Improving Students' Cognitive Abilities in Invasion Games Through Deep Learning: Strengthening Kurikulum Merdeka

Oman Hadiana (Universitas Muhammadiyah Kuningan)
Ribut Wahidi (Universitas Muhammadiyah Kuningan)
Hana Astria Nur (Universitas Muhammadiyah Kuningan)
Reni Reni (Universitas Muhammadiyah Kuningan)

Indonesian Journal of Physical Education and Sport Science (IJPESS)
Vol 6 No 3 (2026): September: Article in Progress

Study purpose. The Merdeka Curriculum promotes deep learning to foster meaningful, mindful, and joyful learning experiences that support higher-order thinking skills. In physical education, however, cognitive development is often overshadowed by psychomotor-oriented instruction. This study examined the effect of a deep learning approach on elementary school students’ cognitive abilities in invasion games. Materials and Methods. The study employed a randomized pretest–posttest controlled experimental design involving 57 fifth-grade students. Participants were randomly allocated to either the intervention group (n = 29), which participated in an invasion games program grounded in deep learning principles, or the comparison group (n = 28), which received instruction through conventional teaching methods. The intervention consisted of eight sessions involving soccer and basketball activities. Data were collected using a 25-item cognitive ability test assessing invasion game concepts, spatial awareness, decision making, tactical understanding, game analysis, and problem solving. The instrument demonstrated high content validity (Aiken’s V = 0.89) and reliability (KR-20 = 0.88). Data analysis involved descriptive statistics, Shapiro–Wilk and Levene’s tests, paired- and independent-samples t-tests, N-Gain analysis, and Cohen’s d effect size estimation. Results. Validity and reliability analyses supported the instrument’s adequacy for data collection. Both groups demonstrated statistically significant improvements from pre-test to post-test (p < .05); however, the magnitude of change differed between them. The experimental group achieved a high normalized gain (N-Gain = 0.71), whereas the control group showed a moderate gain (N-Gain = 0.39). Further comparative analysis confirmed a significant difference in cognitive improvement between the groups (p < .001), accompanied by a large effect size (Cohen’s d = 1.12) Conclusions. The deep learning approach effectively enhances students’ cognitive abilities in invasion games and supports the implementation of the Merdeka Curriculum in physical education. Future studies should investigate its long-term effectiveness across diverse educational contexts and learning domains.