EFFECTIVENESS OF FLIPPED LEARNING WITH DEEP LEARNING FRAMEWORK BASED ON MULTIPLE REPRESENTATIONS IN ACID-BASE MATERIAL TO IMPROVE SCIENTIFIC PROCESS SKILLS

Abstract

This study aims to describe the effectiveness of flipped learning with a deep learning framework based on multiple representations on acid-base material to improve science process skills. The population in this study were all grade XI students of SMA Negeri 1 Bandar Lampung in the 2025/2026 academic year. The research sample was taken using a purposive sampling technique and obtained class F.1 as the experimental class and class F.2 as the control class. In the experimental class, flipped learning with a deep learning framework based on multiple representations was applied and the control class was applied with conventional learning. This study is a quasi-experimental study with a non-equivalent control group design. The instrument used was a science process skills pretest–posttest test. Supporting data on student activities were collected using observation sheets. Data analysis in this study included analysis of primary data, supporting data, and hypothesis testing. Hypothesis testing used was a test for differences in two means with an independent sample t-test. The results showed that the average N-gain of students’ science process skills in the experimental class was significantly higher than the control class, with the average N-gain of students’ science process skills in the experimental class being categorized as moderate. It can be concluded that flipped learning with a deep learning framework based on multiple representations on acid-base material is effective in improving students’ science process skills.