Foundational bilingual education scholarship conceptualizes bilingualism as a cognitive and educational resource rather than a deficit-oriented condition (Cummins, 2000). More recent translanguaging perspectives further frame bilingualism as a dynamic and socially situated process through which learners mobilize their full linguistic repertoires to construct meaning (García, 2009; García & Wei, 2014; Wei, 2018). These developments align with the broader multilingual turn in applied linguistics, which emphasizes the fluid and socially embedded nature of multilingual practices (Cenoz, 2013; Conteh & Meier, 2014).

Despite these advances, a substantial proportion of the bilingual education literature continues to prioritize linguistic outcomes, often overlooking the broader sociocultural and institutional contexts within which learning occurs. In response, recent scholarship has increasingly adopted ecological and sociocultural frameworks that conceptualize education as a system of interrelated components operating across multiple layers of influence. This perspective is theoretically grounded in the ecological systems theory of Bronfenbrenner (1979) and further extended in bilingual education research by Hornberger (Hornberger, 2009). Within this framework, bilingual education is understood not merely as language instruction but as a complex socio-educational process shaped by dynamic interactions among learners, teachers, institutions, and communities.

The ecological orientation is particularly salient in multilingual and minority-language contexts, where language use is deeply embedded in social structures and power relations. As Hornberger (2009) argues, the sustainability of minority languages depends not only on their inclusion in formal education systems but also on their legitimacy within broader sociocultural environments. Consequently, the effectiveness of bilingual education programs must be evaluated in terms of their capacity to align linguistic, pedagogical, and contextual dimensions. This perspective provides a robust conceptual foundation for examining bilingual education through an ecosystem-informed lens, in which implementation-level effectiveness is associated with factors operating across micro, meso, and macro levels. This positioning reframes bilingual education effectiveness as an emergent, system-level outcome rather than a function of isolated linguistic variables. This multi-layer framework further offers policy-relevant insights by identifying strategic leverage points across systemic, instructional, and learner dimensions.

At the micro level, a substantial body of research has consistently emphasized the critical role of learner-related factors in shaping bilingual education outcomes. Internal learner characteristics, including motivation, attitudes, cognitive capacities, and sociocultural integration, are widely recognized as proximal determinants of both language acquisition and academic performance (Dörnyei, 2009; Norton, 2013; Pavia et al., 2019). Learners who exhibit strong intrinsic motivation and maintain positive orientations toward both their heritage and majority languages are more likely to develop balanced bilingualism and achieve sustained academic success. This perspective aligns with the theoretical arguments advanced by Cummins (2000) and further elaborated in translanguaging frameworks proposed by García and Wei (2014), which conceptualize bilingual competence as a dynamic and integrative process.

Language proficiency also plays a central role in bilingual education outcomes. Majority-language proficiency facilitates access to curricular content and academic participation (Thomas & Collier, 2002), while heritage-language competence contributes to cultural continuity, identity formation, and cognitive flexibility (Bialystok, 2011; Fishman, 1991). However, emerging research suggests that the influence of micro-level factors is neither uniform nor deterministic. Rather, their effects are contingent upon and often mediated by broader instructional and institutional conditions. For instance, while learner motivation and language ability are foundational, their impact can be amplified or attenuated by pedagogical practices, access to instructional resources, and the quality of institutional support (Norton, 2013). This perspective underscores the importance of situating learner-level variables within a broader ecosystem framework, where educational outcomes are understood as the result of dynamic interactions across micro, meso, and macro levels rather than isolated individual attributes. Collectively, the literature suggests that micro-level learner factors function as proximal but ecosystem-dependent predictors of bilingual education effectiveness, whose influence is shaped by broader instructional and institutional conditions.

At the meso level, instructional processes and institutional mechanisms play a pivotal role in translating policy objectives into observable educational outcomes. Within bilingual education, teacher-related factors, particularly teacher competence and engagement in professional development, have been consistently identified as central to effective program implementation (Echevarría et al., 2006; Babinski et al., 2018). Teachers who demonstrate both linguistic proficiency and pedagogical expertise are better positioned to design and deliver instruction that integrates language and content, incorporates culturally responsive practices, and accommodates diverse learner needs. In this regard, teacher competence functions not merely as an individual attribute but as a key institutional resource that shapes the quality and consistency of bilingual instruction.

