodes held the most importance in the network map in terms of

6
M. Chowdhury et al.
connecting two or more other nodes. The most prominent leverage point
was found to be the GBL experience code of translate information into an
interactive game (the largest node in Fig. 3). Some other leverage points
were LL experience codes of going back to text and enjoyed reading/
learning about nonfiction and the AFF code of fun.
To answer research question 2c, a three-cluster semantic network
map was generated using the codes for the categories “constructionist
theory,” “dual-coding theory,” and “motivation theory.” Using the
Girvan-Newman algorithm [77], we identified a three-cluster semantic
map (Q = 0.22) using the correlation matrix at the p<0.001 level
(Fig. 4).
The construction-together cluster (hourglasses cluster in Fig. 4)
includes codes of learner agency, authentic audience, situating learners
as designers, and tinkering from the constructionist theory, codes of
learning together, interdependence, and sharing from constructivist
theory, and codes from dual coding theory including multiple modality
and text-graphics. The schema-construction cluster (diamonds cluster
in Fig. 4) includes the schema theory codes of activation of schemas,
relating new vocabulary with prior knowledge, and word associations,
the cognitive constructivist code of transform information into knowl­
edge, and intrinsic motivation from motivation theory. The rounded
square node (right side of Fig. 4) represents one code of making from
constructionist theory; according to Girvan-Newman analysis [77], it is
“clustered” on its own, but has a significant relationship to the
schema-construction cluster.
We also calculated betweenness centrality measures (Fig. 4) and re-
sized the nodes by betweenness to discover the leverage points among
the learning theories and found three leverage points. These include the
constructionist theory code of tinkering (the largest hourglass node in
Fig. 4), the motivation theory code of intrinsic motivation (large node on
the right side of the diamonds cluster), and the schema theory code of
activation of schemas (large diamond node in the center of the map).
Discussion of network analysis findings for research question 2
Research question 2 explored learner experiences. Findings related
to research questions 2a and 2b display the interconnection among the
learners’ experiences in terms of game-based learning, language
learning, and students’ affective experiences. The reading-application
cluster shows us that the students have enjoyed reading/learning about
non-fiction texts, learned new facts, and went back to texts and these lan­
guage learning experiences have helped them to apply their knowledge
as they translate information into an interactive game. They feel confident,
and nice, good, great about this experience overall. This aligns with the
affordance of providing space for sheltered vocabulary learning where
the DGBLL experience has helped the English language learners to apply
their knowledge in a low-anxiety environment with the students feeling
positive and confident [81,82].
The fun-coding cluster displays that the students have a fun and easy
experience in the game-based learning environment. They want to
improve their coding knowledge and they want to practice more. This aligns
with the concept of intrinsic motivation where the learner is self-
Fig. 4. Learning Theories with Leverage Points.
7
Computers and Education Open 6 (2024) 100160
regulated and motivated to practice for their own skill development
due to affective reasons such as interest or enjoyment [47]. The
game-making-learning cluster reinforces this idea of intrinsic motiva­
tion as the students mention that they are motivated to learn more about
the non-fiction topic.
Findings from research question 2b show the leverage points inter­
connecting game-based learning, language learning, and students’ af­
fective experience. The most important leverage point among the
experiences was translating information from the nonfiction text into an
interactive game. DGBLL experience may provide the unique affordance
of contextualized language learning where the students use the language
in a meaningful, goal-focused way [83]. Contextualization is an effective
method of L2 vocabulary learning as it helps the learners to remember
the vocabulary in semantically related groupings [20]. This DGBLL
experience also provided the students with space for sheltered use where
the students progress through steps (Reiser & Tabak, 2014; [36]). For
example, the leverage points included the whole cycle of the workshop
where the students first read the non-fiction text, jotted down the facts
from their reading along with newly learned vocabulary, designed their
game around the facts, went back to the text if necessary, coded and
built their game, and enjoyed the whole process. This suggests that
designing generative learning experiences helps students translate in­
formation into knowledge by constructing meaningful representations
in the DGBLL context. Furthermore, educators should design these
learning experiences such that they include opportunities for reading
and leverage the affordance of DGBLL to facilitate positive affective
experiences (fun) but must take care not to lose the joy of learning.
Findings related to research question 2c reveal that aspects of
constructionist theory are closely related to aspects of the L2 vocabulary
learning process posited by social constructivist theory, schema theory,
dual coding theory, and motivation theory. The construction-together
cluster shows that several aspects of constructionist theory–including
focused tinkering, authentic audience, and situating learners as
designers–are connected to aspects of interdependence and learning
together from social constructivist theory. This aligns with the connec­
tion between the DGBLL and social constructivist theory of vocabulary
