Table 1

Sample size and power under different scenarios

Raw difference on the ADOS scale (SD = 5)

Standardised effect size (Cohen's d)

Power at first interim look1

Test power over all 4 interim looks2


2.5

0.5 (medium effect)

93%

100%


2

0.4

76%

100%


1.5

0.3

47%

99%


1

0.2 (small effect)

20%

80%


Note. Based on 10000 simulations for each scenario. 1Usable N = 235 (300 randomised). 2Maximum usable N = 939.

Geretsegger et al. BMC Pediatrics 2012 12:2   doi:10.1186/1471-2431-12-2

Open Data