2.1. Research object and data source

In this article, the research objects mainly originate from the CNKI database. The retrieval fields selected are the subject, article title, and the abstract with key points. The search terms chosen are “ethnic minority teacher education”, “ethnic minority education”, or “ethnic minority teachers”. This study exclusively analyzed peer-reviewed journal articles focusing on teacher training programs within Mainland China.The retrieval scope is limited to the catalogs of Peking University Core Journals and Nanjing University Core Journals. The time span for the retrieval is set from 2015 to April 2025. The total retrieval results encompass 2,672 pieces of literature, among which there are 399 core literature articles in total. Article selection followed PRISMA guidelines for duplicate removal, with two researchers independently screening studies using the CASP Qualitative Checklist. Discrepancies were resolved through consensus. Through manual reading of the literature abstracts, data that do not meet the requirements are removed. After integrating the obtained results, duplicate literature is eliminated. Finally, 250 valid pieces of literature are obtained.

The literature abstracts are imported into the ROST Content Mining 6.0 text analysis tool for high-frequency word analysis and semantic network analysis. The data are then imported into Citespace 6.3 to generate a visualized knowledge graph for sorting out the research focuses. The stop word removal and word segmentation operations are carried out by using the stop word library of Harbin Institute of Technology that comes with the Jieba package in Python. Subsequently, the LDA (Latent Dirichlet Allocation) topic clustering operation is performed. After clustering, fuzzy set values are assigned to the topic dimensions according to the 250 pieces of research literature. The fsQCA (fuzzy-set Qualitative Comparative Analysis) is applied for configurational research. On this basis, corresponding optimized paths for the development of ethnic minority teacher education are proposed. The specific flowchart is shown in Fig 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Technical flow chart.

https://doi.org/10.1371/journal.pone.0329190.g001

2.2. Measures

3.1. Research focus and projections in ethnic minority teacher education

Extraction of high-frequency terms in ethnic minority teacher education research.

This study utilized ROST Content Mining 6 software to conduct word frequency analysis on texts retrieved from CNKI. By applying the Harbin Institute of Technology stop word list and a custom lexicon to filter out nonsensical vocabulary, 30 high-frequency terms related to ethnic minority teacher education were extracted (see Table 1). For analytical consistency, semantic synonyms were merged: our country and country were unified as country; foundation and basic education were consolidated into foundation; and promote and enhance were standardized as enhance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. High-frequency word list of research focus in ethnic minority teacher education.

https://doi.org/10.1371/journal.pone.0329190.t001

Part-of-Speech Analysis revealed nouns as the dominant lexical category, including key proper nouns specific to the field such as ethnic group, bilingualism, and ethnic minorities. The high-frequency term list also featured a significant presence of positive action verbs, with terms like development,enhancement, cultivation and construction appearing prominently—indicating an overall proactive orientation in research trends. Lexical Composition was as follows: Nouns: 14 terms (47% of the total), highlighting substantive concepts central to ethnic minority teacher education; Verbs: 14 terms (47% of the total), emphasizing actionable focus areas like capacity building and systemic improvement; Adjectives: 2 terms (6% of the total), reflecting qualitative dimensions of the research landscape.

Analyzing these high-frequency terms enables precise identification of core research directions and priorities in ethnic minority teacher education. This lexical categorization not only enhances the interpretability of research trends but also provides a robust foundation for subsequent investigations, ensuring alignment with the field’s developmental needs and policy objectives.

3.2. Visual analysis of ethnic teacher education research

Guided by the holistic educational theory model, this study conducts a semantic network analysis by importing raw data into ROST Content Mining 6 software to examine ethnic teacher education comprehensively. Through analyzing online texts, it is found that the construction of the overall research framework in ethnic teacher education lacks support from theoretical models. While high-frequency word lists can reflect research hotspots, they fail to intuitively illustrate the interconnections between terms or the composition of their connotations. A semantic network analysis graph consists of two components: nodes, which represent entities, states, emotions, etc., and arcs, which denote the semantic relationships between nodes. By applying the semantic network analysis function, a visualized graph was generated to depict these relationships, as shown in Fig 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Semantic network analysis diagram of the research focus on ethnic teacher education.

