# Table 2 Sample size calculations

Outcome ICC (95%CI) Alpha
(two-sided)
Power Minimal effect detectablea Total number schools (pupils) required
Time spent in MVPA per dayb 0.07
(0.00, 0.21)
5% 80% 0.25 SD 51 (1274)
Time spent sedentary per dayb 0.02
(0.00, 0.11)
5% 80% 0.25 SD 51 (1272)
Fruit and vegetable portions per dayc 0.04
(0.00, 0.09)
5% 85% 0.25 SD 51 (1268)
Secondary outcomes
Self-report time spent screen-viewing 0.00
(0.00, 0.03)
1% 90% 0.25 SD 40 (1000)
Snacks portions per dayc 0.03
(0.00, 0.07)
1% 90% 0.30 SD 50 (1254)
High fat food portions per dayc 0.05
(0.00, 0.10)
1% 85% 0.30 SD 51 (1274)
High energy drink portions per dayc 0.02
(0.00, 0.05)
1% 90% 0.30 SD 43 (1070)
Mean BMI 0.00
(0.00, 0.02)
1% 90% 0.30 SD 32 (794)
Mean Waist 0.05
(0.03, 0.20)
1% 80% 0.30 SD 51 (1274)
% overweight or obese (by BMI) 0.03
(0.01, 0.05)
1% 55% 20%d 51 (1274)
% centrally obese (by Waist) 0.06
(0.02, 0.09)
1% 60% 20%d 51 (1274)
% overweight or obese (by BMI) 0.03
(0.01, 0.05)
1% 80% 30%e 51 (1274)
% centrally obese (by Waist) 0.06
(0.02, 0.09)
1% 82% 30%e 51 (1274)
1. a Difference in mean continuously measured outcomes and relative % difference for binary outcomes; b Assessed by accelerometer; c For these outcomes the minimal effect detectable is calculated using SD differences on the log-scale; d Relative difference (i.e. ability to detect an odds ratio of at least 0.8 or 1.2; e Relative difference of larger magnitude (i.e. ability to detect an odds ratio of at least 0.7 or 1.3)
2. SD: standard deviation; Calculations assume similar numbers of schools randomised to control or intervention and ~ 25 pupils per school; In all calculations we have used the ICC upper 95% CI value