Abstract
Learning models that do not actively involve students in the learning process are one of the factors causing low learning outcomes. This research aims to test and analyze how effective the problem-based learning model, which utilizes Educaplay media, is on student learning outcomes. The method used by researchers is a quasi-experimental design with a nonequivalent control group. The subjects of this study were 85 fifth-grade students of Jatisari state elementary School, Semarang City, Indonesia i.e. 28 students of class VB as the experimental class, and 28 students of class VC as the control class. Learning outcomes data were obtained from test techniques and non-test techniques. Data analysis used SPSS 25, consisting of normality, homogeneity, t-test, and N-Gain tests. The results indicated that the significant value in the independent t-test was 0.000 < 0.05, so Ha was accepted and H0 was rejected. This data indicates a significant difference in the learning outcomes between the students in the experimental and control classes. N-Gain percent was 66% (moderately effective) in the experimental class and 39% (ineffective) in the control class. This study indicates that the application of the PBL model based on Educaplay media is effective in improving student learning outcomes.