Table 5.

Maximum percent missing to remain under target bias of .25: centralization, node removal defined by in-degree.

Network	Correlation with centrality	In-degree	Out-degree	Total degree	Bonacich power	Closeness	Betweenness
Interlock	−.75	48	48	48	^a	10	15
Interlock	.75	20	20	20	38	8	12
Prison	−.75	45	60	57	62	^a	10
Prison	.75	24	31	25	40	24	6
Sorority	−.75	55	50	59	52	^a	6
Sorority	.75	28	30	31	35	24	4
6th graders	−.75	35	34	37	65	^a	32
6th graders	.75	21	25	22	57	67	42
Coauthor	−.75	42	42	42	^a	5	12
Coauthor	.75	12	12	12	^a	1	9
Prosper	−.75	40	42	44	^a	52	17
Prosper	.75	23	27	25	63	27	14
Co-citation	−.75	42	42	42	^a	13	15
Co-citation	.75	11	11	11	^a	22	22
Elites	−.75	33	48	34	^a	68	13
Elites	.75	11	20	11	^a	23	6
HS 13	−.75	35	43	39	^a	^a	38
HS 13	.75	21	27	22	^a	51	43
BioTech	−.75	37	37	37	^a	5	15
BioTech	.75	12	12	12	^a	1	11
HS 24	−.75	38	43	42	^a	^a	37
HS 24	.75	21	26	22	^a	45	44
CSprings	−.75	33	33	33	^a	5	15
CSprings	.75	13	13	13	^a	1	11
Mean (Std Dev)	−.75	40.2 (6.6)	43.5 (7.4)	42.8 (8.3)	67.4 (5.5)	42.3 (31.1)	18.8 (10.7)
Mean (Std Dev)	.75	18.1 (5.9)	21.2 (7.5)	18.8 (6.8)	60.2 (14.2)	24.5 (21)	18.7 (15.4)

Cases where percent missing is above 70, our observed maximum. In these cases, 70 is used to calculate overall means. The maximum percent missing was calculated based on a quadratic fit to the data. Note that centralization is measured as the standard deviation of the centrality scores.