Skip to main content
. 2014 Sep 11;11:114. doi: 10.1186/s12966-014-0114-z

Table 8.

Sample size calculations based on detecting a 10 minute difference in weekday MVPA using a cluster size of 24 and 12 per school

Outcome a Cluster size ICC b α Power (β) n required per arm c n schools/arm N per arm inflated for attrition d Total N e
Cluster size = 24
MVPA/weekday (min) 24 0.12977 5% 80% 360 15 450 900
MVPA/weekday (min) 24 0.12977 5% 90% 456 19 570 1140
MVPA/weekday (min) 24 0.12977 1% 90% 624 26 780 1560
ICCf
MVPA/weekday (min) 24 0.06534 5% 80% 240 10 300 600
MVPA/weekday (min) 24 0.06534 5% 90% 312 13 390 780
MVPA/weekday (min) 24 0.06534 1% 90% 408 17 510 1020
Cluster size = 12
Cluster size ICC b α Power (β) n required per arm c n schools/arm N per arm inflated for attrition d Total N g
MVPA/weekday (min) 12 0.12977 5% 80% 204 17 255 510
MVPA/weekday (min) 12 0.12977 5% 90% 276 23 345 690
MVPA/weekday (min) 12 0.12977 1% 90% 384 32 480 960
ICCf
MVPA/weekday (min) 12 0.06534 5% 80% 156 13 195 390
MVPA/weekday (min) 12 0.06534 5% 90% 204 17 255 510
MVPA/weekday (min) 12 0.06534 1% 90% 276 23 345 690

aTarget difference = 10 minutes of MVPA per weekday.

bobserved ICC (95% CI) was 0.06534 (0.00091 to 0.12977); upper limit of 95% CI used.

ccluster size of 24.

destimated attrition = 20% (~6 per cluster).

ebased on 30 pupils per school (i.e. 24 + 6).

fPoint estimate of ICC used.

gbased on 15 pupils per school (i.e. 12 + 3).