INTRODUCTION

English has long been a strategic language in the context of globalisation. It supports participation in international education, research, trade, and mobility. Many education systems therefore position English proficiency as a form of national capacity. As a result, English language education is often framed as both an economic and a social development priority. In Vietnam, English language education has been shaped by large-scale national planning particularly during the last two decades. The National Foreign Languages Project (NFLP) 2020 initiated a system wide reform agenda for foreign language learning, with English as a key focus (Prime Minister, 2008). Policy adjustment continued through Decision No. 2080 in 2017, which extended and revised the national scheme through 2025 (Prime Minister, 2017). This decision re-emphasised a central priority of improving English proficiency among teachers and learners, while also strengthening curriculum innovation, assessment reform, and technology integration.

A further policy shift occurred with Decision No. 2371 in 2025, which advanced an agenda for English to function as a second language in schools and expanded the role of English as a language of instruction in schooling contexts (Prime Minister, 2025). The progression from proficiency targets to expanded instructional use signals a qualitative change in national expectations. Changes in policies are more likely to bring more effective outcomes when they are aligned with the research agendas. Previous reviews show that research priorities sometimes align with national agendas but sometimes reflect broader international movements. For example, a review of 180 articles in Thai Scopus-indexed ELT and applied linguistics journals found that second or foreign language learning and teaching dominated, with vocabulary ranked as the most popular focus, and survey research identified as the most frequent design (Phoocharoensil, 2022).

In the subfield of self-assessment research in the Korean EFL context, co-occurrence analysis identified multiple thematic clusters, including writing assessment, formative assessment, learner autonomy, and listening skills (Lee, 2025). A more recent bibliometric review of artificial intelligence (AI) in language teaching and learning in the global context reports that influential work in this domain has largely focused on technological solutions and empirical studies of AI applications, with particular attention to effectiveness and implementation methods (Ma et al., 2024). The emergence of ChatGPT as a high-frequency keyword in 2023 is documented, signalling an acceleration of attention to generative tools (Ma et al., 2024). The review also highlights inadequate attention from research to academic integrity and related risks. Such findings in global shifts are useful references for interpreting the national research landscape of Vietnam and its policies.

These research trends across various contexts highlight the need for a focused review of TESOL and EFL research trends in Vietnam. Decision No. 2080 places sustained emphasis on proficiency development and improved assessment and evaluation mechanisms, while Decision No. 2371 raises the stakes by framing English as a language of instruction, which implies new demands for teacher capacity, classroom practice, and system readiness (Prime Minister, 2017, 2025). International trend evidence suggests that research communities often concentrate on measurable outcomes, survey-based designs, and skill specific subfields, while technology related agendas can expand rapidly and bring new ethical and pedagogical questions. When the Vietnamese national agenda expands toward English mediated instruction, an updated overview is needed to assess where research attention aligns with that expansion and where potential gaps remain.

This scoping review study responds to that need by analysing the titles and abstracts of submissions accepted for presentation at the VietTESOL International Convention (VIC) across five consecutive years leading up to 2025. This venue is chosen as it is the largest forum for TESOL researchers and scholars in Vietnam. Its conference abstracts offer a timely lens on emerging concerns because they often reflect immediate professional problems and pedagogical strategies. They also capture work before it appears in journal publication cycles. In addition, conferences commonly include both research papers and workshops, which means the corpus can reflect both research-oriented priorities and practice-oriented responses. By mapping trends in titles and abstracts, the current study aims to provide an updated account of what Vietnamese TESOL and EFL scholars and practitioners foreground during a period of policy transition.

METHODOLOGY

This study adopts a corpus design that combines quantitative lexical trend mapping with qualitative interpretation of discourse contexts in titles and abstracts, like the one employed in Smith and Oskoz (2023). Many review studies in TESOL and EFL rely on systematic review procedures that code full texts for themes, settings, and research designs. Others use bibliometric methods that model publication and citation networks. In a similar fashion, the present study uses accepted conference titles and abstracts as the primary unit of evidence. This choice prioritises topical immediacy and practical relevance. It also allows the study to capture early signals of emerging concerns that may not yet appear in journal publications. In addition, the present approach differs from purely automated topic discovery as it retains an interpretive layer. Keywords are not adopted from the keywords section but identified through counting rules. They are then interpreted through close reading of the local contexts in which the keywords occur in the titles and abstracts. This combination supports both pattern detection and substantive meaning making.

