Introduction

The widespread adoption of mobile devices has significantly impacted the way in which young learners acquire language skills. With a significant portion of school-aged children possessing internet-capable smartphones and tablet computers (Lenheart, 2010), mobile technologies have become increasingly prevalent in the daily lives of young students. These technologies not only provide learners with the flexibility to engage in language study anytime and anywhere (Burston, 2015; Reinders & Benson, 2017; Shadiev et al., 2017), but also offer cost-effective solutions for language education (Yu et al., 2022).

Despite the growing popularity of mobile learning apps for dominant languages such as English and Chinese (e.g., Duolingo, ELSA, HelloChinese), less attention has been paid to less commonly taught languages by developers and researchers (Lake & Beisly, 2019; Loewen et al., 2019; Song et al., 2022). Additionally, many language instructors in regions such as Thailand face challenges with inadequate vocabulary and grammar knowledge and technological advancement (Poonpon, 2021). Thus, the development of effective mobile learning solutions to support language teachers and enhance the vocabulary, sentence structure, and grammar learning outcomes of young dual-language learners presents a critical challenge in the current educational context.

This exploratory study aims to address this challenge by developing and implementing the RILCA mobile learning app to enhance vocabulary, sentence structure, and grammar learning outcomes for young dual-language learners in elementary schools. The study will compare the learning outcomes of students who use the RILCA app to those who do not, with the goal of answering the overarching research question: How does the effectiveness of the RILCA mobile learning app compare to traditional modes of learning in terms of improving vocabulary, sentence structure, and grammar learning outcomes?

Literature review

Mobile-assisted language learning (MALL) is a rapidly growing area of research in language education (Booton et al., 2021; Gutiérrez-Colón et al., 2020; Lin & Lin, 2019). Touchscreen devices, such as tablets, smartphones, iPods, and PDAs, are increasingly being used for both formal and informal language education (Yang & Xie, 2013; Yu et al., 2022). Despite the potential limitations of smaller screen sizes, restricted audio-visual quality, and limited battery life (Chinnery, 2006), MALL appeals to the new generation of language learners who have grown up with technology.

The increasing prevalence of mobile devices in children's daily lives is transforming the educational landscape. According to Rideout and Robb (2020), 97% of children living at home have used a mobile device, with 61% of children between the ages of 5 and 8 owning their own tablet. Children as young as 3 years old are capable of independently tapping and swiping on these devices, presenting new opportunities for education (Marsh et al., 2018; Vatavu et al., 2015a, 2015b). In Thailand, data show that there were 95.60 million cellular mobile connections at the start of 2022, increasing by 3.8 million (+ 4.1%) from 2021 to 2022, equivalent to 136.5% of the total population (Kemp, 2022). The widespread use of mobile devices offers exciting prospects for language and literacy development beyond physical learning environments.

During the crucial developmental period between the ages of 3 and 11, young learners undergo substantial advancements in language and literacy skills, including vocabulary, grammar, and pronunciation (Booton et al., 2021). Many children also study a second language during this time, either as a medium of instruction or as a foreign language (Global Education Monitoring Report Team, 2016; Qi, 2016). Mobile devices have the potential to significantly impact the development of language and literacy skills in both native and target languages. The use of specific digital technologies, such as mobile apps, has been positively associated with the acquisition of two languages, for example dual-language dictionary apps (Godwin-Jones, 2011), flashcard apps (Godwin-Jones, 2017), and digital book apps (Rivera et al., 2014).

Empirical evidence supports the positive effects of instructional apps on young learners’ vocabulary, sentence, and grammar comprehension (Coogle et al., 2018; Dore et al., 2019; Dunn, 2015; Hsiao & Chen, 2015; Liao et al., 2013; Rogowsky et al., 2018; Xin & Affrunti, 2019). For instance, Hsiao and Chen (2015) found that students who used eReaders had significantly higher comprehension scores compared to those who read printed texts. Another study showed that the use of a commercial app, such as Book Writer, led to a moderate increase in young learners’ expressive vocabulary (Dennis, 2016). Lin and Lin (2019) also found that mobile-assisted L2 vocabulary retention was more effective than traditional methods.

