<![CDATA[Mathematical critical thinking remains chronically underdeveloped in Indonesian junior secondary classrooms, yet no study has simultaneously integrated socio-emotional differentiation, Problem-Based Learning (PBL), and structured worksheet scaffolding while examining how self-regulated learning (SRL) predicts critical thinking development. This study addressed that gap by evaluating the quality and effectiveness of a socio-emotionally differentiated PBL model assisted by LKS Ceria, and testing SRL as a predictor of critical thinking performance among eighth-grade students. A quasi-experimental nonequivalent control group design with an explanatory sequential mixed-methods component was implemented at SMPN 4 Juwana (2024/2025 academic year; n = 30 per group). Instruments were validated via Aiken's V (mean V = 0.83) and Cronbach's alpha (α = 0.84–0.87). Data were analyzed using Mann-Whitney U, binomial, Chi-square, and bootstrap regression tests. The model satisfied all three quality criteria: instruments were valid and reliable (planning stage), implementation reached 66% with 77.33% positive student responses (implementation stage), and the experimental class achieved 70% classical mastery — significantly exceeding both the control class (0% mastery; U = 108.5, z = −6.21, p < .001, r = .80) and the criterion proportion (binomial p = .041; χ² = 28.46, p < .001). SRL was a significant positive predictor of critical thinking (β = 1.005, SE = 0.313, p = .004, 95% CI [0.387, 1.679]; R² = .42), with performance varying systematically by SRL level. This study contributes an empirically validated, socio-emotionally differentiated PBL framework that simultaneously operationalizes scaffolded worksheet support and SRL-based differentiation as a replicable design for developing mathematical critical thinking in junior secondary mathematics.]]>