Corpus and Sampling

The dataset consists of accepted submissions from VIC across five consecutive years, from 2021 to 2025. Each record includes a title and an abstract. Records represent both research paper submissions and workshop submissions. Titles and abstracts were analysed separately because they serve different communicative functions. Titles are compressed and may be rhetorical. Abstracts are longer and usually more explicit about aims, methods, and findings. This separation reduces the risk of treating title phrasing as a direct proxy for research content. It is necessary to note that VIC is the largest annual forum for teachers and scholars in Vietnam to report their works in TESOL/ELT over the last decade. In 2025, for example, the event attracted approximately 900 participants and presenters.

Counting Principle and Lemmatization

The unit of analysis is the record. Counts are based on document frequency rather than raw frequency because the number of records varies by year. The number of accepted abstracts also depends on the convention venue, as smaller venues accommodate fewer presentations. A word contributes at most one count within a record. This applies separately to the Title field and the Abstract field. A restricted synonym mapping is applied to three concept families, namely teacher (teacher, lecturer, and educator), learn (learn and study), and student (student, pupil, and learner). If a record contains both teacher and educator in the title, it still contributes one count to that concept family for that record. This rule reduces over aggregation and supports comparison across years.

Lemmatization is applied to consolidate inflectional variants into a shared base form. Plural nouns are normalised to singular forms. Regular verb inflections are reduced to base forms. A small set of irregular forms is also normalised. Lemmatization is applied after synonym mapping and before counting. This ensures that the vocabulary used for analysis is already lemmatised, and multiple members of a synonym family within the same record do not increase counts.

All text processing is case insensitive. Punctuation is removed. Numbers and non-alphabetic texts are excluded. Compound expressions are treated as separate words. Words shorter than two characters are excluded to reduce noise. Functional words are removed using an English stopword list. This step ensures the analysis on content words that are more likely to index topical focus and analytic framing. In addition, contextual and abstract move words that do not represent research foci are removed from the counting. These include obvious contextual markers such as Vietnam or foreign, non-topical ordinal terms such as language or instrument, and abstract moves such as purpose and future. This step supports interpretive clarity when summarising what each trend represents. To standardise keyword frequencies across years with different numbers of accepted records, the average counts are calculated. This means that the frequency count was divided by the total number of records for that year to obtain an average per record value. These averages therefore represent the proportion of titles or abstracts in a given year that contained the keyword.

Trend Analysis

Trends are identified using the five annual average values. Each word is assigned to one of four groups, namely upward trend (requiring a consistent increase across the five years), downward trend (requiring strict decreases year by year), constant trend (requiring low dispersion across the five years), and fluctuating trend (including all remaining non monotonic patterns). To interpret Title trends, the synthesis focuses on the most salient keywords within each trend group. For the fluctuating group in Titles, interpretation is based on the top five keywords after contextual word removal. To interpret Abstract trends, the synthesis uses an additional salience threshold after contextual word removal. Only keywords with a mean average of at least 0.05 across the five years are retained for qualitative synthesis. This threshold reduces attention to low frequency items that are more sensitive to sampling variation.

For each trend group, keywords are interpreted through their discourse contexts in titles and abstracts. Keywords are clustered into coherent topic strands where appropriate. The resulting summaries are used to describe dominant topical emphases, shifts in framing, and recurring research concerns. When patterns diverge between Titles and Abstracts, the Abstract based trend interpretation is prioritised because abstracts are more descriptive of study content and less affected by rhetorical compression.

