Teaching games for understanding in school handball: a controlled pre-post study with intact Brazilian physical education classes

Introduction This controlled pre-post study with intact school classes examined whether participation in a TGfU-based handball unit was associated with motivational, cognitive, and technical-tactical changes compared with Direct Instruction and a comparison condition in Brazilian public-school physical education. Methods Sixty-three students aged 11–13 years from three intact classes were allocated to TGfU handball, Direct Instruction handball, or a comparison volleyball unit, with one class per condition. The intervention comprised 20 lessons of 50 min. Outcomes were assessed before and after the instructional unit and included motivation for physical activity participation (MPAM-R), cognitive flexibility assessed exploratorily using the Stroop test, and technical-tactical indicators observed during small-sided handball games. Data were analyzed using nonparametric longitudinal models based on Wald-Type Statistics, with post-hoc comparisons and effect sizes. Results Enjoyment showed a Group × Time effect, WTS(2) = 85.60, p < 0.001, increasing only in the TGfU group (4.69 ± 0.81 to 6.42 ± 0.46; p < 0.001, r = 0.876). Competence also showed a Group × Time effect, WTS(2) = 28.02, p < 0.001, increasing only in the TGfU group (4.77 ± 1.06 to 5.99 ± 1.08; p < 0.001, r = 0.846). Stroop outcomes improved over time, but no Group × Time effects were observed. For technical-tactical outcomes, the TGfU group reduced inaccurate passes (1.71 ± 1.38 to 0.43 ± 0.75; p = 0.002, r = 0.667) and increased occupying open spaces (1.19 ± 1.63 to 3.86 ± 3.65; p = 0.001, r = 0.698), whereas the Direct Instruction group decreased accurate passes (20.76 ± 8.90 to 10.14 ± 5.05; p < 0.001, r = 0.831). Discussion A TGfU-based handball unit was associated with favorable changes in intrinsic motivation-related dimensions and selected offensive tactical behaviors in this school context. No condition-specific changes in cognitive flexibility were observed, suggesting that Stroop performance changes were not specific to TGfU. Because allocation occurred at the class level with one class per condition and without blinding, the findings should be interpreted as contextual associations rather than causal estimates.