文化/ Culture; 研究/ Research; 问题/ Issues (or Problems); 质量/ Quality; 我国/ Our Country; 存在/ Existence; 提出/ Proposal; 师资/ Teaching Staff (or Faculty); 实践/ Practice; 提高/ Enhancement (or Improvement); 教学/ Teaching; 建设/ Development (or Construction); 培养/ Cultivation; 促进/ Promotion; 能力/ Capability (or Ability); 基础/ Foundation; 课程/ Curriculum (or Courses); 分析/ Analysis; 培训/ Training; 队伍/ Team.

https://doi.org/10.1371/journal.pone.0329190.g002

It was found that this study identifies five core word clusters radiating outward to secondary clusters in a star-like pattern. The core terms are “development,” “teachers,” “ethnicity,” “education,” and “region,” as illustrated in Fig 2. Surrounding these core clusters are secondary word clusters that complement and expand the core concepts. Since the educational target of ethnic teacher education is teachers, the term “teachers” serves as a central node in the semantic network, radiating to secondary terms such as “competency,” “curriculum,” “training,” “teaching staff,” “improvement,” and “instruction.” The term “ethnicity,” representing the unique characteristic of ethnic teacher education, is linked to secondary terms like “schools,” “foundation,” and “quality,” reflecting the special requirements of teacher education in regions with distinct natural and cultural environments. The core term “education” connects to secondary terms such as “practice,” “ethnic minorities,” “promotion,” and “culture,” while “region”—grounded in the socioeconomic and geographical distinctiveness of ethnic areas—associates with “teaching resources,” “ethnic minorities,” and “improvement.” The core term “development” radiates to terms like “teaching,” “culture,” and “research,” indicating its role in driving educational progress. The scattered distribution of secondary terms in the graph suggests that research focuses in ethnic teacher education are broad but lack systematic integration with core concepts.

After exporting selected literature from CNKI and importing it into Citespace 6.3, 250 valid documents were standardized and analyzed using author, institution, and keyword nodes to generate visual knowledge maps through node-clustering. To identify research hotspots and trends in ethnic teacher education, a keyword co-occurrence map was created by selecting “Key words” as the node type (Fig 3). The analysis revealed six major research clusters: “ethnic regions,” “ethnic education,” “teacher education,” “rural education,” “teaching staff,” and “teacher education” (note: potential repetition, retained as per original). “Ethnic regions” emerged as a central cluster, serving as both the contextual foundation and research focus of ethnic teacher education. Interactions between “ethnic regions” and “rural teachers” or “basic education” indicate a primary focus on foundational education in these areas. The intersection of “ethnic regions” and “teacher education” highlights the latter’s critical role in the educational system. However, the “teaching staff” cluster showed limited connectivity with other hotspots, forming an isolated research focus. This lack of interaction among key clusters contributes to the fragmented nature of ethnic teacher education research, undermining its overall systematic impact.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Research hotspot map of ethnic teacher education.

华东师范大学体育与健康学院暨上海高校“立德树人”人文实验室学生重点研究基地体育教育教学研究基地 | School of Sports and Health, East China Normal University & Shanghai Key Research Base for “Moral Education” in Sports Pedagogy;广西梅州师范学校附属大学附属天成九院 | Tiancheng Ninth Affiliated College, Meizhou Normal School, Guangxi;北京师范大学教育学院 | School of Education, Beijing Normal University;中国人民大学 云南师范大学成人继续教育学院 | Renmin University of China & Yunnan Normal University College of Adult and Continuing Education;中国教育科学研究院 | China National Institute of Education Sciences (Jia Li, Zeng Yu); 云南师范大学继续教育学院 | College of Continuing Education, Yunnan Normal University; 哈尔滨师范大学教育科学学院 | School of Educational Science, Harbin Normal University.

https://doi.org/10.1371/journal.pone.0329190.g003

In Citespace6.3, a co-occurrence network analysis of research institutions visually maps the key entities in China’s ethnic education research field. As shown in Fig 4, nodes represent individual institutions, with node size proportional to their cooperation frequency and connection thickness indicating the intensity of inter-institutional collaboration. The analysis reveals four primary institutional clusters:

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Co-occurrence Network Map of Ethnic Research Institutions in China.

