Table 4. Comparison of five richness estimators based on 1,000 simulation data sets under a broken-stick model with S = 1,000 and CV = 1.01.

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 (499.8)	${\hat{S}}_{\mathrm{Chao1}}$	755	-245	42.2	42.7	248.6	0
	${\hat{S}}_{\mathrm{Jack1}}$	750.1	−249.9	22.6	22.4	250.9	0
	${\hat{S}}_{\mathrm{CB}}$	991.7	−8.3a	130.1	139.2	130.2a	0.92a
	${\hat{S}}_{\mathrm{LB}}$	907.1	−92.9	143.6	145.6	170.9	0.77
	${\hat{S}}_{\mathrm{GP}}$	861.7	−159.3	80.2	81.6	168.6	0.73
2,000 (666.5)	${\hat{S}}_{\mathrm{Chao1}}$	832	−168	31.9	29.1	171	0.01
	${\hat{S}}_{\mathrm{Jack1}}$	887.5	−112.5	22.1	19.5	114.7	0
	${\hat{S}}_{\mathrm{CB}}$	909.7	−90.3	44.4	43.2	100.6a	0.43
	${\hat{S}}_{\mathrm{LB}}$	923.7	−76.3a	97.5	87.2	123.8	0.67
	${\hat{S}}_{\mathrm{GP}}$	893.8	−114.2	56	53.3	120.1	0.77a
4,000 (799.6)	${\hat{S}}_{\mathrm{Chao1}}$	901.1	−98.9	22.3	21	101.3	0.05
	${\hat{S}}_{\mathrm{Jack1}}$	959.4	−40.6	17.3	16	44.1a	0.31
	${\hat{S}}_{\mathrm{CB}}$	906.3	−93.7	19.8	19.9	95.7	0.02
	${\hat{S}}_{\mathrm{LB}}$	961	−39a	66.2	60.9	76.7	0.75
	${\hat{S}}_{\mathrm{GP}}$	943.2	−56.8	37.8	37	68.2	0.79a
8,000 (889.7)	${\hat{S}}_{\mathrm{Chao1}}$	948.1	−51.9	16.6	15.3	54.4	0.24
	${\hat{S}}_{\mathrm{Jack1}}$	989.6	−10.4a	13.9	12.5	17.4a	0.83
	${\hat{S}}_{\mathrm{CB}}$	937.6	−62.4	12.5	12.1	63.6	0.01
	${\hat{S}}_{\mathrm{LB}}$	985.1	−14.9	46.8	44.4	49	0.81
	${\hat{S}}_{\mathrm{GP}}$	973.9	−26.1	26.9	25.9	37.5	0.87a

Notes.

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