Recurrent and system-specific model-attribution patterns in PISA 2022 science performance across nine education systems
- 1School of Artificial Intelligence in Traditional Chinese Medicine, Zhejiang Chinese Medical University, Hangzhou, China
- 2School of Medical Technology, Zhejiang Chinese Medical University, Hangzhou, China
- 3School of Pharmaceutical Sciences, Zhejiang Chinese Medical University, Hangzhou, China
Abstract
Using science data from the 2022 Programme for International Student Assessment (PISA), we examined survey-weighted descriptive association and model-attribution patterns across a fixed, non-probability cohort of nine education systems (78,685 students). Because PISA science plausible values are drawn from a latent population distribution generated with a conditioning model that uses background information, the analysis describes survey-weighted descriptive associations rather than independent individual predictions. Seven machine-learning pipelines were evaluated descriptively across 20 repeated school-grouped held-out partitions in each system. The genetic algorithm-optimized backpropagation neural network (GA-BPNN) was retained as the focal architecture for Shapley additive explanations (SHAP), without claiming systematic or universal superiority. Alternative training-only median/mode imputation produced system- and metric-specific performance changes. The four sensitivity-supported variables—ESCS, HOMEPOS, MATHEFF, and ST255Q01JA—formed a conservative subset of the eight variables that recurred under the primary top-35 screening procedure. Feature-screening stability was assessed under the high-missingness quartile stress test, whereas held-out performance and SHAP magnitude/rank variation were assessed only under the full alternative median/mode imputation workflow; the sample-restriction analysis was not used to assess changes in SHAP ranks. A cross-architecture check showed substantial but incomplete agreement between GA-BPNN SHAP and XGBoost TreeSHAP. These results describe recurrent and system-specific attribution patterns across the nine education systems. They do not support causal inference, individual diagnosis, or generalization to all PISA systems; missing-data analyses did not resolve missing-not-at-random (MNAR) mechanisms or residual selection bias.
Related
- Marker-specific pixel attribution for single-cell protein quantification with PixAQ
- Recurrent severe pyogenic infections revealing neutrophil-specific granule deficiency caused by a novel CEBPEsplice-region variant: a case report.
- Structural variants from donor-specific assemblies affect recurrent chromatin contact divergence in humans
- Bilingualism and Cognitive Ageing: Domain-Specific Benefits After Controlling for Education and Lifestyle
- Different patterns of eye-specific input are relayed to overlapping populations in the mouse superior colliculus
BibTeX
@article{liu2026recurrent,
title = {{Recurrent and system-specific model-attribution patterns in PISA 2022 science performance across nine education systems}},
author = {Meng Liu and Xin Bai and Xiaohong Li and Wei Jin and Weifeng Jin},
journal = {PLOS One},
year = {2026},
month = sep,
doi = {10.1371/journal.pone.0351534},
url = {https://doi.org/10.1371/journal.pone.0351534},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Recurrent and system-specific model-attribution patterns in PISA 2022 science performance across nine education systems
AU - Meng Liu
AU - Xin Bai
AU - Xiaohong Li
AU - Wei Jin
AU - Weifeng Jin
PY - 2026
DA - 2026/09/18
JO - PLOS One
DO - 10.1371/journal.pone.0351534
UR - https://doi.org/10.1371/journal.pone.0351534
AB - Using science data from the 2022 Programme for International Student Assessment (PISA), we examined survey-weighted descriptive association and model-attribution patterns across a fixed, non-probability cohort of nine education systems (78,685 students). Because PISA science plausible values are drawn from a latent population distribution generated with a conditioning model that uses background information, the analysis describes survey-weighted descriptive associations rather than independent individual predictions. Seven machine-learning pipelines were evaluated descriptively across 20 repeated school-grouped held-out partitions in each system. The genetic algorithm-optimized backpropagation neural network (GA-BPNN) was retained as the focal architecture for Shapley additive explanations (SHAP), without claiming systematic or universal superiority. Alternative training-only median/mode imputation produced system- and metric-specific performance changes. The four sensitivity-supported variables—ESCS, HOMEPOS, MATHEFF, and ST255Q01JA—formed a conservative subset of the eight variables that recurred under the primary top-35 screening procedure. Feature-screening stability was assessed under the high-missingness quartile stress test, whereas held-out performance and SHAP magnitude/rank variation were assessed only under the full alternative median/mode imputation workflow; the sample-restriction analysis was not used to assess changes in SHAP ranks. A cross-architecture check showed substantial but incomplete agreement between GA-BPNN SHAP and XGBoost TreeSHAP. These results describe recurrent and system-specific attribution patterns across the nine education systems. They do not support causal inference, individual diagnosis, or generalization to all PISA systems; missing-data analyses did not resolve missing-not-at-random (MNAR) mechanisms or residual selection bias.
SN - 1932-6203
ER -