FINDINGS

As presented in Table 1, the number of accepted submissions has steadily increased over this five-year period (2021-2025). In 2025, the number of submissions is approximately six times higher than that of 2021. It is necessary to note that abstracts of the Pre-Convention Workshops and Technology Demonstrations are excluded. The Keynote and Featured Presentations are treated as Oral Presentations. This section reports the key trends identified from the title corpus and the abstract corpus. The analysis groups lemmatised content keywords into four trend types, namely upward, downward, constant, and fluctuating, based on changes in average per record frequency over time. Findings are first presented for Titles and then for Abstracts.

Trends in Titles

There are five keywords shown in the upward trend as presented in Table 2. These are integrat, academic, service, rural, and quality. Titles associated with integrat (standing for integrate family) most often included pedagogical integration. The integration was commonly described as integration of skills, integration of tools, or integration of new instructional approaches. Titles associated with academic reflected sustained attention to academic English and academic literacy. Many titles referred to academic writing and academic performance in tertiary settings. The keyword service was mainly associated with preservice teacher education and teacher preparation. Rural indicated a growing visibility of rural settings and rural education constraints. Quality pointed to quality assurance and quality improvement agendas in teaching, assessment, and program evaluation. As a set, these upward keywords foreground a movement toward system-oriented concerns and classroom implementation issues that are directly relevant to the continuing policy emphasis on improved proficiency, improved assessment, and improved evaluation mechanisms in the period covered by Decision No. 2080 in 2017 (Prime Minister, 2017). They also anticipate the expanded instructional expectations implied by Decision No. 2371 in 2025 (Prime Minister, 2025).

Table 2. Keywords Identified in the Upward Trend from Title Corpus

In the downward trend, there were only three keywords identified as presented in Table 3. These were online, covid, and distance. Their decline indicates that titles became less likely to frame studies through pandemic related labels. Early titles frequently foregrounded remote delivery, online learning conditions, and pandemic disruption. Later titles used these labels less often. This pattern matches a wider international movement away from crisis framing as online provision becomes normalised rather than exceptional (Ma et al., 2024).

Table 3. Keywords Identified in the Downward Trend from Title Corpus

Similarly, there were only two keywords identified to be in the constant trend across the five years. These were learn and assessment. The keyword learn includes the synonym set learn and study in the current analysis. The stability of learn signals that learning remains the central construct across shifting topics. The stability of assessment suggests persistent interest in testing, feedback, and measurement. This stability is consistent with the prominence of assessment related clusters reported in other national landscapes, including work on assessment focused subfields in the Korean EFL context (Lee, 2025).

Table 4. Keywords Identified in the Constant Trend from Title Corpus

The fluctuating trend identified a set of top five keywords, consisting of skill, speak, classroom, writ, and major as presented in Table 5. These keywords reflect recurring attention to core language skills and classroom practice, with periodic shifts in emphasis. Speak and writ point to alternating waves of interest in speaking and writing. Major reflects a repeated comparative framing between English majors and non-majors. The fluctuation suggests that the skill focus and participant focus vary by year, rather than moving in a single direction.

Table 5. Keywords Identified in the Fluctuating Trend from Title Corpus

Trends in Abstracts

As the Abstract corpus is considerably larger than the Title corpus, only keywords with a mean average of at least 0.05 were retained for interpretation. This produces more salient and content focused themes than Titles because abstracts provide more explicit study detail.

The keywords identified in the upward trend included qualitative, quantitative, integrat, engagement, proficiency, cultural, plan, and evaluation as described in Table 6. Two patterns were prominent. The first pattern is a rise in explicit research design labelling. Qualitative and quantitative increased across the period. This pattern is compatible with review evidence from other national contexts where quantitative survey designs and structured reporting are prominent, such as Thai Scopus-indexed ELT and applied linguistics journals (Phoocharoensil, 2022). It is also consistent with the growing use of review and bibliometric approaches in the wider field, which foreground explicit categorisation of methods and themes (Lee, 2025).

