A Comparative Analysis of User Reviews and AI-Based Assessments of Educational Applications

Abstract

Evaluation of early childhood learning applications requires a multifaceted approach that considers factors such as educational value, appropriateness for developmental age, and the integration of technology. Recent research empha- sises the necessity for clear assessment instruments and performance scales with the goal of setting the criteria by which educational apps and programmes are evaluated to determine whether their design aligns with the developmental needs of young children. This study compares Google Play user reviews with AI-generated assessments of educational applications for children aged 3–5. User-generated content was analysed through sentiment analysis and topic modelling. The results are then compared with rubric-based assessment that have been generated by AI model which rates educational value, usability, and developmental appropriateness. Results indicate that there is a weak correla- tion between public perception and AI evaluations, where users mainly focus on aspects such as entertainment and interface design, whereas AI places more emphasis on pedagogical criteria. The research emphasises the methodological value of combining user data from the internet with automated analysis tools for the evaluation of digital learning environments. It contributes by illustrat- ing the strengths and limitations of AI-assisted evaluation in educational tech- nology research.