Table 1.

Cross-Language Factorial Invariance Model Testing for the Ethnic Identity Scale Administered to English- and Spanish-Speaking Mexican American Adolescents

Subscale	df	χ2 Value	Comparative fit index	Root mean square error of approximation	Standardized root mean square residual	Akaike information criterion	Δχ2(df) S-B χ2 2Δ (df)b
Exploration
Configural invariance	28	111.38	0.93	0.09	0.04	13,671.30	—
Configural invariancea	26	54.31	0.98	0.06	0.03	13,618.22	—
Weak invariance	32	56.98	0.98	0.05	0.03	13,608.90	2.68(6)
Strong invariance	38	63.16	0.98	0.04	0.04	13,603.08	6.18(6)
Strict invariance	45	72.23	0.98	0.04	0.04	13,598.14	9.07(7)
Resolution
Configural invariance	4	12.62	0.99	0.08	0.02	5,670.55	—
Weak invariance	7	15.04	0.99	0.06	0.02	5,666.98	2.43(3)
Strong invariance	10	15.95	1.00	0.05	0.02	5,661.89	0.91(3)
Strict invariance	14	19.44	1.00	0.03	0.03	5,657.37	3.49(4)
Affirmation
Configural invariance	18	53.01	0.91	0.08	0.04	9,498.46	—
Configural invariancea	16	25.56	0.97	0.04	0.03	9,455.84	—
Weak invariance	21	32.12	0.97	0.04	0.05	9,454.82	5.26(5)
Strong invariance	26	42.12	0.96	0.04	0.07	9,456.42	11.72(5)*
Strict invariance	32	71.10	0.90	0.06	0.15	9,518.91	20.84(6)**
Partial strict invariance	30	44.81	0.96	0.04	0.10	9,463.48	4.35(4)c

^a.Represents an adjusted configural model in which unique factor variances were permitted to correlate in both language groups to improve model fit.

^b.For exploration and resolution, under maximum likelihood estimation, the chi-square difference test is employed. For affirmation, under maximum likelihood restricted estimation, the Sattora–Bentler chi-square difference test is employed.

^c.Compares the partial strict model to the strong invariance model.

^*p < .05.

^**p < .01.