This study examines structural patterns in pedagogical innovation across six Indonesian doctoral dissertations, five in chemistry education and one in physics education, using the Glocal Myths-Facts Analysis Framework (GMFAF) to diagnose whether an innovation is contextually relevant, resistant to adaptation, or likely to last.
A critical cross-case analysis combined rubric-anchored evaluation with fuzzy reasoning (Mamdani inference), modelling relevance, resistance and impact via triangular membership functions to derive a Glocal Fit Compliance Index (GFCI), a 0–1 score expressing structural alignment, not academic merit.
Adaptation resistance is the dominant moderating factor in sustainability, limiting transferability even when relevance and impact are strong. Dissertations incorporating contextual realities (D3, D4, D6) reached higher GFCI values than those dependent on infrastructure and facilitation (D1, D2, D5).
The study analyses a small corpus and does not support statistical generalisation; its contribution is analytic generalisation: adaptation resistance moderates relevance and impact.
Assessing contextual relevance and adaptation resistance at the research-design stage may help doctoral studies produce more sustainable innovations.
Infrastructural dependence, not a lack of pedagogical merit, most often limits the reach of promising innovations; addressing it directly may narrow the gap between policy and classroom practice.
This study introduces GMFAF, a framework combining qualitative evaluation with fuzzy reasoning to diagnose pedagogical innovations' contextual viability, a novel method for evaluating doctoral research design in higher education.
