This mixed-methods study investigates the effectiveness of Khan Academy in improving fifth-grade students’ ability to solve fraction word problems in a remote primary school in East Java, Indonesia. FouTwenty-four students participated in a quasi-experimental pretest-posttest design, supported by qualitative interviews with the classroom teacher. Quantitative results indicated a paradoxical outcome: although 86% of students expressed positive attitudes toward the platform, the class average decreased slightly from 37.14 to 34.29, with an increased standard deviation from 15.67 to 16.47. Only 36% of students improved their scores, while 43% experienced notable declines. These findings suggest that initial technological engagement does not guarantee improved academic performance, particularly in under-resourced settings where digital fluency and mathematical foundations are limited. Qualitative analysis revealed that while students favored features such as audio feedback and step-by-step guidance, many struggled with contextual understanding of word problems. Teacher insights highlighted the importance of scaffolding and cultural adaptation in content delivery. The study concludes that Khan Academy holds potential for enhancing learning but requires thoughtful integration with teacher-led instruction, contextual relevance, and differentiated support. This research contributes to the growing discourse on equity and contextualization in digital education, particularly in remote and underserved communities. It also emphasizes the need for blended pedagogical models that bridge the gap between digital enthusiasm and conceptual understanding.
                        
                        
                        
                        
                            
                                Copyrights © 2025