Wednesday, March 04, 2009

Where does attention fit in the CHC intelligence model?

I just read with great interest (and attention) an excellent article that investigated the relations between the construct of attention and intelligence within the confines of the CHC model of intelligence. A constant source of discomfort with the CHC model, and, in particular, the use of the CHC nomenclature to classify what is measured by the tests in individual intelligence batteries, has been the lack of clarity of the role/presence/validity of AC (attention/concentration) in the model. Carroll (1993) clearly articulated the unknown status of attention in a model of cognitive abilities when he stated:
"...it can be argued that attention is involved, in varying degrees, in all cognitive performances and, thus, in all performances that are regarded as indicating cognitive abilities. One can expect it to be very difficult to separate the attentional components of such performances from those components that represent latent traits of abilities other than the ability to attend. An individual differences factor could often be equally well interpreted either as a factor of some particular cognitive ability or as a factor of attentional ability
(p. 547)"
The current article reference and abstract is listed below:

Burns, N., Nettelbeck, T. & McPherson, J. (2009) Attention and intelligence: A factor analytic study. Journal of Individual Differences, 30(1), 44–57. (click here to view)

  • Abstract: Carroll (1993) found few factor-analytic studies that addressed attentional abilities. We reviewed and reanalyzed some of these studies and concluded that an exploratory approach to the study of the relationships between tests of attention and cognitive abilities was warranted. We sampled N = 147 adults from the general community and administered 17 tests of attention, including well-known neuropsychological tests along with tests drawn from the differential and experimental literatures on attention. We also administered 14 tests of cognitive ability designed to measure constructs described in Carroll’s taxonomy of intelligence, including a higher-order general ability factor. Regression of a general factor from the abilities battery onto a general factor from the attentional battery showed these two latent variables to be near identical (β = .98). Exploratory structural equation modeling, which allowed a model wherein the abilities part of the model was a confirmatory measurement model but the attention variables were modeled by three rotated exploratory factors, clarified this outcome. There were two sustained attention factors and one working-memory capacity factor with differential relationships with the latent abilities variables and with age. Results are discussed in the context of the network of processes that underlies a description of general cognitive ability at the psychological level, which includes mental speed, working memory, and sustained attention.

Although the study suffers from a being a restricted sample of adults (n=147), therefore begging for replication in younger samples, the beauty of the study is the presence of a large number of cognitive variables (14) selected to represent the CHC domains of Gc, Gf, Gv, Gs, and Gy (in the Catell-Horn model this would be combining Glr and Gsm) and 17 attentional variables AND the use of (I need to find out more about this) a combined confirmatory and exploratory factor analysis procedure called ESEM (exploratory structural equation modeling). Of interest were some of the following findings:

  • The latent factor correlaiton between the cognitive g factor and the attention g factor was .977. According to the authors, this suggests the interesting hypohtesis that "there is little that determines performance on the attentional tests but g, or that g is constituted essentially by executive attentional capacities." I find the later hypothesis of interesting in the context of Kane, Engle, Conway et al.s controlled executive attention model of working memory and the working memory = (or is strongly related to) G or Gf. The authors do discuss the relations between these various research findings.
  • Three separate attention factors were identified, two interpreted as reflecting aspects of sustained attention (with one being very similar to Gs abilities) and one working memory (Gsm-MW). Practically this suggests that many speeded cognitive tests on intelligence batteries may be reflecting the strong influence of sustained attention (as suggested in Carroll quote above). The other sustained attention factor might be getting at a more "attentional" construct as it had "less explicit demand for continuous speeded performance but a demand that performance be maintained for longer periods, or with more complex task demands, or both. This attention factor had a near-zero relationship with Gs but a substantial one with the higher-order general factor." Maybe we in the field of applied test development should examine the variables of this second attention factor and experiment with the development of applied psychometric measures for clinical use.
Kudos to the authors for making an important contribution to the evolution of CHC theory, a theory/model that needs continual refinement and exploration (see McGrew, 2009).

Technorati Tags: , , , , , , , , , , , , , , , , , , , , , ,

No comments: