The reliability and stability of verbal working memory measures

Waters, Gloria; Caplan, David · 2003 · OpenAlex-citations

DOI: 10.3758/bf03195534

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the psychometric properties of seven commonly used verbal working memory (WM) measures, specifically assessing their reliability, stability, and factor structure across a wide age range. The research was motivated by the widespread use of WM tasks in cognitive aging literature and the need to determine whether these diverse tasks measure a single underlying construct or distinct cognitive capacities. Previous studies had yielded inconsistent results regarding the correlation between different WM tasks and the stability of subject classification into high- and low-span groups, particularly in elderly populations. The authors tested 139 individuals divided into five age groups (18–30, 50–59, 60–69, 70–79, and 80+ years). Participants were tested twice, approximately six weeks apart, on seven WM span tasks: alphabet span, backward digit span, missing digit span, subtract 2 span, running item span, and sentence span (using both syntactically simple and complex sentences). The study evaluated internal consistency, test–retest reliability, the stability of classifying subjects into discrete memory groups, and the correlations among tasks. Confirmatory factor analysis was also employed to examine the underlying structure of the measures. The results indicated that all measures demonstrated adequate internal consistency. Significant age effects were observed on all tasks except two. Test–retest correlations were significant for all tasks except the missing digit span, with magnitudes ranging from .52 to .81 and remaining consistent across age groups. However, classifying subjects into discrete memory span groups based on a single measure proved highly inconsistent across testing sessions and tasks. Classification into upper and lower quartiles was more stable than using absolute cutoff scores or three-group classifications. Correlational analyses revealed moderate relationships among many span tasks, and confirmatory factor analysis suggested that six of the seven tasks reflected a common factor. Notably, both test–retest reliability and the stability of subject classification improved significantly when a composite measure reflecting performance across several tasks was used, rather than relying on individual task scores. The findings suggest that while individual WM span tasks have adequate internal consistency, their test–retest reliability and the stability of categorical classification are limited. The strong evidence for a common factor across most tasks supports the view that these measures tap into a shared cognitive capacity. Consequently, the authors conclude that using composite measures derived from multiple tasks provides a more reliable and stable assessment of verbal working memory capacity than single-task measures. This has important implications for research in cognitive aging and other fields relying on WM assessments, suggesting that multi-task batteries are preferable for capturing individual differences accurately.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• working memory