. Author manuscript; available in PMC: 2009 Nov 2.

Published in final edited form as: Psychol Med. 2007 Sep 6;38(3):353–363. doi: 10.1017/S0033291707001341

Table 4.

Measurement model construction

	One-factor							Three-factor

Model	χ²	df	CFI	RMSEA	Test	Test difference		χ²	df	CFI	RMSEA	Test	Test difference
1	482.73***	27	<0.01	0.14	–	–		483.87***	27	<0.001	0.14	–	–
2	211.68***	27	0.58	0.09	2 v. 1	329.60^a	<0.0001	197.06***	26	0.62	0.09	2 v. 1	329.60^a	<0.0001
3	52.17**	26	0.94	0.03	3 v. 2	193.24^a	<0.0001	38.86**	25	0.97	0.03	3 v.2	193.24^a	<0.0001
4	46.71**	25	0.95	0.03	4 v. 3	6.64^a	<0.01	33.55**	24	0.98	0.02	4 v.3	6.65^a	<0.01
5	130.18	25	0.76	0.07	5 v. 4	127.70^a	<0.0001	116.27	24	0.79	0.07	5 v. 4	127.70^a	<0.0001
6	38.59*	24	0.97	0.03	6 v. 5	10.32^a	<0.01	25.63*	23	0.99	0.01	6 v. 5	10.32^a	<0.01
7	33.15	23	0.98	0.02	7 v. 6	6.17^a	<0.05	20.31	22	1.00	<0.0001	7 v. 6	6.17^a	<0.05
8	34.00	24	0.98	0.02	8 v. 7	0.72^a	0.39	21.07	23	1.00	<0.0001	8 v. 7	0.72^a	0.39
9	33.27	22	0.97	0.02	9 v. 8	1.91^b	0.59	19.98	21	1.00	<0.0001	9 v. 8	1.91^b	0.59
10	33.83	24	0.98	0.02	10 v. 8	1.53^a	0.22	22.39	23	1.00	<0.0001	10 v. 8	0.08^a	0.77

df, Degrees of freedom; CFI, comparative fit index; RMSEA, root mean square error of approximation.

A 1-df test.

A 2-df test. χ² difference testing was conducted using the difftest option in Mplus.

Model 1: factor loadings within and between depressive and anxiety disorders, threshold parameters within and between depressive and anxiety disorders, and covariance paths between concurrent depressive and anxiety disorders constrained to be equal. Model 2: same constraints as Model 1 except threshold parameters across depressive and anxiety disorders are different. Model 3: same constraints as Model 2 except the threshold parameter for childhood depressive disorder is free to vary. Model 4: same constraints as Model 3 except the threshold parameters for adolescent depressive disorder is free to vary. Model 5: same constraints as Model 4 except the threshold parameters for childhood and adolescent depressive disorder are constrained to be equal. Model 6: same constraints as Model 4 except the threshold parameter for childhood anxiety disorder is free to vary. Model 7: same constraints as Model 6 except the threshold parameters for young adulthood anxiety disorder is free to vary. Model 8: same constraints as Model 7 except the threshold parameters for childhood and young adulthood anxiety disorders are constrained to be equal. Model 9: same constraints as Model 8 except covariance paths between concurrent depressive and anxiety disorders are allowed to vary. Model 10: same constraints as Model 8 except factor loadings between depressive and anxiety disorders are allowed to vary.