Corrigendum

Abstract

‘Sample size requirements to detect a two- or three-way interaction in longitudinal cluster randomized clinical trials with second-level randomization’ by Moonseong Heo, Xiaonan Xue and Mimi Y Kim published in Clinical Trials 2014; 11: 503–507. DOI: 10.1177/1740774514532724.

The contents of Table 2 should be corrected as indicated by the italicized bold-faced numbers.

Table 2.

Comparison of sample sizes, theoretical power and empirical power to detect a three-way interaction effect XZT between two interventions, X and Z, and time T with the following parameters fixed: δ = 0.125, σ² = 1, ρ₁ = 0.5, ρ₂ = 0.05.

Randomization at third level
λ01, λ10, λ11	rτ	N 3 (0,0)	N 2	N 1	Total N	φ	$\tilde{\phi }$
1, 1, 1	0.0	8	15	5	2400	0.865	0.855
		4	30	5	2400	0.865	0.863
	0.1	24	15	5	7200	0.865	0.865
		12	30	5	7200	0.865	0.869
1.5, 1, 1.5	0.0	8	15	5	3000	0.918	0.917
		4	30	5	3000	0.918	0.918
	0.1	20	15	5	7500	0.865	0.866
		10	30	5	7500	0.865	0.879
Randomization at second level
λ01, λ10, λ11	rτ	N 3	N 2 (0,0)	N 1	Total N	φ	$\tilde{\phi }$
1, 1, 1	0.0	8	15	5	2400	0.865	0.862
		4	30	5	2400	0.865	0.865
	0.1	24	15	5	7200	0.865	0.872
		12	30	5	7200	0.865	0.876
1.5, 1, 1.5	0.0	8	15	5	3000	0.918	0.943
		4	30	5	3000	0.918	0.924
	0.1	20	15	5	7500	0.865	0.860
		10	30	5	7500	0.865	0.872