Beyond teacher-related dimensions, recent scholarship has increasingly highlighted the importance of instructional mediation mechanisms, particularly information channels, as critical components of the bilingual education ecosystem. Information channels include formal and informal communication systems such as classroom interaction, digital learning platforms, extracurricular activities, and institutional communication structures. These mechanisms facilitate knowledge dissemination, instructional coordination, and learner engagement across multilingual educational environments (Bahari, 2022; Müller-Spitzer et al., 2018).

Conceptually, these instructional mechanisms function as intermediating structures that link macro-level policy and institutional conditions with micro-level learner processes. For example, digital tools and experiential learning activities can support the integration of linguistic and cultural knowledge, while structured communication systems extend learning beyond the classroom and promote continuity across learning environments. In this sense, meso-level mechanisms serve a translational function, converting institutional intentions into actionable pedagogical practices and learner experiences. This perspective aligns with research emphasizing that language education policies are often negotiated and operationalized through school-level actors and classroom practices rather than implemented uniformly from policy directives alone (Menken & García, 2010).

However, the effectiveness of these mechanisms is contingent upon their alignment with local contexts, institutional capacities, and learner needs. In practice, professional development programs often fail to adequately prepare teachers for the specific pedagogical and sociolinguistic challenges associated with bilingual education, resulting in implementation gaps between policy design and classroom enactment (Tran et al., 2023). This highlights the necessity of context-sensitive approaches to teacher training and instructional design, in which professional development and instructional systems are tailored to the sociocultural and linguistic realities of specific educational settings. Taken together, the literature suggests that meso-level instructional mechanisms, particularly information channels and professional development, function as the operational core through which broader systemic conditions are translated into learner-level bilingual education practices and outcomes.

At the macro level, bilingual education is shaped by a broad constellation of contextual and structural factors, including policy frameworks, resource allocation, community engagement, and sociocultural norms. These macro-level conditions define the institutional and environmental landscape within which bilingual programs are designed, implemented, and sustained. Contextual and institutional conditions, including school support structures, educator coordination, and community engagement, have been identified as important contributors to the implementation quality of bilingual education programs serving linguistically minoritized students (Hornberger, 2009; Hopkins et al., 2023, Hult, 2010). This interpretation is also consistent with Hornberger’s (2009) argument that multilingual education policies operate through layered institutional and sociocultural processes rather than through isolated instructional interventions. In particular, the availability of material and human resources, together with the clarity and consistency of policy directives, plays a decisive role in ensuring the fidelity and continuity of bilingual education implementation.

Beyond institutional conditions, community-level influences represent a vital dimension of the macro context, particularly in settings where formal education systems are closely embedded within local sociocultural structures. In the Mekong Delta, for instance, Khmer pagodas and community-based organizations function as key sites for language use, cultural transmission, and informal learning, thereby complementing formal schooling processes (Tran et al., 2023). Community engagement not only enhances the sociocultural relevance and legitimacy of bilingual education programs but also fosters positive learner attitudes and expands opportunities for authentic language practice. In this sense, community factors serve as a bridging mechanism between institutional policies and lived cultural experiences. Such complexities further reinforce the sociopolitical nature of bilingual education policy and implementation in multilingual societies (Epstein, 2011; Tollefson, 2006). In contrast, family factors may influence bilingual education indirectly through learner motivation, educational engagement, and access to learning resources. However, their effects are often conditioned by broader socioeconomic and institutional constraints (Davis-Kean, 2005).

From this perspective, bilingual education policies cannot be separated from the institutional and sociolinguistic environments in which language practices are regulated and reproduced (Spolsky, 2004). Taken together, these macro-level dynamics underscore that bilingual education effectiveness cannot be attributed to any single structural or contextual variable. Rather, outcomes emerge from the interaction of multiple influences operating across interconnected levels. This reinforces the necessity of adopting a systemic and multi-layer analytical approach, in which macro-level conditions are understood as enabling or constraining forces that shape, and are mediated by, meso-level mechanisms and micro-level learner processes within the broader bilingual education ecosystem. Accordingly, macro-level conditions are conceptualized as distal ecosystem factors that may influence bilingual education effectiveness through meso-level instructional and institutional processes, although these indirect pathways are not directly tested in the present study.

