Digital game-based learning to support language learning

DGBL research, however, is still in its early stages. A meta-analysis by Acquah and Katz (Reference Acquah and Katz2020) found that DGBL research has primarily focused on the implementation of digital games in or in tandem with formal classroom instruction. Butler (Reference Butler2022) points out the need for more studies that investigate how digital learning games work when used outside formal classroom settings. This is an important direction for DGBL research from the perspective of understanding how digital games impact not only language development but also the promotion of learner autonomy. An increasing number of games leverage technology to foster more self-direction for learners in their own learning experiences. Self-direction is generally understood as an autonomous process in which learners are the ones who initiate their own learning experiences, seeking out resources to facilitate their learning and enable them to meet their learning goals. This means that learners who are highly self-directed exhibit high levels of both autonomy and perseverance (Knowles, Reference Knowles1975). In a recent large-scale questionnaire-based study of online L2 learning involving 615 learners residing in 29 countries, Paradowski and Jelińska (Reference Paradowski and Jelińska2023) found that self-direction was an essential component, along with motivation, of successful outcomes for online L2 learning. With self-directed learning playing an increasingly central role in the gameplay of digital games for L2 learning, commercial language learning platforms are integrating components that enable their users to explore virtual environments (e.g., ImmerseMe, https://immerseme.co/, Pearson Education’s Mondly, https://www.mondly.com/vr), select their own material (e.g., Babbel’s flexible lesson models), and stay motivated with in-app prizes and friendly competition (e.g., Duolingo’s learning “streaks” and Friends Quests).

Research on the impact of DGBL outside formal classroom settings, specifically with young FL learners, has explored the relationship between gaming and a range of variables, including vocabulary. For example, Jensen (Reference Jensen2017) found that English-related extramural gaming among 8- and 10-year-old Danish learners of English was significantly related to higher vocabulary scores. In a study of 10- and 11-year-old Swedish learners of English, Sundqvist and Sylvén (Reference Sundqvist and Sylvén2014) explored young learners’ gameplay as well as the relationship between gameplay and factors such as gender, motivation, self-assessed English ability, and self-reported strategies related to speaking English, revealing some positive trends for learners who were frequent gamers. Building on this research, Sundqvist (Reference Sundqvist2019) undertook a large-scale, longitudinal study following 1,340 Swedish adolescent language learners of English in their ninth year of formal schooling (ages 15–16). The study explored the relationship between the types of games adolescent language learners play, the amount of time they played the games, and their L2 English vocabulary development. To investigate the relationship between informal language learning, formal language instruction, and overall vocabulary development, Sundqvist identified four distinct categories within her adolescent learners: (a) nongamer or those who did not report playing any games, (b) single-player (SP) games in which learners only interacted with the game interface, (c) multiplayer (MP) games where multiple players played locally on the same device, and (d) massively multiplayer online (MMO) games where players from around the world can play with one another simultaneously. Data on the learners’ gameplay habits were collected via questionnaires and interviews. The adolescent learners’ language proficiency was measured through two vocabulary tests (receptive and productive), an essay, national test scores, and year-end report cards. The results showed that the two strongest predictors of L2 English vocabulary development were time spent playing games and social interactions while playing games (i.e., those who played MP games or MMO games performed better than those who did not play games or played alone). Lastly, those who played SP games performed better than the nongamers who did not have extracurricular exposure to L2 English language via any type of gameplay.

Additionally, DGBL has been studied as part of a larger, formal, classroom-based setting, such as digital review games commonly used by L2 teachers to review discrete skills or lexical items. For example, in their longitudinal study of collaboration, engagement, and competition investigating language use during in-class L2 vocabulary repetition games (via Kahoot!, a digital quiz platform), Sandlund and colleagues (Reference Sandlund, Sundqvist, Källkvist and Gyllstad2021) demonstrated that learners used embodied actions like response cries, complaints, and assessments to co-construct and display their orientation to L2 vocabulary learning. Over the course of four academic years, the research team collected classroom observation video data from six multilingual English classrooms in Sweden (each group was in year 9 when they participated).

These studies provide evidence in support of what has been called extramural language learning (by Sundqvist, Reference Sundqvist2009), being used to positively supplement formal language education and to encourage games in which interaction with other players is incorporated. This paper will use the term extramural language learning as opposed to informal language learning or naturalistic language learning because, as Sundqvist and Sylvén (Reference Sundqvist and Sylvén2016) put it, this kind of language learning

However, interpreting and synthesizing results of different studies on DGBL is somewhat complicated by the operationalization of terms like “playing games,” as well as high levels of variation in learners’ individual gameplay experiences and logistical difficulties in collecting detailed, reliable data at the individual level in relation to learners’ in-game engagement (see All et al., Reference All, Castellar and Van Looy2016 for a discussion). Furthermore, while adolescent learners are considered young learners in the field of second language acquisition (SLA), it is difficult to generalize results from studies such as Sundqvist’s to the much younger population represented in the current study. Fewer studies have focused on very young learners’ experiences with digital learning games specifically designed to be played outside classroom settings, particularly those games and apps capable of recording individual-level data about learners’ in-game engagement and choices.