Table 6. Keywords Identified in the Upward Trend from Abstract Corpus

The second pattern concerns substantive foci. Integrat and engagement increased together, indicating more frequent attention to integrated pedagogies and learner engagement outcomes. Proficiency also increased, suggesting a stronger emphasis on proficiency as an outcome variable or a framing construct. Evaluation increased, which indicates an expanding focus on evaluating materials, tools, programs, or learning outcomes. Cultural and plan point to a rise in intercultural or cultural concerns and in planning-oriented constructs, such as lesson planning or program planning. These upward abstract trends are consistent with international work that identifies an expanding research footprint for technology and innovation agendas, often framed through effectiveness and implementation concerns (Ma et al., 2024).

The downward trend identified a set of four keywords, namely online, covid, pandemic, and blend. Covid and pandemic declined, reflecting reduced reliance on explicit pandemic framing over time. Online and blend also declined. This suggests a move away from broad online delivery mode labels. It likely reflects a shift toward more specific naming of tools, practices, or contexts, rather than general modality descriptors.

Table 7. Keywords Identified in the Downward Trend from Abstract Corpus

The constant trend identified a set of six keywords, namely learn, skill, classroom, activity, test, and performance as presented in Table 8. This group indicates stability in two domains. First, it confirms the enduring pedagogical core of learning and skills within classroom activity. Second, it confirms stable attention to evaluation through test and performance. The continuing prominence of assessment related reporting is consistent with the sustained visibility of assessment themes in other national contexts, including clusters reported for self-assessment and formative assessment work (Lee, 2025).

Table 8. Keywords Identified in the Constant Trend from Abstract Corpus

The fluctuating trend, as partly presented in Table 9, included interview, questionnaire, survey, practice, strategy, group, structur, technology, knowledge, development, experience, challenge, need, effective, positive, class, and speak. This group shows shifting methodological emphasis and shifting problem solution framing across years. Interview, questionnaire, and survey show that data collection preferences vary over time, with alternating salience for interview based and survey-based studies. Structur (standing for structure family) often co-occurs with interview in the phrase semi structured, indicating fluctuating visibility of interview design conventions. Group is consistent with intermittent waves of experimental or quasi experimental comparison designs. This pattern resonates with the dominance of survey research reported in Thai journal-based trend mapping, while also showing a strong presence of interview-based work in some years (Phoocharoensil, 2022).

Table 9. Keywords Identified in the Fluctuating Trend from Abstract Corpus

The substantive portion of the fluctuating trend combines applied classroom concerns and evaluative claims. Practice, strategy, and technology point to recurring attention to classroom implementation and applied intervention design. Challenge and need signal variable emphasis on problem diagnosis, often linked to learning constraints and contextual limitations. Effective and positive reflect alternating emphasis on outcome claims and perception reporting. Speak appears here as well, suggesting that speaking related outcomes rise and fall over time in abstracts, even when learning and assessment remain stable core themes.

Across the two corpuses, Abstracts provide a more method specific and content direct signal than Titles. Titles concentrate on headline themes such as quality, rural, and academic. Abstracts more consistently reveal research design labelling, instrument reporting, and evaluative framing. The rise of qualitative and quantitative labels, together with the visibility of surveys and questionnaires, is consistent with patterns reported in other national research landscapes where quantitative and survey dominated work remains prominent (Phoocharoensil, 2022). The stable prominence of assessment and performance-oriented terms aligns with continued international attention to assessment related constructs and subfields (Lee, 2025). The growing visibility of integration, evaluation, and technology aligned language reflects broader field level developments that have been documented in reviews of AI and technology-oriented language education research (Ma et al., 2024).

DISCUSSION

This section analyses the findings from the Title corpus and the Abstract corpus. The Title corpus does not show any salient information about the research methodology, but the Abstract corpus does. This is quite conventional. Because the methodological details follow the same trends in the Abstract corpus, these patterns reveal no differences in this respect. When more research is conducted, methodological details certainly appear with greater frequency. They therefore have been excluded from the synthesis. As presented in Table 10, the two corpuses show quite similar themes across the four trends. More research has focused on pedagogical aspects, learning engagement, and learners’ contexts over the five years of 2021-2025. In contrast, less research has focused on generic online learning mode and pandemic context. In addition, learning activities and assessment tended to be stable, regardless of the temporal changes. Meanwhile, the research on productive skills and classroom practice appeared to be mediated by socio-cultural and temporal changes.

