The unique contributions of the facilitation of procedural memory and working memory to individual differences in intelligence C.A. Was a, ⁎, J. Dunlosky a , H. Bailey b , K.A. Rawson a a Kent State University, Kent, OH 44242,United States b Washington University, One Bookings Drive, St. Louis, MO 3130, United States abstract article info Article history: Received 21 July 2011 Received in revised form 21 December 2011 Accepted 31 December 2011 Available online 8 February 2012 PsychINFO classification: 2340 Keywords: Working memory Procedural memory Intelligence Fluid intelligence Comprehension Individual differences Individual differences in working memory account for a substantial portion of individual differences in complex cognitive processes (e.g., comprehension) and fluid intelligence. However, a large portion of the variance in fluid intelligence and comprehension is unexplained. The current investigation was conducted to evaluate whether individual differences in the facilitation of procedural memory accounts for unique var- iance in intelligence not accounted for by working memory. To measure variability in the facilitation of pro- cedural memory, we used a task that required participants to first classify exemplars of two categories; facilitation was then operationalized by subsequent improvements in the speed of classifying new exemplars from those categories (i.e., an operation-specific memory procedure). Three measures of each focal construct (facilitation in procedural memory, working memory, comprehension and fluid intelligence) were adminis- tered to 256 participants. We used structural equation modeling to examine the relationships among these latent variables. Working memory did account for variance in fluid intelligence and comprehension, but most important, individual differences in facilitation of procedural memory accounted for unique variance in fluid intelligence and comprehension. © 2012 Elsevier B.V. All rights reserved. 1. Introduction Working memory (WM) has been touted as a major source of individual differences in learning and problem solving since Baddeley and Hitch (1974) proposed the multiple components model of WM. Measures of WM are related to comprehension, reasoning ability, crys- tallized intelligence (gC) and fluid intelligence (gF). Nevertheless, as we discuss below, WM is not identical to higher-order cognition, and in particular, gF. That is, WM accounts for only a portion of gF, with a large portion of variance left unexplained. Accounting for this unex- plained variance is the focus of our investigation, so we will briefly discuss previous research on the relations between WM and gF that mo- tivate it. A great deal of the research on intelligence and reasoning ability has focused on the relationship between WM and gF (e.g., Kyllonen & Christal, 1990), which continues to demonstrate that these two constructs are highly related. Based on these consistent results, sever- al researchers have argued that WM and gF (or perhaps general intel- ligence) are unitary concepts (for reviews, see Ackerman, Beier, & Boyle, 2005; Kane, Hambrick, & Conway, 2005; Oberauer, Schulze, Wilhelm, & Süß, 2005), but this view is no longer well received. For instance, Heitz et al. (2006) explained that although WM and gF are indisputably related (r = .70), approximately 50% of the variance between the two constructs is not shared. Ackerman et al. (2005) completed a meta-analysis and found the average correlation (r) between WM and g to be .48. Given that the majority of variance between the two constructs is unexplained, the question remains: If WM and gF are not unitary concepts, what other cognitive processes contribute to gF? Another potential contributor to variance in gF was described by Was and Woltz (2007), who investigated the relationship between WM, discourse comprehension, and a new task referred to as the availability of long-term memory (ALTM) task (see also Woltz & Was, 2006). This task in part measures the facilitation of procedural mem- ory, and in particular the facilitation of the procedures involved in classifying exemplars from a specific category. They proposed that individual differences in this facilitation accounted for unique vari- ance in discourse comprehension. To better understand their ratio- nale, we describe the ALTM task in detail next, and then we more fully explore how the construct that it taps (i.e., facilitation of proce- dural memory) differs functionally from WM. The procedure for mea- suring the facilitation of procedural memory (Woltz & Was, 2006, 2007) is illustrated in Fig. 1, which presents an example trial of the original ALTM task (Woltz & Was, 2006). Each trial in the task includes four components. All four trial components were completed before moving on to the next trial. Acta Psychologica 139 (2012) 425–433 ⁎ Corresponding author at: Kent State University, Educational Foundations and Spe- cial Services, 405 White Hall, Kent, OH 44242, United States. Tel.: +1 330 672 2929; fax: +1 330 672 3407. E-mail addresses: cwas@kent.edu (C.A. Was), jdunlosk@kent.edu (J. Dunlosky), hroth@artsci.wustl.edu (H. Bailey), krawson1@kent.edu (K.A. Rawson). 0001-6918/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.actpsy.2011.12.016 Contents lists available at SciVerse ScienceDirect Acta Psychologica journal homepage: www.elsevier.com/ locate/actpsy