The current study aims to fill this gap by shedding light on one such implementation of a mobile app-based digital game—ABCmouse English—designed for young EFL learners to use anywhere at any time. The app was designed based on a research-driven, game-based approach to L2 learning, combining incrementally developmentally sequenced lessons with unstructured, exploration-based environments, giving learners a high degree of control over what, how, and at what pace they learn.

English education of young children in Japan

Educational policy in Japan includes the use of English examination scores for college admissions or job opportunities, making English learning for young children very important. English education in Japan has been compulsory in secondary grades since just after World War II (Shimizu, Reference Shimizu2010). In elementary schools, English language education began in 1992 when the Ministry of Education, Culture, Sports, Science and Technology (MEXT) selected two elementary schools and one junior high school in Osaka to implement experimental English activities in public education (Triana, Reference Triana2017). In 2002, MEXT announced a proposal to have students achieve communicative competence in English and provided new guidelines for teaching English. These included improving the teachers’ skills and having English taught through the medium of English (Shimizu, Reference Shimizu2010). In 2014, timed with the announcement of Tokyo as the seat of the 2020 Olympics, MEXT announced a plan to reform English education so that the instructional goals are aligned with globalization. This included nurturing the foundations for communication skills starting in the third and fourth grades (MEXT, 2014). The 2014 reform plan also included measures to improve the teaching skills of elementary school class teachers and promoted the use of assistant language teachers (ALTs)—foreign nationals with native-level English skills. However, the plan did not include any guidelines for English language education for students in grades 1 and 2, suggesting a lack of guidance for English education for younger children.

In this context, many Japanese parents seek extramural English lessons for their children or self-help bilingual parenting resources to try to increase their children’s exposure to English at home (Nakamura, Reference Nakamura2023). Some of these resources recommend media as the source of English language input for children, recognizing that Japanese parents may not have adequate English language proficiency to provide the linguistic input that would help their children’s English vocabulary and grammatical development (Nakamura, Reference Nakamura2023). The app employed in the current study provides a self-guided game that young Japanese children can use to develop communicative skills in English from early on and, thus, is well suited for use in the Japanese context.

Developing a digital language learning app informed by principles of learning, motivation, and second language acquisition

Studies show that app and game design and learner engagement are critical factors that determine an app’s effectiveness in promoting positive learning outcomes (Berkling & Gilabert Guerrero, Reference Berkling, Gilabert Guerrero, Koivisto and Hamari2019; Chen, Reference Chen2018). In this study, the collaborative team of curriculum specialists, English teachers, second language researchers, and game developers who created ABCmouse English sought to build a game-based app that would keep learners engaged and motivated to continue learning while providing input and output opportunities that matched and drove developmental levels. The team considered theories of motivation, L2 learning, and L2 assessment to create an app that was developmentally appropriate, engaging, and designed for use in informal learning settings (see Arndt, Reference Arndt2023).

ABCmouse English promotes engagement and motivation through an activity-based game environment designed to foster independent, active learning. Learners are encouraged to spend 15 to 20 minutes a day using the app, which offers them relatively compact and highly engaging play sessions. Players (henceforth, “learners,” as we discuss them in relation to their experiences playing to learn English) exercise autonomy by freely choosing activities from three different play environments. One environment is a structured learning path that guides learners along a sequence of activities with incrementally more developmentally advanced language objectives. The interactive games in the app provide linguistic problems of increasing difficulty (e.g., new vocabulary words or sounds that may be challenging to pronounce), keeping the experience within, but at the outer edge of the learner’s competence (Gee, Reference Gee2005). A second environment is an open exploration area that offers opportunities for learners to repeat and review content that they have already encountered on the learning path without being exposed to new language forms and is designed to promote linguistic automatization. The third environment in the app is a free-play area where learners can use tickets that they have earned through gameplay, exploring puzzles, songs, videos, and games based on their interests. Learners can customize their learning by choosing their games and environments and trying out new activities. New language is introduced in contexts that the children have selected based on their interests, increasing the chance that they will be motivated and engaged.

