China's "Double Reduction" policy aims to alleviate the burden on compulsory education, yet significant subject-based discrimination persists in teacher evaluation, with bonus proportions for non-main subject teachers approaching zero. This study integrates critical policy analysis with organizational justice theory to construct the Multi-Source Process Evaluation Model (MSPEM) and conducted a quasi-experimental study involving 277 teachers at a typical secondary school. The research employed mixed methods, integrating AHP-entropy combined weighting and SoftMax nonlinear allocation to realize multi-dimensional process evaluation across six dimensions (core teaching, social-emotional, technology integration, home-school collaboration, professional development, and student outcomes). Results show that MSPEM significantly weakened the association between subject attributes and bonuses (r = -0.116, 95% CI [-0.22, -0.01], p = 0.053), raising the bonus proportion for non-main subject teachers from 0% to 4.25%. The Gini coefficient decreased to 0.1744 (95% CI [0.168, 0.181], below the 0.20 fairness reference value), with an effect size d = -0.162 (small effect). This indicates improvement in distributive justice, though procedural justice and interactional justice still have room for enhancement. Robustness tests demonstrated high model stability. This study provides empirical support for teacher evaluation reform under the ``Double Reduction'' policy and offers a practical, actionable reference framework for educational evaluation practice. Future work can enhance the model's generalizability through multi-center validation.
South Africa's Science, Technology, Engineering, and Mathematics (STEM) education pipeline continues to face critical attrition, particularly during the transition from secondary schooling to post-secondary engineering pathways. This study examines how subject streaming, especially the widespread placement of learners into Mathematical Literacy, affects preparedness for both Technical and Vocational Education and Training (TVET) and university-level engineering studies. Employing a convergent mixed-methods design within an interpretivist paradigm, the research integrated structured surveys, semi-structured interviews, and focus group discussions. Purposive and stratified sampling enabled representation across rural, peri-urban, and urban contexts, engaging 120 qualitative participants (policymakers, principals, and STEM lecturers) and 180 quantitative respondents (secondary learners, STEM teachers, and tertiary students). Thematic analysis identified five systemic challenges: (1) curriculum misalignment as a structural barrier, (2) curriculum tracking and exclusion through Mathematical Literacy, (3) teacher capacity deficits in STEM pedagogy, (4) institutional fragmentation and policy incoherence, and (5) persistent urban-rural inequalities in STEM access. Quantitative data corroborated these findings, with 72% of teachers reporting insufficient training and 63% of tertiary students indicating low preparedness for engineering studies. The study recommends systemic reforms, including vertical curriculum articulation, bridging programs, and sustained STEM teacher development, to reduce attrition and strengthen South Africa's STEM-to-engineering pipeline for the Fourth Industrial Revolution.
A strategy of preparing and reading authentic texts in fully online classes in senior school is suggested. This strategy, based on cloud technology, provides scaffolding and feedback as students read authentic texts during online lessons. Preparing and activity stages are suggested, and possible cloud tools are analysed for each stage. The key idea of our approach is to organise feedback during silent reading with the use of specialised software that allows learners to underline some words by simply clicking them. Some examples of using this feedback are suggested for basic comprehension and then for intensive reading and content discussion. The proposed approach was implemented in the village secondary school, and the results are discussed
In most school subjects a large language model is something a class consumes; in the preparation of informatics teachers it is something the class must be able to open, inspect, adapt, and govern - the very competencies the discipline obliges its graduates to teach. This dual relation makes the openness of free and open-weight models not a cost-driven substitute for proprietary services but the decisive pedagogical affordance. This conceptual, framework-building paper develops that claim from a structured review of a 202-record Web of Science pool, extended by targeted, DOI-verified supplementation. It re-conceptualizes openness as a vector of six affordances: four descend directly from the software freedoms to run, study, modify, and share - inspectability, adaptability, deployability with data sovereignty, and shareability with licensing - and two are reconstructions demanded by a stochastic, versioned artifact: reproducibility and governability. These affordances are arrayed against four enacted teacher roles - designer, facilitator, monitor, and evaluator - as a claim-graded openness-agency matrix whose every cell is marked as empirically evidenced, transferred from adjacent evidence, or argued as a proposition. Four deployment tiers, T0 through T3, locate where each affordance becomes actualizable; the decisive threshold is the T0-to-T1 transition, where a hosted endpoint gives way to a locally run model and three constrained affordances switch on together. A four-axis governance scaffold - configuration, boundaries, logging, and adjudication - makes deployment documentable and doubles as student-authored coursework. Five falsifiable propositions frame the framework for empirical test. The organizing maxim follows: model selection in teacher education is a curriculum decision, not a procurement decision.
The digitalisation of road transport and logistics is reshaping the work of automobile-transport college graduates, yet the pedagogical literature tends to catalogue digital tools rather than the entrepreneurial actions those tools should serve, and no evidence-based taxonomy links practice types to competence outcomes for this sector. This study derives such a taxonomy through a two-stage design: a PRISMA-ScR scoping review that maps a heterogeneous, cross-disciplinary evidence base, followed by best-fit framework synthesis that codes the extracted practices against an a-priori frame of entrepreneurial-competence areas and entrepreneurial-action stages. From an initial 320 records, 63 sources were retained; the 22 practice-exemplifying sources among them were classified into groups by two independent coding passes that agreed closely (Cohen's κ=0.94 for group assignment); as both passes were AI-generated, this indexes coding-scheme consistency rather than independent human-rater reliability. Four groups of digital entrepreneurial practice are distinguished: (1) project-based entrepreneurial, (2) technical-economic, (3) logistical-organizational, and (4) communicative-reflective practices. The first three are each supported by at least three independent sector or pedagogy studies; the fourth is warranted by the completeness of the entrepreneurial-action cycle together with cross-sector reflection and assessment scholarship, an evidential asymmetry we report rather than smooth over. The four groups span a coherent cycle of entrepreneurial action - a structure partly entailed by the analytic frame - and are mapped onto the EU EntreComp and DigComp frameworks and onto experiential and activity-learning theory. Two conceptual moves extend the taxonomy into a framework: a trace-reflection mechanism, in which each action leaves an inspectable digital artefact that mediates assessment and revision, and a two-dimensional practice-design space that positions each group against tiers of environment fidelity and treats immersive technology as a level of that dimension rather than a separate group. Eight testable propositions and a progression-anchored assessment rubric are proposed. The contribution is conceptual and design-oriented; the framework and rubric await empirical validation.
