Invariance in the Reynolds Intellectual Assessment Scales for Black and White referred students.

An important part of validity evidence for a test's scores is that they are invariant-measuring the same constructs the same way across different groups. Although published a decade ago, no independent investigation of invariance has been conducted on the Reynolds Intellectual Assessment Scales (RIAS). In the current study, we investigated invariance of the RIAS's measurement of general cognitive ability (g) between Black and White students who were referred for special education services. Results indicated that the subtest scores showed strict invariance, but g's variance was not the same across groups. Specifically, the White group exhibited a higher mean (d = 0.60) and almost twice the variability in g as the Black group did. Thus, although the between-group mean differences in subtest scores and the Composite Intelligence Index are due to between-group differences in g, the Black group used a narrower range of g than the White group did in answering the RIAS items.