Furthermore, mobile technology has also been linked to improvements in other literacy areas, such as sentence structure and grammar knowledge (Dunn, 2015), writing (Patchan & Puranik, 2016), and phonological skills (Lan et al., 2007). For example, a mobile phone-based phonological skills training programme for elementary students resulted in increased interaction compared to traditional instruction (Lan et al., 2007). Similarly, students in a mobile-supported task-based language teaching (TBLT) group performed better on vocabulary and conversation comprehension assessments compared to a regular TBLT group (Fang et al., 2021). These findings highlight the need for further investigation.

Method

Participants

In this study, 146 students (57 males and 89 females) in Grade 2, who began the study of English as a compulsory subject in grade 1, were divided into three groups: Group A (62 participants with 30 in the control group and 32 in the experimental group) from a Bangkok-based demonstration school, Group B (33 participants with 15 in the control group and 18 in the experimental group) from a Surat Thani mainland public school, and Group C (51 participants with 26 in the control group and 25 in the experimental group) from a Surat Thani public school on Samui Island.

The students in the control group received traditional face-to-face instruction from their teachers, which was conducted in a physical classroom through the use of traditional methods such as lectures, practices, and hands-on activities to present the material. The students in this group did not receive any additional instruction through technology or other means.

In contrast, the experimental groups received supplementary instruction via the beta version of the RILCA mobile app, which was installed through TestFlight. Teachers from three schools volunteered to implement the app with their learners and participated in an orientation to test RILCA's efficacy as a dual-language learning tool. For the 15-week study, the researchers provided each experimental group with two 11-inch iPads and additional peripherals, including HDMI cables, USB converters, sim cards, and Apple pencils.

Materials

RILCA mobile learning app on iPad

RILCA is a revolutionary mobile learning app for enhancing the dual-language learning efficiency of young learners. It is designed for both instructor-led learning, in which the entire class completes each activity or lesson simultaneously, and student-paced learning, in which learners can progress through each activity or lesson at their own pace outside of the class using their own tablets. RILCA is not only an acronym for “Research Institute for Languages and Cultures of Asia,” where this project is being launched, but it is also intended to represent the concept of “Revolution In teaching Languages and Cultures for Active learners", which is central to our endeavour. RILCA was a prototype project, developed by using Unity Pro software as a native mobile app that could only be installed on iOS-powered mobile devices for this study. Therefore, RILCA was attempted to be designed as an iPad-compatible app for elementary school pupils to acquire Thai and English. The app’s features were developed based on the outcomes of previous research (Thumvichit et al., 2023), which allowed to simultaneously build the technical and pedagogical usability of the app. According to this, there are five key components: (1) app attribute, (2) lesson content, (3) immersive environment, (4) learning strategy, and (5) instructional use. These investigation-identified factors serve as a direct reference for us to construct the RILCA for young learners with the goal of boosting their dual-language outcomes. The app also contains listening features for the pronunciation of English vocabulary, allowing pupils to practise their accents.

The present study aimed to advance language proficiency in Thai and English by leveraging the mother tongue as a foundational tool for foreign language learning, as advocated by Premsrirat (2018). This course was crafted to enhance comprehension of vocabulary, sentence structure, and grammar, while fostering language mastery through a repetition-based approach. Students were introduced to relevant vocabulary, sentence structure, and grammar, followed by engaging in comparative activities within each module, progressing to the next session upon reaching a predetermined accuracy rate. To gauge mastery, a game-based evaluation was conducted prior to moving on to the next module. The learning materials, adapted from the “Happy Campers” textbook series (Llanas & Williams, 2021), were transformed into bilingual content for five modules, as illustrated in Table 1, with some app content depicted in Fig. 1.

