Optimal class size and learning outcomes in Morocco: A normative-causal approach using PISA 2022 data

This study addresses optimal class size in Moroccan public lower secondary schools using PISA 2022 microdata (N = 5,962 students, 150 schools). We develop an integrated dual-methodology framework combining normative multi-criteria optimization with simulated fuzzy regression discontinuity design (FRDD) to triangulate evidence on class size effects. The optimization module estimates domain-specific performance functions using natural cubic splines with socioeconomic controls, constructs an economic cost function based on Moroccan budget parameters (teacher salaries, infrastructure), and derives optimal class sizes across eight policy scenarios varying pedagogical-economic trade-offs. Results converge robustly on a global optimum of 23.4 students (95 % CI: [23.35, 23.36]), requiring a 36 % reduction from the current 36.5 students. The simulated FRDD independently corroborates these findings by identifying a "high-performance corridor" at thresholds 23–26 students, where local average treatment effects on pedagogical, multi-criteria, and economic indices remain positive, statistically significant, and robust to covariate adjustment (specification variation <10 %). Beyond 30 students, performance deteriorates sharply (LATE=-0.10 to −0.22, p < 0.001). Moving to the optimal size increases per-student costs by 49.3 % but generates substantial learning gains (+25.1 PISA points), yielding a return on investment of 29.1—classified as "Good" and exceeding typical education interventions in developing countries. Heterogeneity analyses reveal minimal variation across socioeconomic status and performance levels (gaps <0.2 students), supporting a uniform national policy target of 23–24 students. These findings provide evidence-based guidance for Morocco's education reforms while demonstrating methodological transferability to other resource-constrained systems in the MENA region.