learning as it emphasizes the DGBLL affordance of opportunities for
collaboration [83,84]. This affordance is based on the
social-collaborative foundation of second language acquisition (SLA)
which emphasizes aspects of collaboration, cooperation, empathy, and
critical thinking [85]. In this cluster, constructionist theory and
constructivist theory are also connected with multiple modality and
text-graphics connection of dual coding theory of vocabulary learning.
The schema-construction cluster shows that the code of activation of
schemas is connected to the codes of relating new vocabulary with prior
knowledge, and word associations. This aligns with the schema theory of
vocabulary learning which contends that connecting new words with
prior knowledge and making word associations helps the learners to
activate their schema which in turn, results in their vocabulary learning
[86]. Even though the single node of making is not a cluster, it is an
important node to create the dyad with the code of intrinsic motivation.
Without this node, the semantic map becomes weak. This dyad suggests
that the DGBLL context may have the potential to promote intrinsic
motivation.
The construction-together cluster included all the nodes that point to
the active, external learning by making artifacts whereas the schema-
construction cluster included all the nodes that point to the internal,
cognitive process of knowledge construction. This aligns with the
feature of constructionist theory where the learners make artifacts (in
this case, digital games) and these artifacts mirror the embodied
cognition of learners’ minds [42]. This external and internal process
simultaneously helps the learner with knowledge construction.
The three leverage points also reveal that the DGBLL environment
may provide opportunities for focused tinkering [45] which helps the
learners to immerse themselves in an “experimental, iterative style of
engagement” (p. 164), promote intrinsic motivation [47], and activate
M. Chowdhury et al.
their schemas for vocabulary knowledge development [87].
Findings from network analysis for research question 3
To answer research question 3a, a clustered semantic network map
was generated using the codes for the categories “design strengths” and
“design challenges” resulting in a three-cluster semantic map (Q = 0.34)
using the correlation matrix at the p<0.05 level (Fig. 5).
The instructions-glitch cluster (up-triangles cluster in Fig. 5) in­
cludes the codes of technical glitch and need more instruction on Scratch
from the category “design challenges,” and codes of working with colors
and working with images from the category of “design strengths.” The
coding-time cluster (down-triangles cluster in Fig. 5) includes the codes
of coding was challenging, not enough time, and more free play/diversity of
games from the category design challenges and a code of learned coding
from the category “design strength.” The foreground-background clus­
ter (circles cluster in Figure 7) includes three codes of wants to improve
game, background and foreground work, and writing facts on foreground
from the category “design strengths” and a code of no improvement
needed from the category of “design challenges.”
To answer research question 3b, we calculated the betweenness
measures of centrality to find out the leverage points to address the
design challenges. There were two prominent leverage points: technical
glitch (large up-triangle node in the center of Fig. 5) and coding was
challenging (large down-triangle node) from the category “design
challenges.”
Discussions of network analysis findings for research question 3
From the findings of research questions 3a and 3b, we can see the
connections between the design strengths and challenges of the DGBLL
workshop. The instructions-glitch cluster shows that the “design
strengths” of working with images and working with colors is connected to
the “design challenges” of technical glitch and need more instruction on
Scratch. To design powerful DGBLL experiences, the curriculum de­
velopers should pay more attention to ensuring the stability of the
internet and technical issues as well as provide ample instruction on
Scratch.
From the coding-time cluster, learning to code the game was con­
nected to the “design challenges” of coding was challenging, not enough
time, and more free play/diversity of game. Connecting this finding to the
instructions-glitch cluster, even though these students enjoyed working
with colors and images, they would prefer more time to develop their
skill in coding. The foreground-background cluster shows that there is
no improvement suggested by the data when it comes to the “design
strengths” of wants to improve game, background and foreground work, and
writing facts on foreground.
From the findings of research question 3b, the two prominent design
challenges that the learner faced were technical glitches and coding was
challenging. Participants mentioned ensuring a stable internet connec­
tion and “internet” when discussing the technical glitches. One partici­
pant mentioned that her computer restarted occasionally. Another
student mentioned not being able to share her screen during their
sharing time. Because these are significant leverage points, addressing
these issues in the future can be expected to improve the learning
experience. Another design challenge was that coding was challenging
Fig. 5. Design Strength vs Challenges with Leverage Points.
8
Computers and Education Open 6 (2024) 100160
which is connected to lack of time, not enough instruction on Scratch,
and not enough free play. Therefore, to address this design challenge in
future, instructors should not only give ample time and instruction on
Scratch, but also make sure that the students have more autonomy for
free play [88] and are provided with a diversity of games.
Conclusion, limitations, and future research
Findings from this study of DGBLL for L2 vocabulary learning shows
