<![CDATA[Developing critical thinking in physics is a priority in 21st-century education, yet evidence comparing the effectiveness of different instructional approaches—particularly those integrating problem-based learning (PBL) with digital quiz platforms—remains limited. This study compared the effects of three learning models—PBL assisted by Zep Quiz, Zep Quiz alone, and traditional lecture—on 11th-grade students' critical thinking skills in mechanics. A counterbalanced within-subject quasi-experimental design was employed with 102 students from SMAN 4 Tangerang Selatan. Each student experienced all three models across three topics (Newton's Laws, Momentum & Impulse, Rotational Motion) in a rotated order. Critical thinking was measured using a validated 18-item test based on Facione's six indicators (reliability: 0.78–0.79 for two topics, 0.45 for Rotational Motion). Data were analyzed using N-gain, repeated-measures ANOVA, and ANCOVA with pretest scores as covariate. All models improved critical thinking with moderate-to-high N-gain (0.73–0.78). Repeated-measures ANOVA showed significant differences between models (F(2,202)=4.920, p=.008, η²=.046). However, after controlling for pretest scores, ANCOVA revealed no statistically significant differences between models (F(2,202)=2.229, p=.109, η²=.015, small effect). Topic characteristics explained substantially more variance (η²=.642) than model variations (η²=.015). The largest gains occurred in Momentum & Impulse (N-gain=0.89), the lowest in Rotational Motion (N-gain=0.55). The three learning models are comparably effective in improving critical thinking in mechanics. Topic characteristics—particularly conceptual complexity—outweigh model variations in determining learning gains. Physics educators should prioritize alignment between instructional design and topic demands rather than focusing solely on model selection. Future research should extend to other physics topics and investigate longer-term retention.]]>