Table 1 Overview of all modules on the RILCA app
Full size table
Fig. 1
figure 1

Examples of RILCA’s screen on the iPad

Full size image

Once a week for fifty minutes, the class was taught in Thai. Although Thai was primarily used to clarify grammar issues for practical reasons, such as time limits, English was employed to encourage activity participation and classroom discussion. As our participants were young language learners with limited conversational competence, the activities were designed to help them understand basic grammar and vocabulary so that they could construct a variety of sentences (e.g., affirmative, negative, and interrogative). During class time, a teacher’s iPad was connected to a television to create a larger screen on which young learners could view and interact with activities by touching their teacher’s iPad.

Dual language achievement tests

The researchers constructed parallel pre- and post-tests, with 35 multiple-choice items to assess students' dual-language competency by integrating Thai and English. In accordance with the course’s curriculum objectives, students were required to enhance their vocabulary, sentence structure, and grammatical knowledge. Thus, the dual-language achievement test focusses on (1) vocabulary, sentence structure, and grammar knowledge and (2) the functions of vocabulary, sentence, and grammar in everyday conversation. The vocabulary part featured 11 questions with associated pictures or fill-in-the-blanks. The sentence and grammatical parts included 12 questions each. The grammatical comprehension test presented learners with communicative situations in which they chose the expression that most effectively filled a conversational gap. The highest possible score was 35.

To ensure test validity, the pre- and post-tests were administered to 60 second-grade students, divided into two groups of 30 each, with one group taking the pre-test and the other group taking the post-test. The results showed that 17 questions from the pre-test were considered usable, accounting for 48.57% of the items, with 42.86% of the items revised and 8.57% eliminated. In the post-test, 14 questions were considered usable, accounting for 40.00% of the items, while 54.29% of the items were revised and 5.71% were removed.

The questions were selected using criteria set forth by Crocker and Algina (2008) and Lawthong (2004), ensuring that the question difficulty level ranged from difficult to quite easy, with at least one effective decoy and a power of discrimination between low and very good. Any item that did not have the power of recognition, with two ineffective decoys, was discarded.

Procedure

After obtaining signed consent forms, the study adhered to the outlined procedure depicted in Fig. 2. In order to determine students' baseline linguistic abilities, pre-test scores were collected prior to the intervention. For example, the implementation of Module 1 involved the following steps:

Fig. 2
figure 2

Experimental procedure

Full size image

In the experimental group, the RILCA app was linked to a wide-screen television or projector, and the vocabulary and grammar of the week were introduced using the app. The first interactive button, “My Body”, was selected, and students were tasked with identifying various body parts and their corresponding English names. They were encouraged to take notes in personal notebooks and complete a sub-test in which they matched English words to different body parts.

The second interactive button, “Family Members”, was selected, and students were asked to guess the family relationships depicted in various images. They were guided in reading, writing, and spelling both Thai and English words related to the selected image. This process was repeated until all images were covered, followed by a sub-test that required students to match Thai and English words for 15 items.

The final interactive button, “Family Game”, was selected, and students were challenged to match Thai words to corresponding English words and pictures. Worksheet 1 from https://tinyurl.com/rilcasheet was downloaded and distributed to students, providing opportunities for further reinforcement of their knowledge through interactive activities.

For the control group, the teacher initiated a conversation regarding the terminology of various body organs and used PowerPoint displaying Module 1, which aligned with the content of the RILCA app. The content of “My Body” was presented, and students were asked to select a body part and practice reading, spelling, and writing both Thai and English words. A quiz was provided for students to select the correct English terms for various body parts.

The content of “Family Members” was also presented to the control group, and students were asked to guess the family relationships depicted in various images and practice reading, writing, and spelling both Thai and English words. The content of “Family Game” was also presented, and students were challenged to match Thai words to corresponding English words and pictures. Worksheet 1 was distributed for further practice and enhancement of students' knowledge.

The entire experiment was completed over a period of fifteen weeks, with 50-min weekly sessions. At the conclusion of the trial, students from both groups took a parallel dual-language proficiency test and were debriefed and thanked for their participation. The results of this study demonstrate the effectiveness of the RILCA app in enhancing students' linguistic knowledge compared to traditional instructional materials.

