2.3. Data Analysis

IBM SPSS Statistics 27 (IBM Corp., Armonk, New York, NY, USA) was used for all the statistical analyses. First, children’s baseline proficiency levels were determined per test item. The 15% lowest scores on each test item (per grade and gender) were classified as low, with the 15% highest scores as high. All other scores were classified as average. Second, the mean and standard deviation of the test item scores on T0 and T1 were calculated for each test item (both overall and per grade and gender). Third, repeated measures ANOVA was used to elucidate both the main and interaction effects of grade, gender, and baseline proficiency, the three between-subject factors, and the within-subjects variable time per test item. In case of significant main and/or interaction-effects, post hoc analysis was conducted with independent samples t-test, with Bonferroni correction. Fourth, to determine whether the children within the three proficiency level categories at baseline scored significantly differently from each other at T0 and on T1, independent samples t-tests with Bonferroni correction were performed.

The EHC-test item was conducted differently for children in grades 3 and 4 (allowed to catch the ball with two hands) compared to children in grade 5–8 (they had to catch the ball with one hand) [10]. Due to this difference in the protocol, results of the EHC-test cannot be compared between grades. Therefore, only results of the EHC-test item of children in grades 5–8 were used for analysis.

Based on the smallest detectable difference (SDD) of the test items, it was determined whether children showed positive (T1-T0 ≥ +SDD), no (-SDD < T1-T0 < +SDD), or negative (T1-T0 ≤ −SDD) development. The SDDs for the four test items were based on previous test-retest studies within two weeks (2.7 (WB), 2.5 (MS), 5.7 (JS) [21]; and 2.7 (EHC) [22]). Frequencies of positive, no, and negative development were calculated for each subtest per grade, for gender, and for baseline proficiency level.

Alpha was set at 0.05 for statistical significance for all analyses. For the repeated measures ANOVA, the partial eta squared was calculated. The partial eta squared is interpreted as follows: η2 = 0.01 indicates a small effect, η2 = 0.06 indicates a medium effect, and η2 = 0.14 indicates a large effect [23].