Table 2.

Estimated power to detect treatment differences in percent weight loss.

Groups per arm	Patients per group	Intraclass correlation coefficient (%)	Effective sample size	Detectable difference (effect size)	Power (%)
10	18 (180 per arm)	1	154	0.32	80
10	18 (180 per arm)	2	134	0.35	82
10	20 (200 per arm)	1	168	0.31	81
10	20 (200 per arm)	2	145	0.33	80
12	20 (240 per arm)	1	201	0.28	80
12	20 (240 per arm)	2	174	0.31	82