Table 3. Comparison of five richness estimators based on 1,000 simulation data sets under a negative binomial model with S = 1,000 and CV = 0.74.

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 (554.8)	${\hat{S}}_{\mathrm{Chao1}}$	846.9	−153.1	47.2	45.9	160.1	0.19
	${\hat{S}}_{\mathrm{Jack1}}$	843	−157	23.9	24.2	158.8	0
	${\hat{S}}_{\mathrm{CB}}$	1,016.4	16.4a	121.5	128.4	122.4	0.94a
	${\hat{S}}_{\mathrm{LB}}$	966.6	−33.4	137.3	138.5	141.2	0.87
	${\hat{S}}_{\mathrm{GP}}$	931.6	−68.4	89.9	89.8	113.0a	0.87
2,000 (750.5)	${\hat{S}}_{\mathrm{Chao1}}$	916.7	−83.3	30.1	28.3	88.6	0.28
	${\hat{S}}_{\mathrm{Jack1}}$	996.1	−3.9a	21.9	20.4	22.2a	0.92a
	${\hat{S}}_{\mathrm{CB}}$	984.8	−15.2	40.8	40	43.5	0.9
	${\hat{S}}_{\mathrm{LB}}$	982	−18	78.5	73.6	80.5	0.87
	${\hat{S}}_{\mathrm{GP}}$	966.2	−33.8	52.3	51.6	62.3	0.91
4,000 (890.8)	${\hat{S}}_{\mathrm{Chao1}}$	961.5	−38.5	17.3	16.4	42.2	0.5
	${\hat{S}}_{\mathrm{Jack1}}$	1,035.8	35.8	15.8	14.9	39.1	0.31
	${\hat{S}}_{\mathrm{CB}}$	966.2	−33.8	15.1	14.9	37.1	0.4
	${\hat{S}}_{\mathrm{LB}}$	991.2	−8.8a	40.5	39	41.4	0.88
	${\hat{S}}_{\mathrm{GP}}$	986.3	−13.7	28.2	27.9	31.4a	0.92a
8,000 (960.7)	${\hat{S}}_{\mathrm{Chao1}}$	984.5	−15.5	9.7	9.1	18.3	0.72
	${\hat{S}}_{\mathrm{Jack1}}$	1,021.6	21.6	10	9.5	23.8	0.37
	${\hat{S}}_{\mathrm{CB}}$	977.5	−22.5	7	6.8	23.6	0.12
	${\hat{S}}_{\mathrm{LB}}$	997	−3a	23	21.3	23.2	0.86
	${\hat{S}}_{\mathrm{GP}}$	993.8	−6.2	14.7	14.3	16.0a	0.93a

Notes.

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