Table 1.

Fraction of rejected null hypotheses by PERMANOVA with simulated homoscedastic data and varying number of observations in two samples

	Type I error				Power, efficient size = 2
Sample size	5	10	15	20	5	10	15	20
5	0.046	—	—	—	0.072	—	—	—
10	0.041	0.059	—	—	0.096	0.134	—	—
15	0.051	0.049	0.047	—	0.089	0.135	0.172	—
20	0.049	0.047	0.059	0.050	0.106	0.134	0.167	0.225