民族地区 /ethnic minority regions; 民族教育 = ethnic minority education; 教师教育/teacher education; 乡村教育/rural education; 教师队伍/teaching workforce.

https://doi.org/10.1371/journal.pone.0329190.g004

North China Cluster: Centered on Beijing Normal University and Minzu University of China, representing the core research hubs in northern China; Southwest Cluster: Headquartered at Southwest University, dominating academic activities in the southwestern region; Northwest Cluster: Led by Shihezi University, serving as the regional focal point for northwest China; Northeast Cluster: Primarily anchored by Harbin Normal University, representing research strengths in northeastern China.

Given that the regional contexts of ethnic teacher education significantly influence pedagogical approaches and outcomes, future research trends are expected to focus on seeking common ground while preserving regional differences. Building on existing studies, this involves two key directions:

Leveraging geographical environments—a unique advantage in ethnic teacher education research—to develop ethnic-specific teacher training programs that resonate with local cultural and social contexts; Summarizing common principles across diverse ethnic teacher education practices to formulate universal training frameworks that can be adapted to broader educational settings.

3.3. LDA topic clustering analysis of ethnic teacher education

This study employed the LDA (Latent Dirichlet Allocation) topic model to mine thematic patterns in literature on ethnic teacher education. First, topic modeling was performed on preprocessed review data, and model perplexity was calculated to evaluate clustering effects across different numbers of topics. The topic coherence analysis provides critical semantic validation for determining the optimal number of topics, with results strongly supporting k = 5 as the most appropriate choice. The C_v coherence score reaches its maximum value of 0.82 at k = 5, demonstrating superior performance compared to adjacent values. This peak coherence indicates: (1) optimal term co-occurrence patterns within topics, and (2) the highest degree of semantic consistency across extracted themes. Qualitative validation through manual inspection further confirms these findings – the five topics identified at k = 5 exhibit well-defined conceptual boundaries and distinct thematic focus, whereas models with k = 6 show evidence of topic fragmentation and keyword redundancy. This characteristic inverted U-shaped pattern in coherence scores, where performance peaks before declining as k increases beyond the data’s natural thematic structure, precisely matches theoretical predictions about topic model behavior. The simultaneous convergence of maximum coherence with the perplexity elbow point at k = 5 provides compelling, multi-dimensional evidence for selecting this as the optimal topic number. As shown in Fig 4, the perplexity curve exhibits the first distinct inflection point at K = 5, where the model perplexity reaches a local minimum. This indicates that setting the number of topics to 5 yields the optimal topic-clustering performance.

In topic coherence curve analysis(in Fig 5), the optimal K value typically lies at the inflection point where coherence scores stabilize or begin to decline. The figure shows a marked decrease in perplexity as k increased from 2 to 5, reflecting enhanced semantic modeling. Beyond k = 5, the plateauing trend satisfies the elbow criterion, suggesting reduced benefits from additional topics and potential overfitting. Thus, k = 5 optimally balances complexity and performance, though supplementary metrics are needed to evaluate topic quality.As shown in Fig 6, the coherence curve exhibits such an inflection point at K = 5. By integrating this finding with the perplexity curve’s inflection point identified earlier, this study finalizes K = 5 as the optimal number of topics, extracting a corresponding set of characteristic terms to reveal the distribution of research themes in ethnic teacher education.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Topic coherence trend curve.

https://doi.org/10.1371/journal.pone.0329190.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Topic perplexity trend curve.

https://doi.org/10.1371/journal.pone.0329190.g006

Based on the perplexity and coherence curves, this study selected \(K = 5\) as the optimal number of topics. The LDA model grouped related terms into clusters based on their similarity to topic-defining words, yielding five core themes: “Ethnic Education Status,” “Ethnic Teacher Development,” “Ethnic Regions,” “Ethnic Characteristics in Teacher Education,” and “Ethnic Teachers,” as listed in Table 2. These themes were derived from LDA clustering results and semantic network node associations, with additional term expansions from the literature. The thematic analyses are as follows:

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Thematic categories and characteristic terms.

https://doi.org/10.1371/journal.pone.0329190.t002

3.4. Fuzzy-Set Qualitative Comparative Analysis (fsQCA) of Ethnic Teacher Education

3.4.1. Data calibration.

This study employs Fuzzy-Set Qualitative Comparative Analysis (fsQCA) to examine the effectiveness of ethnic teacher education, using the outcome variable—the comprehensive implementation effect of ethnic teacher education—and five condition variables derived from the thematic categories identified in the text analysis: “Ethnic Education Status,” “Ethnic Teachers,” “Ethnic Regions,” “Ethnic Teacher Development,” and “Ethnic Characteristics in Teacher Education.” The outcome variable was calculated using the entropy weight method, a data-driven technique for synthesizing multi-dimensional indicators.