Despite the growing recognition of bilingual education as a multifaceted and context-sensitive phenomenon, the existing body of research remains notably fragmented. Much of the literature has tended to examine isolated variables or focus on single levels of analysis, thereby limiting the development of comprehensive theoretical models and constraining the generation of actionable policy insights. This fragmentation is particularly problematic in multilingual and resource-constrained contexts such as Vietnam, where bilingual education systems operate within complex socio-institutional environments. Crucially, there remains a lack of empirical studies that systematically integrate learner-level, instructional-level, and contextual-level factors within a unified analytical framework. As a result, the mechanisms through which these dimensions interact to shape educational effectiveness remain insufficiently understood.

To address these limitations, the present study adopts a multi-layer ecosystem perspective and proposes an integrated analytical model that captures the interrelationships among key determinants of bilingual education effectiveness. This approach is theoretically grounded in ecological and socio-educational system perspectives, which conceptualize educational processes as nested and interdependent systems operating across multiple layers of influence. By structuring the analysis across macro, meso, and micro levels, the study not only enhances theoretical coherence but also provides a more robust foundation for policy-relevant insights. This multi-layered perspective also reflects language policy scholarship emphasizing that educational language practices are shaped through interactions across multiple institutional scales (Hult, 2010).

At the macro level, the model encompasses structural, institutional, and sociocultural conditions that define the broader environment in which bilingual education is implemented. This conceptualization is rooted in Bronfenbrenner’s ecological systems theory, which posits that distal environmental systems, including policy frameworks, institutional arrangements, and community structures, exert foundational influences on human development (Bronfenbrenner, 1979). In bilingual education contexts, macro-level factors such as policy coherence, resource allocation, and community engagement establish enabling or constraining conditions for program implementation (Hornberger, 2009). Accordingly, constructs such as contextual and institutional factors (CIF), community factors (CF), and family factors (FF) are situated at this level, as they represent systemic forces external to immediate instructional processes yet critically shaping their effectiveness. Contextual and institutional factors (CIF) refer to the broader structural, organizational, and policy-related conditions that facilitate or constrain the implementation of bilingual education programs, including institutional support, infrastructural capacity, technological readiness, and policy coherence. These conditions collectively shape the institutional environment within which bilingual education practices are operationalized and sustained.

At the meso level, the framework captures the mediating mechanisms through which macro-level conditions are translated into pedagogical practices. This level corresponds to what ecological theory conceptualizes as the interactional space where institutional structures are enacted in practice. Within bilingual education, instructional mechanisms, particularly information channels (IC) and engagement in professional development activities (EPDA), function as key intermediaries that operationalize policy directives and institutional resources into classroom-level practices (Echevarría et al., 2006). These mechanisms facilitate knowledge exchange, pedagogical innovation, and communication among stakeholders, thereby serving as a critical translational layer linking systemic conditions with learning processes.

At the micro level, the model focuses on individual learner characteristics and cognitive-linguistic capacities that directly influence educational outcomes. This classification aligns with foundational theories in bilingual education and applied linguistics, particularly the interdependence hypothesis proposed by Cummins (Cummins, 2000), which emphasizes the role of language proficiency and cognitive factors in academic development. Constructs such as internal learner factors (ILF), Vietnamese language ability (VLA), and Khmer language ability (KhLA) are conceptualized as proximal determinants of learning outcomes. However, consistent with socio-cultural theory advanced by Vygotsky (Vygotsky, 1978), these individual processes are inherently embedded within broader social and instructional contexts, reinforcing the necessity of situating micro-level variables within a multi-layer analytical framework.

