Guided inquiry in school science: a mini review of orchestration, assessment, and AI

Abstract

This mini review asks how to orchestrate guidance, assessment, and student agency so that inquiry reliably yields durable understanding and inclusion, moving beyond the inquiry vs. direct instruction debate. It draws on twenty-six peer-reviewed studies from 2017 to 2025 spanning classroom interventions, video studies, large-scale secondary analysis, discipline-specific and teacher-education reviews, flipped and other technology-mediated inquiries, and AI. Studies were coded for inquiry phases, forms of guidance, assessment practices, and outcomes and were synthesized configuratively across four strands: learning and affect, enactment quality, assessment literacy, and technology/AI. The findings revealed that guided, discourse-rich inquiry supported conceptual understanding, higher-order thinking, and motivation, whereas the raw frequency of “inquiry activities” related inconsistently to achievement when quality and guidance were ignored; assessment literacy for scientific practices remained a key bottleneck; and flipped structures and AI copilots helped when they scaffolded prediction-evidence-explanation sequences and preserved learner agency. The review concludes that systems should prioritize phase-specific orchestration; build assessment capabilities for modeling, data analysis, and argumentation; instrument classrooms to capture process quality; and evaluate AI as formative infrastructure rather than as an answer engine.