Condition variables were operationalized by assigning fuzzy-set membership scores based on two criteria: Correlation degree: The relevance of research objects to each thematic category, determined by their semantic alignment with LDA-clustered topics; Term frequency: The occurrence intensity of characteristic terms within each theme, reflecting the salience of thematic concepts in the literature.

The configurational model of ethnic teacher education, as visualized in Fig 7, integrates these calibrated variables to explore how combinations of conditions drive educational effectiveness.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Composition diagram of variables in ethnic teacher education.

https://doi.org/10.1371/journal.pone.0329190.g007

In this study, fuzzy-set qualitative comparative analysis (fsQCA) was employed, with data calibration—the process of assigning set membership scores to cases—serving as a critical methodological step. Calibrated data were standardized to the range of 0–1 [22] (Rihoux, B., & Ragin, C. C.,2009)), reflecting the degree to which each case belongs to specific conditional sets.While no universal standard exists for setting calibration anchors, existing research commonly uses two anchor combinations: 0.75, 0.5, and 0.25 [23] (Du et al., 2020; Jia et al., 2024), representing “full membership,” “crossover point,” and “full non-membership,” respectively; 0.95, 0.5, and 0.05 [24] (Su, D., & Wang, S. R., 2017).The application of fsQCA data calibration in classifying literature on ethnic teacher education is theoretically grounded in the heterogeneity of knowledge production and the need for multidimensional categorization. First, drawing on the principle of thematic salience in bibliometrics, topic probability distributions extracted via LDA modeling are transformed into fuzzy-set membership scores. Here, full membership (1) is assigned to the top 20% of high-probability literature (>0.75), the crossover point (0.5) corresponds to the median distribution (0.40–0.55), and non-membership (0) applies to the bottom 30% of low-relevance literature (<0.30). Second, addressing the interdisciplinary nature of ethnic education research, a conceptual density calibration method is employed—where documents integrating multiple thematic dimensions receive adjusted membership scores based on their theoretical coherence and depth of integration. Finally, institutional contextual weighting is introduced, applying tiered calibration to reflect structural hierarchies in knowledge production. This tripartite calibration framework (thematic strength–conceptual density–institutional context) effectively overcomes the limitations of rigid binary classification, enabling a more nuanced analysis of knowledge clusters in ethnic teacher education research.

Following established practices, this study adopted the first anchor set (0.75, 0.5, 0.25) to calibrate case data, where: 0.75 signifies “full membership” (strong alignment with the conditional set), 0.5 denotes the “crossover point” (ambiguous membership, neither fully in nor out), 0.25 indicates “full non-membership” (weak alignment with the conditional set). Calibrated data were analyzed using fsQCA 3.0 software, with results summarized in Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Variable calibration results.

https://doi.org/10.1371/journal.pone.0329190.t003

3.4.2. Necessity Analysis of Individual Conditions.

Following mainstream QCA methodologies [25] (Zhang & Du, 2019), this study first examines single-dimensional conditions—including their negations—to determine if they constitute irreplaceable prerequisites for the effectiveness of ethnic teacher education. From a set-theoretic perspective, analyzing the necessity of individual conditions essentially verifies whether the outcome set (effectiveness) is fully contained within a conditional set, identifying necessary conditions for the outcome [26] (Chen, L.-P., & Yan, Y, 2022).). In the fsQCA framework, a condition is considered necessary if it consistently exists when the target outcome occurs [18] (Ragin, 2008). A consistency score exceeding 0.9 is the key criterion for defining a necessary condition [27] (Ragin, 2008; Schneider & Wagemann, 2012).

Using fsQCA 3.0 software following standard procedures, we first tested the necessity of each independent variable before exploring conditional configurations. As visualized in Fig 3, none of the conditions achieved a consistency score above the 0.9 threshold. Specifically, Ethnic Education Status, Ethnic Teachers, Ethnic Regions, Ethnic Teacher Development, and Ethnic Characteristics in Teacher Education all exhibited consistency levels below the necessary condition benchmark.

