The heritability of ability tilts

Abstract

Tilts arise from within-subject differences in performance between two distinct cognitive ability measures (e.g., verbal minus quantitative). These are independent of general cognitive ability (GCA) and are likely a function of differential investment of time and other resources into the cultivation of one ability, at the expense of another. There is some debate about the meaning and measurement of tilts among psychometricians, but a body of research is emerging demonstrating that these are predictive of real-world outcomes independent of GCA. An open question concerns the heritability of tilts. Since nearly all phenotypic individual differences are heritable, tilts, if substantive, should not be an exception. It was found that tilts are modestly heritable (after controlling for participant age and residual correlations with GCA) in three samples (US children, Georgia Twin Study; Swedish adults, Swedish Twin Registry; US adults, MIDUS II). AE models better fit the tilt data in all but one case (Verbal - Reasoning, in the GTS, where an ACE model better fit the data). Comparatively large (non-shared) environmentalities were noted in all cases, potentially consistent with models predicting a role for niche-picking and experience-producing-drive dynamics in generating tilts. A Wilson-like effect was observed when the tilt heritabilities in the GTS were compared with their equivalent parameters in the other two (older) samples. The finding that tilts exhibit non-zero heritability in different age ranges and in two countries strengthens their external validity, and weakens claims that they are measurement artifacts, as predisposing genetic and environmental factors are part of their nomological network.