Construct validity of the WISC-IV with a referred sample: direct versus indirect hierarchical structures

Area of Science:

• Psychometrics
• Developmental Psychology
• Educational Psychology

Background:

• The Wechsler Intelligence Scale for Children-Fourth Edition (WISC-IV) is a key tool for assessing children with learning difficulties.
• Previous research generally supported a higher-order structure for the WISC-IV, but alternative models have emerged.
• These alternative models propose general intelligence as a breadth factor rather than a superordinate factor.

Purpose of the Study:

• To examine the construct validity of the WISC-IV by comparing different factor models.
• To determine the best explanation for the WISC-IV factor structure in a sample of referred children.
• To assess the implications for interpreting WISC-IV scores in clinical practice.

Main Methods:

• Confirmatory factor analyses were used to compare first-order factor models and hierarchical models (indirect vs. direct).
• WISC-IV core subtest data from 345 children (224 boys, 121 girls) referred for learning difficulty assessments were analyzed.
• Model fit indices were evaluated to determine the best-fitting structure.

Main Results:

• The correlated four-factor Wechsler model showed good fit to the data.
• A direct hierarchical model provided a statistically significant improvement in fit over the indirect hierarchical and correlated four-factor models.
• The direct hierarchical model was identified as the best explanation of the WISC-IV factor structure, with the general factor explaining 71.6% of the common variance.

Conclusions:

• The findings support the direct hierarchical model as the most appropriate representation of the WISC-IV factor structure in this sample.
• The results align with previous research suggesting that the general intelligence factor plays a dominant role.
• Primary interpretation of the Full Scale IQ is recommended over factor index scores for these referred children.