Preprint
Investigating the heritability of school grades from age 10 to 16 using biometric growth curve models
PsyArXiv preprints
41 pages.
PsyArXiv
created: 18 August 2023; last edited: 19 August 2023 (2023), unpublished
Abstract
School grades are an important signal of academic ability with far-reaching implications, from educational attainment to labor market success. While there is robust cross sectional evidence of the influence of environmental and genetic factors on school grades, less is known about the development of both influences over time. On the one hand, cognitive ability is one central predictor of school grades so that the well-studied pattern of increasing heritability of cognitive ability with age may also apply to school grades. On the other hand, school grades are affected by a number of non cognitive abilities for which heritability estimates decrease or remain stable over time. Using twin data from the German TwinLife survey (52% males, 32% with migration background, 51% with high-educated parents), we employ biometric growth curve models to study how the heritability of math and German grades develops from age 10 to 16. For math grades, environmental influences shared on the family level decrease over time while genetic influences increase as the result of the amplification of already active genetic influences. By contrast, the genetic influences on German grades slightly decrease over time while non-shared environmental influences increase. So, for math grades the development of genetic and environmental influences resembles the pattern of cognitive ability, while for German grades, the results are comparable to the developmental pattern of non-cognitive traits. Our findings suggest that the Wilson effect cannot be simply applied to school grades in general as there are substantial between-subject differences in the developmental etiology of school performance.