Causal ML in Education: Enhancing Teacher pedagogical Strategies and Decision-Making

Abstract

The present paper examines shifts in preservice students understandings about teaching and learning because of developing of a student. This review paper explores how causal ML techniques can support teachers in evaluating and refining pedagogical interventions using observational educational data. Pedagogical decision made by teacher educators construct curriculum development. This explores the application of Causal ML in education, highlighting its potential to enhance teacher decision-making, personalize instruction, and optimize resource allocation. Research indicates that ideas prioritized by teacher educators are presented more explicitly within the curriculum. This report explores how teacher educators shape curriculum through pedagogical decisions, influenced by their purposes and contexts, as evidenced by collaborative self-studies in physical education teacher education. Furthermore, it proposes an educational artificial intelligence framework that integrates neural collaborative filtering with causal reasoning. causal inference foundations, modern machine learning–based causal estimators, and recent applications in educational data mining. It also discusses the role of counterfactual explanations in making AI-driven recommendations interpretable and actionable for teachers. By utilizing a hybrid architecture of generalized matrix factorization and multi-layer perceptron, alongside double machine learning for quantifying individual treatment effects, the framework achieves counterfactual decision optimization. The integration of these technologies enhances long-term learning efficiency while maintaining ethical regulation. Experimental validation demonstrates that this framework significantly enhances the explainability and long-term learning efficiency of teaching decisions while maintaining large-scale exhortation efficiency, providing a solution robust in both technology and ethical regulation. The main goal is to build teacher-centered, fair, and trustworthy causal decision- pedagogical support systems that empower educators to make evidence-based decisions that are technically robust and philosophically aligned with core pedagogical values. This paper reviews the foundations of causal inference and recent machine learning–based causal models used in educational data mining to analyze observational data from learning management systems, online assessments, and intelligent tutoring platforms.