Analysis

To select the particular data analysis programme, we first examined the normality of the data distribution (Tables 2 and 3).

Table 2 Test of normality within experimental groups
Full size table
Table 3 Test of normality within control groups
Full size table

The results of the Kolmogorov–Smirnov and Shapiro–Wilk tests in experimental groups are distinct. According to Table 2, both tests suggest an abnormal distribution (p < 0.05) at the 0.05 significance level for pre- and post-vocabulary, pre- and post-sentence in all groups, pre- and post-grammar in Groups A and B. However, both the tests indicate a normal distribution (p > 0.05) at the 0.05 significance level for pre- and post-grammar in Group C.

Using the Kolmogorov–Smirnov and Shapiro–Wilk tests, the normality of the control groups was determined. Contrasting results are shown in Table 3, which reveals an abnormal distribution (p < 0.05) at the 0.05 significance level for pre-vocabulary, pre-sentence, and pre-grammar, in all groups, post-vocabulary in Groups A and C, and post-sentence and post-grammar in Groups A and B. The Kolmogorov–Smirnov and Shapiro–Wilk tests, nonetheless, demonstrate a normal distribution (p > 0.05) at the 0.05 significance level for post-vocabulary in Group B and for post-sentence and post-grammar in Group C.

These tests yield diverse or contradictory outcomes for other variables. Therefore, we opted to analyse the non-normally distributed data using the Wilcoxon signed-rank test to compare pre- and post-intervention within dependent groups, the Mann–Whitney U test to evaluate outcomes between experimental and control groups, and the Kruskal–Wallis H test to determine whether there are statistically significant differences across independent groups.

Results

In light of the contradictory results of data distribution, we decided to analyse the data using the Mann–Whitney U test instead of the independent t-test, the Wilcoxon test instead of the dependent t-test, which does not require a particular probability distribution of the dependent variable, and the Kruskal–Wallis H test instead of a one-way ANOVA. The research question aimed to compare the effectiveness of the RILCA mobile learning app with traditional modes of learning in terms of improving vocabulary, sentence structure, and grammar learning outcomes. To account for learners' initial disparities in language competency prior to the intervention, the Mann–Whitney U test was conducted using pre-test scores as the covariate and post-test scores on vocabulary, sentence structure, and grammatical comprehension as the dependent variables. As presented in Table 4, the test shows the rank averages and the differences between the two groups for each of the language components before and after the intervention. As comparisons are made, there are differences in mean ranks between the two groups.

Table 4 Mean and sum of ranks
Full size table

The Mann–Whitney U test results illustrated in Table 5 reveal the statistical significance of the mean rank differences between the groups in Table 4. According to the test results, the post-intervention differences in three linguistic components between the two groups are statistically significant (p < 0.05). These data demonstrate the RILCA app has distinct effects on both users and non-users.

Table 5 Mann–Whitney U test results of the differences between groups
Full size table

The Wilcoxon test results reveal a significant difference between the experimental group’s pre- and post-test scores. In addition, there are no statistically significant differences between the pre- and post-tests in the control group. Table 6 displays the differences between the experimental group's pre- and post-test results. The Z value in the table is negative (scores of − 7.101, − 7.125, and − 6.870), and the p value is less than 0.05, indicating that the dual-language accomplishment scores increased after the experimental groups are treated with the RILCA app.

Table 6 Wilcoxon test results of the differences between the pre- and post-test scores of the experimental group
Full size table

When examining the results of comparing the pre- and post-test scores of the control group in Table 7, it is concluded that there are no significant differences (p > 0.05).

Table 7 Wilcoxon test results of the differences between the pre- and post-test scores of the control group
Full size table

As depicted in Table 8, the Kruskal–Wallis H test results reveal that there are statistically significant differences in pre-vocabulary outcomes among the three experimental groups, \(X\) 2(2) = 36.471, p < 0.05, with a mean rank of 54.69 for Group A, 19.03 for Group B, and 30.30 for Group C; there are statistically significant differences in pre-sentence scores among the three groups, \(X\)2(2) = 28.164, p < 0.05, with a mean rank of 52.81 for Group A, 21.89 for Group B, and 30.64 for Group C; there are statistically significant differences in pre-grammar values among the groups, \(X\)2(2) = 26.776, p < 0.05, with a mean rank of 52.77 for Group A, 24.11 for Group B, and 29.10 for Group C.