Each activity throughout ABCmouse English is designed to last about three minutes, with a play session intended to consist of approximately 5 to 6 activities (i.e., a typical play session would last about 18 minutes). This design was to leverage the extramural learning setting to promote learner autonomy in two ways. First, learners have agency over their own learning experiences by choosing what to focus on every few minutes. Second, learners can decide how many sessions to engage in and when to do those sessions. Although the app does not limit how long a learner can engage with the game in one sitting, the recommended usage is 15 to 20 minutes at a time to avoid learner fatigue while maximizing engagement.

Importantly, ABCmouse English was designed based on key principles from L2 learning research. Studies show that learners learn best through technology when doing tasks or activities that use language to accomplish a goal rather than activities aimed solely at learning the grammatical aspects of the language (e.g., Heift et al., Reference Heift, Mackey and Smith2019). Activities in the app used in this study were designed to engage young learners in authentic contexts (i.e., situations that are relevant or familiar to children based on their lives outside of the language learning app) (Long, Reference Long2015; Reference Long2016), which has also been shown to increase engagement and interest (Pinter, Reference Pinter2019). The activities were designed to promote immersive learning experiences involving familiar topics such as family, pets, toys, food, numbers, and colors so that learners need to understand the language to complete an activity. In other words, the focus is on meaning, but linguistic form is important for comprehension (e.g., Doughty & Long, Reference Doughty and Long2003; Long, Reference Long2015; Long & Robinson, Reference Long, Robinson, Doughty and Williams1998). The app used “focus on form,” which, in the L2 research literature, is distinct from a focus on formS, as in grammar drills. Rather, a focus on form, is where the context is meaning, but form understanding is important. Focus on form has been shown to be effective in multiple primary studies (e.g., Gholami, Reference Gholami2022; Rassaei, Reference Rassaei2020) as well as metanalytic work in instructed L2 research (Lee & Lee, Reference Lee and Lee2022; Norris & Ortega, Reference Norris and Ortega2000; see also Doughty & Williams, Reference Doughty and Williams1998).

The activities in the game encourage learners to construct knowledge through trial and error as they produce the L2 while receiving developmentally appropriate types of implicit and explicit feedback when they need it (Mackey, Reference Mackey2012). Learners are also exposed to meaningful, comprehensible input through pedagogical tasks and content (e.g., song-based input, followed by pronunciation games). Learners’ developmental levels are considered throughout the games, which build on the learners’ developing knowledge (Gass, Reference Gass2017). In designing the app, L2 research principles were supplemented with considerations of measurement and assessment. The Common European Framework of Reference (CEFR) standards (Council of Europe, 2020), the American Council on the Teaching of Foreign Languages (ACTFL) guidelines (ACTFL, 2012)—along with developmental sequence work on what is learnable and teachable (e.g., Pienemann, Reference Pienemann1998)—were all considered as underpinnings for the app, including the sequence in which language is presented in the game’s learning path. Figure 1 below provides examples of activities in the app.

Figure 1.

Sample Activities in ABCmouse English.

The app design was also informed by spaced repetition theory in SLA, which posits that the timing of review and practice affects learning, and that practice can be more effective when spaced out over time rather than when grouped together (Nakata & Suzuki, Reference Nakata and Suzuki2019; Rogers, Reference Rogers2017; Seibert Hanson & Brown, Reference Seibert Hanson and Brown2020; Zhang et al., Reference Zhang, Zou and Xie2022). Based on this research and other work on efficient memory-building techniques, the app repeats and reviews content in an algorithm-driven pattern to enhance long-term retention (Tabibian et al., Reference Tabibian, Upadhyay, De, Zarezade, Schölkopf and Gomez-Rodriguez2019). Also incorporated into the design of the app is a consideration of the importance of learning from scaffolded target language input (Bang et al., Reference Bang, Olander and Lenihan2020; Mahan, Reference Mahan2022; Moeller & Roberts, Reference Moeller, Roberts and Dhonau2013), formulaic sequences (Ellis, Reference Ellis2012; Wray, Reference Wray2000; Reference Wray, Siyanova-Chanturia and Pellicer-Sánchez2018), and a focus on communication. All of these are woven into the presentation of the learning path’s lessons, practice, and production skills, with interactive activities designed to foster the development of fluency (Shrum & Glisan, Reference Shrum and Glisan2009).

The current study focuses on understanding the outcomes and experiences of young Japanese L1 children who used a research-driven self-guided language learning app to learn English as a foreign language in nonclassroom settings. Specifically, the research has two main goals: to (a) determine the extent to which a self-guided English learning app used outside school, can effectively help young learners develop English skills and (b) understand the relationship between in-game choices and the development of English skills.