that there were statistically significant gains from pre to post-tests in
both vocabulary learning and reading comprehension. Because the
novel approach of network analysis embraces the complexity of learning
experiences, our findings show that DGBLL contexts have the potential
to provide students with engaging low-anxiety environments for the
facilitation of vocabulary development [81,82]. The prominent
theory-grounded leverage points found in this study point to the
importance of learning designs that promote joyful and generative ex­
periences. These leverage points are relevant throughout the DGBLL
learning experience. Therefore, if researchers want to design a more
powerful DGBLL learning experience for their students, they are
well-advised to design a generative learning experience in which stu­
dents will have the opportunity to translate information into knowledge
by constructing meaningful artifacts. The experience of joy appears to be
facilitated by game-based approaches that center learner agency. We
also found that if educators ground their research on strong learning
theories such as constructionism, motivation theory, and dual coding
theory, and promote a maker experience for the students, it may lead to
powerful learning. Finally, this is a reminder that educators can facili­
tate positive learning experiences for their students but must pay
attention to the aspect of enjoyment. The current study has found
several promising aspects of DGBLL, such as promoting joy, generative
experiences, and learner agency. Future research could explore these
aspects in more depth, for example, by investigating which specific
affordances of digital games contribute to these outcomes and how they
can be enhanced.
DGBLL contexts may promote intrinsic motivation for the learners,
which is a key element of the motivation theory, and provide them with
the opportunity to contextualize language learning where vocabulary is
dual coded through multimodal narrative in the gaming context. The
DGBLL context of this study provided students with social constructivist
opportunities to collaborate and share ideas with each other. This
experience gave them an opportunity to enact the constructionist prin­
ciple of focused tinkering which may lead to vocabulary learning. Our
network analysis of relationships between theoretical principles found
that constructionist theory principles of focused tinkering, authentic
audience, and situating learners as designers were clustered with aspects
of social constructivist theory, schema theory, and dual coding theory.
This clustering emphasized contextual and externalized processes and
relationships. Another cluster focused on internal processes from
cognitive constructivist theory, with intrinsic motivation connecting
these processes with the constructionist principle of learning by making.
Future DGBLL research might explore different ways of conducting
network analysis to further uncover the complex relationships between
different aspects of the learning experience.
The struggles students experienced in this study were connected
through complex interdependent relationships with design strengths
which can be leveraged to address such struggles in the future. Struggles
with technical glitches and feeling unsure about how to use the tech­
nologies were clustered with strengths related to visual aspects of the
learning environment, suggesting that aspects of the visual environment
could be used more effectively as prompts for using the tools. Struggles
with timing and computational thinking were clustered with strengths
related to technical skill development. Therefore, educators may wish to
build more time into the activity structure for skill development. Future
research could explore innovative ways to address these struggles to
improve the effectiveness of the DGBLL approach.
M. Chowdhury et al.
Because this study was not experimental (no control group) and the
number of participants was small, these gains cannot be ascribed to the
DGBLL workshop entirely. Most qualitative research has issues with
researcher bias [72]. To address this, the researchers acknowledged the
issue, ran intra-rater reliability tests, and tried to reflect critically
through all the stages of analysis and interpretation. Furthermore, the
workshop only lasted for eight weeks. More time and progression
through time in a longitudinal study could provide more in-depth
findings regarding DGBLL and digitally mediated language learning
contexts. Finally, because our focus of analysis was on qualitative data
and we used a constructivist research paradigm, generalizability was not
a goal. Even though replicability of protocol and procedures may be
possible, replicability of findings will not be possible. Therefore, future
research should focus on long-term longitudinal data collection, multi­
ple iterations, and implementation of the research design with multiple
data sources for the triangulation of the data. To increase the robustness
of the findings, future studies could involve larger and more diverse
groups of learners. This could include learners of different ages, learning
preferences, and cultural backgrounds.
Appendix A: Pre and Post Tests
Pre-Test
Post-Test
9
Computers and Education Open 6 (2024) 100160
CRediT authorship contribution statement
Mahjabin Chowdhury: Conceptualization, Data curation, Formal
analysis, Funding acquisition, Investigation, Methodology, Project
administration, Resources, Software, Supervision, Validation, Visuali­
zation, Writing – original draft, Writing – review & editing. L．
．Quentin
Dixon: Conceptualization. Li-Jen Kuo: Supervision. Jonan Phillip
Donaldson: Writing – review & editing, Writing – original draft, Visu­
alization, Methodology, Formal analysis, Conceptualization. Zohreh
Eslami: Supervision. Radhika Viruru: Supervision. Wen Luo:
Supervision.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
M. Chowdhury et al. Computers and Education Open 6 (2024) 100160

