• Corpus ID: 151519282

Do the Kaufman Tests of Cognitive Ability and Academic Achievement Display Ethnic Bias for Students in Grades 1 through 12

@inproceedings{Scheiber2016DoTK,
  title={Do the Kaufman Tests of Cognitive Ability and Academic Achievement Display Ethnic Bias for Students in Grades 1 through 12},
  author={Caroline Scheiber},
  year={2016}
}

• C. Scheiber
• Published 2016
• Psychology

Cultural bias in cognitive testing has a long and controversial history. As the demographic profile in the United States continues to change and becomes ethnically more diverse, the need for culturally appropriate test instruments has become a national concern among educators, clinicians, and researchers. The Kaufman Assessment Battery for Children, – 2 Edition (KABC-II) and the Kaufman Test of Educational Achievement – 2 Edition (KTEAII) are two well-known tests of intelligence and achievement. The tests’ popularity for the assessment of children is not only due to the quality of their psychometric properties, but also because they appeal to an ethnically diverse client population. Although the test publishers have put great effort in ensuring the appropriate validity and reliability criteria for these tests, the issue of test bias in terms of the tests’ construct and predictive validity across different ethnic groups has not been addressed. The present study investigated construct and predictive invariance across a nationally representative sample of Caucasian (n = 1313), Black (n = 312), and Hispanic (n = 376) children in grades 1-12. Confirmatory factor analysis and structural equation modeling, using the data from the KABC-II and KTEA-II standardization samples, was used to assess whether increasing sets of equality constraints fit the tests’ underlying theoretical model equally well for all three ethnic groups. Results from the construct invariance analysis showed that factorial invariance of the factor structure, based on seven Cattell-Horn-Carroll (CHC) broad abilities, was met for all three groups. Results from the predictive invariance analysis also demonstrated a lack of ethnic bias in the analysis of slopes; virtually all slopes for the five CHC-based KABC-II did not differ significantly by ethnicity. Thus, these five cognitive Indexes correlated about equally well with reading, math, and writing for Caucasians, Blacks, and Hispanics across three different grade Ethnic Test Bias in Intelligence and Achievement Testing 5 groups (grades 1-4: n = 724; grades 5-8: n = 743; grades 9-12: n = 534). However, the invariance analysis showed bias. Four of the five KABC-II CHC-based Indexes (excluding Knowledge/Gc, often considered the most culturally-loaded Index) demonstrated a persistent overprediction of the minority groups’ achievement across the same three grade groups. The overprediction was especially notable in the areas of relative strength for Blacks (Sequential/Gsm and Learning/Glr) and Hispanics (Simultaneous/Gv and Planning/Gf). One possible interpretation of this pervasive overprediction is that the educational system has failed to be flexible enough to take advantage of ethnic children’s strengths when teaching reading, math, and writing. Another key finding is that the global score of Fluid-Crystallized Index emerged as the fairest predictor of achievement across the age range. Perhaps the most global score, rather than the profile of five CHC Indexes, should be featured when predicting school achievement—a contention consistent with Gary Canivez’s theory and research. Outcomes of this research contribute to a scarce body of literature on ethnic test bias that goes beyond the simple comparison of mean score differences. Results of this study provide the evidence needed to justify continuous use of the KABC-II and KTEA-II in the assessment of children and adolescents for diverse ethnic groups. Furthermore, findings are generalizable beyond the Kaufman tests to other popular tests of intelligence and achievement; this is, because this study is based on the CHC factor structure, a universal theory of cognition that is used as the theoretical underpinning by many well-known tests of intelligence and achievement, including the most recent versions of the Wechsler scales. Ethnic Test Bias in Intelligence and Achievement Testing 6