Importantly, the macro-meso-micro classification provides an analytically coherent framework for examining bilingual education as an interconnected socio-educational ecosystem. Rather than attributing effectiveness to isolated learner variables, this perspective recognizes that bilingual education outcomes emerge through interactions among learner-related, instructional, and institutional processes operating across multiple levels (García & Wei, 2014; Hornberger, 2009; van Lier, 2004). Although the proposed framework conceptually recognizes possible indirect relationships among ecosystem factors, the present study focuses primarily on direct structural associations as an initial empirical step toward examining the broader organizational dynamics of bilingual education. Future research may further extend the framework by investigating mediation and moderation mechanisms among instructional, contextual, and learner-related factors.

ILF positively influences KhVBE effectiveness.

VLA positively influences KhVBE effectiveness.

KhLA positively influences KhVBE effectiveness.

IC positively influences KhVBE effectiveness.

EPDA positively influences KhVBE effectiveness.

CIF positively influences KhVBE effectiveness.

CF positively influences KhVBE effectiveness.

FF positively influences KhVBE effectiveness.

This study employed a cross-sectional survey design to examine the determinants of KhVBE in the Mekong Delta, Vietnam. A purposive sampling strategy was adopted to recruit participants who possessed relevant professional expertise and were directly involved in the implementation of bilingual education programs, thereby ensuring the substantive relevance of the data to the research objectives (Hair et al., 2019). Prior to completing the survey, all participants were required to review and provide written informed consent through a statement presented at the beginning of the electronic questionnaire. The consent form clearly explained the objectives of the study, emphasized the voluntary nature of participation, and outlined key ethical considerations, including the confidentiality and anonymity of responses, as well as participants’ right to withdraw from the study at any time without penalty or adverse consequences. Only individuals who provided informed consent were permitted to proceed with the survey. Ethical clearance for the study was obtained from the Ethics Review Board of the University of Social Sciences and Humanities, Vietnam National University Ho Chi Minh City, ensuring that all procedures adhered to established standards for research involving human participants.

The final sample consisted of 269 respondents, including educational administrators and primary school teachers from three provinces, An Giang, Soc Trang, and Tra Vinh, where Khmer-Vietnamese bilingual education programs are actively implemented. In terms of gender distribution, male participants accounted for 36.4% of the sample, while females represented 63.6%, reflecting the demographic composition commonly observed in primary education contexts. Regarding professional roles, teachers comprised the majority of respondents (53.2%), with educational administrators accounting for 46.8%. The sample was overwhelmingly composed of bachelor’s degree holders (95.2%), with only a small proportion holding master’s degrees (0.7%), and the remainder reporting other qualifications. The majority of respondents (88.1%) had more than 10 years of professional experience, reflecting a highly experienced sample.

All participants were directly engaged in instructional delivery, student assessment, and/or program management, thereby constituting an appropriate unit of analysis for examining implementation-level dynamics of bilingual education effectiveness. Given the study’s focus on institutional and instructional processes, the use of practitioner-based respondents aligns with an implementation-oriented analytical perspective rather than a learner-outcome design. While the purposive sampling approach may limit the statistical generalizability of the findings, it enhances contextual relevance and interpretive validity, particularly in specialized educational settings where domain-specific expertise is required (Echevarría et al., 2006; Tran et al., 2023). Because the respondents were administrators and teachers rather than students, the study captures implementation-level perceptions of bilingual education effectiveness rather than direct learner outcomes. This distinction is important when interpreting learner-related constructs, which reflect practitioner assessments of learner conditions rather than students’ self-reported experiences or measured academic performance.

Data were collected using a structured questionnaire developed with reference to established theoretical frameworks in bilingual education and sociocultural learning (Cummins, 2000; García & Wei, 2014). All constructs were operationalized as reflective multi-item measures and assessed using a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). The dependent variable, KhVBE, was conceptualized with measurement indicators capturing multiple implementation-related aspects, including educational objectives, curriculum content, teaching approaches, and assessment practices. Independent variables were organized according to a multi-layer ecosystem framework, including micro-level factors (ILF, VLA, and KhLA), meso-level mechanisms (IC and EPDA), and macro-level conditions (CIF, CF, FF).