Policy Alignment in the Research Landscape

The upward and constant patterns show a reasonable alignment with the Vietnamese national agenda of proficiency improvement and strengthened evaluation mechanisms. Decision No. 2080 extended the NFLP and re-emphasised priorities that include curriculum innovation and reform in assessment and evaluation. The upward trend in proficiency and evaluation directly echoes these policy emphases. The constant presence of learn and assessment further suggests that learning outcomes and outcome measurement remain stable anchors in the research landscape. This continuity is consistent with the long-standing proficiency orientation of Vietnam that has characterised the reform trajectory since the launch of the NFLP. It also mirrors trends in other contexts where assessment related themes form salient clusters (Yang & Wang, 2025). For example, in the Korean EFL self-assessment literature, bibliometric review identifies clusters around writing assessment, formative assessment, and skill related assessment concerns (Lee, 2025).

At the same time, the national policy trajectory has shifted in 2025 toward English functioning more broadly in schools, including expanded expectations for English as a language of instruction. The present corpuses show only partial evidence of this shift. The upward trend in integrat may indirectly support instructional expansion, because these terms often index curriculum design and pedagogical integration. However, the trends do not strongly foreground English medium instruction as an explicit research objective. A similarly limited research trend on English as a medium of instruction has also been reported in the context of Latin America (Aliaga Salas & Pérez Andrade, 2023), Algeria (Sahki, 2025), and elsewhere. This suggests a potential lag between the 2025 policy escalation and research attention, at least within this convention set of corpuses. This pattern is plausible given the well documented implementation constraints that have historically affected reform outcomes, including teacher preparedness, uneven resources, and challenges in contextual adaptation (Le & Barnard, 2020).

Limited Salience of Important Research Focuses

A critical observation from the trend results is that technology does not emerge as a salient and stable research topic in this set of corpuses, although it has been a major concern in English language education in recent years (Nguyen & Tran, 2022; Nguyen et al., 2023; Santosa et al., 2022), particularly with the rapid emergence of AI tools and large language models (Dang, 2025a). This limited salience is notable because technology has also been emphasised in the international landscape and has been positioned as an important component in Decision No. 2080. It is also noteworthy in light of the convention context itself, where technology related themes have been in two consecutive years.

The decline of research relevant to online, covid, and pandemic indicates that technology was strongly tied to emergency remote teaching and pandemic framing in the earlier years, but that this framing gradually faded. This decline should not be interpreted as a simple disappearance of technology from research. Rather, it suggests a shift away from broad modality labels that were characteristics of the early pandemic period, such as online and blended. In principle, such a shift is expected as digital pedagogy becomes normalised and as research moves from general modality descriptions to more specific investigations. However, in the present corpuses, this move toward specificity does not translate into technology becoming a stable, high salience thematic signal. Instead, technology remains largely dispersed, appearing within fluctuating patterns rather than forming an upward or constant trend.

One plausible explanation is lexical fragmentation. As research becomes more tool specific, attention may spread across different platforms and applications, so that no single technology label accumulates enough frequency to function as a stable keyword. This interpretation is consistent with international evidence showing rapid expansion of AI related work but also a wide range of focal technologies and implementation concerns, which can diffuse keyword prominence across a corpus (Ma et al., 2024). The current findings therefore suggest that the research community may be engaging with technology in more diverse and localised ways, but that this engagement is not consolidated into a coherent keyword signature at the abstract level.

Another notable absence in these corpuses is the lack of salient keywords that explicitly index self-directed learning, lifelong learning, learner autonomy, or learner agency. This gap is important because these constructs are widely recognised as core capacities for sustained language development, particularly when learners must continue learning beyond classroom time (Dang, 2025b). In the context of Vietnam, the NFLP and its extension place pressure on learners to develop higher levels of communicative competence and to sustain progress across schooling levels (Prime Minister, 2008, 2017). The later policy shift toward broader and more demanding uses of English in schools further increases the need for learners to manage learning independently over time. Yet, in the current corpuses, lifelong learning and related constructs do not consolidate into a high salience thematic signature.