Appendix B: Journal Prompt and Interview Protocol

Please write about your game-design today

What is working well? Needs Improvement

Player role/character
Game goals/rules
Is it clear what you are supposed to do?
What ways could the goals/rules work better?
Challenges
Which part was too hard/boring/easy/just right?
Language
Do you understand the words/language/story?
Did you learn a new word?What was it?
Can you write a sentence with it?
Style
What images/color choices work well?

Game-Design Journal Prompt

Interview Protocol for Study 1

1. What did you think about the program?

2. How did you feel about making the games?
3. What was your game about?
4. What did you like about working on your game in Scratch?
5. How do you think your game turned out?
6. What steps did you follow to make your game after reading a nonfiction text?
7. How did you work with any other students on your game?
8. Do you think you will show your video game to other people?
9. Would you be interested in designing video games about other stories?
10. Do you feel more confident about reading non-fiction texts now (after this program)?
11. Do you feel more confident about learning new words?

10
M. Chowdhury et al. Computers and Education Open 6 (2024) 100160

Appendix C

Example Codes and Coded Segments

Code Description Example Quote

SOCIAL CONST - coding together
SOCIAL CONST - interdependence
SCHEMA - activation of schemas
SCHEMA - relating new vocabulary
with prior knowledge
COG CONST - relevance exploration
(what it means to the lrnr)
COG CONST - transform information
into knowledge
DCT - text - graphics
DCT - multiple modality
CONSTRUCTIONIST - Authentic
audience
CONSTRUCTIONIST - learner agency
(also in motivation theory)
CONSTRUCTIONIST - Making
MT - intrinsic motivation
EXP GBL - Enjoyed coding
EXP AFF - confident
EXP AFF - Fun
EXP AFF - motivated to learn more
about the topic
DESIGN CHALL - technical glitch
DESIGN CHALL - not enough time
EXP LL - enjoyed reading/learning
about non-fiction
EXP LL - context clues
DESIGN STR - Background and
Foreground work
The social constructivist principle of co-constructing
knowledge through coding together.
The social constructivist principle of interdependence.
The schema theory principle of activation of schemas.
The schema theory principle of integrating new vocabulary
with existing knowledge.
The cognitive constructivist approach of exploring the
relevance of new information to the learner’s own life and
context.
The cognitive constructivist process of transforming newly
acquired information into personal knowledge.
The dual-coding theory’s emphasis on integrating text and
graphics to enhance learning.
The dual-coding theory’s principle of using multiple
modalities (such as visual and verbal) for effective learning.
The constructionist theory’s focus on creating for an
authentic audience, enhancing the meaningfulness of the
learning experience.
The constructionist perspective on learner agency,
emphasizing the role of the learner in directing their own
learning process.
The constructionist approach of learning through making
and creating tangible outputs.
The motivation theory’s concept of intrinsic motivation,
where the learner is driven by internal desire and interest.
The experience of enjoyment in coding, reflecting a positive
engagement with the learning material.
The affective experience of feeling confident, indicating a
positive emotional response to the learning process.
The affective experience of finding the learning process fun,
contributing to engagement and interest.
The affective experience of increased motivation to learn
more about a specific topic.
Challenges encountered in design, specifically technical
glitches that can hinder the learning process.
The challenge of insufficient time for design and learning
activities.
The experience of enjoyment in reading and learning about
non-fiction, reflecting a positive engagement with textual
material.
The language learning strategy of using context clues to
understand new vocabulary or concepts.
The strength in design, particularly in effectively using
background and foreground elements to enhance learning
materials.
"I feel pretty happy that my friends can join me in coding."
"I worked by doing like helping. We made it. We helped each other with the
scratch, and we made sure everyone was doing okay and they were making
the right codes, and they were also making sure that everything was okay"
"I went back to read the text again when I was in Scratch so that I can
remember where what I read"
"like the other day I was eating chocolate and I remember that I know how
they make chocolate from cocoa bean."
"because it’s very much like I like math that I want to do"
"Then we went to Scratch and coded what we read to make the game"
"I remember the text and I was like, if I didn’t understand something I would
just try to remember it from the text, then do that game on Scratch"
"Then we went to Scratch and coded what we read to make the game"
"I would like to show it to my friends and I want to work on mine more"
"because it’s very much like I like math that I want to do"
"We made a "game" of a African Elephant."
"we learn like I do math on [redacted]. He has this coding section, and really
into that so I like coding and making animations and drawings"
"I learned facts about watercycle, elephant and chocolate and made games
with my coding"
"I feel more confident in reading nonfiction texts because I can, I can explain
all of the main stuff that are in the nonfiction passages and text"
"Scratch, I don’t know how but I....I think helped me to remember more of
what I read…It was fun, and easy and exciting, and I made games"
"Scratch, I don’t know how but I....I think helped me to remember more of
what I read…It was fun, and easy and exciting, and I made games"
"It was one problem because one of them more. on this side, um, people’s
names are coming out. That was distracting"
"I needed a lot of help it was confusing and hard and difficult, and I don’t
know any other words to describe but I wanted to work on it and not enough
time"
"I feel more confident in reading nonfiction texts because I can, I can explain
all of the main stuff that are in the nonfiction passages and text"
"like if you don’t understand a word, try to make sense of it from words
around it"
"The background "forest" works because it has a tree and cocoa beans grow
on trees."