Data were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM). PLS-SEM was considered appropriate for three reasons. First, the study is prediction-oriented and seeks to explain variance in implementation-level KhVBE effectiveness rather than reproduce a covariance matrix for confirmatory theory testing. Second, the model includes multiple latent constructs and a relatively complex set of structural paths, making variance-based SEM suitable for estimating direct associations among ecosystem-informed determinants (Hair et al., 2019). The analytical procedure followed a two-stage approach. In the first stage, the measurement model was evaluated to assess the reliability and validity of the constructs. In the second stage, the structural model was examined to test the hypothesized relationships among constructs. Statistical significance was determined using a bootstrapping procedure, which is widely recommended for PLS-SEM analysis.

Convergent validity was assessed through the examination of standardized factor loadings, Cronbach’s alpha, rho_A, Composite Reliability (CR), and Average Variance Extracted (AVE). In accordance with the PLS-SEM guidelines, factor loadings and Cronbach’s alpha values above 0.60, together with rho_A and CR values exceeding 0.70 and AVE values greater than 0.50, were considered indicative of satisfactory convergent validity (Hair et al., 2019). Discriminant validity was evaluated using the Fornell-Larcker criterion, whereby the square root of the AVE for each construct was required to exceed its correlations with other constructs, thereby confirming that each construct represented a distinct conceptual domain (Fornell & Larcker, 1981).

The structural model was evaluated based on three principal criteria: explanatory power (R²), effect size (f²), and predictive relevance (Q²) (Cohen, 1988; Hair et al., 2019). Predictive relevance (Q²) was examined using the blindfolding procedure, whereby values greater than zero indicate that the model possesses adequate predictive capability. To minimize the potential influence of common method bias, the full collinearity diagnostics were conducted. The collinearity values (VIF) must be evaluated to ensure no multicollinearity issues among predictors.

Reliability of the measurement model was evaluated using Cronbach’s alpha coefficients, rho_A, Composite Reliability (CR), and standardized factor loadings. The findings demonstrated a high level of internal consistency across all constructs. Specifically, the Cronbach’s alpha coefficients were 0.961 for KhVBE, 0.983 for ILF, 0.974 for VLA, 0.980 for KhLA, 0.963 for IC, 0.930 for EPDA, 0.979 for CIF, 0.979 for CF, and 0.972 for FF. Furthermore, all constructs reported Cronbach’s alpha, rho_A, and Composite Reliability values exceeding the recommended threshold of 0.70, indicating satisfactory levels of internal reliability and construct stability. In addition, all standardized factor loadings were above the acceptable cut-off value of 0.60, confirming that the indicators adequately represented their corresponding latent constructs. Collectively, these results provide strong evidence for the reliability, stability, and consistency of the measurement scales used in the study.

Convergent validity was assessed through the examination of Average Variance Extracted (AVE) values. As presented in Table 1, all constructs achieved AVE values ranging from 0.579 to 0.925, thereby exceeding the recommended threshold of 0.50 recommended in partial least squares structural equation modeling research. These AVE values indicate that a substantial proportion of variance in the observed indicators is explained by their corresponding latent constructs. Collectively, the results provide strong evidence that the measurement items adequately capture the intended theoretical constructs, thereby confirming satisfactory convergent validity of the measurement model.

Discriminant validity was evaluated using the Fornell-Larcker criterion. The results revealed that the square root of the AVE for each construct exceeded its correlations with all other constructs, suggesting that each construct shared greater variance with its own indicators than with indicators associated with other constructs. This finding confirms that the constructs are empirically distinct and adequately represent unique theoretical domains within the proposed bilingual education ecosystem framework. As shown Table 2, each construct shares greater variance with its own observed indicators than with any other construct in the model. In addition, all the VIF values were below the threshold of 5.0, suggesting that severe multicollinearity was not present. However, several predictors approached this threshold, indicating that the interpretation of non-significant effects should be made with appropriate caution (as shown in Table 3). Collectively, these findings substantiate the model’s robust discriminant validity and establish a solid foundation for confidently proceeding with the evaluation of the structural model.

