Table 5 Control ICC = 0.1, 24 clusters of size 25 per arm.
From: Use of the bootstrap in analysing cost data from cluster randomised trials: some simulation results
How does the ICC change as a result of the intervention? | Between-cluster distribution combination1 | Rejection Rate (%)2 | ||||||
|---|---|---|---|---|---|---|---|---|
| Â | Â | Â | Huber-White | Single Bootstrap | Double Bootstrap | |||
| Â | Â | Â | Lower | Upper | Lower | Upper | Lower | Upper |
No change | - | N, N | 2.70 | 2.72 | 3.25 | 3.34 | 3.00 | 3.16 |
| Â | Â | LN, LN | 2.47 | 2.38 | 4.95 | 4.92 | 4.53 | 4.47 |
| Â | Â | N, LN | 1.48 | 3.64 | 3.23 | 4.04 | 2.85 | 4.05 |
Double |
| N, N | 2.66 | 2.53 | 3.12 | 3.19 | 2.85 | 3.13 |
| Â | Â | LN, LN | 2.28 | 2.36 | 4.99 | 5.07 | 4.60 | 4.63 |
| Â | Â | N, LN | 1.59 | 3.70 | 3.49 | 4.25 | 3.04 | 4.25 |
| Â |
| N, N | 2.43 | 2.77 | 2.94 | 3.38 | 2.71 | 3.05 |
| Â | Â | LN, LN | 1.42 | 3.90 | 4.45 | 5.94 | 3.99 | 5.49 |
| Â | Â | N, LN | 1.10 | 4.97 | 3.38 | 5.00 | 2.91 | 4.95 |
Halve |
| N, N | 2.38 | 2.72 | 3.17 | 3.30 | 3.18 | 3.03 |
| Â | Â | LN, LN | 2.25 | 2.50 | 4.48 | 4.61 | 4.15 | 4.14 |
| Â | Â | N, LN | 1.65 | 3.77 | 3.47 | 4.02 | 3.10 | 3.89 |
| Â |
| N, N | 2.73 | 2.48 | 3.33 | 3.16 | 3.10 | 3.09 |
| Â | Â | LN, LN | 3.34 | 1.69 | 5.26 | 4.06 | 4.91 | 3.64 |
| Â | Â | N, LN | 2.20 | 2.95 | 3.61 | 3.56 | 3.17 | 3.49 |
