Table 5.

Standardized Parameter Estimates From the MIMIC Model, Visit 2 (N = 395, 2017–2019).

Step 1a: Differences in scores across groups using education in years
Group	Estimate	Standard error	p value	Cohen’s d
Male sex	−.069	.051	.176	0.019
Age	−.106	.052	.039	0.316
Education	.324	.048	<.001	0.096
Language	−.086	.052	.095	0.024
Step 1b: Differences in scores across groups using WRAT scores
Male sex	−.060	.048	.212	0.016
Age	−.077	.049	.112	0.256
WRAT	.477	.042	<.001	1.24
Language	−.061	.049	.213	0.017
Step 2: Language group differences across letter tasks
Letter ‘‘F’’	.005	.035	.881	—
Letter ‘‘A’’	−.005	.034	.873	—
Letter ‘‘S’’	.001	.035	.986	—

Note.‘‘—’’ indicates not applicable. Step 2 was only conducted once across the two models because education and WRAT scores were not involved in assessing differential test effects. The effective sample size was 395, given that nine cases had missing values. WRAT = Wide Range Achievement Test.