
192 Sample Size Calculations for Clustered and Longitudinal Outcomes
TABLE 6.1
Sample size and power for detecting a main effect δ
(3)
in model (6.4) when
randomizations occur at the third level with ρ
2
= 0.05 (two-sided significance
level α = 0.05)
λ
3
∆
(3)
ρ
1
N
1
N
2
N
(0)
3
Total N ϕ
(3)
1 0.3 0.1 5 4 19 760 0.805
5 8 14 1120 0.807
0.2 5 4 23 920 0.813
5 8 16 1280 0.813
0.4 0.1 5 4 11 440 0.816
5 8 8 640 0.813
0.2 5 4 13 520 0.815
5 8 9 720 0.813
1.5 0.3 0.1 5 4 16 800 0.809
5 8 12 1200 0.818
0.2 5 4 19 950 0.810
5 8 13 1300 0.803
0.4 0.1 5 4 9 450 0.809
5 8 7 700 0.832
0.2 5 4 11 550 0.821
5 8 8 800 0.837
Note: λ
3
represents the ratio of the number of level 3 units in the interv ...