Further insights on the French WISC-IV factor structure through Bayesian structural equation modeling

Area of Science:

• Psychometrics
• Developmental Psychology
• Cognitive Science

Background:

• The Wechsler Intelligence Scale for Children--Fourth Edition (WISC-IV) interpretation relies on a 4-factor model, partially aligning with the Cattell-Horn-Carroll (CHC) intelligence theory.
• Classical confirmatory factor analysis (CFA) often imposes zero cross-loadings, potentially causing model fit issues and biased results, not fully reflecting theoretical constructs.

Purpose of the Study:

• To investigate the structure of the WISC-IV using a novel statistical approach, Bayesian structural equation modeling (BSEM).
• To compare the fit of a direct hierarchical CHC-based model against the WISC-IV's standard 4-factor model and higher-order models.

Main Methods:

• Employed Bayesian structural equation modeling (BSEM) with approximate zero cross-loadings using informative priors.
• Analyzed data from 249 French-speaking Swiss children aged 8-12 years.
• Estimated the influence of latent variables on WISC-IV subtest scores.

Main Results:

• A direct hierarchical CHC-based model with 5 factors plus a general intelligence factor demonstrated superior fit to the WISC-IV data compared to the 4-factor structure and higher-order models.
• The general intelligence factor was better conceptualized as breadth rather than a superordinate factor.
• BSEM enabled detailed estimation of latent variable influences on individual subtest scores.

Conclusions:

• The 5-factor CHC model provides a more adequate representation of the WISC-IV structure.
• Reconceptualizing the general intelligence factor as breadth has implications for test interpretation.
• BSEM enhances the understanding of intelligence test structures and facilitates improved clinical interpretation of subtest scores.