. Within the investigation of memory span tasks, there are two basic theoretical issues presumably germane to the understanding of individual differences in cognitive abilities (Miyake, 2001). First, to what extent are individual differences in memory span tasks domain specific or domain general? Friedman and Miyake (2000) and Mackintosh and Bennett (2003) highlighted the domain-specific components of complex span tasks, whereas Turner and Engle (1989), Kane et al. (2004), andShih (2004) emphasized the domain-general component. Nevertheless, a quantification of the importance of those domain-general and domain-specific components is strongly needed. To our knowledge, this quantification has not yet been attempted. Engle, Kane, and Tuholski (1999) proposed a hierarchical view based on both domain-general and domain-specific components. A general factor can be equated with the domain-general component of span tasks, whereas specific factors can be equated with domain-specific components. This hierarchical view is explicitly tested in the present article. In addition, the relative importance of those components is specifically quantified. This quantification can help to illuminate the causes of the association usually found between memory span tasks and cognitive abilities. Moreover, it could shed light on the factors underlying individual differences in these tasks. Thus, the first goal of the present article is met by this quantification.The second theoretical issue concerns which factors underlie individual differences in memory span tasks. Surely, there are several relevant factors (Miyake, 2001). However, here we focus on the influential theoretical framework proposed by Engle, Kane, and Tuholski (1999)-namely, WM ϭ STM ϩ controlled attention. Thus, the second goal of the present article is to test the likelihood of the controlled attention view of the WM system in its relationship to cognitive abilities.Interestingly, Engle and Kane (2004) distinguished the microanalytic approach from the macroanalytic approach. The former approach is based on the selection of participants according to their scores on a given WM measure. Participants with high and with low scores are grouped and compared in elementary cognitive tasks such as the Stroop, the antisaccade, or the flanker task. The macroanalytic approach, in contrast, tests a large number of participants through several different tasks intended to represent the constructs of interest, such as short-term storage or WM. Structural equation modeling (SEM) analyses are usually performed to test the likelihood of a given theoretical view. Engle and Kane stated that the macroanalytic approach "gives a much cleaner and clearer picture [than The research referred to in this article was supported by Grant BSO-2002-01455 from the Ministerio Español de Ciencia y Tecnología. We thank Gerald Tehan, Mike Kane, and one anonymous reviewer for their helpful comments. Correspondence concerning this article should be addressed to R. Colom, Facultad de Psicología, Universidad Autóno...