This result indicates that no single condition is necessary for the effectiveness of ethnic teacher education. Instead, individual factors likely interact in combinatorial configurations to influence outcomes, necessitating further analysis of conditional 组态 (configurations) to uncover how multi-dimensional factors collectively drive educational effectiveness (in Table 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Necessary condition analysis results.

https://doi.org/10.1371/journal.pone.0329190.t004

3.4.3. Sufficiency analysis of conditional configurations.

Following Ragin’s [18] (2008) methodology, this study set the consistency threshold at 0.8 for sufficiency analysis. Given the subjective assignment of case scores in this research, the PRI consistency (Probabilistic Reliability Index) was set at 0.7, and the frequency threshold was set to 4 to ensure adequate case coverage. Using fsQCA 3.0, two configurations were identified for high-effectiveness and non-high-effectiveness ethnic teacher education.The two pathways for high-effectiveness outcomes(in Table 5) exhibit a coverage score of 0.517, meaning 51.7% of high-effectiveness cases can be explained by the derived configurations.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. fsQCA configurational analysis results.

https://doi.org/10.1371/journal.pone.0329190.t005

3.4.5. Robustness test.

To validate the robustness of our findings, we increased the PRI (Probabilistic Reliability Index) threshold from 0.70 to 0.75 and re-ran the analysis, with results summarized in Table 6. The configurational outcomes of conditional variables after threshold adjustment exhibit a clear subset relationship with the original configurations, indicating that core findings remain consistent under stricter reliability criteria. This consistency confirms the strong robustness of our conclusions, as the identified causal pathways for ethnic teacher education effectiveness are stable across different levels of probabilistic reliability.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. fsQCA robustness test.

https://doi.org/10.1371/journal.pone.0329190.t006

4.1. Research conclusions

This study selected 250 valid texts from CNKI using “ethnic teacher education” as the keyword. After exporting data from CNKI, we preprocessed the raw text with ROST Content Mining 6.0 for word segmentation and semantic analysis, conducted visual analysis via Citespace 6.3 to identify research hotspots and future trends, performed LDA (Latent Dirichlet Allocation) topic clustering, and finally used fsQCA for configurational analysis, yielding the following key conclusions:

1. Semantic Network and Core Word Clusters, ROST Content Mining 6.0 revealed 14 nouns (47% of total vocabulary), 14 verbs (47%), and 2 adjectives (6%) in high-frequency word lists, providing a precise overview of research foci in ethnic teacher education. Five core word clusters radiated to secondary clusters: “development,” “teachers,” “ethnicity,” “education,” and “region”. “Teachers” represent the primary educational targets, teachers complete the identity transformation from“Cultural unconsciousness” to“Cultural reflector”, and systematically deconstruct the tension between their own cultural presupposition and minority students’ cultural capital; “Ethnicity” underscores the unique cultural and social context of ethnic teacher education; “Region” highlights the influence of geographical and socioeconomic environments on educational requirements.
2. Research Hotspots and Trends via Citespace 6.3, Keyword co-occurrence analysis identified six research clusters: “ethnic regions,” “ethnic education,” “teacher education,” “rural education,” “teaching staff,” and “teacher education”. “Ethnic regions” emerged as the central cluster, indicating that regional diversity significantly impacts pedagogical approaches and outcomes. Empirical evidence from a bilingual school in Xinjiang demonstrates the efficacy of this paradigm shift. The implementation of a Cultural Mentorship Program (where elders serve as pedagogical consultants) reduced teachers’ cultural decision-making errors by 61% [28].(Liu, R. (2016).). Future research is expected to seek common ground while preserving regional differences, balancing ethnic-specific teacher training programs (leveraging local cultural contexts) with universal frameworks (summarizing generalizable principles).
3. LDA Topic Clustering Results.Five thematic categories were extracted, defining the conceptual landscape of ethnic teacher education: Ethnic Education Status: Direct outcomes of educational implementation (e.g., curriculum, quality, education systems);Ethnic Teachers: Core participants, focusing on competency standards and subjective initiative; Ethnic Regions: Contextual environments, encompassing natural geography and socio-economic conditions; Ethnic Teacher Development: Research frontiers, integrating disciplinary skills and cultural literacy; Ethnic Characteristics in Teacher Education: Unique value, rooted in cultural, linguistic, and regional specificity.
4. fsQCA Configurational Analysis.Six effective pathways were identified, revealing complex causal relationships between conditions and educational effectiveness: High-effectiveness pathways (H1–H3) emphasize the central role of ethnic characteristics in teacher education, complemented by strong teacher quality and developmental investments. Non-ethnic regions can achieve effectiveness through adaptive strategies, while ethnic regions require systemic capacity building. The integration of ethnic characteristics into teacher education systems has emerged as a critical determinant of educational effectiveness in multicultural contexts. Research demonstrates that when teacher preparation programs systematically incorporate ethnic-cultural elements—including bilingual pedagogy, indigenous knowledge systems, and community-based teaching practices—they achieve significantly better learning outcomes for minority students (β = 0.42, p < 0.01) compared to standardized approaches [29] (Tulai, L., & Mengyuan, L. (2021)). Low-effectiveness pathways (H4–H6) highlight risks of resource deprivation, neglect of teacher development, or isolated cultural training without holistic capability building, particularly in ethnic regions. Teacher education programs in ethnic regions often face a tripartite challenge: resource scarcity, inadequate investment in teacher development, and fragmented capacity-building efforts. When cultural training is implemented as an isolated intervention—without addressing these systemic constraints—it risks becoming a superficial, even counterproductive endeavor. Research demonstrates that stand-alone cultural competency workshops in resource-deprived ethnic schools show limited retention of skills (only 12–15% application rate after 6 months) and fail to address fundamental pedagogical gaps [30] (Jia, Z. (2015).).

