Working memory capacity and resistance to interference

Single-task and dual-task versions of verbal and spatial serial order memory tasks were administered to 120 students tested for working memory capacity with four previously validated measures. In the dual-task versions, similarity between the memory material and the material of the secondary processing task was varied. With verbal material, three additional words had to be read aloud in the retention interval, and their phonological and semantic similarity to memory list words was varied orthogonally. With spatial material, choice RT tasks in the retention interval used stimuli from either the same or a different kind as the memory stimuli. Similarity had little effect on dual-task costs. For correlational analyses, individual dual-task costs were measured in various ways, which varied as to their direction of correlation with working memory capacity. In general, these correlations were low. Dual-task costs, although measured reliably, did not correlate across verbal vs. spatial tasks. The results lend little support to theories identifying working memory capacity with the ability to resist interference, or the ability to coordinate two concurrent tasks. 2004 Elsevier Inc. All rights reserved. Working memory capacity is often measured by tasks combining memory for serial order with a processing demand. For example, the reading span task designed by Daneman and Carpenter (1980) requires reading of a series of sentences, often followed by some judgment, together with memory for the last words of the sentences in their order of presentation. Likewise, operation span (Turner & Engle, 1989) and counting span (Case, Kurland, & Goldberg, 1982) combine arithmetic operations with retention of words or digits. These so-called ‘‘complex span tasks’’ can be described as dual tasks This work was supported by a TransCoop grant from the Alexander v. Humboldt Foundation to K. Oberauer and R.W. Engle, and by Grant OB 121/3-2 of Deutsche Forschungsgemeinschaft (DFG). We thank Petra Gr€ uttner for her invaluable help by organizing the data collection. We also thank Katrin G€ othe, Christina K€ orner, and Nadine Meyer for their help in scoring the tests. * Corresponding author. E-mail address: [email protected] (K. Oberauer). 0749-596X/$ see front matter 2004 Elsevier Inc. All rights reserv doi:10.1016/j.jml.2004.03.003 requiring simultaneous short-term retention of some information and processing of other, often unrelated, information, thereby matching the definition of working memory as a system for simultaneous storage and processing (Baddeley, 1986). Several studies have shown that complex span tasks can be dissociated from simple span tasks (i.e., serial recall tasks without additional processing demand) by factor analysis (Cantor, Engle, & Hamilton, 1991; Conway, Cowan, Bunting, Therriault, & Minkoff, 2002; Engle, Tuholski, Laughlin, & Conway, 1999; Kail & Hall, 2001; Oberauer, S€ uß, Wilhelm, & Wittmann, 2003) and by neuroimaging (Fletcher & Henson, 2001; Postle, Berger, & D Esposito, 1999; Smith, Geva, Jonides, Miller, Reuter-Lorenz, & Koeppe, 2001; Smith & Jonides, 1999). Moreover, complex spans are better predictors than simple spans of performance in complex cognitive tasks such as language comprehension (for a review see Daneman & Merikle, 1996) and reasoning (Conway et al., 2002; Engle et al., 1999; Kyllonen & Christal, 1990; S€ uß, Oberauer, Wittmann, Wilhelm,