Purpose This study investigates the challenges faced by Vietnamese master's students enrolled in bilingual and English-Medium Instruction (EMI) programs in Taiwan. It examines how institutional, linguistic, cultural, financial, and health-related factors shape their academic and social adaptation within the bilingualization framework of Taiwan's Bilingual 2030 Policy (BN2030). By doing so, it highlights implementation gaps and the sociolinguistic tensions embedded in Taiwan's push for bilingual higher education. Design/methodology/approach A mixed-methods design was employed. Six semi-structured interviews with Vietnamese master's students were analyzed using thematic analysis to capture lived experiences. To complement these qualitative insights, Latent Dirichlet Allocation (LDA) topic modeling was applied to the interview transcripts, enabling computational validation of emerging themes. The research adapts Cheung's (2013) Transactional Acculturation Stress and Adjustment Model (TASAM) to interpret the dynamic process of adaptation. Findings Four major challenges emerged: socio-cultural adjustment, academic challenges, financial constraints, and health and well-being challenges. Students reported uneven EMI delivery, linguistic hierarchies, limited institutional support, and reliance on peer networks. LDA analysis confirmed these patterns and revealed hidden subthemes of identity negotiation and coping strategies. Research limitations/implications This exploratory study is based on a small sample of six Vietnamese master's students in Taiwan; therefore, the findings are not statistically generalizable. The results identify emerging patterns of adaptation rather than causal relationships. Future research may expand to other Southeast Asian student groups and employ comparative or longitudinal designs to examine institutional responses across disciplines, thereby strengthening generalizability and further refining the TASAM framework in bilingual and international higher-education contexts. Practical implications This study demonstrates how combining qualitative thematic analysis with Latent Dirichlet Allocation (LDA) can provide institutions with richer, evidence-based insights into international student challenges. Practically, universities can adopt similar mixed-method approaches to systematically analyze student feedback and detect hidden patterns that may not emerge from surveys alone. Beyond methodological innovation, the findings emphasize the need for consistent EMI practices, Mandarin language support, and culturally inclusive services. By leveraging computational tools like LDA alongside traditional methods, administrators can make data-driven decisions that directly improve student well-being and institutional competitiveness. Social implications This study highlights the social dimensions of bilingualization and international student inclusion in Taiwan. Vietnamese master's students' experiences reveal how language policies intersect with identity, belonging, and equity in multicultural academic settings. Strengthening institutional trust and intercultural communication can reduce social isolation and promote empathy among local and international communities. The LDA–TASAM integration also demonstrates how data-informed qualitative research can uncover hidden social mechanisms of adaptation, providing actionable evidence for creating linguistically inclusive environments that foster cross-cultural understanding and contribute to Taiwan's broader goals of social sustainability and educational globalization. Originality/value This study contributes by combining qualitative and computational approaches under a single framework (TASAM–LDA). It provides new insights into how Southeast Asian students experience Taiwan's bilingualization in practice and suggests directions for more inclusive EMI implementation.
Vo et al. (Wed,) studied this question.