Deep Learning–Based QuizWhizzer for Innovative IPAS Assessment in Elementary School

Abstract

This classroom action research aims to improve students’ learning outcomes and engagement through an innovative learning evaluation based on Deep Learning using the QuizWhizzer digital platform in the Natural and Social Sciences (IPAS) subject. The initial problem identified was the low level of student participation in evaluation activities and the limited use of varied and engaging digital assessment media aligned with 21st-century learning demands. Integrating Deep Learning principles into the evaluation process was expected to help students develop deeper conceptual understanding, contextual reasoning, and reflective learning abilities. The study was conducted in two cycles following the Kemmis and McTaggart model, consisting of planning, action, observation, and reflection stages. The research subjects were 30 fifth-grade students. Data were collected through observations, interviews, documentation, and learning outcome tests, and analyzed using descriptive qualitative and quantitative approaches. The findings indicate a significant improvement in both student activity and learning outcomes. Mastery learning increased from 58% in the pre-cycle to 76% in Cycle I and reached 91% in Cycle II. Students also demonstrated higher motivation, stronger collaboration, and improved critical-thinking skills during the gamified evaluation process. The use of QuizWhizzer created an interactive, enjoyable, and competitive learning atmosphere that supported mindful, meaningful, and joyful learning. This study concludes that the Deep Learning–based evaluation using QuizWhizzer is an effective and innovative assessment strategy that enhances IPAS learning quality in elementary schools and can be recommended as an alternative digital evaluation method within the Merdeka Curriculum.