The significance of the hypothesized direct relationships was examined using the bootstrapping procedure with 5,000 resamples. As presented in Table 4, the structural model results reveal a differentiated pattern of relationships among micro-level, meso-level, and macro-level determinants of Khmer-Vietnamese bilingual education effectiveness (KhVBE), indicating that the proposed multi-layer ecosystem framework received partial empirical support rather than uniform confirmation. The figure 2 illustrates the structural model of the current study.

At the micro level, ILF did not demonstrate a statistically significant effect on KhVBE (β = 0.086, t = 1.233, p = 0.218). Similarly, VLA showed a positive but non-significant relationship with KhVBE (β = 0.129, t = 1.231, p = 0.219). These findings suggest that learner-related conditions, although theoretically relevant, may not directly explain variations in bilingual education effectiveness when broader institutional and instructional factors are simultaneously considered in the model. In contrast, KhLA exerted a statistically significant positive effect on KhVBE (β = 0.203, t = 2.369, p = 0.018), indicating that Khmer language competence plays a meaningful role in strengthening perceptions of bilingual education effectiveness, particularly within minority-language educational settings where heritage language maintenance remains a core pedagogical objective.

At the meso level, IC emerged as the strongest predictor in the model. IC demonstrated a statistically significant positive effect on KhVBE (β = 0.343, t = 6.109, p < 0.001), highlighting the central role of information channels, communication systems, school-based knowledge exchange, and instructional coordination mechanisms in enhancing the effectiveness of Khmer-Vietnamese bilingual education. The magnitude of this coefficient suggests that IC functions as a key implementation mechanism within the ecosystem framework. EPDA also demonstrated a statistically significant positive effect on KhVBE (β = 0.140, t = 2.026, p = 0.043), although the effect size was comparatively smaller. This finding indicates that engagement in professional development activities contributes positively to bilingual education implementation, particularly through strengthening pedagogical practices, instructional alignment, and teacher preparedness.

At the macro level, CIF did not exert a statistically significant direct effect on KhVBE (β = 0.005, t = 0.051, p = 0.959). Similarly, CF demonstrated a positive but non-significant relationship with KhVBE (β = 0.144, t = 1.229, p = 0.220), while FF also failed to demonstrate a statistically significant effect (β = −0.007, t = 0.068, p = 0.946). These findings suggest that broader contextual, institutional, community, and family-related conditions may not directly shape implementation-level perceptions of Khmer-Vietnamese bilingual education effectiveness within the present model. Instead, their influence may potentially operate indirectly through meso-level instructional systems, communication infrastructures, or teacher-related implementation mechanisms, although such indirect relationships were not examined in the current study.

The explanatory power of the model was assessed using the coefficient of determination (R²). The R² value for KhVBE was 0.411, while the adjusted R² value was 0.392. These results indicate that the proposed model explains approximately 41.1% of the variance in Khmer-Vietnamese bilingual education effectiveness, representing a moderate level of explanatory power according to commonly accepted PLS-SEM benchmarks. This suggests that the proposed ecosystem model demonstrates satisfactory capability in accounting for variations in bilingual education effectiveness within the Mekong Delta context. Effect size (f²) values were further examined to assess the relative contribution of each predictor to KhVBE. The results indicate that IC demonstrated the largest effect size (f² = 0.156), representing a moderate effect and confirming its role as the most influential predictor in the structural model. EPDA (f² = 0.021), KhLA (f² = 0.028), FF (f² = 0.008), and CIF (f² = 0.005) exhibited relatively small effects, while ILF (f² = 0.000) and CF (f² = 0.000) showed negligible contributions to the explained variance of KhVBE. Overall, these findings reinforce the dominant role of instructional communication mechanisms relative to other contextual and learner-related predictors. Predictive relevance was evaluated using the Stone-Geisser Q² statistic obtained through the blindfolding procedure. The Q² value for KhVBE was 0.220, which is substantially above zero, indicating satisfactory predictive relevance and confirming that the model possesses acceptable out-of-sample predictive capability.

