The claim of the author of the NNAT is that it produces equal proportions of Black, White, Hispanic, and low-income students scoring in the gifted range. The present study involved 392 second grade students from a rural Florida school district who were administered the NNAT in a group format. The testing by the school district was for the purpose of increasing the number of low-income and minority students in their Gifted Program. A cut-score of 115 was used as the criterion for convening a child study team meeting to discuss if a formal intellectual evaluation by the district school psychologist was warranted. Of the total sample, only the Black, White, and low-income subsamples had sufficient students to allow for statistically meaningful comparisons. Mean IQ scores for Blacks (89.0) and Whites (98.1) differed by 9.1 standard score points. Students receiving a free or reduced price lunch (used as a proxy measure for low socioeconomic status) were compared with students not receiving this service. Mean IQ scores for all students on free or reduced lunch status (93.5) and all students not receiving this service (100.6) differed by 7.1 standard score points. Mean IQs of White students receiving a free or reduced price lunch (95.1) and White students not receiving this service (100.6) differed by 5.5 standard score points. Matching Black and White students on free and reduced lunch status reduced the Black-White difference by approximately one-third, or a difference of 6.0 standard score points (89.1 and 95.1, respectively). The results of this study do not support Naglieri's claim.
Key Words: IQ; NNAT; Blacks; Whites; Minorities; Low income groups; Naglieri; Gifted; Disproportionality.
The use of nonverbal tests of intelligence for the identification of intellectually gifted low- income and minority students has been advocated by those who consider such tests to be less biased towards culturally, linguistically and racially diverse populations. A further impetus for the use of such tests is the underrepresentation of the aforementioned groups in public school programs for the intellectually gifted and the concomitant belief that the use of such tests will ameliorate the problem of "disproportionality." Tests such as the Kaufman Assessment Battery for Children (KABC) and the more recent Kaufman Assessment Battery for ChildrenSecond Edition (KABGII), the Universal Test of Nonverbal Intelligence (UNIT), the Comprehensive Test of Nonverbal Intelligence (CTONI), and the Naglieri Nonverbal Ability Test (NNAT) are examples of nonverbal tests that are touted as less biased towards minority and lowincome groups than are the more verbally weighted intellectual measures. Such tests show reduced group differences, in contrast to traditional intelligence tests, such as the Wechsler series and the Stanford-Binet: Fourth Edition (SB:FE), which show an approximately one standard deviation difference between Blacks and Whites at the adult level (Gottfredson, 1997).
Proponents of nonverbal tests tend to define bias not in the conventional statistical sense, where bias is defined in terms of how uniformly a test predicts some outcome variable (e.g., school achievement), for different groups, but in terms of mean group differences, with the test showing a lesser group difference defined as the less biased test. Jensen (1980), for example, has characterized this view (that mean differences between groups are evidence of bias) as the "egalitarian fallacy." However, the more verbally and quantitatively weighted tests are better predictors of school success than nonverbal measures, particularly nonverbal measures that are limited in content, such as the NNAT, which consists only of figurai reasoning items (Lohman, 2005). In this sense, comprehensive measures of intelligence are less biased than strictly nonverbal measures of intelligence because they are better at predicting school success for members of all groups to which they are applied, and success in a Gifted program in particular. …