Table 8 Results of Kruskal–Wallis tests for the experimental group
Full size table

The sub-scales of dual-language achievement also reveal significant differences in post-vocabulary scores (\(X\) 2(2) = 18.708, p < 0.05, with a mean rank of 50.38 for Group A, 28.53 for Group B, and 28.98 for Group C), post-sentence (\(X\) 2(2) = 24.371, p < 0.05, with a mean rank of 51.55 for Group A, 33.22 for Group B, and 24.10 for Group C), and post-grammar (\(X\)2(2) = 10.000, p < 0.05, with a mean rank of 42.22 for Group A, 45.44 for Group B, and 27.24 for Group C).

Similar to the experimental group, the Kruskal–Wallis H finding confirms statistically significant differences among the three control groups (Table 9). There are significant distinctions in pre-vocabulary values among the groups, \(X\)2(2) = 14.615, p < 0.05, with a mean rank of 46.65 for Group A, 28.93 for Group B, and 27.79 for Group C; there are significant differences in pre-sentence values among the three groups, \(X\)2(2) = 15.117, p < 0.05, with a mean rank of 46.83 for Group A, 30.03 for Group B, and 26.94 for Group C; there are significant differences in pre-grammar scores among the groups, \(X\)2(2) = 22.461, p < 0.05, with a mean rank of 49.33 for Group A, 29.93 for Group B, and 27.58 for Group C.

Table 9 Results of Kruskal–Wallis tests for the control group
Full size table

The sub-scales of young learners’ dual-language accomplishment also demonstrate significant discrepancies in post-vocabulary outcomes (\(X\)2(2) = 19.917, p < 0.05, with a mean rank of 47.63 for Group A, 21.00 for Group B, and 31.23 for Group C), post-sentence (\(X\)2(2) = 15.717, p < 0.05, with a mean rank of 45.03 for Group A, 19.63 for Group B, and 35.02 for Group C), and post-grammar (\(X\)2(2) = 26.643, p < 0.05, with a mean rank of 50.17 for Group A, 21.23 for Group B, and 28.17 for Group C).

Discussion

Despite the increasing body of research that emphasises the efficacy of MALL, there is still a need to investigate the impact of mobile technologies on the dual-language development of young learners. To address this research gap, we developed the RILCA app, which aims to facilitate Thai-English language learning for young pupils. Our experimental results demonstrate that the use of the RILCA app, in conjunction with face-to-face instruction, significantly improved the vocabulary, sentence structure, and grammar comprehension of learners in all experimental groups compared to learners in control groups who used traditional materials (Dunn, 2015; Fang et al., 2021; Hwang et al., 2014; Li & Hegelheimer, 2013; Liu, 2016; Liu & Chen, 2015; Loewen et al., 2019; Rivera et al., 2014; Song et al., 2022). Specifically, in experimental groups, learners in Group A outperformed those in Groups B and C. Therefore, we discussed the impact of the RILCA app on Group A's language learning outcomes separately from the other groups.

Our findings suggest that the utilisation of mobile devices, particularly the RILCA app, significantly impacted the growth of language and literacy skills in both native and target languages. The app's activities enhanced the vocabulary, sentence structure, and grammar comprehension of young dual-language learners and made face-to-face teaching and learning more engaging, especially for young children who require high levels of stimulation. Mobile-supported settings offered possibilities for learners to interact with virtual objects, aiding in language processing (Lan et al., 2015). The app was also structured to emphasise form and provide opportunities for learners to apply newly acquired vocabulary and structures to a variety of sentences, which facilitated the formation of form-meaning associations and long-term memory retention (Ellis, 2005).