Taken together, the findings indicate that KhVBE effectiveness is not uniformly shaped by all proposed predictors. Instead, the results reveal a selective and layered pattern of influence across the ecosystem model. IC functions as the strongest meso-level mechanism influencing KhVBE effectiveness, while KhLA provides meaningful micro-level support through heritage language competence. EPDA contributes positively, though to a lesser extent, reflecting the importance of professional learning and instructional capacity building. In contrast, ILF, VLA, CF, and FF do not demonstrate statistically significant direct effects within the present model. These findings suggest that the effectiveness of Khmer-Vietnamese bilingual education is shaped primarily by instructional communication systems, Khmer language competence, and targeted professional development mechanisms rather than by broad contextual or family-related conditions alone. This pattern further reinforces the importance of conceptualizing bilingual education as a multi-layer ecosystem in which different determinants operate with varying degrees of direct influence on implementation effectiveness.

The dataset supporting the findings of this study is available in Figshare: Determinants of Khmer-Vietnamese Bilingual Education Effectiveness, https://doi.org/10.6084/m9.figshare.32402682 (Vu Thanh Tam Nguyen, 2026).

The project contains the following underlying data:

Data are available under the terms of the Creative Commons Attribution 4.0 International License (CC BY 4.0).

The authors would like to express their sincere gratitude to the anonymous reviewers and editors for their valuable comments and suggestions, which substantially improved the quality of this manuscript. This research is funded by Vietnam National University Ho Chi Minh City under grand number B2024-18b-03.

