A comparison of sex, age, and culture on mental processing scales of the Kaufman Assessment Battery for Children, English (K-ABC) and Spanish (K-ABC-S) versions

Abstract

The Kaufman Assessment Battery for Children purports to measure intelligence based on Luria-Das designated sequential and simultaneous processes, that are separate from academic achievement. It proposes to discriminate between levels of intelligence and to yield process profiles for educational diagnoses and prescriptions of special populations. Presently there is no consensus that the processes exist as proposed in the model, nor, through factor analysis, that the test accurately measures the proposed processes. In this study the Mental Processing Scales of the Kaufman Assessment Battery for Children in English (K-ABC) and in Spanish (K-ABC-S) were compared for ten ageblock years (3-12), for sex (males and females), and for culture (U.S.-Hispanic, U.S.-white, and Mexican). Using the standardization samples' raw scores on the K-ABC for 151 English speaking U.S.-Hispanics, and 1372 U.S.-whites, and using raw scores on the Spanish language K-ABC-S for 1065 Spanish-speaking Mexicans, age was found to be a main effect on 10 overall ANOVAs. Subtest raw scores were found to be linearly increasing with age, and a Tukey post hoc comparison indicated an apparent strong age effect for 3-8 year olds that diminished for 10-12 year olds. Age-related patterns emerged in the 78 ageblock ANOVAs, analyzed within cultural group. Only a few ANOVAs yielded significant differences on the sex main effect. A cultural group effect was significant on all 10 overall ANOVAs. The Mexican sample was lower than either English language cultural group on all but one subtest; there was no difference between U.S.-Hispanics and U.S.-whites on 7 of the 10 overall ANOVAs. On 41 of the 78 ageblock ANOVAs the Mexican sample had lower mean raw scores than either U.S. English language sample; 69 of the 78 ageblock ANOVAs revealed no difference between U.S.-Hispanics and U.S.-whites. Examination of the subtests' scores showed that no U.S.-Hispanic on the K-ABC, or Mexican on the K-ABC-S, obtained the maximum number possible right answers for five subtests. Caution was suggested in interpreting scores and constructing profiles. Questions still exist about the value of educational prescriptions based on test results. Recommendations were suggested for future research.