Development and Evaluation of an Online, Multicomponent Working Memory Battery

Abstract

Research has demonstrated strong connections among working memory (WM), higher-level cognition, and academic achievement. Despite the importance of WM, currently available WM tests have practical limitations and lack comprehensive coverage of multiple WM components. The Working Memory Battery (WOMBAT) includes nine subtests measuring multiple content domains and processing demands, in accordance with contemporary WM theoretical frameworks. The current study evaluated the WOMBAT factor structure and identified misfitting items using confirmatory factor analysis and Rasch modeling with scores from 125 adolescents and 177 adults (N = 302). Overall, results indicated the WOMBAT measures separate Verbal, Static Visual-Spatial, and Dynamic Visual-Spatial dimensions, and that more than 98% of items contribute to measurement of those dimensions. This provides support for the theoretical organization of WM into three distinct content domains in the WOMBAT. Misfitting items were identified using infit and outfit indices for further review to improve reliability and stability. Results also demonstrated adequate person separation and Rasch person reliability and item reliability. Test-retest reliability and internal consistency coefficients suggest adequate reliability for early-stage research, but further refinement is needed before the WOMBAT can be used for individual decision making. Implications for future test development and research on the WM construct are provided.