Table 7. Comparison of five richness estimators based on 1,000 simulation data sets under a power decay model p_i ∼ 1/i^0.9 with S = 1,000 and CV = 4.

See Table 1 for the notations of the estimators.

Size n (Observed richness)	Estimator	Average estimate	Bias	Sample s.e.	Average estimated s.e.	Sample RMSE	95% CI coverage rate
1,000 (388.5)	${\hat{S}}_{\mathrm{Chao1}}$	786	−214	75.5	72.1	226.9	0.31
	${\hat{S}}_{\mathrm{Jack1}}$	629.2	−370.8	23.6	24.5	371.6	0
	${\hat{S}}_{\mathrm{CB}}$	2,127.4	1,127	13,983	261,663.4	14,014	0.8
	${\hat{S}}_{\mathrm{LB}}$	1,171.9	171.9	381.1	351.9	417.7	0.94a
	${\hat{S}}_{\mathrm{GP}}$	944.3	−55.7a	139.8	134.8	150.3a	0.94a
2,000 (574.1)	${\hat{S}}_{\mathrm{Chao1}}$	900.3	−99.7	52.2	51.4	112.5	0.58
	${\hat{S}}_{\mathrm{Jack1}}$	861.9	−138.1	24.7	24.7	140.3	0
	${\hat{S}}_{\mathrm{CB}}$	2,529.5	1,529	2,294.9	3,729.1	2,756	1
	${\hat{S}}_{\mathrm{LB}}$	1,074.8	74.8	178.3	175.1	193.2	0.95a
	${\hat{S}}_{\mathrm{GP}}$	996.7	−3.3a	102.6	101.8	102.7a	0.9
4,000 (771.6)	${\hat{S}}_{\mathrm{Chao1}}$	971.2	−28.8	31.9	31.8	42.9a	0.86
	${\hat{S}}_{\mathrm{Jack1}}$	1,039.9	39.9	22	21.5	45.5	0.52
	${\hat{S}}_{\mathrm{CB}}$	1,159.2	159.2	68.6	68.4	173.3	0.29
	${\hat{S}}_{\mathrm{LB}}$	1,030.5	30.5	79.6	80.5	85.2	0.96a
	${\hat{S}}_{\mathrm{GP}}$	1,002.7	2.7a	56.3	58	56.4	0.9
8,000 (923.1)	${\hat{S}}_{\mathrm{Chao1}}$	994.2	−5.8	15.2	15.2	16.3a	0.95a
	${\hat{S}}_{\mathrm{Jack1}}$	1,080.5	80.5	14.6	14.9	81.8	0
	${\hat{S}}_{\mathrm{CB}}$	1,015.3	15.3	15.3	16.2	21.6	0.89
	${\hat{S}}_{\mathrm{LB}}$	1,005.3	5.3	30.7	29.4	31.1	0.95a
	${\hat{S}}_{\mathrm{GP}}$	1,000.6	0.6a	23.6	23.1	23.6	0.93

Notes.

^adenotes the smallest bias, smallest RMSE, and figure closest to 95% coverage.