Limitations and avenues for future research

The current study set out to better understand whether and how a game-based app impacted second language learning in young children in a foreign language learning context outside the classroom. Pre- and posttest data were collected using Zoom, and one of the recruitment criteria was access to a desktop or laptop computer to conduct the language tests, in addition to access to a mobile device on which children could use the app. This inevitably eliminated a segment of the population and may have introduced some bias in the sample. Moreover, Zoom itself, at the time of data collection, was not a very familiar tool for many of the young children and thus may have inhibited some of their responses given that they were interacting with adults unknown to them via a medium with which they were not very familiar. Although their adult caretakers were close by, they were asked not to intervene in providing answers for the children, their abilities during both pre- and posttests may have been impacted by the medium.

This study involved collecting in-game tracking data of children’s play behavior along with pre- and posttest data. Linking what children are doing in games to consequential differences in learning patterns has clear implications for both theory and practice. Understanding the patterns in the children’s play behaviors (i.e., the sequence of activities and the proportion of time children spent in different types of structured versus unstructured learning activities) can be helpful for recommendations about how app-based games can be designed to encourage more children to adopt specific play behaviors conducive to learning. Our future work focuses on these aspects of the study. We also note that while this particular app has been examined in China and Japan, the findings might not hold true in different cultural contexts, and thus, tests in other settings are warranted. It is also important to note, as discussed in the methodology section, children were either given ABCmouse English and asked to continue with what they usually did (or didn’t do) or asked to continue whatever they usually did (or didn’t) do in relation to English language learning and were not asked to use the app. In other words, some children in both groups might have been taking English classes. This was a conscious design choice made to investigate the effect of an app on the regular lives of children. However, it could have impacted the results. Despite these limitations, the current study has shown that the app facilitated learning and that the different ways in which children engaged with the app were associated with different learning patterns. There is currently a great deal of interest in apps of all kinds, along with cautions from groups and countries about overreliance on technology and the potentially negative impact of screen time. This study has pointed to the learning benefits that can be derived from carefully designed apps such as the one tested here, which was designed based on L2 research findings and SLA principles.

Additionally, because the app (and this research project as a whole) is aimed at identifying helpful educational tools to improve learning outcomes, it is possible that despite the research team’s best efforts to avoid any coercion to play the game, some parents—eager to maximize exposure to educational content for their children—may have pushed their children to use the app, regardless of instructions from the research team not to do so. This may have contributed to the consistently high level of usage of the app throughout the 16-week period of the study, in contrast to other studies on language learning apps that have documented more sporadic use (e.g., García Botero et al., Reference García Botero, Questier and Zhu2019). Future research could explicitly ask parents at the end of the study if this was the case.

This study provides an early attempt at examining how young language learners may benefit from using educational apps outside of classroom settings. Since there are a limited number of studies with young language learners, mostly with small samples, it would be beneficial to conduct this type of study with larger samples to identify and share more generalizable patterns. Similarly, as this study investigated second language learning of English, future research could explore how the app-driven language development of young learners of languages other than English compares to those in this study and similar studies. Additionally, this study tested just one app, but a variety of apps of this nature exist, not all of them designed based on L2 research, and examining how different apps perform can help us identify the relative importance of features that are associated with learning.

Also, while the current study employs a randomized control design with the control group being “business-as-usual,” without the app, future research might include a second control group comprised of children who receive similar amounts and types of English input in a self-guided context, but without the resources and affordances of the application. While it may not be possible to achieve a truly balanced design, balancing—to the extent possible—variables such as language exposure, input, and output would help us better understand the effectiveness of the application versus that of other resources outside the classroom.

As mentioned in the discussion, an extension of this study could include a stimulated recall interview to help better understand why and when young learners decide to replay (or not) activities within the application. With one of the affordances of game-based learning being self-directed features, it would be illuminating to ask young learners (in their L1s) about their reasoning for the choices they made when using the different features found within the language learning application. Obviously, answers would be limited by the ages and cognitive abilities of the children, but from the perspective of educational app developers, it is vital to understand how these data, as well as in-game tracking data of the children’s play behavior, might be linked to consequential differences in learning gains. Analyzing the patterns in children’s play behavior (e.g., the sequence of activities and the proportion of time children spent in different types of structured versus unstructured learning activities, the choices the children make in choosing between structured and unstructured activities) would be helpful in developing game mechanics in apps aimed at encouraging more children to adopt specific play behaviors that are conducive to positive language development.

Another positive affordance of technology-assisted language learning is that technology can be more accessible than language classes or teachers for many. However, while educational technology is becoming increasingly more accessible every day, there are still populations who do not have access to the necessary technology to participate in the study.

Notwithstanding these inevitable limitations in the current study, it sheds light on an underresearched population (i.e., young learners) in the growing field of technology-assisted language learning. Given the findings in favor of app-based learning, continuing to investigate how language learning applications can best be utilized for language development will likely benefit the field of second language acquisition.