Table 7.

Standardized Parameter Estimates, χ² Test, and Model Fit Indices for Confirmatory Factor Analysis Final Models for the Whole Sample and Each Group

	The whole sample	Cognitively unimpaired	Dementia / MCI	Male	Female	URG	Non URG	< 65 years	≥ 65 years	Low education	High education
n	411	317	82	171	240	90	314	165	152	144	267
Factor loadings
Flanker (F1, FA)	0.85	0.70	0.87	0.87	0.83	0.83	0.86	0.67	0.63	0.80	0.87
DCCS (F1, FA)	0.82	0.78	0.53	0.80	0.82	0.80	0.82	0.82	0.69	0.78	0.84
Processing Speed (F1, FA)	0.78	0.62	0.81	0.70	0.83	0.57	0.81	0.52	0.54	0.78	0.76
Working Memory (F1, FB)	0.79	0.45 (0.24)a	0.69	0.82	0.77	0.74	0.68 (0.19)a	0.67	0.68	0.78	0.78
Episodic Memory (F1, FB)	0.60	0.36	0.41	0.68	0.55	0.42	0.63	0.42	0.27	0.66	0.55
Vocabulary (F2, FC)	0.85	0.79	0.65	0.85	0.85	0.85	0.81	0.78	0.91	0.87	0.81
Reading (F2, FC)	0.90	0.93	1.00	0.87	0.92	0.89	0.87	0.91	0.86	0.91	0.82
Test intercepts
Flanker	6.04	9.94	4.01	5.86	6.18	6.43	6.05	11.62	9.40	5.44	6.54
DCCS	4.74	8.20	2.98	4.44	5.02	4.33	4.89	10.96	7.26	3.90	5.55
Processing Speed	4.03	5.58	2.65	4.17	3.96	4.72	4.00	6.71	5.47	3.82	4.22
Working Memory	4.17	6.33	2.77	4.16	4.19	4.29	4.22	7.40	6.00	3.75	4.52
Episodic Memory	−0.98	−0.79	−4.57	−1.32	−0.80	−1.50	−0.88	−0.52	−1.19	−1.40	−0.82
Vocabulary	3.03	3.57	2.50	3.32	2.86	2.23	3.74	3.50	3.66	2.46	3.82
Reading	2.35	3.01	1.62	2.24	2.43	1.26	3.23	3.24	2.81	1.55	3.57
Test residual variances
Flanker	0.29	0.52	0.25	0.24	0.31	0.31	0.27	0.55	0.61	0.36	0.24
DCCS	0.34	0.39	0.73	0.36	0.33	0.35	0.32	0.33	0.52	0.39	0.30
Processing Speed	0.40	0.62	0.34	0.51	0.32	0.67	0.35	0.73	0.70	0.39	0.42
Working Memory	0.38	0.65	0.53	0.33	0.40	0.45	0.38	0.56	0.54	0.39	0.39
Episodic Memory	0.64	0.87	0.83	0.54	0.70	0.82	0.60	0.82	0.93	0.57	0.69
Vocabulary	0.28	0.38	0.58	0.28	0.28	0.28	0.34	0.40	0.18	0.25	0.35
Reading	0.19	0.13	0b	0.24	0.15	0.21	0.24	0.17	0.26	0.17	0.33
Factor covariance c
F1 with F2	0.56	0.42	0.22	0.53	0.57	0.72	0.51		0.61	0.63	0.44
FA with FB								0.40
FA with FC								0.34
FB with FC								0.57
χ2 test
χ2	29.014	16.945	10.452	26.686	17.082	23.978	18.566	26.796	22.133	20.249	27.342
df	13	12	14	13	13	13	12	11	13	13	13
p-value	0.007	0.152	0.729	0.014	0.196	0.031	0.100	0.005	0.053	0.089	0.011
Model fit indices
CFI	0.986	0.991	1.000	0.973	0.994	0.952	0.992	0.939	0.969	0.983	0.979
RMSEA (90% CI)	0.055 (0.028, 0.082)	0.036 (0.000, 0.073)	0.000 (0.000, 0.080)	0.078 (0.034, 0.121)	0.036 (0.000, 0.078)	0.097d (0.029, 0.157)	0.093d (0.000, 0.077)	0.093 (0.049, 0.139)	0.068 (0.000, 0.115)	0.062 (0.000, 0.112)	0.064 (0.030, 0.098)
SRMR	0.031	0.031	0.044	0.043	0.036	0.054	0.025	0.041	0.042	0.035	0.034

Note. For the whole sample and each group except for the group with age < 65, a two-factor model fit well, including (1) fluid cognition (F1) measured by Flanker, DCCS, Processing Speed, Working Memory, and Episodic Memory, and (2) crystalized cognition (F2) measured by Vocabulary and Reading. For the group with age < 65, the two-factor model did not fit well, χ2(df=13) = 52.761, p-value < 0.001, CFI = 0.846, RMSEA (90% CI) = 0.136 (0.099, 0.175), SRMR = 0.077. However, a three-factor model fit well, including (A) executive function / processing speed (FA) measured by Flanker, DCCS, and Processing Speed, (B) memory (FB) measured by Working Memory and Episodic Memory, and (C) language (FC) measured by Vocabulary and Reading. Factor variances are one under standardized solution. Factor means are fixed at zero to satisfy model identification. Insignificant parameter estimates (p > .05) are underscored. CFI = comparative fit index; RMSEA = root mean squared error of approximation; SRMR = standardized root mean squared residual. Model fit is considered adequate by meeting the following criteria: CFI ≥ 0.95, RMSEA ≤ 0.08, SRMR ≤ 0.08.

^aFactor loadings in “()” are crossloadings on the crystalized cognition factor (F2).

^bFixed at zero to satisfy model identification.

^cThe factor covariance equals correlation because factor variances are one under standardized solution.

^dRMSEA was slightly higher than the cutoff value. With small sample sizes (n < 200), RMSEA tends to be too high and over reject the true population model (Curran et al, 2003; Hu & Bentler, 1998). Given this limitation of RMSEA and the other model fit indices CFI and SRMR both being adequate, the factor model was still concluded to be acceptable.