Table 4.

Assumptions of the sample size calculation

Sample size calculation assumption
100 participants per school
2 full intervention, 2 partial intervention and 4 control schools
Participants of cohorts F, G, H, and I will be included (Table 1)
A significance level (alpha) of 0.05
A power of 80 %
Independent-samples t-test on difference in absolute BMI Z-scores
An intraclass correlation coefficient of 0.01, as based on Amorim et al. [64]
Unequal cluster sizes with a relative efficiency of 90 % [65]
A dropout rate of 20 % (including both study drop-out and natural drop-out such as migration)
An average absolute BMI Z of 0.76 and an SD of 0.60 in the population (these values were calculated over all children aged between 4 and 11 living in Parkstad region who visited the Youth Health Services in 2013).