The interplay of g and mathematical abilities in large-scale assessments across grades

- We analyzed the interplay between mathematical ability, g, and Gf in six age cohorts.
- We compared three different models of cognitive abilities for two LSA test types.
- The two-correlated factors model and nested-factor models fitted the data best.
- Across three grade levels, LSAs measured math-specific ability and g.
- We related mathematical ability, g, and Gf to different student characteristics.

This study investigates the interplay between general cognitive abilities and domain-specific mathematical ability by applying three different models. The first model is a g factor model, the second model reflects two correlated factors, and the third model is a nested-factor model that comprises a g factor, a domain-specific mathematical factor and cognitive ability. Furthermore, we analyzed the relation between student characteristics such as self-concept in mathematics, gender, SES, and grade in mathematics. Using data from six different German LSAs of the three cohorts, Grade 5, 9 and 13 (sample sizes range from NGrade5 = 730 to NGrade9 = 3893), and two different frameworks (i.e., literacy and curricular), we confirmed that LSAs do test mathematical ability beyond g. The two cognitive factors correlated differently with the student characteristics considered; this calls for future longitudinal approaches to the issue.