Penerapan Model Kooperatif Tipe TGT dengan Pendekatan Deep Learning dalam Upaya Meningkatkan Hasil Belajar Gerakdasar Lokomotor, Non Lokomotor, dan Manipulatif pada Peserta Didik Kelas IV Jenjang Sekolah Dasar (SD)

Muhammad Nurdin Syaputra (Universitas Lambung Mangkurat)
Mashud (Universitas Lambung Mangkurat)
Syamsul Arifin (Universitas Lambung Mangkurat)

Gelanggang Olahraga: Jurnal Pendidikan Jasmani dan Olahraga (JPJO)
Vol. 9 No. 4 (2026): Gelanggang Olahraga: Jurnal Pendidikan Jasmani dan Olahraga

This study aims to improve the learning outcomes of combining locomotor, non-locomotor, and manipulative basic movements through the Teams Games Tournament (TGT) cooperative model with a Deep Learning (joyful learning) approach. The subjects were 21 fourth-grade elementary school students. Following the Kemmis and McTaggart model, this study was conducted in two cycles. Data were collected through teacher and student observation sheets, deep learning implementation checklists, and movement skill performance tests. The results showed significant improvements. In the pre-cycle, classical completeness was 0%. After Cycle I, completeness increased to 47.61% (10 students passed). Following reflections and improvements in Cycle II, classical completeness jumped significantly to 85.71% (18 students passed), supported by optimal learning activities and a joyful classroom atmosphere (deep learning score reached 98.44). In conclusion, the combination of the TGT model and Deep Learning approach is proven effective in improving students' basic movement learning outcomes actively and joyfully. Keywords: Deep Learning, Basic Movements, Physical Education, Elementary School, Teams Games Tournament (TGT)