Time and resource limits on working memory: Cross-age consistency in counting span performance Sarah Ransdell * and Steven Hecht Division of Psychology, Florida Atlantic University, 2912 College Avenue, Ft. Lauderdale, FL 33314, USA Received 26 February 2003; revised 14 August 2003 Abstract This longitudinal study separated resource demand effects from those of retention interval in a counting span task among 100 children tested in grade 2 and again in grades 3 and 4. A last card large counting span condition had an equivalent memory load to a last card small, but the last card large required holding the count over a longer retention interval. In all three waves of assessment, the last card large condition was found to be less accurate than the last card small. A model predicting reading comprehension showed that age was a significant pre- dictor when entered first accounting for 26% of the variance, but counting span accounted for a further 22% of the variance. Span at Wave 1 accounted for significant unique variance at Wave 2 and at Wave 3. Results were similar for math calculation with age accounting for 31% of the variance and counting span accounting for a further 34% of the variance. Span at Wave 1 explained unique variance in math at Wave 2 and at Wave 3. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Counting span; Memory capacity; Memory decay; Working memory Children who score higher on measures of complex working memory (WM) span perform better on measures of reading comprehension (Daneman & Carpenter, 1980; Daneman & Merikle, 1996), arithmetic (Turner & Engle, 1989) and speed of processing (Case, Kurland, & Goldberg, 1982; Kail & Park, 1994) among other basic * Corresponding author. Fax: 1-954-236-1119. E-mail address: ransdell@fau.edu (S. Ransdell). 0022-0965/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.jecp.2003.08.002 J. Experimental Child Psychology 86 (2003) 303–313 www.elsevier.com/locate/jecp Journal of Experimental Child Psychology