Table 4.

Model fit (multigroup)

Model	χ2 (df)	CFI	BIC	RMSEA (90 % CI)	SRMR	χ2 Δ (df)	p	ΔCFI
1	315.466 (73)	0.977	126128.2	.062 (.055 to .069)	0.033	–	–
	166.316S
	149.150T
2	458.865 (84)	0.964	126189.5	.071 (.065 to .078)	0.131	143 (11)	0.0001	–
	216.519S							0.013
	242.346T
3	325.970 (75)	0.976	126123.8	.062 (.055 to .069)	0.048	10 (2)	0.04	–
	169.479S							0.001
	156.491T
4	375.655 (76)	0.971	126166	.067 (.060 to .074)	0.095	60 (3)	0.0001	–
	181.051S							0.006
	194.604T
5	375.703 (76)	0.971	126166	.067 (.060 to .074)	0.106	60 (3)	0.0001	–
	192.830S							0.006
	182.872T
6	320.676 (76)	0.977	126111	.061 (.054 to .068)	0.044	5 (3)	ns	0.000
	168.493S
	152.183T
7	331.689 (78)	0.976	126107.1	.061 (.054 to .068)	0.054	16 (5)	ns	–
	171.801S							0.001
	159.897T
8	355.477 (81)	0.974	126108.5	.062 (.056 to .069)	0.055	23 (3)	0.01	–
	192.932S							0.003
	162.545T
9	363.673 (83)	0.973	126101.8	.062 (.056 to .069)	0.051	31 (5)	0.01	–
	187.293S							0.004
	176.380T
10	386.363 (86)	0.971	126102	.063 (.057 to .070)	0.057	54 (8)	0.001	–
	210.075S							0.006
	176.288T

Models 1 has no restrictions; Model 2 fixes all factor loadings to be equal across groups; Model 3 fixed only decoding factor loadings, Model 4 fixed only fluency factor loadings, Model 5 fixed only comprehension factor loadings, and Model 6 fixed only comprehension/fluency factor loadings; Model 7 fixes decoding and comprehension/fluency factor loadings, and freed comprehension and fluency factor loadings. Model 8 fixed the intercepts of the decoding and comprehension/fluency factors, Model 9 fixed the residual variances of the indicators of these factors, and Model 10 fixed both the intercepts and residual variances. All χ² values are significant, p<.0001. χ² Δ values in Models 2 through 7 are relative to Model 1; those in Models 8, 9, and 10 are relative to Model 7

CFI comparative fit index, BIC Bayesian information criteria, RMSEA root mean square error of approximation, SRMR standardized root mean square residual