References [9] Yudintseva A. Game-enhanced second language vocabulary acquisition strategies:

a systematic review. Open J. Soc. Sci. 2015;3(10):101.
[10] Papert S. Does easy do it? Children, games, and learning. Game Developer 1998;5
[1] Acquah EO, Katz HT. Digital game-based L2 learning outcomes for primary
(6):88–9.
through high-school students: a systematic literature review. Comput. Educ. 2020;
[11] York J, deHaan J, Childs M, Collins M. How is gamification like being trapped in
143:103667.
the matrix? And what is the ‘real-world’ of game-based learning? Digit. Cult. Educ.
[2] Ishaq K, Rosdi F, Zin NAM, Abid A. Serious game design model for language
2022;14(3).
learning in the cultural context. Educ. Info. Technol. 2022. https://doi.org/
[12] Hopkins, I., & Roberts, D. (2015). ‘Chocolate-covered broccoli’? Games and the
10.1007/s10639-022-10999-5.
teaching of literature. Changing English, 22(2), 222–36. https://doi.org/10.1
[3] Enyinnah CO. The impact of esl professional development on teacher practice.
080/1358684X.2015.1022508. doi:10.1145/2157136.2157264. doi:10.1598/
(Doctoral dissertation); 2014.
RT.64.1.1.
[4] Kelley, J.G., Lesaux, N.K., Kieffer, M.J., & Faller, S.E. (2010). Effective academic.
[13] Prensky M. Digital game-based learning. Paragon House; 2007.
[5] Govender T, Arnedo-Moreno J. An analysis of game design elements used in digital
[14] Al-Azawi R, Al-Faliti F, Al-Blushi M. Educational gamification vs. game based
game-based language learning. Sustainability 2021;13(12):6679.
learning: comparative study. Int. J. Innov. Manage. Technol. 2016;7(4):132–6.
[6] Xu Z, Chen Z, Eutsler L, Geng Z, Kogut A. A scoping review of digital game-based
[15] Krath J, Schürmann L, von Korflesch HFO. Revealing the theoretical basis of
technology on English language learning. Educ. Technol. Res. Develop. 2020;68:
gamification: a systematic review and analysis of theory in research on
877–904.
gamification, serious games and game-based learning. Comput. Human Behav.
[7] Khalili N. Student designed science games: a study of the design process, artifacts,
2021;125:106963. https://doi.org/10.1016/j.chb.2021.106963.
and attitudes in a constructivist and constructionist learning environment. Doctoral
[16] Butler YG. Motivational elements of digital instructional games: a study of young
dissertation; 2014.
L2 learners’ game designs. Lang. Teach. Res. 2017;21(6):735–50.
[8] Gee JP. Digital games and libraries. Knowledge Quest 2012;41(1):60–4.

