Table 3.

Estimated coefficients of overlapping for distributions of statistics in Scenario 2.

Statistic	Scheme 1	Scheme 2
		Reciprocal-proportional	Uniform
Percent error	Uniform	0.78 (0.935)
	Proportional	0.84 (0.671)	0.98 (0.87)
Percent bias	Uniform	0.44 ($2.6\times {10}^{-11}$)
	Proportional	0.22 ($4.8\times {10}^{-14}$)	0.76 (0.504)
Adjusted percent error	Uniform	0.8 (0.178)
	Proportional	0.74 (0.74)	0.78 (0.554)
HPD size	Uniform	0.64 ($4.61\times {10}^{-4}$)
	Proportional	0.54 ($9.18\times {10}^{-6}$)	0.9 (0.649)

Each entry in each table compares a statistic between two sampling schemes. Numbers in parentheses are P values from post-hoc (Nemenyi) tests for the null hypothesis that the data used to estimate each KDE came from the same distribution, where these are <0.05 the coefficient of overlapping is given in boldface.