Optimizing working memory assessment: development of shortened versions of complex spans, updating, and binding tasks.

Abstract

Given the lengthy administration of most working memory (WM) tasks, some researchers have developed reduced versions of these tests. However, they have focused primarily on complex spans. Recent studies suggested that estimating working memory capacity (WMC) using multiple tasks from different paradigms enhances measurement accuracy by isolating WMC variation from task- and paradigm-specific influences. Considering this, we evaluated whether complex spans, updating, and binding tasks could be shortened while maintaining robust psychometric properties. Participants completed full-length versions of tests from these paradigms, which were then segmented into early, intermediate, and later trial blocks. The shortened WM tasks were based on the early trial blocks. They accounted for most of the variance in a set of factor scores derived from the full-length versions of the WM tests (R² = 0.90). Additionally, the shortened and full-length versions presented a similar ability to predict fluid intelligence (Gf). The shortened tasks reduced administration time by 35%, saving around 30 min. To help researchers select the most suitable combination of shortened and/or full-length tasks, we calculated the Gf and WMC variance predicted by every possible task combination and the respective administration time. We believe that the shortened WM tasks will be highly valuable to researchers, as they provide reliable and valid WMC estimates in a time-efficient manner. We also examined whether using tests from different paradigms provides better WMC estimates than employing collections of tasks from the same class. Our results confirmed this hypothesis, highlighting the importance of diverse task selection to accurately assess WMC.