Table 2.

Reweighting observations can correct for sample bias.

$\mathit{\beta }$ Range	Ideal	Biased Sample
		KL		Bias against Minority
		Corrected	Naive	Corrected	Naive
0.01–0.125 (dispersed)	0.13	0.16	0.16	−0.2%	−14%
0.125–0.25 (ordered)	0.34	0.45	0.46	−0.1%	−40%
0.25–0.5 (near-critical)	0.43	0.55	0.60	9.6%	−51%
0.5–1.0 (critical)	0.48	0.57	0.71	0.1%	−65%