. Author manuscript; available in PMC: 2009 Aug 18.

Published in final edited form as: J Am Stat Assoc. 2008 Dec 1;103(484):1648–1658. doi: 10.1198/016214508000001057

Table 2.

Comparison of ranks of the true model M1 from various model selection criteria for MAR covariates in GMls

	(I)															(II)
	n = 100					n = 300					n = 500					n = 100					n = 300					n = 500
Rk	1	2	3	4	5	1	2	3	4	5	1	2	3	4	5	1	2	3	4	5	1	2	3	4	5	1	2	3	4	5
IC _H̃ ₍₀₎_,Q with 2d
1	294	22	2	1	0	317	29	2	1	0	316	32	6	0	0	267	34	12	1	0	298	46	6	2	0	305	47	12	4	0
2	8	70	17	1	0	15	53	13	5	0	8	50	19	0	0	12	62	29	5	0	8	37	23	3	0	6	29	36	3	0
3	0	2	61	4	0	0	1	51	11	0	0	0	44	9	0	0	2	47	8	0	0	5	49	8	0	0	6	33	6	0
4	0	0	2	13	2	0	0	2	0	0	0	0	1	15	0	0	0	2	15	2	0	0	2	13	0	0	2	1	10	0
5	0	0	0	1	0	0	0	0	1	0	0	0	0	0	0	0	0	1	0	1	0	0	0	0	0	0	0	0	0	0
IC _H̃ ₍₁₎_,Q with 2d
1	290	24	2	1	0	307	26	3	0	0	296	32	5	1	0	265	35	13	1	0	280	37	9	3	0	269	45	14	3	0
2	11	68	13	1	0	22	49	12	5	0	26	44	17	1	0	10	60	21	4	1	24	41	14	3	0	36	27	27	6	0
3	1	2	63	2	0	3	8	47	11	0	1	4	44	11	0	4	3	52	11	0	2	10	50	10	0	5	10	38	6	0
4	0	0	4	14	2	0	0	6	0	0	1	2	4	11	0	0	0	4	11	1	0	0	7	10	0	1	2	3	8	0
5	0	0	0	2	0	0	0	0	12	0	0	0	0	0	0	0	0	1	12	1	0	0	0	0	0	0	0	0	0	0

NOTE: Two cases of missing fractions for x_i were included. Three different sample sizes n = 100, 300, and 500 simulated data sets were used for each case. The columns represent the results from AIC_Q.