A Preliminary Investigation of the Psychometric Properties of the WISC-V and KABC-II for South African Children Experiencing Barriers to Learning and Development

Abstract

With a paucity of emic intelligence tests, South African psychologists frequently employ etic intelligence tests, such as the Wechsler Intelligence Scale for Children – Fifth Edition (WISC-V) and the Kaufman Assessment Battery for Children – Second Edition (KABC- II) to assess children’s intellectual functioning. As important decisions are made based on the results of these tests, investigations into the reliability, validity, and cross-cultural applicability of etic intelligence tests regularly used in South Africa are deemed imperative. With this in mind, the present study aimed to conduct a preliminary investigation into the reliability and validity of the WISC-V and KABC-II in a sample of South African children who experience barriers to learning and development. In addition, it aimed to investigate the cross-cultural applicability of these two measures in a South African sample through item analysis and a comparison of the performance of English First Language (EFL) and English Additional Language (EAL) participants. Situated in a post-positivist research paradigm, correlational and causal-comparative research designs were employed. A sample of 90 children with a variety of barriers to learning was recruited from three private remedial schools in the greater Johannesburg area. Quantitative statistical methods were used to investigate the psychometric properties of the two intelligence tests. These included descriptive statistics as well as an investigation into the internal consistency reliability of the two intelligence tests’ subtests through the Spearman-Brown split-half reliability method. Construct validity was investigated using Pearson’s correlation coefficients and second-order confirmatory factor analyses (CFA). To investigate the cross-cultural suitability of the two intelligence tests, the global and index/scale scores of 56 EFL and 32 EAL participants were compared through a repeated measures General Linear Model. Item analysis of the WISC-V and KABC-II subtests’ items was conducted through a Graded Response Model for all items and differential item functioning (DIF) between the two language groups investigated. The reliability coefficients of the subtests of the WISC-V and KABC-II indicated that the tests are consistent in assessing the cognitive abilities and processes they were designed to assess. Moreover, construct validity was supported by intercorrelations which highlighted that the WISC-V and KABC-II subtests demonstrated stronger correlations with their respective indices and scales than they did with any of the other indices or scales. Furthermore, for the majority of the WISC-V and KABC-II subtests, stronger correlations were observed between subtests within an index or scale, than with subtests from other indices or scales. Exceptions included the WISC-V Digit Span subtest, and the KABC-II Story Completion and Pattern Reasoning subtests. In addition, subtests of the KABC-II Learning Scale, Atlantis and Rebus, showed a weak correlation. While the results of this study supported the five-factor model proposed by the WISC-V development team for the primary indices and subtests through second-order CFA, the factor structure of the KABC-II Fluid Crystallised model, proposed by the development team, could not be fitted as suggested. Participants’ English language status had a significant impact on the WISC-V Full Scale Intelligence Quotient, General Ability Index, Verbal Comprehension Index, Visual-Spatial Index, and Fluid Reasoning Index. In contrast to this, only one KABC-II scale was significantly influenced by English language status, this being the Knowledge Scale. Based on level of difficulty, the item analysis revealed that the sequence of the items in the KABC-II and WISC- V subtests, differed to some extent from the ordering observed in the current sample. Although, cognisance must be taken that item order is usually only established on large standardisation samples. DIF between the two language groups was nonetheless observed in all of the WISC-V indices, while the KABC-II only demonstrated DIF in three of its scales. With very limited published research findings available on the WISC-V and KABC-II in samples of South African children, this research study contributed to our understanding of the psychometric properties of these measures in the South African context. In addition, the investigations conducted into the impact of English language proficiency on the performance of the participants provided evidence that the KABC-II is culturally more suitable than the WISC-V. Moreover, with very limited published research findings using item analysis to explore the discriminatory ability, difficulty, and DIF of the KABC-II and WISC-V subtests’ items, this study further contributed to our understanding of the cross- cultural suitability of these two measures, not only at a global level but also at an item level.