Table 8.

Summary of results for Simulation II: Mean cluster recovery precision for each method and each level of each design feature: Part 3 – the average of the absolute deviation of the true number of clusters minus the algorithmically selected number of clusters.

Design feature	Level	LCM (AIC2)	LCM (AIC3)	LCM (BIC)	K-means (CH)	K-median (MRPC)
Sample size	N = 100	0.38	0.55	0.92	1.28	0.79
	N = 200	0.31	0.27	0.55	1.23	0.63
	N = 400	0.31	0.11	0.27	1.18	0.53
Number of clusters	K = 2	0.16	0.01	0.00	0.43	0.07
	K = 3	0.20	0.06	0.14	0.59	0.14
	K = 4	0.26	0.20	0.43	1.27	0.50
	K = 5	0.41	0.47	0.89	1.60	0.96
	K = 6	0.63	0.83	1.44	2.26	1.56
Number of variables	V = 6	0.34	0.54	0.97	1.49	1.11
	V = 9	0.34	0.29	0.57	1.32	0.54
	V = 12	0.31	0.11	0.20	0.88	0.29
Cluster sizes	Equal	0.23	0.17	0.41	0.91	0.26
	60%	0.51	0.55	0.87	1.57	1.25
	10%	0.25	0.21	0.46	1.22	0.43
Error level	5%	0.19	0.18	0.28	0.37	0.26
	10%	0.29	0.26	0.53	1.36	0.63
	15%	0.52	0.50	0.93	1.96	1.06
Overall		0.33	0.31	0.58	1.23	0.65