•  Babinski LM, Amendum SJ, Knotek SE, et al.: Improving young English learners’ language and literacy skills through teacher professional development: A randomized controlled trial. Am. Educ. Res. J. 2018; 55(1): 117–143. Publisher Full Text
•  Bahari A: Affordances and challenges of technology-assisted language learning for motivation: A systematic review. Interact. Learn. Environ. 2022; 31: 5853–5873. Publisher Full Text
•  Baker C: Foundations of bilingual education and bilingualism. Multilingual Matters; 5th ed2011.
•  Bialystok E: Reshaping the mind: The benefits of bilingualism. Can. J. Exp. Psychol. 2011; 65(4): 229–235. Publisher Full Text
•  Bronfenbrenner U: The ecology of human development: Experiments by nature and design. Harvard University Press; 1979. Publisher Full Text
•  Bui TH: Challenges in implementing bilingual education for ethnic minority students in Vietnam. Journal of Asian Education and Development Studies. 2022; 11(3): 456–470. Publisher Full Text
•  Cenoz J: Defining multilingualism. Annu. Rev. Appl. Linguist. 2013; 33: 3–18. Publisher Full Text
•  Cohen J: Statistical power analysis for the behavioral sciences. Routledge; 2nd ed1988. Publisher Full Text
•  Conteh J, Meier G: The multilingual turn in languages education. Multilingual Matters; 2014. Publisher Full Text
•  Cummins J: Language, power and pedagogy: Bilingual children in the crossfire. Multilingual Matters; 2000.
•  Darling-Hammond L, Hyler ME, Gardner M: Effective teacher professional development. Learning Policy Institute; 2017. Publisher Full Text
•  Davis-Kean PE: The influence of parent education and family income on child achievement: The indirect role of parental expectations and the home environment. J. Fam. Psychol. 2005; 19(2): 294–304. PubMed Abstract | Publisher Full Text
•  DeJaeghere J, Dao V, Duong B-H, et al.: Learning inequities in Vietnam: Teachers’ beliefs about and classroom practices for ethnic minorities. Compare: A Journal of Comparative and International Education. 2021; 53: 399–416. Publisher Full Text
•  Dörnyei Z: The psychology of second language acquisition. Oxford University Press; 2009.
•  Echevarría J, Vogt M, Short DJ: Making content comprehensible for English learners. The SIOP model. Pearson; 3rd ed.2006.
•  Epstein JL: School, family, and community partnerships: Preparing educators and improving schools. Routledge; 2nd ed2011. Publisher Full Text
•  Fishman J: Reversing language shift: Theory and practice of assistance to threatened languages. Clevedon: Multilingual Matters; 1991.
•  Fornell C, Larcker DF: Evaluating structural equation models with unobservable variables and measurement error. J. Mark. Res. 1981; 18(1): 39–50. Publisher Full Text
•  Fullan M: Leading in a culture of change. Jossey-Bass; 2001.
•  Fullan M: The new meaning of educational change. Teachers College Press; 4th ed2007.
•  García O: Bilingual education in the 21st century: A global perspective. Wiley-Blackwell; 2009.
•  García O, Wei L: Translanguaging: Language, bilingualism and education. Palgrave Macmillan; 2014. Publisher Full Text
•  General Statistics Office of Vietnam: The 2019 Viet Nam population and housing census.2019. Reference Source
•  Hair JF, Hult GTM, Ringle CM, et al.: A primer on partial least squares structural equation modeling (PLS-SEM). Sage; 2nd ed2019. Publisher Full Text
•  Ho DT: Language barriers and academic achievement among ethnic minority students in Vietnam. Asia Pac. Educ. Rev. 2019; 20(3): 471–482. Publisher Full Text
•  Hopkins M, Nowak-Żółty E, Kurowska-Pysz J, et al.: Dimensions of bilingual education program implementation in schools serving linguistically minoritized students in the Czech-Polish border region. Frontiers in Education. 2023; 8: 1277575. Publisher Full Text
•  Hornberger NH: Multilingual education policy and practice: Ten certainties (grounded in indigenous experience). Lang. Teach. 2009; 42(2): 197–211. Publisher Full Text
•  Hult FM: Analysis of language policy discourses across the scales of space and time. Int. J. Sociol. Lang. 2010; 2010: 7–24. Publisher Full Text
•  Lambert WE: Culture and language as factors in learning and education.Wolfgang A, editor. Education of immigrant students. Ontario Institute for Studies in Education; 1975; pp. 55–83.
•  May S: Language and minority rights: Ethnicity, nationalism and the politics of language. Routledge; 2nd ed2012. Publisher Full Text
•  Menken K, García O: Negotiating language policies in schools: Educators as policymakers. Routledge; 2010. Publisher Full Text
•  Müller-Spitzer C, Koplenig A, Wolfer S: Using digital tools for language learning: A corpus-based perspective. ReCALL. 2018; 30(2): 1–20. Publisher Full Text
•  Nguyen T, Hamid M: Bilingualism as a resource: language attitudes of Vietnamese ethnic minority students. Current Issues in Language Planning. 2018; 19(4): 343–362. Publisher Full Text
•  Nguyen VD, Tran VD, Tran DN, et al.: Managing the quality assurance of education of students on ethnic boarding schools: A case study on the Mekong Delta, Vietnam. Educational process: International Journal. 2025; 17: e2025383. Publisher Full Text
•  Nguyen VTT: (DATASET) Determinants of Khmer-Vietnamese Bilingual Education Effectiveness.. [Data set]. Figshare. 2026. Publisher Full Text
•  Norton B: Identity and Language Learning: Extending the Conversation. Bristol: Multilingual Matters; 2nd ed2013. Publisher Full Text
•  Pavia N, Webb S, Faez F: Incidental vocabulary learning through listening to songs. Stud. Second. Lang. Acquis. 2019; 41(4): 745–768. Publisher Full Text
•  Spolsky B: Language policy. Cambridge University Press; 2004.
•  Thomas WP, Collier VP: A national study of school effectiveness for language minority students’ long-term academic achievement. Center for Research on Education, Diversity & Excellence; 2002.
•  Tollefson JW: Critical theory in language policy.Ricento T, editor. An introduction to language policy: Theory and method. Blackwell Publishing; 2006; pp. 42–59.
•  Tran VH, Nguyen TH, Pham LT: Community engagement and bilingual education in the Mekong Delta. J. Multiling. Multicult. Dev. 2023; 44(2): 123–140. Publisher Full Text
•  van Lier L : The ecology and semiotics of language learning: A sociocultural perspective. Kluwer Academic Publishers; 2004. Publisher Full Text
•  Vygotsky LS: Mind in society: The development of higher psychological processes. Harvard University Press; 1978.
•  Wei L: Translanguaging as a practical theory of language. Appl. Linguis. 2018; 39(1): 9–30. Publisher Full Text