Do non-g factors of cognitive ability tests align with specific academic achievements? A combined bifactor modeling approach

- Estimation of bifactor (BF) models of intelligence and scholastic aptitude
- Successful application of BF modeling to educational achievement (grades)
- Convergent validities for general factors of intelligence, aptitude and grades
- Evidence for convergent (i.e., numeric-math) group factor relations
- Evidence for divergent (e.g., verbal-math) group factor relations

Bifactor models can clarify how general and specific intelligence factors relate to general and specific academic achievements. By modeling specific group factors that are orthogonal to the general factors one can establish systematic correlations between corresponding non-g residuals of general intelligence and achievement factors. Using data from the National Longitudinal Survey of Youth 1997 (NLSY97; n = 2155), bifactor models were estimated for cognitive ability tests (intelligence, scholastic aptitude) and high school grade point averages (GPA), presuming group factors for numeric/mathematics and for verbal/language content, respectively. The three general factors (intelligence, scholastic aptitude, GPA) were highly correlated. The group factors evidenced convergent validity for numeric abilities and mathematics achievement; the remaining group factor correlations were small. The results demonstrate that besides substantial correlations of the general factors, specific non-g abilities can be successfully linked to specific group factors of academic achievement.

264