This absence cannot be explained by a lack of attention to learners’ circumstances. In fact, learner related contexts do surface. Rural is visible in the Title corpus, and cultural rises in the Abstract corpus. However, contextual visibility is not the same as conceptual engagement with independent learning. Research can describe constraints in rural settings or cultural influences without directly examining how learners develop self-regulation, learner autonomy, and sustained learning practices. This distinction is critical in Vietnam because culture and context are not peripheral. They shape what learner autonomy can look like in practice. A sociocultural discussion of learner autonomy in Vietnam argues that autonomy is mediated by community constraints and can be fostered or hindered within the local educational ecology (Dang, 2010). More recently, empirical study also shows that students’ autonomous learning behaviours are strongly shaped by cultural and situational factors, including familial expectations, available learning opportunities, exam schemes, and teacher influence (Dang, 2024). These findings show that culture is not simply background information. It is a central condition that structures whether independent learning can be enacted.

The absence of independent learning as a consolidated keyword theme in the current corpuses can be interpreted in a couple of ways. One possibility is that learner autonomy research is present but dispersed across different labels, or embedded implicitly within studies framed around engagement, proficiency, or classroom practice (Dang, 2026). In such cases, autonomy may be treated as a secondary consideration rather than the primary construct under investigation. Another possibility is that, within this convention venue, research attention is directed more toward instructional practices and evaluative outcomes than toward learners’ capacity to sustain learning beyond formal instruction.

Comparative evidence suggests that learner autonomy can become a visible construct when a subfield consolidates around it. For example, in the Korean EFL context, a bibliometric analysis of self-assessment research identifies learner autonomy as a distinct thematic cluster (Lee, 2025). This contrast underscores the interpretive weight of its absence in the corpuses of the current study. In short, while learners’ contexts such as rural conditions and culture are visible in the trends, they do not translate into a salient research focus on learner autonomy and independent learning. This represents a meaningful gap in relation to both long standing reform ambitions and the intensified expectations that follow the 2025 policy shift (Prime Minister, 2008, 2017, 2025).

Finally, the analysis demonstrates a gap concerning sustainability related themes. Keywords that would typically label inclusive and equitable education, critical thinking, adaptability, and social emotional learning did not emerge as salient in the corpuses of the current study. The only partial signal is the recurring visibility of rural as a contextual keyword, which points to place-based constraints and educational inequalities. However, this signal does not directly indicate a research focus on sustainability-oriented learner attributes, such as adaptability, socio emotional competence, or critical global awareness.

This pattern contrasts with the wider international agenda, where education is explicitly positioned as central to Sustainable Development Goal 4, including inclusive and equitable quality education. UNESCO’s guidance on learning objectives for the SDGs highlights competency development for sustainability and global citizenship as a core educational expectation (UNESCO, 2017). In ELT, sustainability has increasingly been framed as both curricular content and pedagogical purpose. Systematic reviews show a growing body of studies integrating sustainability and SDG oriented themes into English language teaching, and they highlight common emphases on global issues, classroom tasks, and teacher capacity for implementation (Antari et al., 2024; Micalay-Hurtado & Poole, 2022). Empirical work also illustrates how sustainability-oriented ELT can be linked to outcomes such as learner engagement, deeper thinking skills, and ecological literacy, which aligns closely with critical thinking and adaptability aims often associated with SDG oriented education (Alharbi, 2026).

The limited salience of these themes is also notable because international scholarship increasingly frames inclusion and wellbeing as integral to educational quality. Work on inclusive and equitable learning environments explicitly positions equity and inclusion as a research priority connected to SDG 4 (Heinz et al., 2025). In language education, social emotional learning has likewise been argued to support English language learners’ participation, relationships, and classroom adjustment, which links the wellbeing agenda to equity and learning outcomes. These external patterns suggest that sustainability-oriented topics are not absent from the global ELT research agenda. Rather, they appear to be under consolidated in the current corpuses. Some sustainability related work may exist in the dataset, but it likely remains dispersed across specific issues and tool or task labels. Thus, it does not accumulate into stable, high salience keywords.

