Confirmatory factor analysis of the Wechsler Memory Scale-III

Area of Science:

• Neuroscience
• Psychology
• Psychometrics

Background:

• The Wechsler Memory Scale-III (WMS-III) is a widely used neuropsychological assessment tool.
• Understanding the underlying factor structure of the WMS-III is crucial for accurate interpretation of memory performance.
• Previous research has explored various models for the WMS-III's latent structure.

Purpose of the Study:

• To analyze the latent structure of the Wechsler Memory Scale-III (WMS-III) using confirmatory factor analysis.
• To evaluate multiple theoretical models of memory constructs within the WMS-III.
• To determine the best fitting model for the WMS-III's subtests based on standardization sample data.

Main Methods:

• Confirmatory factor analysis (CFA) was employed to test five different models of the WMS-III's latent structure.
• The analysis utilized the covariance matrix derived from the 11 primary subtests of the WMS-III standardization sample.
• Model fit was assessed using standard fit indices (NNFI, RMSEA, GFI) and the Bayesian Information Criterion (BIC).

Main Results:

• A three-factor model, encompassing working memory, auditory memory, and visual memory, demonstrated the best fit to the data (NNFI = .98, RMSEA = .05, GFI = .98).
• Models proposing distinct immediate and delayed memory constructs yielded inadmissible parameter estimates, indicating model misspecification.
• The Faces subtest exhibited low commonality with the visual memory construct across all tested models.

Conclusions:

• The findings support a three-factor structure for the Wechsler Memory Scale-III, integrating working, auditory, and visual memory.
• The results question the construct validity of separate immediate and delayed memory factors in the WMS-III.
• The Faces subtest may not adequately represent the visual memory construct as measured by the WMS-III.