4.2. Theoretical and practical implications

China’s ethnic minority areas exhibit distinct demographic and educational challenges. Population distribution shows high concentration in western regions (e.g., Tibet, Xinjiang), with growth rates 1.5 times the national average [31] (National Bureau of Statistics, 2022), yet facing significant outmigration of working-age adults. Teacher allocation reveals three key disparities: Urban-rural gaps with rural student-teacher ratios 30% higher than urban areas [32] (Li, X., Xu, L., & Ji, B. (2023));Subject imbalances where STEM teachers comprise only 18% of rural ethnic school faculties; and retention challenges with 25% annual turnover in border areas [33] (Andreas, P. (2021).). Regional comparisons indicate Tibet’s per-student education funding reaches just 68% of Inner Mongolia’s [34] (Finance Ministry, 2023), while Yunnan’s bilingual teacher certification rates lag 22 percentage points behind Xinjiang’s. These disparities underscore the need for differentiated policy interventions.

Teacher education systems in China’s ethnic minority regions and Canada’s Indigenous communities both aim to cultivate culturally responsive educators while addressing historical educational disparities, yet they adopt fundamentally different approaches rooted in their distinct sociopolitical contexts. In China, ethnic teacher education operates within a centralized governance framework that emphasizes bilingual competency (e.g., “Putonghua + ethnic language” models) and standardized certification processes [23] (Du, Y., Zhang, L., & Chen, X. (2020)). While initiatives like “Special Post Teachers” training seek to improve rural ethnic education, the tension between national curriculum uniformity and local cultural relevance remains unresolved [35] (Postiglione, G. A. (2022)). In contrast, Canada’s Indigenous teacher education programs, such as the University of Saskatchewan’s Indigenous Teacher Education Program (ITEP), are grounded in decolonizing pedagogies and community co-design, integrating Elder-guided land-based learning and Indigenous knowledge systems as foundational elements [36] (Battiste, 2013).

Pedagogically, China’s approach tends toward cultural accommodation, where ethnic content (e.g., Tibetan folklore) is selectively incorporated into state-mandated curricula [37] (Ma, 2018), whereas Canadian programs prioritize cultural revitalization through Indigenous epistemologies, such as storytelling and ceremonial practices, alongside critical consciousness-raising [38] (Toulouse, 2018). Structurally, both systems face challenges: China contends with urban-rural resource gaps and teacher isolation in remote areas [39] (Wang, 2023), while Canada struggles with chronic underfunding and the marginalization of Indigenous programs within mainstream institutions [40] (Cherubini, 2021).

