Table 1

Sample sizes required to provide nominal statistical power for 50% treatment effect within subject.

ARES

PSG

Assumption

1-β = .80

1-β = .90

1-β = .80

1-β = .90

No bias, no reliability

24

31

57

76

No bias, observed reliability

10

13

27

35

Observed bias, no reliability

26

34

3,499

4,684

Observed bias, observed reliability

11

14

1,541

2,062

Assumed α = .05 *Nearest whole number sample size to generate, at minimum, the nominal value.

Levendowski et al. International Archives of Medicine 2009 2:2   doi:10.1186/1755-7682-2-2