Assessing Factor Structure and Measurement Invariance Of WISC-IV Among Almajiris Attending Quranic Schools and Pupils Attending Mainstream Public Primary Schools in Northern Nigeria
We examined the factor structure and measurement invariance of the Weschler Intelligence Scale for Children (WISC-IV)- among 11-17-year-old Almajiris (street children attending Quranic school) in Northern Nigeria, and age-sex-matched mainstream public primary school boys. Across 200 Almajiris and 201 public school children, we tested multiple plausible models to confirm the factor structure of WISC-IV, with the four-factor model (reflecting the four WISC domains) and the 2nd-order model showing adequate fit for both groups. We then tested the four-factor model’s measurement invariance using multi-group confirmatory factor analysis, which revealed partial metric invariance between the two groups with the Vocabulary sub-test being the most problematic, and partial scalar invariance, with sub-tests of the Working Memory Index (WMI) and Processing Speed Index (PSI) showing insufficient equivalence (internal bias). Further, examination of the latent means revealed that on the remaining two domains VCI and PRI Almajiris as a group showed poorer performance than PSCs not attributable to the bias of the subtests (external bias). Our study highlights within-country sub-population differences in the functioning of western tools in cross-cultural contexts, while offering cautious support for measurement invariance in certain domains of the WISC-IV in this non-Western population.
Steps to reproduce