Questionnaire-Based Analytics for the Cultivation of “Double-Qualified” Teachers: An Empirical Study Using PCA, K-Means Clustering, and Random Forest
Year: 2026
Type: conference-paper
Study design: Observational Study
Abstract: “Double-qualified” teachers are TVET instructors who combine pedagogical credentials with industry-recognized occupational competence or substantial workplace experience, linking teaching to authentic tasks and enterprise standards. Against the backdrop of quality improvement in vocational education, double-qualified teachers are widely regarded as pivotal for strengthening practical instruction and school–enterprise collaboration; however, student-perceived heterogeneity and the key drivers for the cultivation of double-qualified teachers remain insufficiently quantified. This study aims to use student questionnaire data to uncover the latent structure of double-qualified teaching experiences, identify heterogeneous learner segments, and determine the major qualities of teachers associated with students’ overall satisfaction and growth. Methodologically, we implement an integrated PCA–K-Means–Random Forest machine-learning framework: PCA reduces the dimensionality of highly correlated Likert-scale items and mitigates multicollinearity; K-Means clustering is performed in the PCA score space to segment students into distinct experiential profiles, with clustering quality assessed using metrics such as the Silhouette Score; and a Random Forest regression model is trained with dimension scores and background variables (e.g., grade level, major category, dual-teacher class exposure, and practical training experience) to estimate feature importance and identify key drivers. The results indicate a clearly stratified segmentation, with systematic differences across perceived teaching competence, practical guidance experience, school–enterprise interaction, and student satisfaction and growth; moreover, school–enterprise interaction and practical guidance emerge as the most influential components, suggesting that strengthening authentic task participation and the quality of collaboration should be prioritized. This work contributes a reproducible, data-driven “reduce–segment–explain” pathway that provides actionable quantitative evidence for cultivating and optimizing double-qualified teaching in a differentiated manner.
Institution: Beijing Information Technology College
Language: English
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W7155625331