Further analysis revealed that Group A, regardless of whether they were in the experimental or control group, was more effective than other groups, possibly due to the learners in Group A coming from demonstration schools, while the other groups came from public schools in Surat Thani. Apart from regional differences, there were also differences in other contexts, such as in the management of bilingual education, including Thai and English, as well as other important languages. Demonstration schools are managed by faculties of education of universities that have both local and international teachers, experts, media, and various learning resources. Moreover, the learners themselves have higher levels of knowledge than the average learners (Lianyang et al., 2015), which may be attributed to the willingness of parents to support additional learning for learners outside the classroom. As a result, learners in demonstration schools are more proficient in using English in their daily lives and are confident enough to engage in effective English conversations with others (Sha'ar & Boonsuk, 2021).

Regardless, when comparing the scores of each group, the young learners in Group A, which consist of the experimental group and the control group from the demonstration school, showed differences in their proficiency in vocabulary, sentence structure, and grammar after learning through the RILCA app. This is because the demonstration school has established a vision and mission to develop students who aim to excel in university entrance exams and value academic excellence and competition to receive academic awards that lead to academic success (Tongaht, 2010). In addition, the demonstration school also promotes the production of academic materials using new innovative technology to provide academic knowledge that corresponds with the changes in the current world. This includes the use of technology to facilitate the daily life processes of students, such as computer skills, English, and other languages (Uppamaiathichai et al., 2017). Therefore, it is not surprising that the scores of the young learners in Group A were higher before and after learning compared to those of the learners in Groups B and C.

As mentioned earlier, Groups B and C in this study consist of students from small rural schools with limited resources for education, including budget, personnel, teachers, teaching materials and equipment, facilities, and learning resources. Resource constraints in small schools have been widely acknowledged and found to adversely impact the quality of education for students (Chaiheng et al., 2013). In addition, previous reports have emphasised that small schools have lower quality than large schools (Office of Education Reform, 2001). Educational resource readiness is a key factor in determining the quality of education in schools, with high resource readiness being associated with higher-quality education. Nevertheless, the study found that the RILCA app can help small schools access educational resources, particularly to enhance language proficiency among students. It is obvious that the experimental group exhibited significantly better post-test scores in vocabulary, sentence structure, and grammar skills compared to the control group, indicating that the RILCA app can be instrumental in improving language learning outcomes in young learners. These findings are significant for mobile technology integration in language learning contexts and can inform future research in this area.

Our study's results have significant implications for language learning contexts, especially for young learners, as mobile technologies like the RILCA app can enhance language learning outcomes by providing engaging and interactive activities that promote vocabulary acquisition, sentence structure, and grammar comprehension. Nevertheless, the impact of the app on language learning outcomes may vary depending on the context, as demonstrated by our findings.

Conclusion and limitations

The present study offers valuable insights for both language teachers and mobile learning researchers by providing empirical evidence and design specifications for the use of mobile-supported activities in dual-language classes. The results of this study indicate that the use of the RILCA app led to significant improvements in vocabulary, sentence structure, and grammar among young learners. Furthermore, the study highlights the potential of RILCA to alleviate the difficulties of teaching and learning a dual language in Thai elementary classrooms where English is not the medium of instruction, which can be a significant challenge for monolingual teachers in Thailand.

The study not only examines the role of mobile technology in supporting classroom instruction, but it also makes a valuable contribution to the field of MALL by identifying potential solutions to the challenges faced by dual language classrooms. The findings of this study have important implications for language teachers and educational policymakers as they suggest the potential of mobile technology to enhance language learning outcomes and support monolingual teachers in non-native language instruction.

However, these findings should be interpreted with caution due to the study’s limitations. The investigation was conducted in three classroom settings at a specific time and on a specific topic, which resulted in some variability in the RILCA-based activities provided by each teacher. This variability, combined with participant attrition, may have influenced the study outcomes. Further qualitative data collected through interviews with teachers will shed light on their experiences in managing mobile-supported classrooms and overcoming these challenges. Additionally, the results of this study are limited to Thai and English contexts and may not be generalisable to other settings.