11
M. Chowdhury et al.
[17] Wichadee S, Pattanapichet F. Enhancement of performance and motivation
through application of digital games in an English language class. Teach. English
Technol. 2018;18(1):77–92.
[18] Melo C, Madariaga L, Nussbaum M, Heller R, Bennett S, Tsai CC, van Braak J.
Editorial: educational technology and addictions. Comput. Educ. 2020;145:
103730. https://doi.org/10.1016/j.compedu.2019.103730.
[19] Ishak SA, Hasran UA, Din R. Media education through digital games: a review on
design and factors influencing learning performance. Educ. Sci. 2023;13(2):102.
https://www.mdpi.com/2227-7102/13/2/102.
[20] Nation P. Learning vocabulary in another language. Cambridge University Press;
2001. https://doi.org/10.1017/CBO9781139524759.
[21] James E, Currie NK, Tong SX, Cain K. The relations between morphological
awareness and reading comprehension in beginner readers to young adolescents.
J. Res. Read. 2021;44(1):110–30.
[22] Moody S, Hu X, Kuo LJ, Jouhar M, Xu Z, Lee S. Vocabulary instruction: a critical
analysis of theories, research, and practice. Educ. Sci. 2018;8(4):180. https://doi.
org/10.3390/educsci8040180.
[23] Zuo F, Yan X. Incidental English vocabulary acquisition through reading: a review
in the last two decades. English Lang. Teach. 2019;12(12):39–45. https://doi.org/
10.5539/elt.v12n12p39.
[24] Hennebry M, Rogers V, Macaro E, Murphy V. Direct teaching of vocabulary after
listening: is it worth the effort and what method is best? Lang. Learn. J. 2017;45
(3):282–300. https://doi.org/10.1080/09571736.2013.849751.
[25] Ahmadi D, Reza M. The use of technology in English language learning: a literature
review. Int. J. Res. English Educ. 2018;3(2):115–25.
[26] Belo N, McKenney S, Voogt J, Bradley B. Teacher knowledge for using technology
to foster early literacy: a literature review. Comput. Human. Behav. 2016;60:
372–83. https://doi.org/10.1016/j.chb.2016.02.053.
[27] Sabirli ZE, Çoklar AN. The effect of educational digital games on education,
motivation and attitudes of elementary school students against course access.
World J. Educ. Technol. 2020;12(4):325–38.
[28] Hanghøj T, Kabel K, Jensen SH. Digital games, literacy and language learning in L1
and L2: a comparative review. L1-Educ. Stud. Lang. Literature 2022:1–44. https://
doi.org/10.21248/l1esll.2022.22.2.363.
[29] Tsai YL, Tsai CC. Digital game-based second-language vocabulary learning and
conditions of research designs: a meta-analysis study. Comput. Educ. 2018;125:
345–57. https://doi.org/10.1016/j.compedu.2018.06.020. vocabulary instruction
in the urban middle school. The Reading Teacher, 64(1), 5-14.
[30] Zou D, Huang Y, Xie H. Digital game-based vocabulary learning: where are we and
where are we going? Comput. Assist. Lang. Learn. 2021;34(5–6):751–77. https://
doi.org/10.1080/09588221.2019.1640745.
[31] Lasley E. Twenty-first century literacy, game-based learning, project-based
learning. J. Literacy Technol. 2017;18(3):38–55.
[32] Moon AL, Francom GM, Wold CM. Learning from versus learning with technology:
supporting constructionist reading comprehension learning with iPad applications.
TechTrends 2021;65(1):79–89. https://doi.org/10.1007/s11528-020-00532-1.
[33] Collins A, Joseph D, Bielaczyc K. Design Research: theoretical and Methodological
Issues. J. Learn. Sci. 2004;13(1):15–42. https://doi.org/10.1207/
s15327809jls1301_2.
[34] Donaldson JP, Allen-Handy A. What is learning? A complex conceptual systems
analysis of conceptualizations of learning. Int. J. Educ. Res. Open 2023;4:100254.
https://doi.org/10.1016/j.ijedro.2023.100254.
[35] Piaget J. The origins of intelligence in children. W W Norton & Co; 1952.
[36] Vygotsky L. Interaction between learning and development. Read. Develop.
Children 1978;23(3):34–41.
[37] Bruner J. The culture of education. Harvard University Press; 1996.
[38] O’Donnell AM. Constructivism. In: Harris KR, Graham S, Urdan T, McCormick CB,
Sinatra GM, Sweller J, editors. APA educational psychology handbook, vol 1:
theories, constructs, and critical issues. American Psychological Association; 2012.
p. 61–84. https://doi.org/10.1037/13273-003.
[39] Papert S. The children’s machine: rethinking school in the age of the computer.
BasicBooks; 1993.
