<![CDATA[Indonesian students face persistent challenges in developing critical thinking skills in mathematics, as reflected in low international assessment performance and national reports linking weak reasoning to classrooms dominated by one-way, conventional teaching. Accordingly, this study sought to identify the underlying learning problems and propose an evidence-based instructional design. Employing a mixed-method, design-oriented approach that combined a qualitative Design Thinking needs analysis with quantitative meta-analytic synthesis, the researchers gathered data from 85 Grade XI students and 6 mathematics teachers across three senior high schools in Central Java through questionnaires, interviews, and classroom observation. Empathize-phase data showed over 85% of students reporting mathematics anxiety, 70% fearing mistakes, and calculus (functions, limits, derivatives, integrals) as the hardest topic; empathy maps for students and teachers converged on a five-part problem profile: critical thinking held back by mathematics anxiety, monotonous teaching media, weak contextual grounding, slow feedback, and unaddressed socio-emotional needs. These findings point toward the Empathic-PBL model, supported by DECABOT (Desmos Classroom Activities and Chatbot AI), as the resulting design solution. To ground this proposal empirically without directly validating DECABOT, a random-effects meta-analysis of 13 analogous-component studies (not direct trials of DECABOT) covering PBL, SEL integration, interactive digital media, and AI chatbot assistance produced a large pooled effect size (Hedges' g = 0.859, 95% CI: 0.825–0.892, I² = 11.3%, p < .001) via a random-effects model. DECABOT thus stands as a theoretically grounded, conceptual instructional design candidate for senior high school calculus, supported by evidence from analogous interventions rather than a tested product, and still awaiting experimental confirmation.]]>
