Balancing homework and video instruction: a repeated cross-sectional analysis of CAD course outcomesVasileska, Ivona (Avtor) Vukašinović, Nikola (Avtor) Čok, Vanja (Avtor) mechanical engineeringCAD coursestudent performancehomeworkinstructional videosComputer-aided design (CAD) courses require students to develop both procedural skills and higher-order modeling competencies. This 10-year repeated cross-sectional study (N = 1952 students, 2013–2023) examined how replacing mandatory graded homework and the structured practice and feedback mechanisms it provided with optional short instructional videos affected performance. Using descriptive statistics, one-way ANOVA, interrupted time-series models, and student-level regression with clustered standard errors, we analyzed outcomes across pre-intervention (homework), post-intervention (video), and COVID remote-instruction periods. Results show that although homework and video instruction produced equivalent performance on basic modeling tasks (PE1), the homework period was associated with significantly higher performance on complex modeling tasks (PE2) and modestly higher final grades. The video-based period was associated with elevated failure rates, particularly for complex exams, whereas COVID period introduced temporary grade inflation that masked these effects. Findings indicate that while instructional videos effectively support basic procedural learning, structured homework with feedback remains essential for developing advanced modeling competencies. The results highlight the importance of aligning self-learning methods with task complexity and suggest that a combined approach may best support student success in CAD education.20262026-05-06 12:09:54Članek v reviji182289sl