[40] Ackermann E. Piaget’s constructivism, Papert’s constructionism: what’s the
difference? Fut. Learn. Group Publication 2001;5(3):438.
[41] Kafai YB. Constructionism (editor). In: Sawyer RK, editor. The Cambridge
handbook of the learning sciences. Cambridge University Press; 2006. p. 35–46.
[42] Papert S, Harel I. Situating constructionism. In: Papert S, Harel I, editors.
Constructionism. Basic Books; 1991. p. 1–11.
[43] Papert S. Eight big ideas behind the constructionist learning lab (editor). In:
Stager GS, editor. Constructive technology as the key to entering the community of
learners. Philadelphia, PA: National Educational Computing Conference (NECC);
1999. p. 4–5. 2005.
[44] Glazer K. Imagining a constructionist game-based pedagogical model: using
tabletop role-playing game creation to enhance literature education in high school
English classes [Ed.D., Pepperdine University]. ProQuest Dissertations & Theses
Global; 2015.
[45] Resnick M, Rosenbaum E. Designing for tinkerability. In: Honey M, Kanter DE,
editors. Design, make, play: growing the next generation of stem innovators.
Routledge; 2013. p. 163–81.
[46] Holbert N, Berland M, Kafai YB. Introduction: fifty years of constructionism. In:
Holbert N, Berland M, Kafai YB, editors. Designing constructionist futures: the art,
theory, and practice of learning designs. The MIT Press; 2020. p. 1–20.
[47] Deci E, Ryan R. Intrinsic motivation and self-determination in human behavior.
Springer Science & Business Media; 2013.
12
Computers and Education Open 6 (2024) 100160
[48] Karakaya Cirit D. Coding in Preschool Science and Mathematics Teaching: analysis
of Scratch Projects of Undergraduate Students. Int. J. Contemporary Educ. Res.
2022;9(3):460–75.
[49] Lin Y, Weintrop D. The landscape of block-based programming: characteristics of
block-based environments and how they support the transition to text-based
programming. J. Comput. Lang. 2021;67:101075. https://doi.org/10.1016/j.
cola.2021.101075.
[50] Resnick M, Maloney J, Rusk N, Kafai Y. Scratch: programming for all. Commun.
ACM 2009;52:60–7. https://doi.org/10.1145/1592761.1592779.
[51] Burke Q, Kafai YB. The writers’ workshop for youth programmers: digital
storytelling with scratch in middle school classrooms. In: Proceedings of the 43rd
ACM technical symposium on Computer Science Education; 2012. p. 433–8.
[52] Gutiérrez KD. Developing a sociocritical literacy in the third space. Read. Res. Q.
2008;43(2):148–64. https://doi.org/10.1598/RRQ.43.2.3.
[53] Lincoln YS, Lynham SA, Guba EG. Paradigmatic controversies, contradictions, and
emerging confluences, revisited. In: Denzin NK, Lincoln YS, editors. The Sage
handbook of qualitative research. 4th editor. Sage; 2011. p. 97–128.
[54] DeVere Wolsey T, Smetana L, Grisham DL. Vocabulary plus technology: an after
reading approach to develop deep word learning. Read. Teach. 2015;68(6):
449–58. https://doi.org/10.1002/trtr.1331.
[55] Maxwell JA. Qualitative research design: an interactive approach. Sage
Publications, Inc; 2005.
[56] Fielding N, Fielding J. Linking data: the articulation of qualitative and quantitative
methods in social research. Sage Publishing; 1986.
[57] Scholastic. (2023). Guided reading leveling resource chart. Retrieved from: http://t
eacher.scholastic.com/products/guidedreading/leveling_chart.htm.
[58] MagiCore. (2020 a). Main idea and supporting details reading comprehension
passages anchor chart. Teachers Pay Teachers. https://www.teacherspayteachers.
com/Product/Main-Idea-and-Supporting-Details-Reading-Comprehension-Passage
s-Anchor-Chart-2009167.
[59] MagiCore. (2020 b). Nonfiction text features reading passages anchor charts
worksheets RI.2.7 RI.3.7. Teachers Pay Teachers. https://www.teacherspa
yteachers.com/Product/Nonfiction-Text-Features-Reading-Passages-Anchor-
Charts-Worksheets-RI27-RI37-2058493.
[60] MagiCore. (2020 c). Context clues reading comprehension passages and questions
RI.2.4 RI.3.4. Teachers Pay Teachers. https://www.teacherspayteachers.com/Pr
oduct/Context-Clues-Reading-Comprehension-Passages-and-Questions-RI24-RI3
4-1988742.
[61] Nunnally JC. Psychometric theory. 2nd ed. McGraw-Hill; 1978.
[62] Creswell JW. Educational research: planning, conducting, and evaluating
