168 Chapter 2 Notes. David et al. (2019b) report a Cohen’s d of 0.66 when comparing the applied game to the passive condition. David & Fodor (2023) report a Cohen’s d of 0.42 and Perry et al. (2017) report a Cohen’s d of 0.29, with 95% CI [0.09, 0.49]. The studies of Axford et al. (2020), Kuosmanen et al. (2017), Maden et al. (2022), Perry et al. (2017), Shum et al. (2019) and Yunus et al. (2020) used cluster randomisation. For Axford et al. (2020), the adjusted effect size was calculated based on an average cluster size of 153.05 and an ICC of 0.097 (i.e., design effect = 15.75), which gave a Cohen’s d of 0.05, with 95% CI [-0.23, 0.32]. For Kuosmanen et al. (2017), the adjusted effect size was calculated based on an average cluster size of 6.00 and an ICC of 0.02 (i.e., design effect = 1.10), which gave a Cohen’s d of 0.50, with 95% CI [-0.01, 1.02]. In contrast to the unadjusted effect size, the adjusted effect size was nonsignificant. For Maden et al. (2022), the adjusted effect size was calculated based on an average cluster size of 2.00 and an ICC of 0.02 (i.e., design effect = 1.02), which gave for the applied game vs. active condition comparison a Cohen’s d of 0.01, with 95% CI [-0.72, 0.75], and for the applied game vs. passive condition comparison a Cohen’s d of 0.37, with 95% CI [-0.36, 1.09]. For Perry et al. (2017), the adjusted effect size was calculated based on an average cluster size of 54.00 and an ICC of 0.017 (i.e., design effect = 1.90), which gave a Cohen’s d of 0.22, with 95% CI [-0.05, 0.49]. In contrast to the unadjusted effect size, the adjusted effect size was nonsignificant. For Shum et al. (2019), the adjusted effect size was calculated based on an average cluster size of 114.75 and an ICC of 0.02 (i.e., design effect = 3.28), which gave a Cohen’s d of 0.09, with 95% CI [-0.28, 0.46]. For Yunus et al. (2020), the adjusted effect size was calculated based on an average cluster size of 2 and an ICC of 0.02 (i.e., design effect = 1.02), which gave the same effect size as the unadjusted effect size reported in the forest plot. CI = confidence interval, ICC = intracluster/intraclass correlation coefficient.
RkJQdWJsaXNoZXIy MTk4NDMw