Table 2 Simulation method 2a. Estimates (percent bias) of effect (difference at the fourth time point) and variance components, where k = number of clusters per arm and m = cluster size
k | m | ICC = 0.01 | ICC = 0.1 | ||||
|---|---|---|---|---|---|---|---|
effect = 5 | \( {\sigma}_c^2=1 \) | \( {\sigma}_w^2=99 \) | effect = 5 | \( {\sigma}_c^2=10 \) | \( {\sigma}_w^2=90 \) | ||
5 | 10 | 4.94 (− 1.2) | 2.4 (137.2) | 97.4 (− 1.6) | 5.02 (0.5) | 10.5 (4.5) | 90.4 (0.4) |
20 | 4.94 (− 1.1) | 1.6 (58.5) | 98.8 (− 0.2) | 5.09 (1.7) | 9.7 (− 2.7) | 89.8 (− 0.2) | |
50 | 5.01 (0.3) | 1.1 (8.6) | 99.0 (0.0) | 4.95 (− 0.9) | 9.7 (− 2.8) | 89.7 (− 0.4) | |
10 | 10 | 5.01 (0.3) | 1.9 (85.6) | 98.1 (− 0.9) | 4.94 (− 1.2) | 9.8 (− 2.0) | 89.6 (− 0.4) |
20 | 5.03 (0.5) | 1.3 (30.1) | 98.6 (− 0.4) | 4.99 (− 0.1) | 10.0 (0.2) | 89.8 (− 0.2) | |
50 | 5.02 (0.4) | 1.0 (4.8) | 98.9 (− 0.1) | 4.98 (− 0.5) | 9.9 (− 1.0) | 90.0 (0.0) | |
20 | 10 | 5.05 (1.1) | 1.4 (41.4) | 98.6 (− 0.4) | 4.98 (− 0.4) | 10.0 (− 0.2) | 90.2 (0.2) |
20 | 4.97 (− 0.5) | 1.1 (7.1) | 98.6 (− 0.4) | 4.96 (− 0.8) | 10.1 (0.6) | 90.1 (0.1) | |
50 | 5.00 (0.0) | 1.0 (− 0.8) | 99.0 (0.0) | 4.99 (− 0.2) | 10.0 (0.5) | 90.1 (0.1) | |
50 | 10 | 5.02 (0.4) | 1.2 (20.2) | 99.1 (0.1) | 5.00 (0.0) | 10.0 (0.4) | 90.0 (0.0) |
20 | 5.02 (0.4) | 1.0 (1.5) | 98.9 (− 0.1) | 5.02 (0.4) | 10.0 (0.3) | 90.0 (0.0) | |
50 | 5.00 (0.0) | 1.0 (3.0) | 98.9 (− 0.1) | 5.03 (0.5) | 10.0 (0.1) | 90.0 (0.0) | |