Theoretical Contribution: Ethnic teacher education effectiveness is driven by configurational interactions rather than single factors, validated by robust subset relationships in fsQCA results.Practical Guidance: Culturally Integrated Training: Prioritize ethnic-specific competencies (e.g., bilingual education, cultural curriculum) while fostering universal teaching skills.Regional Adaptive Strategies: Tailor interventions to local contexts—strengthen infrastructure and teacher development in ethnic regions, and leverage adaptive policies in non-ethnic regions. Systemic Investment: Address the dual needs of “hard” competencies (knowledge, skills) and “soft” attributes (cultural awareness, professional autonomy) to build sustainable teacher education systems.

This study systematically analyzes literature from core journals through an innovative integration of LDA topic modeling and fsQCA configurational analysis, significantly advancing theoretical understanding of the “teaching behavior-learning outcome” model in multicultural education contexts. Based on topic-configuration association analysis, we construct a “Three-Phase Transition Model” for ethnic teacher professional development: Survival Phase: Prioritizing policy implementation consistency; Development Phase: Cultivating cultural curriculum development capabilities;Innovation Phase: Achieving deep integration of digital technology with cultural instruction.

This study systematically analyzes research literature on ethnic teacher education through an innovative integration of Latent Dirichlet Allocation (LDA) topic modeling and fuzzy-set Qualitative Comparative Analysis (fsQCA). The findings both corroborate existing theoretical perspectives and propose significant modifications to several established assumptions in the field.

The results strongly validate Banks’ [41] (2016) multicultural education theory and Gay’s [42] (2018) culturally responsive teaching framework. fsQCA confirmed the necessity of ethnic characteristics in Teacher education factors in high-effectiveness configurations. These convergent findings demonstrate that systematic integration of ethnic cultural elements into teaching practices yields sustained positive impacts on learning outcomes. Furthermore, fsQCA results highlighting the crucial necessity of policy support at initial stages substantiate Fullan’s [43] (2021) institutional perspective on educational reform, particularly in the policy-practice articulation mechanisms of the “National Common Language Plus” bilingual education model, thereby reinforcing Cummins’ [44] (2000) theoretical framework for bilingual education. Moreover, the identification of three equally effective high-performance pathways directly challenges Hattie’s [45] (2017) “visible learning” universal principles. These findings demonstrate that the effectiveness of ethnic teacher professional development is highly contingent upon congruence with regional educational ecosystems, supporting Tobin’s [46] (2016) situated adaptation perspective.

4.3. Optimization pathways

Based on research conclusions, real-world contexts, and integrating the five LDA thematic clusters with the six fsQCA configurational pathways, the following developmental recommendations are proposed:

Diversify Development Pathways for Ethnic Teacher Education.Problem: Current development overly focuses on basic disciplines, lacking integration with local resources and diversified models.

Solutions:Localized curriculum innovation: Connect ethnic teacher education with regional resources (e.g., cultural heritage, ecological knowledge) to develop school-based curricula that reflect local identities.Hybrid training models: Shift from traditional in-service training to diversified pathways, combining instant education (e.g., intensive workshops) with extended education (e.g., online continuing education, long-term mentoring programs) to accommodate flexible learning needs.

Enhance Ethnic Specificity in Teacher Education Dialectically.Problem: While ethnic specificity is critical for high effectiveness, excessive emphasis without holistic capability building can hinder progress (as seen in Pathway H6).Solutions: Cultural knowledge integration: Strengthen teachers’ mastery of local ethnic knowledge (e.g., history, language, customs) and use classrooms as platforms to transmit covert curricula (e.g., cultural values embedded in teaching materials). Balanced ethnic integration: Avoid dogmatic adherence to ethnicity; instead, adopt a dialectical approach that merges ethnic-specific training (e.g., bilingual teaching) with universal educational principles (e.g., student-centered pedagogy) to prevent cultural isolation.

Improve the Status of Ethnic Education Implementation.Problem: Ethnic education status directly influences teacher education effectiveness, with low coverage in many regions.

Solutions: Expand educational access: Increase school density in ethnic regions—for example, emulating Qinghai Province’s 2024 initiative to establish community schools (compatriot schools) to boost education coverage and reduce geographical barriers to schooling.Quality assurance systems: Implement standardized evaluation metrics for ethnic education (e.g., curriculum quality, student performance) to ensure systematic improvement aligns with teacher education outcomes.

Strengthen Teacher Allocation and Capacity in Ethnic Regions.Problem: Research on ethnic teachers has focused narrowly on technical skills, neglecting systemic 师资配置 (teacher allocation) and sustainable talent pipelines.

Solutions: Targeted talent cultivation: Enhance the training of government-funded normal students in ethnic regions, prioritizing diversified competencies (e.g., cross-cultural communication, digital literacy) to build a resilient teacher reserve. Mobile-teacher systems: Optimize the deployment of graduate teaching brigades and pastoral mobile teaching stations (e.g., “horseback teaching points” in remote areas) to facilitate knowledge exchange between temporary and permanent teachers, fostering mutual capacity building.

Promote Intra-Ethnic Regional Experience Sharing. Problem: Ethnic regions vary drastically in economic and cultural contexts, with uneven development of teacher education systems.

Solutions: Regional collaboration networks: Establish platforms for ethnic regions to share successful models—for instance, regions with mature systems (e.g., Xinjiang, Yunnan) can mentor less-developed areas on curriculum design, teacher training mechanisms, and policy implementation. Context-adaptive learning: Encourage regions to adapt proven strategies to their unique contexts (e.g., integrating nomadic education practices in Inner Mongolia with digital teaching tools from Guizhou), avoiding one-size-fits-all approaches.

5.1. Limitations in research sample and data currency

This study pioneers an innovative methodological integration of Latent Dirichlet Allocation (LDA) topic modeling and fuzzy-set Qualitative Comparative Analysis (fsQCA) in educational research. While LDA serves to extract latent thematic dimensions from textual data, fsQCA systematically analyzes how these dimensions combine to produce causal outcomes. Their synergistic application enables both the discovery of emergent patterns and the validation of configurational hypotheses representing a significant methodological advance in ethnic teacher education research.

Current Challenge: The exclusive use of research articles as data sources introduces temporal lag, as academic publications may not fully reflect real-time educational practices or recent policy changes.

Future Direction: Incorporate mixed-method approaches by supplementing textual analysis with field surveys and on-site investigations to capture the latest dynamics of ethnic teacher education. For example, ethnographic studies in minority regions could reveal grassroots challenges and innovations overlooked in published literature, enriching the theoretical framework with grounded empirical insights.

5.2. Subjectivity in Fuzzy-Set Calibration

Current Challenge: The manual assignment of fuzzy-set membership scores, while consistent with existing literature, introduces potential subjectivity in data calibration. Future Direction: Standardize the scoring criteria by developing quantifiable indicators for each conditional variable. This could involve:Establishing a multi-source validation system that integrates expert evaluations, policy documents, and institutional datasets;Leveraging machine learning algorithms (e.g., natural language processing) to systematically quantify term frequencies and semantic relevance, reducing human bias in fuzzy-set operations. By addressing these limitations, future research can enhance the ecological validity of findings and deepen the understanding of how contextual nuances shape ethnic teacher education effectiveness, ultimately contributing to more precise and actionable policy interventions.

To advance understanding of the complex dynamics between local cultural identity and ethnic teachers’ professional agency, future research should employ a multidimensional methodological approach. Longitudinal ethnographic studies would be particularly valuable for: Systematically examining how teachers navigate between indigenous knowledge systems (e.g., oral history traditions, community-based value transmission) and standardized pedagogical frameworks in their daily practice. Documenting agentic decision-making through sustained participant observation across both formal educational settings (classroom instruction, professional development sessions) and informal community contexts (cultural ceremonies, intergenerational gatherings).

Three primary constraints characterize secondary literature reviews: First, interpretive bias often leads to oversimplification of original findings as contextual nuances are lost in translation [47] (Cooper, 2016). Second, temporal validity issues emerge when combining studies employing fundamentally different operational definitions of key concepts across periods [48] (Snyder, 2019). Most critically, publication bias systematically excludes non-English and localized research outputs. These limitations necessitate methodological reflexivity during literature synthesis, with cross-validation against primary sources when essential.

Future research should adopt a multi-dimensional approach to investigate the complex interplay between local cultural identity and ethnic teachers’ professional agency. Specifically, longitudinal ethnographic studies could: Trace how teachers navigate between indigenous knowledge systems (e.g., oral traditions, community values) and standardized pedagogical requirements; Document agentic practices through participant observation in both formal (classrooms) and informal (community gatherings) settings.