Table 3.

Power for each method, scenario, and trial design

Period Effects	ICC	Sequences	Clusters per Sequence	Power
				Non‐parametric within‐period method	Parametric within‐period method	Cluster model	Cluster‐period model
Common	0.02	3	3	26.9	51.7	48.7	47.1
	0.02	3	11	72.8	77.4	91.8	90.1
	0.02	11	3	78.1	93.9	96.4	96.4
	0.02	11	11	99.8	100	100	100
	0.08	3	3	11.3	32.9	35.5	33.9
	0.08	3	11	29.6	35.4	73.1	70.5
	0.08	11	3	32.3	74.1	93.4	93.4
	0.08	11	11	79.5	95.4	100	100
Varying	0.02	3	3	22.4	48.7	45.0	39.9
	0.02	3	11	79.5	85.3	89.0	85.2
	0.02	11	3	79.9	93.0	95.4	94.5
	0.02	11	11	100	100	100	100
	0.08	3	3	10.5	32.4	45.3	22.1
	0.08	3	11	27.9	35.5	63.2	36.3
	0.08	11	3	35.7	73.0	80.4	67.9
	0.08	11	11	87.0	98.0	99.8	99.2