The concept of STEM integration remains ambiguous, which hampers both empirical research and curriculum design. Drawing on general systems theory and Shannon’s information theory, this paper defines integration in terms of learners’ performance selection. Let U denote the set of all verbs that could be used in a STEM Integrated Learning Environment and let O denote the set of verbs observed in learners’ artefacts. The degree of integration is quantified as the entropy reduction. Because verb frequencies are highly non‑uniform, the traditional Jaccard coefficient is inappropriate to describe the degree. We therefore employed a weighted Jaccard similarity and showed that the measure is numerically identical to the entropy‑reduction measure. Operational procedures for counting verb selections and computing the degree are described. Theoretical and practical implications for assessment, curriculum design, and teacher professional development are discussed. The paper concludes with a statement of limitations and directions for future empirical validation.

This study reports on the pilot implementation of an asynchronous teacher training course developed by an international Community of Practice (COP) of teacher educators across Asia. Responding to UNESCO's Greening Education Partnership and the recognised gap in Climate Change Education provision within teacher education, the COP drew on regional expertise to design a course grounded in cross-curricular and action-oriented approaches. The pilot involved 21 pre- and in-service science teachers from Indonesia, Japan, and Mongolia. Data from pre/post-tests, structured reflections, and focus group discussions were analysed through thematic analysis. Results showed that participants moved toward more integrated and student-centred conceptions of CCE, though many continued to situate climate topics within science subject boundaries. The findings offer early evidence for how international COPS can contribute to implementation of CCE in science education.