<![CDATA[Conceptual understanding in integrated natural and social science (IPAS) may remain limited when instruction emphasizes lectures and memorization. This study examined the effect of deep learning, based on mindful, meaningful, and joyful learning, on fifth-grade students’ conceptual understanding of “Weather Changes and Their Impacts” at SD Negeri 7 Pulau Morotai, Indonesia. A quantitative quasi-experimental study used a nonequivalent control-group pretest–posttest design. Participants were 60 fifth-grade students, comprising an experimental class (n = 30) taught through deep learning and a control class (n = 30) taught conventionally. Data were collected using a 15-item conceptual-understanding test, an observation sheet, and documentation. Descriptive statistics, normality tests, paired-samples t-tests, and analysis of covariance (ANCOVA) were used. The experimental class’s mean score increased from 40.23 to 53.77, a statistically significant gain, t(29) = 4.58, p < .001. The control class also improved significantly, from 34.50 to 47.57, t(29) = 7.48, p < .001. Implementation of deep learning was rated good (M = 3.20 on a four-point scale). However, after controlling for pretest scores, the adjusted difference between groups was not statistically significant, F(1, 57) = 1.572, p = .215, partial η² = .027. These findings indicate improved conceptual understanding in both classes but do not establish a distinct advantage of deep learning over conventional instruction. Further studies should involve larger samples, multiple classes, and stronger research designs.]]>
