Modeling Working Memory Tasks on the Item Level

Area of Science:

• Cognitive Psychology
• Psychometrics

Background:

• Working memory (WM) is crucial for cognitive function.
• Existing models may not fully capture the complexity of WM tasks like Digit Span and Letter-Number Sequencing.

Purpose of the Study:

• To develop a more refined model of working memory tasks.
• To investigate the latent structure underlying Digit Span and Letter-Number Sequencing.

Main Methods:

• Finite mixture (FM) modeling was employed to analyze item responses.
• Ordinal latent traits were compared with continuous traits.

Main Results:

• Ordinal latent trait models better explained task variability.
• Distinct factors identified: short-term storage (STS), general attention control (GAC), mental sorting (MS), and backward ordering (BO).
• STS and GAC significantly correlated with the general ability factor (G); MS and BO roles were less clear.

Conclusions:

• The refined FM model provides a better understanding of working memory's latent structure.
• Working memory factors substantially explain general ability, highlighting WM's importance.
• A discontinuity in GAC suggests potential implications for understanding cognitive variations, particularly in mild mental retardation.