BRIDGING BLOOM'S 2 SIGMA PROBLEM THROUGH WEB-BASED INTERACTIVE LEARNING MULTIMEDIA AND ADAPTIVE TUTORING: A SYSTEMATIC REVIEW OF GEOMETRY LEARNING

Title: BRIDGING BLOOM’S 2 SIGMA PROBLEM THROUGH WEB-BASED INTERACTIVE LEARNING MULTIMEDIA AND ADAPTIVE TUTORING: A SYSTEMATIC REVIEW OF GEOMETRY LEARNING Authors: Tomi Purwanto¹, Majid², Hasan S. Panigoro³ ¹ Magister of Mathematics Education Study Programme, Postgraduate Programme, Universitas Negeri Gorontalo, Indonesia ² Department of Mathematics, Faculty of Mathematics and Natural Sciences, Universitas Negeri Gorontalo, Indonesia Executive Summary This manuscript presents a systematic literature review following the PRISMA 2020 guidelines to investigate how web-based Interactive Learning Multimedia (ILM) and Adaptive Tutoring Systems (ATS) can address the challenge posed by Bloom’s 2 Sigma Problem—specifically within the context of geometry education. The study synthesizes empirical findings across 19 core sources (including R&D/ADDIE development studies, quasi-experiments, randomized controlled trials [RCTs], and recent meta-analyses) to evaluate learning gains, strength of causal inference, and pedagogical mechanisms. Key Information & Research Highlights Theoretical Framework: Uses Bloom’s (1984) finding—that 1-on-1 human tutoring produces learning outcomes 2 standard deviations (2D) above standard classroom instruction—as a historical benchmark to evaluate technology-enhanced personalization in geometry. Methodology & Corpus Structure: Primary Search: Scopus database (424 records; 54 Open Access subset screened). Supplementary Search: Scopus & Consensus academic semantic engine (23 records identified, 13 full-text assessed). Final Synthesis Corpus (19 Core Sources): 7 initial supporting articles, 1 primary Scopus candidate, 11 supplementary empirical studies, and 2 corroborating meta-analyses. Core Findings: Web-Based Multimedia (Strand A): Tools like 3D interactive visualizations and AR websites consistently improve spatial visualization and solid geometry understanding (N-Gain = 0.45-0.89 in development/R&D studies; posttest Cohen's d = 1.17 in quasi-experiments). Adaptive Feedback & Tutoring (Strand B): Systems providing real-time, response-based adaptivity deliver significant added value beyond media features alone (effect sizes d = 0.35--0.73 in comparative and RCT settings). The 2-Sigma Reality Gap: Pooled effect sizes from recent meta-analyses of AI and adaptive learning in mathematics range from g = 0.34 to 0.70. While demonstrating solid, statistically significant learning gains, current adaptive technology achieves roughly 1/4 to 1/3 of Bloom’s original 2.0 SD claim. Main Conceptual Contribution (The 5-Layer Model): The article synthesizes a layered framework for designing web-based ILM for junior-secondary solid geometry: Representation Layer: Interactive 3D models, nets, dimension labeling, visual manipulations. Task Layer: Tiered practice, conceptual questions, real-world scenarios. Diagnosis Layer: Log tracking of answer patterns, error types, attempt history, and time. Adaptation Layer: Automated hints, difficulty scaling, and tailored feedback based on diagnostic data. Evaluation Layer: Measurement of conceptual understanding, retention, and transfer. Subject & Focus Keywords Primary Focus: Geometry Education / Solid Geometry (Junior Secondary / SMP level) Technology Domain: Web-based Interactive Learning Multimedia (ILM), Adaptive Tutoring Systems (ITS), Real-time Adaptive Feedback, AI in Education Methodology Keywords: PRISMA 2020 Systematic Review, Thematic Synthesis, Effect Size Analysis, Comparative Causal Inference Intended Audience & Journal Suitability This manuscript is structured for submission to high-impact, peer-reviewed international journals in educational technology and mathematics education