LIMITATIONS

First, the corpus is limited to accepted titles and abstracts from a single venue, and some of them may not be directly about Vietnam as it is an international forum. While this design provides a timely overview of what researchers and practitioners consider publishable and relevant at the point of submission, it does not represent the full TESOL and EFL research landscape in Vietnam. Studies disseminated through peer-reviewed journals, theses, and institutional outlets may emphasise different topics, apply different methodological standards, or respond to policy agendas in different ways. Tighter process of selecting records relevant to Vietnam only also needs to be employed. As a result, the trends reported here should be interpreted as conference-based signals rather than comprehensive national patterns.

Second, the time window and corpus size are uneven across years. Although the study normalised counts by the number of records per year to reduce the influence of annual volume differences, unequal denominators can still affect the apparent stability of trends, particularly for mid frequency keywords. In such cases, modest absolute changes can produce more visible relative movement. This limitation is important in interpreting strict trend groupings, especially when trends are classified as upward or downward based on year-to-year comparisons.

Third, the analysis relies on titles and abstracts rather than full text articles. Abstracts provide only a condensed representation of a study and may not capture the full essence of the research, including the depth of evidence, the rigor of analysis, or the validity of claims. In some cases, an abstract may be accepted because it presents an excellent rationale or timely topic, even when the study is not strongly empirical or the subsequent analysis is not rigorous. Similarly, the quality of contributions within a conference program is not uniform, yet the frequency-based approach treats all accepted records as equivalent units of evidence. This means that higher weight contributions, such as keynote and featured presentations, are counted in the same way as parallel session abstracts. As a result, the trends reported here reflect the distribution of lexical signals across accepted submissions, rather than a weighted representation of research quality or influence. This limitation is particularly relevant when interpreting alignment with policy directions under Decision No. 2080 and Decision No. 2371. Policy relevant evidence depends not only on topical coverage but also on the strength and reliability of the underlying research (Le & Barnard, 2020; Prime Minister, 2017, 2025).

FUTURE DIRECTIONS

Future research should extend the dataset beyond a single conference. Incorporating peer reviewed journal articles, theses, and other national conferences would enable stronger generalisation and would allow triangulation of whether the conference-based signals identified in the current study persist in more formal publication channels. Importantly, such expansion would help address the gaps highlighted in this research, particularly the limited salience of constructs linked to lifelong learning, learner autonomy, agency, and sustainability-oriented agendas. A broader and more diverse corpus would make it possible to confirm whether these constructs are genuinely under researched in the TESOL and EFL landscape of Vietnam, or whether they are simply less visible in conference abstracts. This would provide a more comprehensive and policy relevant overview of alignment and mismatch between national reform priorities and the evolving research evidence base.

Another further direction concerns the relationship between policy priorities and the research ecosystem. As the national agenda shifts toward the more ambitious goals articulated in Decision No. 2371, policy measures could explicitly encourage research that targets priority areas needed for successful implementation, rather than relying on spontaneous alignment. This includes research on lifelong learning capacities, learner autonomy, and sustainability-oriented competencies that support equitable and long-term learning outcomes, especially as English expands beyond the language classroom. Targeted research funding, dissemination incentives, and practice-linked research programs could strengthen the evidence base for implementation and help identify feasible models across diverse school contexts. Periodic review studies should also be conducted to monitor whether these policy interventions shape research attention over time and to provide updated feedback on emerging gaps, thereby supporting iterative and evidence informed policy refinement.

Finally, cross country comparisons would strengthen interpretation by situating Vietnam within wider regional trajectories. As neighbouring contexts such as Thailand, Indonesia, Korea, and Malaysia, have also pursued large scale policies to expand English capacity, matched corpuses from comparable conferences or journals in these countries could be analysed using the same document frequency, lemmatisation, and trend classification method. This approach would allow direct comparison of topical signals and methodological reporting patterns.