According to the Ability Differentiation Hypothesis (Spearman, 1927), high ability people are less constrained by a cognitive bottleneck and more constrained by specific abilities, compared with lower ability people. ... [ view full abstract ]
According to the Ability Differentiation Hypothesis (Spearman, 1927), high ability people are less constrained by a cognitive bottleneck and more constrained by specific abilities, compared with lower ability people. Theoretical and empirical work in this domain have generally focused more on mechanisms underlying lower differentiation at the lower end of ability: Those with lower cognitive abilities may be more constrained by a cognitive bottleneck, with greater constraints on lower level processes like memory and speed, and thus are more similarly constrained in their levels of performance across a variety of cognitive tests (Anderson, 1992; Detterman and Daniel, 1989). However, the processes underlying higher differentiation at the higher end of ability have received less attention. One possible process contributing to ability differentiation is de-differentiation of specific cognitive abilities with increasing general ability, perhaps because prior knowledge and motivation within a specific domain have larger effects on performance in that same domain for individuals with higher general cognitive ability. According to this hypothesis, those constrained less by lower level cognitive bottlenecks may show larger effects of different, downstream, content-specific bottlenecks.)
Using a large, nationally representative dataset of individuals throughout the lifespan (WJ-III standardization subsample), Tucker-Drob (2009) found strong evidence for ability differentiation using nonlinear factor analysis. We used a similar approach to analyze a large, nationally representative dataset of US school children. We used direct assessments of children’s working memory, executive function, and math, reading, and science ability assessed at six waves from Kindergarten entry to 3rd grade from the ECLS-K:2011 cohort (n=17,703).
We first replicated Tucker-Drob’s (2009) finding of higher g loadings among lower ability children. We then tested the hypothesis that children’s specific abilities (performance on the same measure at different waves) de-differentiated at higher levels of general cognitive ability. The method is similar to testing for ability differentiation. Ability differentiation is inferred from negative factor loadings of tests on a g x g factor. De-differentiation of specific abilities by general cognitive ability implies positive factor loadings of tests of specific ability s on a s x g factor. We find support for de-differentiation of mathematics achievement at higher levels of g. Follow-up analyses will test whether this extends to other specific abilities, and whether transfer of learning is a plausible explanation for specific ability de-differentation at higher levels of general cognitive ability.
Education , Cognition and Attention , Group differences
