Thematic categories and characteristic terms.

Previous studies have rarely examined ethnic teacher education from a configurational perspective. This study addresses that gap.This study analyzes 250 peer-reviewed articles from CNKI’s PKU/CSSCI journals (2015–2025), identified using ‘ethnic teacher education/education/teachers’ keywords, refined from 2,672 initial records through manual screening.This study employs a mixed-methods approach to investigate ethnic teacher education in China, combining computational text analysis with configurational methods. This study employs ROST Content Mining 6.0 for semantic network analysis and CiteSpace 6.3 for knowledge mapping. The analysis reveals five core conceptual clusters (development, teachers, ethnicity, education, and region) and five research hotspots, with ‘ethnic regions’ emerging as the dominant focus. Through LDA topic modeling, five thematic dimensions were extracted, ranging from the status of ethnic education to cultural integration in teacher training. These qualitative insights were subsequently quantified and analyzed via fsQCA, yielding six distinct optimization pathways. The findings suggest a dual trajectory for future research: while emphasizing geo-cultural specificity in teacher training program design, the study also identifies universal principles applicable across ethnic education contexts. Accordingly, five policy recommendations are proposed, including diversifying development channels, enhancing cultural responsiveness in pedagogy, and facilitating inter-regional knowledge exchange among ethnic communities. This study introduces the novel integration of LDA topic modeling and fsQCA in educational research. While LDA uncovers latent themes and fsQCA examines causal complexity, their combined application enables simultaneous discovery and validation of configurations a previously unexplored approach in ethnic teacher education.These findings make dual contributions: theoretically, by advancing a novel conceptual framework; and practically, by yielding actionable policy implications for ethnic teacher education development.以往研究极少从组态视角（configurational perspective）审视民族教师教育领域，本研究填补了这一研究空白。本研究以中国知网（CNKI）的北大核心/中文社会科学引文索引（CSSCI）期刊2015—2025年刊发的文献为初始检索对象，以“民族教师教育/教育/教师”为关键词进行检索，共得到2672条初始记录，经人工筛选后最终纳入250篇同行评议（peer-reviewed）论文作为分析样本。本研究采用混合研究方法探究中国民族教师教育现状，将计算文本分析（computational text analysis）与组态研究方法相结合：使用ROST Content Mining 6.0开展语义网络分析（semantic network analysis），使用CiteSpace 6.3绘制知识图谱（knowledge mapping）。分析结果显示，该领域存在五大核心概念集群（发展、教师、民族、教育与地域），并形成五大研究热点，其中“民族地区”为核心研究主题。通过潜在狄利克雷分配（Latent Dirichlet Allocation, LDA）主题建模，本研究提取出五大主题维度，涵盖从民族教育现状到教师培训中的文化融合等多个范畴。随后通过模糊集定性比较分析（fuzzy-set Qualitative Comparative Analysis, fsQCA）对上述质性结论进行量化分析，最终得到六条差异化的优化路径。研究结果表明，未来民族教师教育研究可遵循双重路径：一方面需强调教师培训项目设计中的地域文化特异性，另一方面也需提炼适用于各类民族教育场景的通用原则。据此，本研究提出五项政策建议，包括拓展多元化发展渠道、提升教学中的文化回应性、推动民族社群间的跨区域知识交流等。本研究首次将LDA主题建模与fsQCA整合应用于教育研究领域：LDA可挖掘潜在主题，fsQCA则用于分析因果复杂性，二者结合能够同时实现组态的发现与验证，这在民族教师教育研究中尚属首次。本研究兼具双重学术贡献：理论层面，构建了全新的概念分析框架；实践层面，为民族教师教育发展提供了可落地的政策启示。