“Consistent g-Factor as well as Consistent Verbal-Factor, Numerical-Factor and Figural-Factors in Nested Factor Models? Confirmatory Factor Analyses Using 3 Test Batteries”, 2014 (; similar):
We analyzed data of n = 562 students who took 26 ability tests from 3 batteries.
Higher-order factor, nested-factor & general-factor models fitted at least acceptably.
Test-battery-specific g-factors in nested-factor models correlated highly (r ≥ 0.91).
Verbal content & numerical content factors evidenced convergent-divergent validity.
Concerning the correlational structure of intelligence, there is a broad consensus regarding hierarchical models with a general factor at the apex (g), and less consensus regarding the number, content, and structure of more specific ability-factors hierarchically below g. Previous studies revealed very high correlations of test-battery-specific g-factors, whereas the consistency of more specific ability-factors has been neglected.
In order to investigate this, current data stemming from n = 562 high school students who took 26 mental ability tests from independently developed test-batteries were analyzed. Regarding the intelligence-structure, nested-factor models revealed a (relatively) better fit than higher-order models and general-factor-models. The test-battery-specific g-factors of the nested-factor models were substantially correlated (r ≥ 0.91); the correlations of the test-battery-specific verbal and numerical factors evidenced convergent and discriminant validity (convergent correlations: verbal—r = 0.83; numerical—r = 0.46; figural—r = 0.22).
These results provided evidence that some group factors (besides the g-factors) of different test-batteries are largely similar.
[Keywords: intelligence, g-factor, domain-specific ability, confirmatory factor analysis, nested-factor modeling]