quantitative and qualitative research. Pearson; 2008.
[63] Merriam SB, Tisdell EJ. Qualitative research: a guide to design and
implementation. 4th ed. Jossey-Bass; 2016.
[64] Corbin JM, Strauss AL. Basics of qualitative research: techniques and procedures
for developing grounded theory. 4th ed. SAGE; 2015.
[65] Thornberg R. Informed grounded theory. Scandinavian J. Educ. Res. 2012;56(3):
243–59. https://doi.org/10.1080/00313831.2011.581686.
[66] Kuckartz U, Rädiker S. Analyzing qualitative data with maxqda. Switzerland:
Springer International Publishing; 2019. https://doi.org/10.1007/978-3-030-
15671-8.
[67] Donaldson JP. Conceptualizations of learning in the learning sciences and stem
education. Philadelphia, PA: [Ph.D. Dissertation, Drexel University]; 2019.
[68] McHugh ML. Interrater reliability: the kappa statistic. Biochemia Medica 2012;22
(3):276–82.
[69] Brennan RL, Prediger DJ. Coefficient kappa: some uses, misuses, and alternatives.
Educ. Psychol. Meas. 1981;41(3):687–99. https://doi.org/10.1177/
001316448104100307.
[70] Krippendorff K. Content analysis: an introduction to its methodology. SAGE; 2004.
https://doi.org/10.1177/1094428108324513. Second editor.
[71] Doerfel, M.L. (1998). What constitutes semantic network analysis? A comparison of
research and methodologies. Connections, 21(2), 16–26. https://doi.org/
10.1.1.211.9021.
[72] Creswell JW, Poth CN. Qualitative inquiry and research design: choosing among
five approaches. Sage publications; 2016.
[73] Lavrakas PJ. Encyclopedia of survey research methods (Vols. 1-0). Sage
Publications, Inc; 2008.
[74] Hanneman RA, Riddle M. Introduction to social network methods. University of
California; 2005. https://faculty.ucr.edu/~hanneman/nettext/C1_Social_Network_
Data.html. accessed October 25, 2021.
[75] Wäsche H, Dickson G, Woll A, Brandes U. Social network analysis in sport research:
an emerging paradigm. Eur. J. Sport Soc. 2017;14(2):138–65. https://doi.org/
10.1080/16138171.2017.1318198.
[76] Freedman D, Pisani R, Purves R. Statistics (international student edition). 4th
editor. WW Norton & Company; 2007.
[77] Girvan M, Newman ME. Community structure in social and biological networks.
Proc. Natl. Acad. Sci. 2002;99(12):7821–6. https://doi.org/10.1073/
pnas.122653799.
[78] Freeman LC. A set of measures of centrality based on betweenness. Sociometry
1977:35–41. https://doi.org/10.2307/3033543.
[79] Lam DP, Martin-Lopez B, Horcea-Milcu AI, Lang DJ. A leverage points perspective
on social networks to understand sustainability transformations: evidence from
Southern Transylvania. Sustainability Sci. 2021;16(3):809–26. https://doi.org/
10.1007/s11625-020-00881-z.
[80] Meadows DH. Leverage points: places to intervene in a system. The Sustainability
Institute; 1999.
M. Chowdhury et al.
[81] Pu M, Zhong Z. Development of a situational interaction game for improving
preschool children’ performance in English-vocabulary learning. In: Proceedings of
the 2018 International Conference on Distance Education and Learning. ACM;
2018. p. 88–92. https://doi.org/10.1145/3231848.3231851.
[82] Rahmi Z. Translating missions in James Worral’s game “grand theft auto’s
missions”. Adv. Lang. Literary Stud. 2018;9(1):44–7. https://doi.org/10.7575/
aiac.alls.v.9n.1p.44.
[83] Reinhardt J. Gameful second and foreign language teaching and learning: theory,
research, and practice. Springer; 2018.
[84] Sykes J, Reinhardt J, Thorne S. Multi-user gaming as sites for research and practice
(editor). In: Hult F, editor. Directions and prospects for educational linguistics.
Springer; 2010. p. 117–35.
13
Computers and Education Open 6 (2024) 100160
[85] Hickey DT, Filsecker M. Participatory learning assessment for organizing inquiry in
educational videogames and beyond. In: Littleton K, Scanlon E, Sharples M,
editors. Orchestrating inquiry learning. Routledge; 2012. p. 146–74.
[86] Rasinski T, Padak N, Newton J. The roots of comprehension. Educ. Leadership
2017;74(5):41–5.
[87] Rasti-Behbahani A. Why digital games can be advantageous in vocabulary
learning. Theory Practice Lang. Stud. 2021;11(2).
[88] Jensen SH. Gaming as an English language learning resource among young
children in Denmark. CALICO J. 2017;34(1):1–19.